the study of cancer research

Select "Patients / Caregivers / Public" or "Researchers / Professionals" to filter your results. To further refine your search, toggle appropriate sections on or off.

American Association for Cancer Research (AACR)

AACR grantee Star Chemtai Mungo, MD, MPH, is developing an innovative self-treatment strategy for cervical precancer.

Support Lifesaving Cancer Research. Donate Now.

Cancer is not a single disease, but rather a collection of diseases all characterized by the uncontrolled proliferation of cells.

In the latest episode of Believe in Progress, we are joined by John Kincade, a prominent voice in sports broadcasting and a three-time cancer survivor, who talks with Mitch about life, sports, and family.

More than 299,000 men in the U.S. are expected to be diagnosed with prostate cancer this year. Read about prevention, screening, and treatment options for this type of cancer.

Get updates on the AACR Annual Meeting 2024 recently held in San Diego on the AACR Blog.

Making Treatment Decisions: Abigail Johnston shares how she has approached decisions about her cancer care since she was diagnosed with metastatic breast cancer in 2017.

Whether honoring a special person or a special day, a donation to the American Association for Cancer Research has a lasting impact.

The official news website of the AACR Annual Meeting 2024. Stay up to date on the latest developments from the most important cancer meeting in the world.

The AACR Cancer Progress Report 2023 provides a comprehensive overview of the latest research-driven advances against the collection of devastating diseases called cancer.

The AACR and its more than 58,000 members worldwide are advancing a scientifically bold agenda against the collection of diseases we call cancer.

A new wave of research-driven discoveries and technological innovations are delivering – and will propel additional – transformative advances to save more lives from cancer..

By the Numbers

percent decrease of the overall age-adjusted cancer death rate in the U.S. from 1991 to 2020

therapeutics were approved for new or expanded uses by the FDA from Aug. 1, 2022, to July 31, 2023

million cancer survivors in the U.S. are living with, through, and beyond their disease thanks to research

cancer diagnoses in the United States are associated with preventable risk factors

Your donation to the American Association for Cancer Research helps our more than 58,000 members worldwide drive progress against cancer.

• The Progression of Cancer
• AACR Project GENIE®: Powering Precision Medicine
• Patient Advocacy

Skip to Content

• Conquer Cancer
• ASCO Journals
• Types of Cancer
• Navigating Cancer Care
• Coping With Cancer
• Understanding the Publication and Format of Cancer Research Studies
• Understanding Cancer Research Study Design and How to Evaluate Results
• How Are Cancer Drugs Discovered and Developed?
• Drug Approval and Labeling in the United States
• Evaluating Cancer Information on the Internet
• How Patient Advocates Help Cancer Research: An Expert Q&A
• Journals and Magazines
• Clinical Trials
• ASCO Annual Meetings
• Health Disparities and Cancer
• For Patient Advocates
• Public Policy Advocacy
• Cancer Awareness Dates
• Survivorship

Cancer research studies look for and find better ways to prevent, diagnose, and treat cancer. Doctors and scientists can design these studies in many ways to answer the questions they have. No study design is perfect. Each has strengths and limits.

When you are researching information about your or your loved one's cancer, it is important to understand how studies are designed. This can help you figure out what the results mean to you. Some research is very preliminary or "early", which means it will be a long time before the results affect patient care. Other research results may have an immediate impact on how doctors care and treat people with cancer.

There are 3 main types of cancer research studies:

Preclinical studies. This type of study is used in a laboratory to test whether a change in approach, called an intervention, may be useful to treat a cancer, and whether it appears to be safe. These studies are often done in cancer cells either in a petri dish or in an animal like a mouse. These results are very important for deciding which interventions to eventually test in people. Often, a lot of time will go by between testing interventions in the laboratory and having them available to treat people with cancer.

Experimental studies, called clinical trials. This type of study gives a group of volunteers an intervention. The intervention is the focus of the study. For instance, it may be a new treatment, medical device, or process. Researchers often compare the results of the intervention group to a group that does not get the intervention. This is known as the control group. For many studies, who does and does not get the intervention is selected at random, called randomization. In other studies, every person gets the intervention. Finally, there are some studies that use a specific selection process to make decisions about treatment.

Experimental studies and clinical trials help researchers learn more about how cancer starts or spreads. These studies can also test new imaging methods and ways to improve quality of life. Learn more about clinical trials .

Observational studies. This type of study observes groups of people in a natural setting. Researchers do not give an intervention. Instead they study the results of interventions already in place. For instance, researchers may find out whether a group of people has more cancer diagnoses than another group.

Observational research helps with the study of epidemiology. Epidemiology looks at how different risks influence, cause, or spread a disease in a community.

What are the types of experimental cancer research studies?

Experimental studies are more reliable than observational studies. This is because people in these studies are placed in the intervention group or control group, usually at random. Randomization lowers the chance that what they or the researchers assume or prefer will change the study results. These assumptions or preferences are called bias.

This type of study also helps researchers to better find and control for such features as age, sex, and other factors that can affect the results of the study.

Researchers may create specific rules, called eligibility criteria , when they ask people to join an experimental study. This often includes the type of cancer and the stage of cancer. This is to make sure the study's participants have specific things in common so the results are helpful for similar patients in the future.

Clinical trials and experimental studies test:

The effectiveness or safety of a new drug or combination of drugs

A new way of giving a kind of treatment, such as radiation therapy or surgery

A new treatment or way to prevent cancer

A new way to lower the risk of cancer coming back, called recurrence

A new way to relieve a side effect of cancer or its treatment

Researchers do clinical trials in segments called phases. Each phase of a clinical trial gives different answers about the intervention being tested. There are 4 phases of clinical trials .

In a phase 3 clinical trial, people in the study are usually randomly placed in either the intervention group or control group. Researchers can prevent bias in a clinical trial by keeping those people and themselves, or just those people, from knowing who is in each group. This is a process called "blinding."

The types of experimental studies include:

Double-blind randomized trial. The people in the study and the researchers do not know who belongs to the intervention group or control group. They find out only when the study ends. Most researchers feel this type of clinical trial produces the best study data.

Single-blind randomized trial. The people in the study do not know whether they belong to the intervention group or control group, but the researchers know.

Open or unblinded trial. The people in the study and the researchers know who belongs to each test group. This approach is used when it is not possible to use blinding. For instance, the study may be comparing a surgical treatment to a drug.

What are the types of observational cancer research studies?

In observational studies, researchers have less control over the people in the study. This means that certain factors could affect the results. However, these studies provide data that can help guide future research.

Types of observational studies include:

Case-control studies. These studies compare 2 groups of people. For instance, researchers could compare information about people with cancer and those who do not have cancer. People who have cancer are the case group. People who do not have cancer are then the control group. Researchers may look for lifestyle or genetic differences between the groups. By doing this, they hope to find out why one group has a disease and the other group does not. These studies are called retrospective. That means researchers study an event that has already occurred.

Cohort studies. These studies are prospective. That means researchers study an event as it occurs. They watch a group of volunteers for a long period of time and track something. For instance, they could track any new cancer diagnoses. This type of study can look at whether certain nutrients, exercise, or other action can prevent cancer. This approach can also find cancer risk factors. For instance, cohort studies have looked at whether postmenopausal hormone replacement therapy increases the risk of breast cancer.

Case reports and case series. Case reports are detailed reports of one person's medical history. If many people receive a similar treatment, the reports may be grouped into a case series. The results of case series studies are descriptions of patients' histories within a specific group. As such, doctors should not use them to choose treatment options. But case reports can help doctors think of new ideas for research studies in the future.

Cross-sectional studies. These studies look at how diseases interact with other factors within a specific group at one point in time. For example, a study might evaluate how patients who took a certain medication are doing ten years later to see who has and has not developed cancer. Because these studies only measure a point in time, they cannot prove that something causes cancer, but they can help scientists with future research.

What are cancer research review articles?

Medical journals publish many cancer research studies each year. This is good for adding to the scientific knowledge about cancer that lead to better treatment and care. However, the fast pace makes it hard for doctors, people with cancer, and caregivers to keep up with all of the new advances. Research studies are always shaping and reshaping the scientific understanding of cancer. But no single study is the final word on a type of cancer treatment. As a result, review articles are very helpful. Review articles study and sum up the findings of already published research on a certain topic.

Types of review articles include:

Systematic reviews. These articles summarize the best existing research on a specific topic at that time. Researchers use an organized method to find, gather, and review a number of research studies on a topic. By combining the findings of these studies, researchers can draw more reliable conclusions.

Meta-analyses. These articles combine data from several research studies on the same topic. This lets researchers find trends that are hard to see in single studies.

How do I know if cancer research results are reliable?

These questions can help you evaluate the quality of research study results:

Was the study peer-reviewed by the journal that published it? Peer review means that researchers who are not a part of the study looked over and approved the study's design and methods. Results from a study are more reliable if they are peer-reviewed. Learn more about peer review .

How long did a study last and how many people took part? A study is more useful and credible if the same results occur in many people over time. However, this rule does not apply to studies of rare types of cancer or cancer that is hard to treat. This is because there are fewer people to study. Also, cancer prevention trials are often much longer than treatment clinical trials. This is because it often takes longer to find out if a prevention strategy works compared to treatment of an existing cancer.

What is the phase of a new treatment study? Phase I and phase II clinical trials often tell researchers more about the safety of a treatment than how well it works. These studies tend to have less people than phase III clinical trials. Phase III clinical trials compare a new treatment with the standard of care. Doctors consider phase III clinical trials to be the most reliable.

Does the study overstate or oversimplify its results? Each study is a small piece of the cancer research puzzle. Medical practice rarely changes because of the results of 1 study. New results are exciting. But other researchers must confirm the results before the medical field accepts them as fact. Review articles are of special interest.

Questions to ask your health care team

Some patients decide to gather research study findings that may apply to their type and stage of cancer. Always talk with your health care team about how a specific study may or may not relate to your treatment plan. It is important to not stop or change your current treatment based only on something you read.

Consider asking your health care team these questions:

I saw a study about a new treatment. Is this treatment related to my type and stage of cancer? Were people like me included in that study?

What publications can I read to learn more about my cancer type ?

Are there medical journals that focus on the type of cancer I have?

Should I think about joining a clinical trial?

What clinical trials are open to me? Where are they located, and how do I find out more about them?

Related Resources

Medical News: 8 Ways to Separate Fact from Fiction What to Know When Searching for Cancer Information Online: An Expert Perspective Evaluating Cancer Information on the Internet

What Is a Cancer Research Advocate?

Research and Advocacy

More in this section.

Office of Academic Clinical Affairs

Masonic Cancer Center

• Diversity, Equity, and Inclusion
• Meet Our Leaders and Staff
• Our Partners
• In the Media
• Publications and Resources
• Our Patients
• Our Research
• Membership & Resources
• Membership Directory
• Proposal Development
• MN Cancer InFocus
• Education and Training
• Community Engagement
• Community Matters Newsletter
• Cancer Survivorship

What are the types of cancer research?

In any battle, knowing and understanding your adversary is crucial. This is true for our fight against cancer as well, and research is the key to our combat strategy. Cancer research plays a vital role in gaining insights into the nature of the disease. Through systematic investigation and the creation of new knowledge, we can enhance our understanding of cancer. Armed with this knowledge, we can then develop more effective strategies for the prevention and treatment of cancer.

Researchers study every stage of the cancer journey, from causes (called etiology) and prevention to screening, diagnosis, and treatment (therapy) as well as survivorship and quality-of-life (palliative) care. This is called the cancer continuum. 

What we know about cancer—how to reduce cancer risk, how it develops, how to treat it, and how to help people cope with it—all depends on different types of research and what is discovered as a result. 

The Masonic Cancer Center, University of Minnesota, is a community of more than 600 researchers who study cancer. The things they study are broken down into six research areas that are organized around specific themes spanning the cancer continuum. The programs interact with one another and with groups throughout the world to uncover better ways to treat and prevent cancer. These six areas, and the portions of the cancer continuum they span, are shown below. 

Within these research programs, MCC scientists conduct research at different levels and settings. We use these different levels and settings to group cancer research into four main types: basic, clinical, population-based, and translational. Below, we take a closer look at each of the four types.

Basic research

The first level of research is called basic research, also known as laboratory research or bench science. Basic researchers study the cells, molecules, and genes that are the building blocks of life, working to understand how healthy cells grow and then identifying the differences between those healthy cells and cancer cells. 

This approach allows researchers to control and test for many different factors, for example, turning specific genes on or off, or exposing cells to a specific substance, condition, or possible treatment—and then measuring the effects of whatever is tested. Researchers can even use the cells from healthy volunteers to do this testing—and the volunteers definitely don’t have to be human. Cells from animals like mice, or even lab grown cells, are commonly used in cancer research.

After finding a promising idea that works in cells, researchers need to take that idea to the next level. But, it’s not safe or practical to move directly from cells to people. That’s where animal research comes into play—because animal models have certain similarities to humans, researchers can carry out and repeat important experiments that would be practically or ethically impossible to test initially in people. Scientists will often use mice, fruit flies, or even zebrafish to try out an idea, test, or treatment! 

Findings from laboratory, or basic, research are an essential starting point for informing future tests and treatments. However, even the best lab research has its limitations. That’s because humans are complex creatures, and no animal model can perfectly predict how a specific type of cancer will progress, or how a particular treatment will work in patients. This is where another level of research, translational research, becomes crucial. We’ll tackle translational research a bit later in this blog.

Translational research

During translational research, researchers take what they have learned in the lab and apply it in patient care. The knowledge gained from this work then goes back to the lab to inform even further investigations. Many people say that translational research “bridges the gap” between basic and clinical research by bringing together a number of different specialists to refine and advance the application of a discovery. 

Translational research seeks to produce more meaningful, applicable results that directly benefit human health—in other words, that directly benefit patients. The goal of translational research is to move basic science discoveries into practice with patients more quickly and efficiently. 

Masonic Cancer Center researchers play a key role in designing and developing medicines that inform future treatment strategies, thanks in large part to our Cancer Research Translational Initiative (CRTI) . For example, CRTI has facilitated the translation of TriKE GTB-3550—a cancer therapy that uses special killer cells to attack cancer cells—led by MCC’s Dr. Jeff Miller. The first generation TriKE, or the initial design, was developed to study its effects on a specific set of drug-resistant leukemias in a first-in-human trial led by Dr. Mark Juckett. Thanks to that trial, MCC’s Dr. Martin Felices and team developed a more potent second-generation TriKE, and MCC’s Dr. Nicholas Zorko and team have used lessons from this process to create a third special TriKE dedicated to examining the response of drug-resistant solid tumors such as prostate cancer and sarcoma. 

And the benefits of translational research don’t stop with patients—this research provides a crucial pivot point after clinical trials are conducted as well. That’s because researchers can explore how the trial’s resulting treatment or guidelines can be implemented by physicians in their practice. And, the clinical outcomes might also motivate basic researchers to re-evaluate their original assumptions or find new things to test that their original research hadn’t yet explored. 

Clinical research

In clinical research, promising treatments or tools are carefully studied in people. Clinical research studies, also known as clinical trials , explore whether new treatments, medications, and diagnostic techniques are safe and effective. 

Clinical research includes more than new drug development—it can be used to test anything that helps prevent, find, predict, treat, or manage cancer. That could include testing a phone app to monitor symptoms, an exercise program to help patients stay active, or a questionnaire to help doctors and nurses monitor potential health issues like pain. 

Clinical research is a critical step in the research pipeline because it ensures that what is being tested is safe and will work well for large groups of people. Clinical trials are often designed to learn if a new treatment is more effective or has less harmful side effects than existing treatments. At MCC, our Clinical Trials Office (CTO) tackles this process in partnership with physician researchers. The CTO is a large team of cancer center professionals who are dedicated to meeting the needs of researchers and their patients by providing exceptional trial management services. Clinical research teams prioritize patient safety, and are focused on ensuring the highest level of data integrity and regulatory compliance. 

Today’s clinical trials often become tomorrow’s new standard of care, boosting many patients' quality of life now, and helping ensure that future patients have continuously higher standards of care. Participation in a clinical trial provides qualified patients with early access to cutting-edge therapies. Rigorous regulatory standards ensure that patient care while on these trials is as good and often better than standard treatments. 

Want to read more about clinical trials and the phases they go through? Check out our explainer blog to learn more about clinical trials and why they’re so important.  

Population-based research

While cancer affects all population groups, studies show it often has a larger or more severe impact on some groups over others. So, understanding and addressing these health disparities is crucial. 

Population-based research explores the causes of cancer, cancer trends, and factors that affect the delivery and outcomes of cancer care in specific populations. This field of cancer research brings together scientists whose research focuses on cancer prevention, early detection, health outcomes, and how to best share with people, especially diverse communities, the information uncovered. 

Many of our researchers focus on cancer risks for vulnerable populations—children, people with severe mental illness, medically-underserved communities, and patients from across Greater Minnesota who are geographically isolated due to cost of treatment and long travel distances. This is evidenced by initiatives like the 10,000 Families Study , a study of family health in Minnesota that invites families from across the state to participate over time with the purpose of understanding the influences of genetics, lifestyle, and environment on health and illness, including cancer.

At the end of the day, the reason our scientists and doctors are able to provide the most cutting-edge cancer treatments, prevention education, and diagnosis tools is because all of these types of research—from the very beginning stages in a lab all the way to the clinic—work together. Solving the problem of cancer is a collaborative effort, and we’re proud and humbled to have such a strong legacy of support from partners, community members, and beyond to continue writing cancer’s last chapter.

Want to get updates on what our researchers and doctors are up to? Sign up today for our Community Matters e-newsletter to receive a monthly round-up of MCC’s top stories and events . And, if you’re looking for education and training opportunities, get on the list for Career Connections, a monthly collection of all career-related offerings at the cancer center. 

Stanford Cancer Institute

Search stanford cancer institute.

Access to the resources of Stanford University -- including the Schools of Engineering, Law, Business, Humanities & Sciences and Education

Excellence in Scientific Discovery

The Institute's comprehensive investigations extend from the earliest phase of basic discovery to the development of new cancer diagnostics, treatment protocols and prevention strategies

Research Programs

Funding Opportunities

• Shared Resources

Key Initiatives

In addition to the six research programs identified in our Cancer Center Support Grant, the Stanford Cancer Institute supports a number of key initiatives designed to foster discovery, application and translation of scientific knowledge. Inter-disciplinary teams of collaborative investigators partner to solve some of the most challenging questions in cancer research. Some of our key initiatives include:

The Stanford Brain Metastases Consortium

The Stanford Brain Metastases Consortium is a unique partnership between scientists, physicians and surgeons from across the Stanford community, bringing together experts in research and clinical care of brain metastases. The goal of the Consortium is to increase the number of available clinical trials and collaborative publications, and to enhance patients' experience through a team-based approach to care.

Stanford Pancreas Cancer Research Group (SPCRG)

The members of this group are committed to advancing research and care for pancreatic cancer.  SPCRG is comprised of an inter-disciplinary team of collaborative investigators from multiple Schools at Stanford.

After Cancer: Stanford Cancer Survivorship Program

After Cancer's mission is    to improve the experience, and outcomes, of patients and caregivers throughout all phases of their cancer journey by advancing survivorship research, clinical care and education. 

Wipe Out Melanoma - California

Wipe Out Melanoma - California is a project focused on primary and secondary prevention research with a particular emphasis on Hispanic and low socioeconomic status Caucasian groups who are at the highest risk for advanced disease. It is a joint partnership with Samuel Oschin Comprehensive Cancer Institute at Cedars-Sinai, and part of a larger nationwide War on Melanoma- Federation of States.

Nasopharyngeal Cancer Program

The SCI offers specialized expertise in clinical trials for multiple rare cancers, including nasopharyngeal cancer. 

• Patient Care
• Clinical Trials
• Health Equity

Stanford Medicine

Health care.

©2024 Stanford Medicine

Advertisement

About the Journal

Cancer Research publishes impactful original studies, reviews, and opinion pieces of high significance to the broad cancer research community. Cancer Research seeks manuscripts that offer conceptual or technological advances leading to basic and translational insights into cancer biology.   Read More About the Journal

Editor-in-Chief

Christine A. Iacobuzio-Donahue

Impact Factor

Featured articles.

Fibroblasts in the Aged Pancreas Drive Pancreatic Cancer Progression

Daniel J. Zabransky, et al.

Exploiting Tertiary Lymphoid Structures to Stimulate Antitumor Immunity and Improve Immunotherapy Efficacy

Giulia Petroni; Serena Pillozzi; Lorenzo Antonuzzo

ASCL1 Drives Tolerance to Osimertinib in EGFR Mutant Lung Cancer in Permissive Cellular Contexts

Bomiao Hu, et al.

Trending - Altmetric

The Twitter live feed is not currently active.

Keep up to date with Cancer Research

Follow @CR_AACR

• Online First
• Collections
• Online ISSN 1538-7445
• Print ISSN 0008-5472

AACR Journals

• Blood Cancer Discovery
• Cancer Discovery
• Cancer Epidemiology, Biomarkers & Prevention
• Cancer Immunology Research
• Cancer Prevention Research
• Cancer Research
• Cancer Research Communications
• Clinical Cancer Research
• Molecular Cancer Research
• Molecular Cancer Therapeutics
• Info for Advertisers
• Information for Institutions/Librarians

• Privacy Policy
• Copyright © 2023 by the American Association for Cancer Research.

This Feature Is Available To Subscribers Only

Sign In or Create an Account

Premium Content

• MIND, BODY, WONDER

Can aspirin help protect against colorectal cancers?

A new study details how a daily dose could prevent or delay the progression of the world’s third most common type of cancer.

Aspirin is well known for its ability to ease pain from muscle aches and headaches; it reduces fevers; and low doses can thin blood, reducing the chance of clots that cause strokes and heart attacks. Now a new study suggests it may also play a role in colorectal cancer prevention.

Colorectal cancer, a cancer of either the large intestine or rectum, is the third most common type of cancer, and the second most common cause of death from cancer , worldwide. There were 1.9 million new cases diagnosed across the globe in 2020, according to the World Health Organization, and these numbers are expected to grow. In the United States, the rates of colorectal cancers have been rising in people younger than age 50 since the 1990s, which includes more young people dying from the disease, according to the National Cancer Institute .  

Now a new study published in the journal Cancer shows that colorectal cancer patients who took a daily dose of aspirin had a lower rate of metastasis to the lymph nodes and stronger immune response to their tumors. The research suggests that aspirin may be boosting the ability of the immune system to hunt for cancer cells.  

“It is a rather unexpected effect, because aspirin is mainly used as an anti-inflammatory drug,” says Marco Scarpa, a researcher at the University of Padova, and one of the authors of the study. As Scarpa notes, this study suggests that aspirin may be playing a slightly different role by stimulating the immune system’s surveillance response, which can then prevent or delay the progression of colorectal cancer.  

Your immune system is always surveilling the body for cells that just aren’t right. When they find such cancer cells, they will kill them just as they would kill invading bacteria or viruses, says Cindy Kin , a surgeon at Stanford University, who specializes in colon and rectal surgery.  

“The data about aspirin and cancer is really evolving,” says Maen Abdelrahim , an oncologist at Houston Methodist Hospital, who specializes in treating colorectal cancers. However, there are still a lot of unanswered questions about how aspirin can prevent and delay the progression of these cancers, as well as which subset of patients would benefit from a daily aspirin.  

( Colon cancer is rising among young adults. Here are signs to watch for. )

People who take a consistent use of aspirin have a lower risk of colorectal cancer, “but it has to be balanced with the risks,” which includes the possibility of bleeding in the gastrointestinal tract, says Jeff Meyerhardt , an oncologist and co-director of the Colon and Rectal Cancer Center at the Dana-Farber Cancer Institute, in Boston.  

Aspirin protects against colorectal cancer

There are several studies that suggest a link between aspirin and colorectal cancer prevention and delay. However, the mechanism by which aspirin does this is still unknown. That makes it hard to predict which patients will benefit the most.  

In a 2020 meta-analysis , which analyzed the results of 45 observational studies, researchers found that regular aspirin use was associated with less incidence of colorectal cancer.  

A low dose, between 75 and 100 milligrams, was associated with a 10 percent reduction in the risk of developing colorectal cancer; a regular dose of 325 milligrams was associated with a 35 percent decline.  

Other studies have shown a link between daily aspirin and a delayed progression , including a lower risk of dying in patients who had already been diagnosed with colorectal cancer.  

“What has been seen in multiple studies for colorectal cancers is that having a more robust immune reaction does seem to have a better outcome,” Meyerhardt says. “This is looking at how aspirin may interact with that.”  

Study suggests mechanism

In the study, researchers obtained tissue samples from 238 patients who had undergone surgery for the colorectal cancers. Of these patients, 12 percent were taking a daily low dose of aspirin for the prevention of heart disease. When compared to patients who were not taking aspirin, the researchers found a lower rate of metastasis to the lymph nodes, and higher numbers of immune cells that had infiltrated the tumors.  

This higher level of infiltration is thought to be linked to slower cancer progression—including the lower rate of spread to the lymph nodes—by allowing immune cells to enter the tumor mass and fight the cancerous cells more effectively.  

( Why is stomach cancer rising in young women? )

The researchers also found higher levels of immune markers that are responsible for triggering the immune system surveillance response. “It’s boosting the immune system, and it’s helping the immune system inside the tumor,” Abdelrahim says.  

In recent years, the immune system’s role in protecting against the development of cancer has become recognized.  

Patients with suppressed immune systems are at higher risk for developing cancers , compared to patients with a fully functional immune system. As these results suggest, aspirin may increase the vigilance of the immune system when it comes to the detection of colorectal cancers.  

“Your immune system is doing all of these things in the background, that you’re not even aware of,” Kin says. “It’s not just the tumor’s behavior and how aggressive it’s going to be, but it’s your body versus the tumor.”  

Related Topics

• INFLAMMATION
• DIGESTIVE SYSTEM

You May Also Like

Why is stomach cancer rising in young women?

These 4 medicinal herbs may help keep men healthy

How exercise can help—or hurt—your digestion

These microbes found in tumors promote cancer. What if we just kill them?

Turkey tail. Chaga. Lion’s mane. Do these mushroom supplements really help you?

• Environment
• Perpetual Planet
• History & Culture

History & Culture

• History Magazine
• Mind, Body, Wonder
• Paid Content
• Terms of Use
• Privacy Policy
• Your US State Privacy Rights
• Children's Online Privacy Policy
• Interest-Based Ads
• About Nielsen Measurement
• Do Not Sell or Share My Personal Information
• Nat Geo Home
• Attend a Live Event
• Book a Trip
• Inspire Your Kids
• Shop Nat Geo
• Visit the D.C. Museum
• Learn About Our Impact
• Support Our Mission
• Advertise With Us
• Customer Service
• Renew Subscription
• Manage Your Subscription
• Work at Nat Geo
• Sign Up for Our Newsletters
• Contribute to Protect the Planet

Copyright © 1996-2015 National Geographic Society Copyright © 2015-2024 National Geographic Partners, LLC. All rights reserved

Immunotherapy shows promise for aggressive breast cancer

• April 17, 2024

University of Hawaiʻi Cancer Center researchers have conducted a phase II clinical trial to evaluate the safety and effectiveness of pembrolizumab maintenance therapy, a type of immunotherapy in aggressive forms of breast cancer. The study aims to determine if it could offer a less toxic and more targeted alternative to traditional chemotherapy, thus improving patients’ quality of life.

The trial was led by Naoto T. Ueno, UH Cancer Center director, and Toshiaki Iwase, assistant professor and medical director, and was conducted at their former institution, The University of Texas MD Anderson Cancer Center.

The study produced significant findings that were recently published in Clinical Cancer Research .

“Undergoing pembrolizumab maintenance therapy made a huge difference in my cancer treatment,” said Deborah Sumulong, a cancer patient who underwent chemotherapy and struggled with allergic reactions, as well as hair loss and lingering fatigue. “I experienced less fatigue; the only side effect was mild colitis (colon inflammation). My scans also showed the cancer declined, with the recent one showing it was completely gone.”

“While the findings are promising, further research, including a large prospective clinical trial, is necessary to validate the identified biomarkers and refine patient selection criteria for pembrolizumab therapy,” said Iwase.

Read more on the UH Cancer Center website .

Related Posts:

• Improving breast cancer outcomes for Native…
• Naoto T. Ueno announced as new UH Cancer Center director
• New UH Cancer Center director shares vision for future
• previous post: Rising seas to eat away at underground infrastructure in Hawaiʻi, globally
• next post: Could fishponds help with Hawaiʻi’s food sustainability?

If required, information contained on this website can be made available in an alternative format upon request. Get Adobe Acrobat Reader

About Calendar COVID-19 Updates Directory Emergency Information For Media MyUH Work at UH

Gagana Samoa

Kapasen Chuuk

Kajin Majôl

ʻŌlelo Hawaiʻi

• Administrative

• Skip to main content
• Keyboard shortcuts for audio player

• Your Health
• Treatments & Tests

Health Inc.

• Public Health

Oncologists' meetings with drug reps don't help cancer patients live longer

Sydney Lupkin

Drug companies often do one-on-one outreach to doctors. A new study finds these meetings with drug reps lead to more prescriptions for cancer patients, but not longer survival. Chris Hondros/Getty Images hide caption

Drug companies often do one-on-one outreach to doctors. A new study finds these meetings with drug reps lead to more prescriptions for cancer patients, but not longer survival.

Pharmaceutical company reps have been visiting doctors for decades to tell them about the latest drugs. But how does the practice affect patients? A group of economists tried to answer that question.

When drug company reps visit doctors, it usually includes lunch or dinner and a conversation about a new drug. These direct-to-physician marketing interactions are tracked as payments in a public database, and a new study shows the meetings work. That is, doctors prescribe about five percent more oncology drugs following a visit from a pharmaceutical representative, according to the new study published by the National Bureau of Economic Research this month.

But the researchers also found that the practice doesn't make cancer patients live longer.

"It does not seem that this payment induces physicians to switch to drugs with a mortality benefit relative to the drug the patient would have gotten otherwise," says study author Colleen Carey , an assistant professor of economics and public policy at Cornell University.

For their research, she and her colleagues used Medicare claims data and the Open Payments database , which tracks drug company payments to doctors.

While the patients being prescribed these new cancer drugs didn't live longer, Carey also points out that they didn't live shorter lives either. It was about equal.

The pharmaceutical industry trade group, which is known as PhRMA, has a code of conduct for how sales reps should interact with doctors. The code was most recently updated in 2022, says Jocelyn Ulrich, the group's vice president of policy and research .

"We're ensuring that there is a constant attention from the industry and ensuring that these are very meaningful and important interactions and that they're compliant," she explains.

The code says that if drug reps are buying doctors a meal, it must be modest and can't be part of an entertainment or recreational event. The goal should be education.

Ulrich also points out that cancer deaths in the U.S. have declined by 33 percent since the 1990s , and new medicines are a part of that.

A comparison study of dynamic [ 18 F]Alfatide II imaging and [ 11 C]MET in orthotopic rat models of glioblastoma

• Open access
• Published: 22 April 2024
• Volume 150 , article number  208 , ( 2024 )

Cite this article

You have full access to this open access article

• Yue Pan 1 , 2   na1 ,
• Haodan Dang 2   na1 ,
• Haoxi Zhou 1 ,
• Huaping Fu 2 ,
• Shina Wu 2 ,
• Huanhuan Liu 1 ,
• Jinming Zhang 2 ,
• Ruimin Wang 2 ,
• Yuan Tian 3   na2 &
• Baixuan Xu 2   na2  

To investigate and compare the dynamic positron emission tomography (PET) imaging with [ 18 F]Alfatide II Imaging and [ 11 C]Methionine ([ 11 C]MET) in orthotopic rat models of glioblastoma multiforme (GBM), and to assess the utility of [ 18 F]Alfatide II in detecting and evaluating neoangiogenesis in GBM.

[ 18 F]Alfatide II and [ 11 C]MET were injected into the orthotopic GBM rat models ( n  = 20, C6 glioma cells), followed by dynamic PET/MR scans 21 days after surgery of tumor implantation. On the PET image with both radiotracers, the MRI-based volume-of-interest (VOI) was manually delineated encompassing glioblastoma. Time-activity curves were expressed as tumor-to-normal brain ratio (TNR) parameters and PET pharmacokinetic modeling (PKM) performed using 2-tissue-compartment models (2TCM). Immunofluorescent staining (IFS), western blotting and blocking experiment of tumor tissue were performed for the validation.

Compared to 11 C-MET, [ 18 F]Alfatide II presented a persistent accumulation in the tumor, albeit with a slightly lower SUVmean of 0.79 ± 0.25, and a reduced uptake in the contralateral normal brain tissue, respectively. This resulted in a markedly higher tumor-to-normal brain ratio (TNR) of 18.22 ± 1.91. The time–activity curve (TACs) showed a significant increase in radioactive uptake in tumor tissue, followed by a plateau phase up to 60 min for [ 18 F]Alfatide II (time to peak:255 s) and 40 min for [ 11 C]MET (time to peak:135 s) post injection. PKM confirmed significantly higher K 1 (0.23/0.07) and K 3 (0.26/0.09) in the tumor region compared to the normal brain with [ 18 F]Alfatide II. Compared to [ 11 C]MET imaging, PKM confirmed both significantly higher K 1 /K 2 (1.24 ± 0.79/1.05 ± 0.39) and K 3 /K 4 (11.93 ± 4.28/3.89 ± 1.29) in the tumor region with [ 18 F]Alfatide II. IFS confirmed significant expression of integrin and tumor vascularization in tumor region.

[ 18 F]Alfatide II demonstrates potential in imaging tumor-associated neovascularization in the context of glioblastoma multiforme (GBM), suggesting its utility as a tool for further exploration in neovascular characterization.

Avoid common mistakes on your manuscript.

Introduction

Glioblastoma multiforme (GBM), the most malignant form of astrocytoma, represents a paramount challenge in neuro-oncology due to its aggressive spreading growth pattern and complex pathophysiology (Burko et al. 2023 ; Wang et al. 2023 ). The current standard treatment modalities for GBM include surgical resection followed by adjuvant radiation therapy and chemotherapy with temozolomide (Clavreul et al. 2022 ; Raue et al. 2023 ). Despite these interventions, the clinical outcome for GBM patients remains poor (Petkovic et al. 2023 ). It ranks as a leading cause of mortality associated with brain tumors in the adult population, characterized by a median survival duration of approximately 12 months post-diagnosis and a five-year survival rate lingering around 5%, but the recurrence rate is close to 100% (Schaff and Mellinghoff 2023 ).

Recent advancements in imaging techniques have significantly enhanced our understanding and management of GBM. Magnetic resonance (MR) imaging, especially with advanced modalities like functional MRI sequences include and diffusion kurtosis imaging (DKI) and perfusion weighted imaging (PWI), has provided deeper insights into the tumor’s anatomy and physiology (Zhuang et al. 2022 ). Positron emission tomography (PET) imaging, coupled with novel radiotracers, has also made significant strides in GBM research. Tracers like [ 11 C]Methionine ([ 11 C]MET), an amino acid PET tracer, have shown promise in better identifying GBM extent and heterogeneity, as well as in monitoring treatment response (Galldiks et al. 2022 ; Nakajo et al. 2022 ). Despite advances in [ 11 C]MET PET imaging, it is important to acknowledge its inherent limitations. The relatively short half-life of [ 11 C], approximately 20 min, which limits the widespread availability and practical use of [ 11 C]MET (Michaud et al. 2020 ). Additionally, the increased uptake of [ 11 C]MET can also be observed in areas of inflammation, infection, and even in some benign brain lesions, which may lead to false positive diagnosis of GBM (Kim et al. 2022 ). This is also a challenge in accurately assessing treatment response and identifying tumor recurrence, particularly in radiation encephalopathy or pseudoprogression (Dang et al. 2022 ; Pessina et al. 2021 ; van Dijken et al. 2022 ).

On the other hand, [ 18 F]Alfatide II, a radiolabeled peptide based on the molecular structure of arginine–glycine–aspartic acid (RGD), which is a key component for targeting integrins. Integrins, specifically the αvβ3 subtype, are cell surface receptors that play a crucial role in cell adhesion, migration, and angiogenesis. The expression of integrin αvβ3 is significantly upregulated in various tumor types, including glioblastoma (GBM), and is particularly associated with tumor angiogenesis and metastatic potential (Bao et al. 2016 ). Although there has been research on [ 18 F]Alfatide II in the context of glioblastoma multiforme (GBM), studies focusing on dynamic PET semi-quantitative analysis and the kinetic parameters of [ 18 F]Alfatide II remain limited (Zhang et al. 2016 ). Additionally, comparative studies diagnosing GBM with [ 18 F]Alfatide II and [ 11 C]MET are scarce.

In our study, we aimed to explore the comprehensive comparison of dynamic PET imaging with [ 18 F]Alfatide II and [ 11 C]MET in orthotopic rat models of GBM, and to assess the appropriateness of [ 18 F]Alfatide II PET imaging for GBM.

Materials and methods

All animal experiments were approved by the ethics committee and institutional review board of our hospital. All SPF grade male Wistar rats from Charles River Laboratories (Beijing, China) aged 6–8 weeks were fed at the temperature of 18–25 ℃, with relative humidity of 35–70 ℃ and a 12-h day-to-day cycle to provide the rats with adequate food and drinking water and environmental enrichment. The animals were allowed a 7-day acclimatization period before the experiments.

Orthotopic animal model of GBM

C6 glioma cells from GuYan Biotech Co., Ltd. (Shanghai, China) were cultured in DMEM at 37 °C and 5% CO2. After two generations of cell culture, C6 cells were trypsinized and resuspended at the concentration of 1 × 10 7  cells/mL to prepare GBM model. Rats ( n  = 23) were anesthetized with 2% pentobarbital sodium (0.2 ml/100 g), and then their heads were fixed on a brain stereotactic instrument. After the hair on the top of the head was cut off, a small hole with a diameter of about 2 mm was drilled at the intersection of the left side of sagittal suture 3 mm and the front side of coronal suture 1 mm. Next, 10 μL of C6 cell suspension with a concentration of 1 × 10 7  cells /mL was injected into the brain with a 0.25 mL microinjector at a speed of 1 μL/min. After administering the injection, the syringe was carefully withdrawn. The burr hole was then sealed with bone wax, followed by suturing the skin closed.

Preparation of radiotracers

The prodrug NOTA-PEG4-E [c (RGDfK)2] kit was purchased from Ruida Fuming Technology Co., Ltd. The 18 F-fluoride in O-18 water was obtained from a cyclotron (HM-20 cyclotron, Sumitomo Heavy Machinery Co., Ltd., China People’s Liberation Army General Hospital). The radiolabeling of NOTA-PEG4-E [c (RGDfK)2] and [ 18 F]AlF was performed following the previously published procedure with some modifications (Guo et al. 2014b ). The final product was designated [ 18 F]Alfatide II ([ 18 F]-ALF-NOTA-E [PEG 4-c (RGDFK)] 2). [ 11 C]methionine was synthesized by online [ 11 C]methylation of L-homocysteine on GE TRACERlab FX-C Pro module. The purity of [ 18 F]Alfatide II and [ 11 C]methionine was determined by HPLC.

PET imaging of GBM rat model

The rats, following tumor cell implantation for 14 days, were anesthetized and scanned on a whole-body PET/MR scanner (SIGNA™ PET/MR, GE Healthcare, Chicago, USA). The animals were imaged with 3T MR imaging using a mouse brain radiofrequency coil by applying T1-weighted imaging and T2-weighted imaging (fast spin echo, FSE). Simultaneous dynamic PET and MRI was performed with 150 mm field of view for whole-body mouse imaging. For [ 18 F]Alfatide II imaging, the rat models received the intravenous bolus injection of 13.0 ± 2.1 MBq of the radiotracer. This was followed by a comprehensive 60-min dynamic scan, executed in list mode. For the [ 11 C]MET imaging procedure, a similar approach was adopted with the intravenous bolus injection of 15.0 ± 2.7 MBq and a 40-min dynamic scan. We conducted [ 11 C]MET and [ 18 F]Alfatide II PET/MR scans on 20 rats, with a time interval of no more than 48 h between the scans using the two tracers.

The list-mode data were reconstructed into 1 × 5 s, 1 × 25 s, 9 × 30 s, 5 × 60 s, 5 × 120 s and 10 × 240 s for [ 18 F]Alfatide II and 1 × 5 s, 1 × 25 s, 9 × 30 s, 5 × 60 s, 5 × 120 s and 5 × 240 s for [ 11 C]MET, respectively, using 3D ordered subset expectation maximization with 1 iteration, 32 subsets, and a VOXEL size of 0.42 mm, applying correction for random coincidences, decay, deadtime, and scatter correction.

PET data analysis

On the PET image with both radiotracers, the MRI-based volume-of-interest (VOI) was manually delineated encompassing glioblastoma. A spherical VOI of constant size (diameter, 10 mm) was positioned in normal brain tissue, including grey and white matter, in the contralesional hemisphere, liver and muscle. The mean standardized uptake value (SUVmean) of the VOI was measured, based on static PET/MR images taken 60 min for [ 18 F]Alfatide II or 40 min for [ 11 C]MET after injection. Time-activity curve (TACs) was obtained by VOI in each time frame of the whole 60 min ([ 18 F]Alfatide II) or 40 min ([ 11 C]MET) dynamic image sequence. The following parameters were extracted from TACs: time to peak (TTP) and slope of curve.

Kinetic modeling

The Two-tissue-compartment models (2TCM) was used to describe tracer kinetics of [ 18 F]Alfatide II and [ 11 C]MET, where C p denotes the concentration of tracer in arterial plasma, C t represents the interstitial and intracellular free or non-specific binding components, and C m signifies the specific components of the tracer. The transport and binding rates of the tracer are described as follows: The transfer from arterial plasma to tissue (K 1 [mL/g/min]), tissue clearance (K 2 [1/min]), the on-rate of specific binding (K 3 [1/min]), and the target dissociation rate (K 4 [1/min]).

Model equations are illustrated as:

The values of K 1 –K 4 are determined by fitting the model to the time–activity curve. The model equation was calculated quantitatively using the left ventricular region of interest (ROI) as C P , representing the arterial blood input function. To optimize the fit, Akaike information criterion (AIC) was employed for evaluating its efficiency (Watabe et al. 2006 ).

In terms of pharmacokinetics, the volume of distribution (V T ) can be categorized into two components: specific volume of distribution (V S ) and non-displaceable volume (V ND ), which represents the nonspecific distribution:

All calculations were performed by the analysis software (PMOD v.4.3, PMOD technologies).

Histological analysis

For validation of the PET image-based results immunofluorescent staining, western blotting and Biodistribution study were performed as described in the Supplemental Information. The blocking studies were carried out in 3 Orthotopic rat models of GBM. After injecting [ 18 F]Alfatide II for 60 min, PET imaging was performed. Two days later, imaging was performed 60 min after co-injection of unlabeled RGD (2 mg/kg) and [ 18 F]Alfatide II (13.0 ± 3.1 MBq).

Statistical analysis

Numerical data are presented as mean ± standard deviation (SD). GraphPad (version 5.03, GraphPad Software, San Diego, Calif) was used for the paired two-sample t test and was performed to compare SUV values. P  < 0.05 was considered to indicate a significant difference.

Radiosynthesis of the radiotracers

For [ 18 F]Alfatide II, the purity of the synthesized product is 99.9%. For [ 11 C]MET, the purity of the synthesized product is 99.0%. (Supplement Fig.  1 ).

Semi‑quantitative analysis of dynamic PET imaging

Both the radiotracers of [ 18 F]Alfatide II and [ 11 C]MET demonstrated enhanced delineation of the tumor tissue visually (Fig.  1 ). The tumor demonstrated a notably higher uptake of [ 11 C]MET, reaching an SUVmean of 1.07 ± 0.15(0.83, 1.30) at the 40 min. Similarly, [ 18 F]Alfatide II showed a persistent accumulation in the tumor, albeit with a slightly lower SUVmean of 0.79 ± 0.25(0.45, 1.63) observed after 60 min (Fig.  2 A). Crucially, compared to [ 11 C]MET, [ 18 F]Alfatide II presented a reduced uptake in the contralateral normal brain tissue, registering values of 0.39 ± 0.12(0.22, 0.71) (11C-MET) and 0.14 ± 0.05(0.07, 0.27) ([ 18 F]Alfatide II), respectively. This resulted in a markedly higher tumor-to-normal brain ratio (TNR) of 6.0 ± 1.91(3.05, 11.12) for [ 18 F]Alfatide II, compared to [ 11 C]MET (2.91 ± 0.81)(1.62, 3.96), as detailed in Fig.  2 . There was a high correlation (Pearson’s r  = 0.799) between [ 18 F]Alfatide II and [ 11 C]MET uptake in the tumor region (Fig.  2 F).The time–activity curve (TACs) showed a significant increase in radioactive uptake in tumor tissue, followed by a plateau phase up to 60 min for [ 18 F]Alfatide II and 40 min for [ 11 C]MET post injection. The TACs of the tumor-to-background revealed that the peak uptake of [ 18 F]Alfatide II occurred at 255 s, whereas that of [ 11 C]MET was observed at 135 s (Fig.  2 E).

PET/MR imaging of the GBM rat model with 11 C-MET and [ 18 F]Alfatide II at various time points. A PET images of C6 tumor-bearing rat at 10, 20, 30, and 40 min after injection of 15 MBq of 11 C-MET. Tumors are indicated by arrows. B PET images of C6 tumor-bearing rat at 15, 30, 45, and 60 min after injection of 13 MBq of [ 18 F]Alfatide II. Tumors are indicated by arrows. The MR imaging shows that the tumor areas are clearly depicted as high signal intensity on Cor T2 ( C ), T2 FSE ( D ), and T2 FLAIR ( F ), while appearing as low signal intensity on T1 FSE ( E )

Dynamic uptake (SUVmean ± SD) following injection of 11 C-MET and [ 18 F]Alfatide II ( n  = 20) after 14 days of tumor growth in the tumor region ( A ), liver ( B ), muscle ( C ) and the contra lateral hemisphere ( D ). E TBR based on SUVmean using the contra lateral hemisphere as reference. F Scatter plot of Pearson’s correlation ( r ) comparing tumor uptake of 11 C-MET and [ 18 F]Alfatide II

Quantitative analysis of dynamic PET imaging

Table  1 presented all the rate constants of both radiotracers. For [ 18 F]Alfatide II, in the tumor area, both K 1 (0.23 ± 0.16) and K 3 (0.26 ± 0.26) were significantly higher than in the normal brain region. Similar features were also shown in the [ 11 C]MET imaging. Compared to [ 11 C]MET imaging, PKM confirmed both significantly higher K 1 /K 2 (1.24 ± 0.79/1.05 ± 0.39) and K 3 /K 4 (11.93 ± 4.28/3.89 ± 1.29) in the tumor region with [ 18 F]Alfatide II.

Validation of results

Fig.  3 showed that the uptake of [ 18 F]Alfatide II by the tumor in the blocking group was significantly reduced. As shown in Fig.  4 B, VEGF, Ki-67, integrin αv and integrin β3 expression increased quantitatively in GBM tissues. Western blotting confirmed integrin αv and integrin β3 expression in GBM tissues but not in normal control tissues (Fig.  4 C). Furthermore, immunofluorescence staining of tumor tissues in the GBM rat brain confirmed the high expression of integrin αvβ3, tumor angiogenesis marker CD31 and astrocyte activation marker GFAP in glioblastoma region (Fig.  5 ), with low expression observed in healthy brain tissue. At the same time, the co-expression of the integrin αvβ3, vascular endothelium marker CD31 and GFAP were observed on neovascular endothelial cells in glioblastoma area.

Results of blocking in three rat models of glioblastoma: A The PET images of non-blocking group and blocking group 60 min after injection of [ 18 F]Alfatide II. B The tumor and other tissues uptake of [ 18 F]Alfatide II in the blocking group was significantly reduced, compared with the non-blocking group. The SUVmean of the tumor was reduced from 1.02 ± 0.09 to 0.55 ± 0.07. The two tumors are indicated with white arrows and the color scale is expressed in SUV values. * p  < 0.05 for paired comparisons

A Diagram of glioblastoma brain section showing the location of tumor (red area) and normal control region (blue area). B With several immunohistochemical staining applied on contiguous slices for a morphological analysis (hematoxylin–eosin) and for analyzing the expressions of VEGF, Ki-67, integrin αv and integrin β3. C Tumor expression of integrin αv and integrin β3 in glioblastoma models

Immunofluorescent microscopy of C6 brain tumor tissue (blue, DAPI/nuclei; red, Integrin αvβ3; green, CD31; orange, GFAP)

In our study, we focused on the comprehensive comparison of dynamic PET imaging with [ 18 F]Alfatide II and [ 11 C]MET in orthotopic rat models of GBM. We studied the biodistribution and kinetics of both the radiotracers and validated the findings by using immunofluorescent imaging, western blot, blocking and the immunohistological analysis, which has confirmed αvβ3 integrin expression in GBM. Our results indicate that [ 18 F]Alfatide II PET/MR imaging demonstrated superior targeting in glioblastoma, in line with previous findings on variable targeting of primary tumors by [ 68  Ga]RGD (Isal et al. 2018 ). Notably, we observed a higher tumor-to-normal brain ratio achieved with [ 18 F]Alfatide II imaging enhanced the detection and quantification of tumors, compared with that of [ 11 C]MET. This suggests a significant potential for [ 18 F]Alfatide II in the precise imaging and characterization of glioblastoma, especially in the context of its angiogenic activity. This approach aligns with the broader goal of understanding the complex nature of tumor neovascularization but also potentially enhances the assessment of treatment efficacy and monitoring of disease progression at the microvascular level.

Angiogenesis, the development of new blood vessels, is a critical process in the progression and metastasis of glioblastoma (Verdugo et al. 2022 ). Extensive research has identified various signaling pathways involved in glioblastoma growth and metastasis, presenting potential targets for therapeutic intervention (Duan et al. 2023 ). In this context, integrins, particularly involved in cell signaling during angiogenesis, emerge as crucial biomarkers and therapeutic targets. Among these, αvβ3 integrin, predominantly expressed on activated endothelial cells during abnormal tissue growth, plays a significant role in the progression and metastatic spread of glioblastoma (Echavidre et al. 2022 ; Mezu-Ndubuisi and Maheshwari 2021 ). This integrin, present on both tumor cells and tumor-associated neovasculature, but more selectively targeted in glioblastoma neovasculature, offers vital insights into the tumor microenvironment (Cheng et al. 2021 ; Roth et al. 2013 ). Our study’s use of [ 18 F]Alfatide II, a radiotracer targeting αvβ3 integrins in PET imaging, provides noninvasive, real-time evaluation of angiogenesis and its dynamics within the glioblastoma microenvironment (Shao et al. 2020 ). The radiosynthesis process of [ 18 F]Alfatide II is convenient and simple, and can be completed within 30 min, which has great advantages in clinical application (Yu et al. 2015 ). The results of biological distribution of [ 18 F]Alfatide II show that the radioactive accumulation in brain and lung area is relatively low, which also proves the successful application of [ 18 F]Alfatide II in patients with lung cancer, breast cancer and brain metastases (Wu et al. 2022 ; Yu et al. 2015 ). [ 18 F]Alfatide II showed high stability in vivo, and could be quickly removed from blood pool and kidney, and radioactive ligands were slowly metabolized into hydrophilic metabolites (Liu 2015 ). The time–activity curves (TACs) for the GBM in our study displayed an initial significant increased uptake followed by a phase of plateau. This pattern suggests the absence of trapping or irreversible binding of [ 18 F]Alfatide II (Guo et al. 2014a ). Reflecting the kinetics of radiotracer uptake in our study, the 2-tissue-compartment model (2TCM) revealed the transport rate of the tracer from blood to tissue, denoted as K 1 , was notably higher in GBM than in the normal brain tissue, suggesting the involvement of processes like increased angiogenesis, heightened tumor permeability, and blood–brain barrier (BBB) disruption (Grkovski et al. 2017 ). The kinetic parameters K 3 , related to radiotracer binding and cellular internalization or dissociation, was also significantly increased in the tumor (Lindemann et al. 2023 ), given that ligands like [ 18 F]Alfatide II are known to be internalized (Guo et al. 2012 ).

GBM, characterized by its aggressive nature, has been shown to compromise the BBB’s integrity through a cascade of tumor-induced vascular modification. The blood–brain barrier in GBM exhibits higher permeability compared to healthy brain tissue, attributed to deficient formation, abnormal neovascularization, upregulated transporters, and downregulated tight junction proteins (Liebner et al. 2018 ; Wu et al. 2021 ). Furthermore, the C6 glioblastoma model is acclaimed for its fidelity in replicating human GBM pathology, including the pivotal aspect of inducing BBB breakdown (Pournajaf et al. 2024 ). Research elucidates that the C6 tumor facilitates this breakdown via the secretion of factors like vascular endothelial growth factors (VEGF) and matrix metalloproteinases (MMPs), which collectively undermine the endothelial tight junctions and degrade the basal lamina, thereby compromising the barrier function (Feng et al. 2022 ; Matsuno et al. 2022 ). This mechanistic insight into BBB disruption by the C6 model underpins the observed efficacy of [ 18 F]Alfatide II in our study, suggesting that its enhanced imaging capability is in part attributable to the increased BBB permeability in GBM. The literature also suggests that [ 18 F]Alfatide II possesses the potential to traverse the compromised blood–brain barrier in GBM owing to its specific binding affinity towards integrin αvβ3, which governs migration and invasion of vascular endothelial cells in tumor area, and its cyclic RGD peptides structure that exhibits a heightened attraction for integrins (Liolios et al. 2021 ). This is particularly relevant for imaging applications, as it implies that [ 18 F]Alfatide II could provide enhanced visualization of brain tumors by accumulating in areas where the BBB is disrupted due to tumor growth. Our study’s findings, which indicate a high tumor-to-normal brain ratio for [ 18 F]Alfatide II, support this notion.

In our study, 11 C-MET demonstrated good morphological representation of GBM. However, based on the time–activity curve, the distribution of radioactivity in normal brain tissue was higher for [ 11 C]MET than [ 18 F]Alfatide II. As a result, its TNR values were significantly lower than those of [ 18 F]Alfatide II. This suggests that [ 11 C]MET might be less effective than [ 18 F]Alfatide II in differentiating between low-grade malignant or benign brain tumor and normal brain tissues. Although the pivotal amino acid imaging agents including [ 11 C]MET and [ 18 F]Fluoroethyl-L-tyrosine ([ 18 F]FET) have shown significant potential in the diagnosis of GBM, their application is still constrained by several important limitations (Lundy et al. 2020 ; Steidl et al. 2021 ). Firstly, [ 11 C]MET faces significant logistical challenges in clinical applications due to its short half-life (approximately 20 min), limiting its widespread use in medical centers without onsite radioisotope production facilities. Secondly, both [ 11 C]MET and [ 18 F]FET struggle to differentiate between tumor recurrence and post-radiation changes, as both conditions may show enhanced uptake, increasing the risk of misdiagnosis (Dang et al. 2022 ). Moreover, [ 11 C]MET may also demonstrate increased uptake in non-tumorous lesions, such as inflammation or infection, further complicating its interpretation (Hotta et al. 2019 ; Mattoli et al. 2021 ). As for [ 18 F]FET, it is slightly less specific as it may accumulate in certain non-malignant brain lesions, such as low-grade gliomas or non-glial pathologies (Fuenfgeld et al. 2020 ; Maurer et al. 2020 ). Overall, while [ 11 C]MET and [ 18 F]FET provide valuable biochemical insights, a more comprehensive assessment and additional diagnostic tools are still required for accurate diagnosis and therapeutic monitoring of GBM.

A notable limitation of our current study is the restricted sample size of the orthotopic Rat Models of GBM utilized, which may not comprehensively represent the diverse heterogeneity characteristic of glioblastomas. This diversity includes a range of biological behaviors and microenvironmental variations. Consequently, the effects observed using [ 18 F]Alfatide II in our study might not fully encapsulate the complexity and variability present in a broader spectrum of glioblastoma cases. Further investigations are required to explore these variances in detail, especially regarding the uptake patterns and distribution of [ 18 F]Alfatide II in different glioblastoma subtypes. Such studies would be instrumental in understanding the variability in tracer uptake and the expression patterns of target biomarkers like integrin αvβ3 across a more diverse GBM population.

In conclusion, our study presents valuable insights into the application of [ 18 F]Alfatide II in the imaging of glioblastoma, demonstrating its potential superiority over [ 11 C]MET in terms of tumor-to-normal brain ratio. The significant uptake of [ 18 F]Alfatide II demonstrates potential in imaging tumor-associated neovascularization in the context of GBM. Ultimately, this research advances more precise imaging for glioblastoma diagnosis and monitoring, potentially improving the personalized therapies.

Data availability

No datasets were generated or analysed during the current study.

Bao X, Wang MW, Luo JM, Wang SY, Zhang YP, Zhang YJ (2016) Optimization of early response monitoring and prediction of cancer antiangiogenesis therapy via noninvasive PET molecular imaging strategies of multifactorial bioparameters. Theranostics 6:2084–2098. https://doi.org/10.7150/thno.13917

Article   CAS   PubMed   PubMed Central   Google Scholar  

Burko P, D’Amico G, Miltykh I et al (2023) Molecular pathways implicated in radioresistance of glioblastoma multiforme: what is the role of extracellular vesicles? Int J Mol Sci 24:4883. https://doi.org/10.3390/ijms24054883

Cheng TM, Chang WJ, Chu HY et al (2021) Nano-Strategies targeting the integrin alphavbeta3 network for cancer therapy. Cells 10:1684. https://doi.org/10.3390/cells10071684

Clavreul A, Autier L, Lemee JM et al (2022) Management of recurrent glioblastomas: what can we learn from the French glioblastoma biobank? Cancers (Basel) 14:5510. https://doi.org/10.3390/cancers14225510

Article   CAS   PubMed   Google Scholar  

Dang H, Zhang J, Wang R et al (2022) Glioblastoma recurrence versus radiotherapy injury: combined model of diffusion kurtosis imaging and 11C-MET using PET/MRI may increase accuracy of differentiation. Clin Nucl Med 47:e428–e436. https://doi.org/10.1097/RLU.0000000000004167

Article   PubMed   Google Scholar  

Duan W, Xia S, Tang M, Lin M, Liu W, Wang Q (2023) Targeting of endothelial cells in brain tumours. Clin Transl Med 13:e1433. https://doi.org/10.1002/ctm2.1433

Echavidre W, Picco V, Faraggi M, Montemagno C (2022) Integrin-alphavbeta3 as a therapeutic target in glioblastoma: back to the future? Pharmaceutics 14:1053. https://doi.org/10.3390/pharmaceutics14051053

Feng D, Zhou J, Liu H et al (2022) Astrocytic NDRG2-PPM1A interaction exacerbates blood-brain barrier disruption after subarachnoid hemorrhage. Sci Adv 8:abq2423. https://doi.org/10.1126/sciadv.abq2423

Article   CAS   Google Scholar  

Fuenfgeld B, Machler P, Fischer DR et al (2020) Reference values of physiological 18F-FET uptake: implications for brain tumor discrimination. PLoS ONE 15:e0230618. https://doi.org/10.1371/journal.pone.0230618

Galldiks N, Lohmann P, Langen KJ (2022) The role of 11C-methionine PET in patients with newly diagnosed WHO grade 2 or 3 gliomas. Neuro Oncol 24:1557–1558. https://doi.org/10.1093/neuonc/noac120

Article   PubMed   PubMed Central   Google Scholar  

Grkovski M, Emmas SA, Carlin SD (2017) (18)F-Fluoromisonidazole kinetic modeling for characterization of tumor perfusion and hypoxia in response to antiangiogenic therapy. J Nucl Med 58:1567–1573. https://doi.org/10.2967/jnumed.117.190892

Guo N, Lang L, Gao H et al (2012) Quantitative analysis and parametric imaging of 18F-labeled monomeric and dimeric RGD peptides using compartment model. Mol Imaging Biol 14:743–752. https://doi.org/10.1007/s11307-012-0541-7

Guo J, Guo N, Lang L et al (2014a) (18)F-alfatide II and (18)F-FDG dual-tracer dynamic PET for parametric, early prediction of tumor response to therapy. J Nucl Med 55:154–160. https://doi.org/10.2967/jnumed.113.122069

Guo J, Lang L, Hu S et al (2014b) Comparison of three dimeric 18F-AlF-NOTA-RGD tracers. Mol Imaging Biol 16:274–283. https://doi.org/10.1007/s11307-013-0668-1

Hotta M, Minamimoto R, Miwa K (2019) 11C-methionine-PET for differentiating recurrent brain tumor from radiation necrosis: radiomics approach with random forest classifier. Sci Rep 9:15666. https://doi.org/10.1038/s41598-019-52279-2

Isal S, Pierson J, Imbert L et al (2018) PET imaging of (68)Ga-NODAGA-RGD, as compared with (18)F-fluorodeoxyglucose, in experimental rodent models of engrafted glioblastoma. EJNMMI Res 8:51. https://doi.org/10.1186/s13550-018-0405-5

Kim JW, Kang YK, Paeng JC et al (2022) Diverse patterns and clinical significance of 11C-methionine PET in dysembryoplastic neuroepithelial tumors. Clin Nucl Med 47:1040–1047. https://doi.org/10.1097/RLU.0000000000004400

Liebner S, Dijkhuizen RM, Reiss Y, Plate KH, Agalliu D, Constantin G (2018) Functional morphology of the blood-brain barrier in health and disease. Acta Neuropathol 135:311–336. https://doi.org/10.1007/s00401-018-1815-1

Lindemann M, Oteiza A, Martin-Armas M et al (2023) Glioblastoma PET/MRI: kinetic investigation of [(18)F]rhPSMA-7.3, [(18)F]FET and [(18)F]fluciclovine in an orthotopic mouse model of cancer. Eur J Nucl Med Mol Imaging 50:1183–1194. https://doi.org/10.1007/s00259-022-06040-z

Liolios C, Sachpekidis C, Kolocouris A, Dimitrakopoulou-Strauss A, Bouziotis P (2021) PET diagnostic molecules utilizing multimeric cyclic RGD peptide analogs for imaging integrin alpha(v)beta(3) receptors. Molecules 26:1792. https://doi.org/10.3390/molecules26061792

Liu S (2015) Radiolabeled cyclic RGD peptide bioconjugates as radiotracers targeting multiple integrins. Bioconjug Chem 26:1413–1438. https://doi.org/10.1021/acs.bioconjchem.5b00327

Lundy P, Domino J, Ryken T et al (2020) The role of imaging for the management of newly diagnosed glioblastoma in adults: a systematic review and evidence-based clinical practice guideline update. J Neurooncol 150:95–120. https://doi.org/10.1007/s11060-020-03597-3

Matsuno H, Tsuchimine S, O’Hashi K et al (2022) Association between vascular endothelial growth factor-mediated blood-brain barrier dysfunction and stress-induced depression. Mol Psychiatry 27:3822–3832. https://doi.org/10.1038/s41380-022-01618-3

Mattoli MV, Trevisi G, Scolozzi V et al (2021) Dynamic (11)C-methionine PET-CT: prognostic factors for disease progression and survival in patients with suspected glioma recurrence. Cancers (Basel) 13:4777. https://doi.org/10.3390/cancers13194777

Maurer GD, Brucker DP, Stoffels G et al (2020) (18)F-FET PET imaging in differentiating glioma progression from treatment-related changes: a single-center experience. J Nucl Med 61:505–511. https://doi.org/10.2967/jnumed.119.234757

Mezu-Ndubuisi OJ, Maheshwari A (2021) The role of integrins in inflammation and angiogenesis. Pediatr Res 89:1619–1626. https://doi.org/10.1038/s41390-020-01177-9

Michaud L, Beattie BJ, Akhurst T et al (2020) (18)F-Fluciclovine ((18)F-FACBC) PET imaging of recurrent brain tumors. Eur J Nucl Med Mol Imaging 47:1353–1367. https://doi.org/10.1007/s00259-019-04433-1

Nakajo K, Uda T, Kawashima T et al (2022) Maximum 11C-methionine PET uptake as a prognostic imaging biomarker for newly diagnosed and untreated astrocytic glioma. Sci Rep 12:546. https://doi.org/10.1038/s41598-021-04216-5

Pessina F, Navarria P, Clerici E et al (2021) Role of 11C methionine positron emission tomography (11CMETPET) for surgery and radiation therapy planning in newly diagnosed glioblastoma patients enrolled into a phase II clinical study. J Clin Med 10:2313. https://doi.org/10.3390/jcm10112313

Petkovic M, Henis M, Heese O, Relogio A (2023) Chronotherapy in glioblastoma: state of the art and future perspectives. EBioMedicine 89:104470. https://doi.org/10.1016/j.ebiom.2023.104470

Pournajaf S, Afsordeh N, Pourgholami MH (2024) In vivo C6 glioma models: an update and a guide toward a more effective preclinical evaluation of potential anti-glioblastoma drugs. Rev Neurosci 35:183–195. https://doi.org/10.1515/revneuro-2023-0067

Raue KD, Duffy JT, Babak MV, Balyasnikova IV (2023) Modeling glioblastoma complexity with organoids for personalized treatments. Trends Mol Med 29:282–296. https://doi.org/10.1016/j.molmed.2023.01.002

Roth P, Silginer M, Goodman SL et al (2013) Integrin control of the transforming growth factor-beta pathway in glioblastoma. Brain 136:564–576. https://doi.org/10.1093/brain/aws351

Schaff LR, Mellinghoff IK (2023) Glioblastoma and other primary brain malignancies in adults: a review. JAMA 329:574–587. https://doi.org/10.1001/jama.2023.0023

Shao T, Chen Z, Belov V et al (2020) [(18)F]-Alfatide PET imaging of integrin alphavbeta3 for the non-invasive quantification of liver fibrosis. J Hepatol 73:161–169. https://doi.org/10.1016/j.jhep.2020.02.018

Steidl E, Langen KJ, Hmeidan SA et al (2021) Sequential implementation of DSC-MR perfusion and dynamic [(18)F]FET PET allows efficient differentiation of glioma progression from treatment-related changes. Eur J Nucl Med Mol Imaging 48:1956–1965. https://doi.org/10.1007/s00259-020-05114-0

van Dijken BRJ, Ankrah AO, Stormezand GN, Dierckx R, Jan van Laar P, van der Hoorn A (2022) Prognostic value of 11C-methionine volume-based PET parameters in IDH wild type glioblastoma. PLoS ONE 17:e0264387. https://doi.org/10.1371/journal.pone.0264387

Verdugo E, Puerto I, Medina MA (2022) An update on the molecular biology of glioblastoma, with clinical implications and progress in its treatment. Cancer Commun (Lond) 42:1083–1111. https://doi.org/10.1002/cac2.12361

Wang F, Liu X, Li S et al (2023) Resolving the lineage relationship between malignant cells and vascular cells in glioblastomas. Protein Cell 14:105–122. https://doi.org/10.1093/procel/pwac006

Watabe H, Ikoma Y, Kimura Y, Naganawa M, Shidahara M (2006) PET kinetic analysis–compartmental model. Ann Nucl Med 20:583–588. https://doi.org/10.1007/BF02984655

Wu W, Klockow JL, Zhang M et al (2021) Glioblastoma multiforme (GBM): an overview of current therapies and mechanisms of resistance. Pharmacol Res 171:105780. https://doi.org/10.1016/j.phrs.2021.105780

Wu J, Tian J, Zhang Y et al (2022) (18)F-Alfatide II for the evaluation of axillary lymph nodes in breast cancer patients: comparison with (18)F-FDG. Eur J Nucl Med Mol Imaging 49:2869–2876. https://doi.org/10.1007/s00259-021-05333-z

Yu C, Pan D, Mi B et al (2015) (18)F-Alfatide II PET/CT in healthy human volunteers and patients with brain metastases. Eur J Nucl Med Mol Imaging 42:2021–2028. https://doi.org/10.1007/s00259-015-3118-2

Zhang H, Liu N, Gao S et al (2016) Can an (1)(8)F-ALF-NOTA-PRGD2 PET/CT scan predict treatment sensitivity to concurrent chemoradiotherapy in patients with newly diagnosed glioblastoma? J Nucl Med 57:524–529. https://doi.org/10.2967/jnumed.115.165514

Zhuang D, Zhang H, Hu G, Guo B (2022) Recent development of contrast agents for magnetic resonance and multimodal imaging of glioblastoma. J Nanobiotechnology 20:284. https://doi.org/10.1186/s12951-022-01479-6

Download references

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Youth Science Fund Project, 2020, 82001859).

This work was supported by the National Natural Science Foundation of China (Youth Science Fund Project, 2020, 82001859).

Author information

Yue Pan and Haodan Dang contributed equally to this work.

Yuan Tian and Baixuan Xu contributed equally to this work.

Authors and Affiliations

Chinese PLA General Hospital, Chinese PLA Medical School, Beijing, China

Yue Pan, Haoxi Zhou & Huanhuan Liu

Department of Nuclear Medicine, First Medical Center of Chinese PLA General Hospital, Fuxing Road 28, Beijing, 100853, China

Yue Pan, Haodan Dang, Huaping Fu, Shina Wu, Jinming Zhang, Ruimin Wang & Baixuan Xu

Department of Radiology, The 8th Medical Center of Chinese PLA General Hospital, Beijing, China

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by HZ, HF, SW, HL, JZ and RW. The first draft of the manuscript was written by YP and HD all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Baixuan Xu .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Ethics approval

All animal experiments were approved by the ethics committee and institutional review board of Chinese PLA General Hospital.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 174 KB)

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Pan, Y., Dang, H., Zhou, H. et al. A comparison study of dynamic [ 18 F]Alfatide II imaging and [ 11 C]MET in orthotopic rat models of glioblastoma. J Cancer Res Clin Oncol 150 , 208 (2024). https://doi.org/10.1007/s00432-024-05688-4

Download citation

Received : 10 January 2024

Accepted : 05 March 2024

Published : 22 April 2024

DOI : https://doi.org/10.1007/s00432-024-05688-4

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Glioblastoma multiforme (GBM)
• [ 18 F]Alfatide II
• [ 11 C]Methionine([ 11 C]MET)
• Positron emission tomography (PET)
• Find a journal
• Publish with us
• Track your research

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• 10 April 2024

How to supercharge cancer-fighting cells: give them stem-cell skills

• Sara Reardon 0

Sara Reardon is a freelance journalist based in Bozeman, Montana.

You can also search for this author in PubMed   Google Scholar

A CAR T cell (orange; artificially coloured) attacks a cancer cell (green). Credit: Eye Of Science/SPL

You have full access to this article via your institution.

Bioengineered immune cells have been shown to attack and even cure cancer , but they tend to get exhausted if the fight goes on for a long time. Now, two separate research teams have found a way to rejuvenate these cells: make them more like stem cells .

Both teams found that the bespoke immune cells called CAR T cells gain new vigour if engineered to have high levels of a particular protein. These boosted CAR T cells have gene activity similar to that of stem cells and a renewed ability to fend off cancer . Both papers were published today in Nature 1 , 2 .

The papers “open a new avenue for engineering therapeutic T cells for cancer patients”, says Tuoqi Wu, an immunologist at the University of Texas Southwestern in Dallas who was not involved in the research.

Reviving exhausted cells

CAR T cells are made from the immune cells called T cells, which are isolated from the blood of person who is going to receive treatment for cancer or another disease. The cells are genetically modified to recognize and attack specific proteins — called chimeric antigen receptors (CARs) — on the surface of disease-causing cells and reinfused into the person being treated.

But keeping the cells active for long enough to eliminate cancer has proved challenging, especially in solid tumours such as those of the breast and lung. (CAR T cells have been more effective in treating leukaemia and other blood cancers.) So scientists are searching for better ways to help CAR T cells to multiply more quickly and last longer in the body.

Cutting-edge CAR-T cancer therapy is now made in India — at one-tenth the cost

With this goal in mind, a team led by immunologist Crystal Mackall at Stanford University in California and cell and gene therapy researcher Evan Weber at the University of Pennsylvania in Philadelphia compared samples of CAR T cells used to treat people with leukaemia 1 . In some of the recipients, the cancer had responded well to treatment; in others, it had not.

The researchers analysed the role of cellular proteins that regulate gene activity and serve as master switches in the T cells. They found a set of 41 genes that were more active in the CAR T cells associated with a good response to treatment than in cells associated with a poor response. All 41 genes seemed to be regulated by a master-switch protein called FOXO1.

The researchers then altered CAR T cells to make them produce more FOXO1 than usual. Gene activity in these cells began to look like that of T memory stem cells, which recognize cancer and respond to it quickly.

The researchers then injected the engineered cells into mice with various types of cancer. Extra FOXO1 made the CAR T cells better at reducing both solid tumours and blood cancers. The stem-cell-like cells shrank a mouse’s tumour more completely and lasted longer in the body than did standard CAR T cells.

Master-switch molecule

A separate team led by immunologists Phillip Darcy, Junyun Lai and Paul Beavis at Peter MacCallum Cancer Centre in Melbourne, Australia, reached the same conclusion with different methods 2 . Their team was examining the effect of IL-15, an immune-signalling molecule that is administered alongside CAR T cells in some clinical trials. IL-15 helps to switch T cells to a stem-like state, but the cells can get stuck there instead of maturing to fight cancer.

The team analysed gene activity in CAR T cells and found that IL-15 turned on genes associated with FOXO1. The researchers engineered CAR T cells to produce extra-high levels of FOXO1 and showed that they became more stem-like, but also reached maturity and fought cancer without becoming exhausted. “It’s the ideal situation,” Darcy says.

Stem-cell and genetic therapies make a healthy marriage

The team also found that extra-high levels of FOXO1 improved the CAR T cells’ metabolism, allowing them to last much longer when infused into mice. “We were surprised by the magnitude of the effect,” says Beavis.

Mackall says she was excited to see that FOXO1 worked the same way in mice and humans. “It means this is pretty fundamental,” she says.

Engineering CAR T cells that overexpress FOXO1 might be fairly simple to test in people with cancer, although Mackall says researchers will need to determine which people and types of cancer are most likely to respond well to rejuvenated cells. Darcy says that his team is already speaking to clinical researchers about testing FOXO1 in CAR T cells — trials that could start within two years.

And Weber points to an ongoing clinical trial in which people with leukaemia are receiving CAR T cells genetically engineered to produce unusually high levels of another master-switch protein called c-Jun, which also helps T cells avoid exhaustion. The trial’s results have not been released yet, but Mackall says she suspects the same system could be applied to FOXO1 and that overexpressing both proteins might make the cells even more powerful.

Nature 628 , 486 (2024)

doi: https://doi.org/10.1038/d41586-024-01043-2

Doan, A. et al. Nature https://doi.org/10.1038/s41586-024-07300-8 (2024).

Article   Google Scholar  

Chan, J. D. et al. Nature https://doi.org/10.1038/s41586-024-07242-1 (2024).

Download references

Reprints and permissions

Related Articles

• Medical research

Bacteria deploy umbrella toxins against their competitors

News & Views 17 APR 24

Brain endothelial GSDMD activation mediates inflammatory BBB breakdown

Article 17 APR 24

ROS-dependent S-palmitoylation activates cleaved and intact gasdermin D

Article 10 APR 24

AI traces mysterious metastatic cancers to their source

News 17 APR 24

Biological age surges in survivors of childhood cancer

Research Highlight 11 APR 24

FOXO1 is a master regulator of memory programming in CAR T cells

AI’s keen diagnostic eye

Outlook 18 APR 24

Refining the impact of genetic evidence on clinical success

Analysis 17 APR 24

Postdoctoral Position - Synthetic Cell/Living Cell Spheroids for Interactive Biomaterials

Co-assembly of artificial cells and living cells to make co-spheroid structures and study their interactive behavior for biomaterials applications.

Mainz, Rheinland-Pfalz (DE)

University of Mainz

2024 Recruitment notice Shenzhen Institute of Synthetic Biology: Shenzhen, China

The wide-ranging expertise drawing from technical, engineering or science professions...

Shenzhen,China

Shenzhen Institute of Synthetic Biology

Recruitment of Global Talent at the Institute of Zoology, Chinese Academy of Sciences (IOZ, CAS)

The Institute of Zoology (IOZ), Chinese Academy of Sciences (CAS), is seeking global talents around the world.

Beijing, China

Institute of Zoology, Chinese Academy of Sciences (IOZ, CAS)

Research Associate - Brain Cancer

Houston, Texas (US)

Baylor College of Medicine (BCM)

Senior Manager, Animal Care

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Milestones in Cancer Research and Discovery

During the past 250 years, we have witnessed many landmark discoveries in our efforts to make progress against cancer, an affliction known to humanity for thousands of years. This timeline shows a few key milestones in the history of cancer research.

1775: Chimney Soot & Squamous Cell Carcinoma

Percivall Pott identifies a relationship between exposure to chimney soot and the incidence of squamous cell carcinoma of the scrotum among chimney sweeps. His report is the first to clearly link an environmental exposure to the development of cancer.

1863: Inflammation & Cancer

Rudolph Virchow identifies white blood cells (leukocytes) in cancerous tissue, making the first connection between inflammation and cancer. Virchow also coins the term "leukemia" and is the first person to describe the excess number of white blood cells in the blood of patients with this disease.

1882: The First Radical Mastectomy to Treat Breast Cancer

William Halsted performs the first radical mastectomy to treat breast cancer. This surgical procedure remains the standard operation for breast cancer until the latter half of the 20th century.

1886: Inheritance of Cancer Risk

Brazilian ophthalmologist Hilário de Gouvêa provides the first documented evidence that a susceptibility to cancer can be passed on from a parent to a child. He reports that two of seven children born to a father who was successfully treated for childhood retinoblastoma, a malignant tumor of the eye, also developed the disease.

1895: The First X-Ray

Wilhelm Roentgen discovers x-rays. The first x-ray picture is an image of his wife's hand.

1898: Radium & Polonium

Marie and Pierre Curie discover the radioactive elements radium and polonium. Within a few years, the use of radium in cancer treatment begins.

1899: The First Use of Radiation Therapy to Cure Cancer

Swedish physicians Tor Stenbeck and Tage Sjogren describe the first cases of basal cell carcinoma of the skin and squamous cell carcinoma of the skin cured by X-ray therapy.

1902: Cancer Tumors & Single Cells with Chromosome Damage

Theodor Boveri proposes that cancerous tumors arise from single cells that have experienced chromosome damage and suggests that chromosome alterations cause the cells to divide uncontrollably.

1909: Immune Surveillance

Paul Ehrlich proposes that the immune system usually suppresses tumor formation, a concept that becomes known as the "immune surveillance" hypothesis. This proposal prompts research, which continues today, to harness the power of the immune system to fight cancer.

1911: Cancer in Chickens

Peyton Rous discovers a virus that causes cancer in chickens (Rous sarcoma virus), establishing that some cancers are caused by infectious agents.

1915: Cancer in Rabbits

Katsusaburo Yamagiwa and Koichi Ichikawa induce cancer in rabbits by applying coal tar to their skin, providing experimental proof that chemicals can cause cancer.

1928: The Pap Smear

George Papanicolaou discovers that cervical cancer can be detected by examining cells from the vagina under a microscope. This breakthrough leads to the development of the Pap test, which allows abnormal cervical cells to be detected and removed before they become cancerous.

1932: The Modified Radical Mastectomy for Breast Cancer

David H. Patey develops the modified radical mastectomy for breast cancer. This surgical procedure is less disfiguring than the radical mastectomy and eventually replaces it as the standard surgical treatment for breast cancer.

1937: The National Cancer Institute (NCI)

Legislation signed by President Franklin D. Roosevelt establishes the National Cancer Institute (NCI).

1937: Breast-Sparing Surgery Followed by Radiation

Sir Geoffrey Keynes describes the treatment of breast cancer with breast-sparing surgery followed by radiation therapy. After surgery to remove the tumor, long needles containing radium are inserted throughout the affected breast and near the adjacent axillary lymph nodes.

1941: Hormonal Therapy

Charles Huggins discovers that removing the testicles to lower testosterone production or administering estrogens causes prostate tumors to regress. Such hormonal manipulation—more commonly known as hormonal therapy—continues to be a mainstay of prostate cancer treatment.

1947: Antimetabolites

Sidney Farber shows that treatment with the antimetabolite drug aminopterin, a derivative of folic acid, induces temporary remissions in children with acute leukemia. Antimetabolite drugs are structurally similar to chemicals needed for important cellular processes, such as DNA synthesis, and cause cell death by blocking those processes.

1949: Nitrogen Mustard

The Food and Drug Administration (FDA) approves nitrogen mustard (mechlorethamine) for the treatment of cancer. Nitrogen mustard belongs to a class of drugs called alkylating agents, which kill cells by chemically modifying their DNA.

1950: Cigarette Smoking & Lung Cancer

Ernst Wynder, Evarts Graham, and Richard Doll identify cigarette smoking as an important factor in the development of lung cancer.

1953: The First Complete Cure of a Human Solid Tumor

Roy Hertz and Min Chiu Li achieve the first complete cure of a human solid tumor by chemotherapy when they use the drug methotrexate to treat a patient with choriocarcinoma, a rare cancer of the reproductive tissue that mainly affects women.

1958: Combination Chemotherapy

NCI researchers Emil Frei, Emil Freireich, and James Holland and their colleagues demonstrate that combination chemotherapy with the drugs 6-mercaptopurine and methotrexate can induce partial and complete remissions and prolong survival in children and adults with acute leukemia.

1960: The Philadelphia Chromosome

Peter Nowell and David Hungerford describe an unusually small chromosome in the cancer cells of patients with chronic myelogenous leukemia (CML). This chromosome, which becomes known as the Philadelphia chromosome, is found in the leukemia cells of 95% of patients with CML.

1964: A Focus on Cigarette Smoking

The U.S. Surgeon General issues a report stating that cigarette smoking is an important health hazard in the United States and that action is required to reduce its harmful effects.

1964: The Epstein-Barr virus

For the first time, a virus—the Epstein-Barr virus (EBV)—is linked to a human cancer (Burkitt lymphoma). EBV is later shown to cause several other cancers, including nasopharyngeal carcinoma, Hodgkin lymphoma, and some gastric (stomach) cancers.

1971: The National Cancer Act

On December 23, President Richard M. Nixon signs the National Cancer Act, which authorizes the NCI Director to coordinate all activities of the National Cancer Program, establish national cancer research centers, and establish national cancer control programs.

1976: The DNA of Normal Chicken Cells

Dominique Stehelin, Harold Varmus, J. Michael Bishop, and Peter Vogt discover that the DNA of normal chicken cells contains a gene related to the oncogene (cancer-causing gene) of avian sarcoma virus, which causes cancer in chickens. This finding eventually leads to the discovery of human oncogenes.

1978: Tamoxifen

FDA approves tamoxifen, an antiestrogen drug originally developed as a birth control treatment, for the treatment of breast cancer. Tamoxifen represents the first of a class of drugs known as selective estrogen receptor modulators, or SERMs, to be approved for cancer therapy.

1979: The TP53 Gene

The TP53 gene (also called p53), the most commonly mutated gene in human cancer, is discovered. It is a tumor suppressor gene, meaning its protein product (p53 protein) helps control cell proliferation and suppress tumor growth.

1984: HER2 Gene Discovered

Researchers discover a new oncogene in rat cells that they call “neu.” The human version of this gene, called HER2 (and ErbB2), is overexpressed in about 20% to 25% of breast cancers (known as HER2-positive breast cancers) and is associated with more aggressive disease and a poor prognosis.

1984: HPV 16 & 18

DNA from human papillomavirus (HPV) types 16 and 18 is identified in a large percentage of cervical cancers, establishing a link between infection with these HPV types and cervical carcinogenesis.

1985: Breast-Conserving Surgery

Results from an NCI-supported clinical trial show that women with early-stage breast cancer who were treated with breast-conserving surgery (lumpectomy) followed by whole-breast radiation therapy had similar rates of overall survival and disease-free survival as women who were treated with mastectomy alone.

1986: HER2 Oncogene Cloning

The human oncogene HER2 (also called neu and erbB2) is cloned. Overexpression of the protein product of this gene, which occurs in about 20% to 25% of breast cancers (known as HER2-positive breast cancers), is associated with more aggressive disease and a poor prognosis.

1993: Guaiac Fecal Occult Blood Testing (FOBT)

Results from an NCI-supported clinical trial show that annual screening with guaiac fecal occult blood testing (FOBT) can reduce colorectal cancer mortality by about 33%.

1994: BRCA1 Tumor Suppressor Gene Cloning

The tumor suppressor gene BRCA1 is cloned. Specific inherited mutations in this gene greatly increase the risks of breast and ovarian cancer in women and the risks of several other cancers in both men and women.

1995: BRCA2 Tumor Suppressor Gene Cloning

The tumor suppressor gene BRCA2 is cloned. Similar to BRCA1, inheriting specific BRCA2 gene mutations greatly increases the risks of breast and ovarian cancer in women and the risks of several other cancers in both men and women.

1996: Anastrozole

FDA approves anastrozole for the treatment of estrogen receptor-positive advanced breast cancer in postmenopausal women. Anastrozole is the first aromatase inhibitor (a drug that blocks the production of estrogen in the body) to be approved for cancer therapy.

1997: Rituximab

FDA approves rituximab, a monoclonal antibody, for use in patients with treatment-resistant, low-grade or follicular B-cell non-Hodgkin lymphoma (NHL). Rituximab is the first monoclonal antibody approved for use in cancer therapy. It is later approved as an initial treatment for these types of NHL, for another type of NHL called diffuse large B-cell lymphoma, and for chronic lymphocytic leukemia.

1998: NCI-Sponsored Breast Cancer Prevention Trial

Results of the NCI-sponsored Breast Cancer Prevention Trial show that the antiestrogen drug tamoxifen can reduce the incidence of breast cancer among women who are at increased risk of the disease by about 50%. FDA approves tamoxifen to reduce the incidence of breast cancer in women at increased risk.

1998: Trastuzumab

FDA approves trastuzumab, a monoclonal antibody that targets cancer cells that overexpress the HER2 gene, for the treatment of women with HER2-positive metastatic breast cancer. Trastuzumab is later approved for the adjuvant (post-operative) treatment of women with HER2-positive early-stage breast cancer.

2001: Imatinib Mesylate

Results of a clinical trial show that the drug imatinib mesylate, which targets a unique protein produced by the Philadelphia chromosome, is effective against chronic myelogenous leukemia (CML). Imatinib treatment changes the usually fatal disease into a manageable condition. Later, it is also shown to be effective in the treatment of gastrointestinal stromal tumors (GIST).

2003: NCI-Sponsored Prostate Cancer Prevention Trial (PCPT)

Results of the NCI-sponsored Prostate Cancer Prevention Trial (PCPT) show that the drug finasteride, which reduces the production of male hormones in the body, lowers a man's risk of prostate cancer by about 25%.

2006: NCI's Study of Tamoxifen and Raloxifene (STAR)

Results of NCI's Study of Tamoxifen and Raloxifene (STAR) show that postmenopausal women at increased risk of breast cancer can reduce their risk of developing the disease if they take the antiestrogen drug raloxifene. The risk of serious side effects is lower with raloxifene than with tamoxifen.

2006: Gardasil

FDA approves the human papillomavirus (HPV) vaccine Gardasil, which protects against infection by the two HPV types (HPV 16 and 18) that cause approximately 70% of all cases of cervical cancer and two additional HPV types (HPV 6 and 11) that cause 90% of genital warts. Gardasil is the first vaccine approved to prevent cervical cancer. NCI scientists made technological advances that enabled development of Gardasil and subsequent HPV vaccines.

2009: Cervarix

FDA approves Cervarix, a second vaccine that protects against infection by the two HPV types that cause approximately 70% of all cases of cervical cancer worldwide. 

2010: The First Human Cancer Treatment Vaccine

FDA approves sipuleucel-T, a cancer treatment vaccine that is made using a patient's own immune system cells (dendritic cells), for the treatment of metastatic prostate cancer that no longer responds to hormonal therapy. It is the first (and so far only) human cancer treatment vaccine to be approved.

2010: NCI-Sponsored Lung Cancer Screening Trial (NLST)

Initial results of the NCI-sponsored Lung Cancer Screening Trial (NLST) show that screening with low-dose helical computerized tomography (CT) reduced lung cancer deaths by about 20% in a large group of current and former heavy smokers.

2011: Ipilimumab

FDA approves the use of ipilimumab, a monoclonal antibody, for the treatment of inoperable or metastatic melanoma. Ipilimumab stimulates the immune system to attack cancer cells by removing a "brake" that normally controls the intensity of immune responses.

2012: NCI-Sponsored PLCO Cancer Screening Trial

Results of the NCI-sponsored PLCO Cancer Screening Trial confirm that screening people 55 years of age and older for colorectal cancer using flexible sigmoidoscopy reduces colorectal cancer incidence and mortality. In the PLCO trial, screened individuals had a 21% lower risk of developing colorectal cancer and a 26% lower risk of dying from the disease than the control subjects.

2013: Ado-Trastuzumab Emtansine (T-DM1)

FDA approves ado-trastuzumab emtansine (T-DM1) for the treatment of patients with HER2-positive breast cancer who were previously treated with trastuzumab and/or a taxane drug. T-DM1 is an immunotoxin (an antibody-drug conjugate) that is made by chemically linking the monoclonal antibody trastuzumab to the cytotoxic agent mertansine, which inhibits cell proliferation by blocking the formation of microtubules.

2014: Analyzing DNA in Cancer

Researchers from The Cancer Genome Atlas (TCGA) project, a joint effort by NCI and the National Human Genome Research Institute to analyze the DNA and other molecular changes in more than 30 types of human cancer, find that gastric (stomach) cancer is actually four different diseases, not just one, based on differing tumor characteristics. This finding from TCGA and other related projects may potentially lead to a new classification system for cancer, in which cancers are classified by their molecular abnormalities as well as their organ or tissue site of origin.

2014: Pembrolizumab

FDA approves pembrolizumab for the treatment of advanced melanoma. This monoclonal antibody blocks the activity of a protein called PD1 on immune cells, which increases the strength of immune responses against cancer.

2014: Gardasil 9

FDA approves Gardasil 9, a vaccine that protects against infection with the same four HPV types as Gardasil plus five more cancer-causing HPV types that together account for nearly 90% of cervical cancers. It is now the only HPV vaccine available in the United States.

2015: NCI-MATCH Clinical Trial

NCI and the ECOG-ACRIN Cancer Research Group launch the NCI-MATCH (Molecular Analysis for Therapy Choice) clinical trial to test more than 20 drugs and drug combinations based on molecular analysis of tumors in people with cancer. The study is designed to determine whether targeted therapies for people whose tumors have specific gene mutations will be effective regardless of their cancer type.

2015: Talimogene Laherparepvec

FDA approves talimogene laherparepvec (T-VEC) for the treatment of some patients with metastatic melanoma that cannot be surgically removed. T-VEC, the first oncolytic virus approved for clinical use, works by infecting and killing tumor cells and stimulating an immune response against cancer cells throughout the body. 

2016: Cancer Moonshot℠

Congress passes the 21st Century Cures Act, which provides funding for the Cancer Moonshot, a broad program to accelerate cancer research by investing in specific research initiatives that have the potential to transform cancer care, detection, and prevention.

2017: Pediatric MATCH

NCI and the Children’s Oncology Group launch Pediatric MATCH, an effort to extend molecular analysis and targeted treatment to children and adolescents with cancer. Like NCI-MATCH, Pediatric MATCH seeks to determine if treating tumors with molecularly targeted drugs based on the tumor’s genetic characteristics rather than the type of cancer or cancer site will be effective.

2017: CAR T-Cell Therapies

FDA approves tisagenlecleucel to treat a form of acute lymphoblastic leukemia in certain children and young adults. FDA subsequently approves axicabtagene ciloleucel for patients with large B-cell lymphomas whose cancer has progressed after receiving at least two prior treatment regimens. Both treatments are chimeric antigen receptor (CAR) T-cell therapies that are personalized for each patient. To create these therapies, T cells are removed from the patient, genetically altered to recognize cancer-specific antigens, grown to large numbers in the lab, and then infused back into the patient to stimulate their immune system to attack cancer cells.

2017: Tumor-Agnostic Approval for Pembrolizumab

FDA extends approval of pembrolizumab to treat metastatic and inoperable solid tumors that have certain genetic changes, wherever they occur in the body , that have progressed following prior treatment and that have no alternative treatment options. With this tissue-agnostic approval, pembrolizumab becomes the first cancer treatment based solely on the presence of a genetic feature in a tumor, rather than a person’s cancer type.

2017: Genomic Profiling Tests

FDA clears two products to test tumors for genetic changes that may make the tumors susceptible to treatment with FDA-approved molecularly targeted drugs. In November, FDA authorizes the MSK-IMPACT test developed and used by Memorial Sloan Kettering Cancer Center to analyze tumors for potentially actionable changes in 468 cancer-related genes. In December, FDA approves the FoundationOne CDx test, which evaluates genetic changes in 324 genes known to fuel cancer growth. The FoundationOne test serves as a companion diagnostic for several FDA-approved drugs targeting five common types of cancer.

2018: TCGA PanCancer Atlas

NIH-funded researchers with TCGA complete an in-depth genomic analysis of 33 cancer types. The PanCancer Atlas provides a detailed genomic analysis of molecular and clinical data from more than 10,000 tumors that gives cancer researchers an unprecedented understanding of how, where, and why tumors arise in humans. 

2018: NCI-Sponsored TAILORx Clinical Trial

Results from the NCI-sponsored Trial Assigning IndividuaLized Options for Treatment (Rx), or TAILORx, clinical trial show that most women with early-stage breast cancer do not benefit from having chemotherapy after surgery. The trial used a molecular test that assesses the expression of 21 genes associated with breast cancer recurrence to assign women with early-stage, hormone receptor–positive, HER2-negative breast cancer that hasn’t spread to the lymph nodes to the most appropriate and effective post-operative treatment. It is one of the first trials to examine a way to personalize cancer treatment

2018: Larotrectinib

FDA approves larotrectinib, the first drug that targets tumors with NTRK gene fusions. The approval is for pediatric or adult patients with metastatic or inoperable solid tumors that have worsened after previous treatment anywhere in the body driven by an NTRK gene fusion without a known acquired resistance mutation. Larotrectinib is the second drug approved to treat cancer with specific molecular features regardless of where the cancer is located.

2020: International Pan-Cancer Analysis of Whole Genomes

A consortium of international researchers analyzes more than 2,600 whole genomes from 38 types of cancer and matching normal tissues to identify common patterns of molecular changes. The Pan-Cancer Analysis of Whole Genomes study, which used data collected by the International Cancer Genome Consortium and TCGA, uncovers the complex role that changes throughout the genome play in cancer development, growth, and spread. The study also extends genomic analyses of cancer beyond the protein-coding regions to the complete genetic composition of cells.  

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Clin Transl Radiat Oncol
• v.27; 2021 Mar

Statistical fundamentals on cancer research for clinicians: Working with your statisticians

a Department of Biostatistics, The Princess Margaret Cancer Centre/University of Toronto, Canada

b Biostatistics Division, Dalla Lana School of Public Health, University of Toronto, Canada

Shao Hui Huang

c Department of Radiation Oncology, The Princess Margaret Cancer Centre/University of Toronto, Canada

d Department of Otolaryngology-Head & Neck Surgery, The Princess Margaret Cancer Centre/University of Toronto, Canada

Shivakumar Gudi

Brian o'sullivan.

• • Different trial types require different sample size calculation and interpretation.
• • Statistical modeling may help to minimize confounding effects and bias.
• • Statistical pitfalls can be avoided by following correct statistical guidance.

To facilitate understanding statistical principles and methods for clinicians involved in cancer research.

An overview of study design is provided on cancer research for both observational and clinical trials addressing study objectives and endpoints, superiority tests, non-inferiority and equivalence design, and sample size calculation. The principles of statistical models and tests including contemporary standard methods of analysis and evaluation are discussed. Finally, some statistical pitfalls frequently evident in clinical and translational studies in cancer are discussed.

We emphasize the practical aspects of study design (superiority vs non-inferiority vs equivalence study) and assumptions underpinning power calculations and sample size estimation. The differences between relative risk, odds ratio, and hazard ratio, understanding outcome endpoints, purposes of interim analysis, and statistical modeling to minimize confounding effects and bias are also discussed.

Proper design and correctly constructed statistical models are critical for the success of cancer research studies. Most statistical inaccuracies can be minimized by following essential statistical principles and guidelines to improve quality in research studies.

1. Introduction

Cancer research is the soundest tool to generate new knowledge to advance oncology practice. Broadly, there are two types of clinical studies: experimental and observational. Observational studies are undertaken without a specific intervention and can be prospective or retrospective [1] . Experimental studies involve an intervention and studying its subsequent effects, often tested in phase I/II/III/IV clinical trials [2] , [3] , [4] . While carefully designed and well-conducted randomized controlled trials (RCTs) provide the highest quality evidence regarding efficacy and safety of a particular intervention, they also have limitations, often related to practical or ethical considerations, that represent the tension between “ideal” trial settings and the “real world” environment [5] . Although with important caveats, observational and non-randomized comparative studies could provide a cost-saving and practical alternative.

An important research principle should be reproducibility with high validity, applicability to the target population of interest, and transferability to clinical practice. While preliminary concept envisioning is expected, it is desirable for a clinician to quickly engage experienced biostatistician colleagues to minimize bias, improve statistical power, and provide robust estimations of effect size and other model parameters [6] . An optimal design, especially for RCTs, should address: (1) relevant primary/secondary/exploratory objectives, (2) clinical endpoints and hypothesis testing, (3) a target study population with inclusion/exclusion criteria, (4) rigorous procedures, including randomization, monitoring and quality control, and plans for possible extension or premature termination, (5) a statistical analysis plan (SAP) with model selection and justification, and (6) planned sensitivity analysis for relevant subgroups.

This paper provides practical insights for clinicians about fundamental statistical concepts and methodologies used in oncology research, especially for phase III trials. Some examples from the head and neck cancer (HNC) perspective are provided, generally in the curative setting. However, the principles are equally applicable to other oncology domains. Other types of trials (e.g., phase I/II trials and umbrella protocols) or emerging methods (e.g., machine learning) are not addressed due to the intended scope of this paper. Also, while not addressed further, we encourage caution at the design phase of trials addressing radiotherapy combined with novel agents since there may be unique toxicities including temporal occurrence and character that may not be anticipated [7] . Interested readers should research this important area separately [8] , [9] .

2. Study design

2.1. superior, non-inferiority and equivalence trials.

Most oncology studies focus on superiority to evaluate whether an intervention is “better” (e.g. higher efficacy, lower toxicity) compared to a control group, using the null hypothesis (H0) that the interventional and control groups are equally effective with an alternative hypothesis (H1, i.e. the clinical “hypothesis”) that they are not equal (i.e. the experimental arm can be either more or less effective than the control arm, which is commonly referred to as “two-sided”) [10] . A nonsignificant result implies insufficient evidence to reject H0. It is critical for H1 to be based on sound clinical judgement and updated knowledge, otherwise it risks exposing patients to unnecessary and/or inferior treatment. For example, the H1 for the recently published ARTSCAN III trial [11] posited a 10% higher 5-year OS for cetuximab versus cisplatin which was based on one trial of cetuximab compared to radiotherapy-alone [12] without considering the effect from concurrent chemotherapy (CCRT) [13] . However, after the trial initiation, the authors responded to emerging evidence showing inferior outcome of cetuximab-radiation versus chemoradiation [14] , prompting an unplanned interim analysis that resulted in early trial termination due to inferior outcomes in the intervention (cetuximab-radiotherapy) arm.

In recent years, non-inferiority studies, such as treatment de-intensification trials in HPV-positive (HPV + ) oropharyngeal cancer (OPC) [15] , or withholding neck surgery following favorable response to radiotherapy [16] , have gained popularity. In contrast to superiority studies addressing effectiveness, non-inferiority studies evaluate whether a less intensive or less costly intervention is not unacceptably less efficacious compared to standard-of-care (SOC) [17] . The H0 is that SOC is better than the experimental intervention, and the H1 is that the experimental intervention is at least as effective as SOC. A non-inferiority study is always one-sided, thus addressing the chance of observing a difference as large as, and in the same direction, as that observed. The margin to be detected is usually also smaller (e.g., 5% in 5-year overall survival [OS] in the recent RTOG-1016 trial) and, therefore, a larger sample size is usually required [18] , [19] . Another example of a non-inferiority trial is NRG HN-002 ( {"type":"clinical-trial","attrs":{"text":"NCT02254278","term_id":"NCT02254278"}} NCT02254278 ) which hypothesized that two treatment arms (reduced dose IMRT with or without weekly cisplatin) were non-inferior to the SOC of high-dose CCRT in low-risk minimal smoking HPV + OPC, where effectiveness was defined as 2-year progression-free survival (PFS) of ≥ 85% with a margin of 6% compared to SOC (assuming 2-year PFS for SOC is 91%), and without unacceptable swallowing toxicity at 1-year. Notably, there was no SOC arm and the comparator PFS value is based on recent historical data with the understanding that the winning arm would be taken forward to compare against SOC in phase III.

Another design to prove absence of a significant difference between treatment interventions is an equivalence study. “Noninferiority” and “equivalence” are often used interchangeably to test whether a new treatment is as effective as the SOC. However, there are subtle difference. To prove clinical equivalence, a margin (Δ) is chosen by identifying the clinically acceptable difference in the justification for equivalence [20] , which is “two-sided”: addressing the chance of seeing a difference in either direction. If two treatments are equivalent to each other (i.e., the difference is within a pre-defined acceptable margin), the 95% confidence interval (CI) of the parameter that assesses the treatment effect must lie within this margin [21] , [22] . For example, the trial by Garrel et al. [23] is an equivalence trial, which compared “equal” effectiveness of sentinel node biopsy versus neck dissection (SOC) with a delta of 10% in operable T1-T2N0 oral and oropharyngeal cancer.

In summary, superiority, non-inferiority, and equivalence studies are three study types with different assumptions about treatment effects [24] [ Table 1 ]. They require different sample size calculations and interpretation. When a superiority study shows a non-significant p value, it is also important not to conclude that the two arms are similar (i.e., non-inferiority or equivalence).

Null Hypothesis and Alternative Hypothesis of Superiority, Equivalent and Non-inferiority Studies.

* Two-sided test means bi-directional (either better or worse effect) on the performance of the primary endpoint.

** One-sided test means uni-directional (i.e., better effect) on the performance of the primary endpoint.

Traditional trials often employ frequentist approaches which require an H0 and use “fixed” input (e.g., effect size, toxicity reduction) at the design phase. However, this may be challenging when data are sparse, especially for novel technologies (e.g., protons). Bayesian adaptive trial design is exploring this uncertain domain, which can allocate more patients with updated information to the more beneficial treatment arm if a difference is observed during a trial as recently used when evaluating protons in lung cancer [25] , [26] . However, it is not being used in four ongoing Phase III proton trials in HNC ( {"type":"clinical-trial","attrs":{"text":"NCT04607694","term_id":"NCT04607694"}} NCT04607694 , {"type":"clinical-trial","attrs":{"text":"NCT01893307","term_id":"NCT01893307"}} NCT01893307 , {"type":"clinical-trial","attrs":{"text":"NCT02923570","term_id":"NCT02923570"}} NCT02923570 , TORPEdO trial-ISRCTN16424014). Nonetheless, a similar philosophy to streamline eligibility to only include patients who may benefit from protons by pre-screening using NTCP modelling is a component of one trial (DAHANCA 35, {"type":"clinical-trial","attrs":{"text":"NCT04607694","term_id":"NCT04607694"}} NCT04607694 ), which has been validated to be feasible [27] .

2.2. Study population, sample size calculations and power analysis

Attention to the study population is critical, including how patients will be selected and informed, who will be excluded, and when following diagnosis will they enter the study. Careful attention to case assembly will reduce variability and maintain power to detect differences. However, selection criteria must not be overly narrow to ensure the generalizability of the results. The case assembly should consider important prognostic factors (e.g., disease stage or important biological factors) that influence disease behaviour/response/tolerance to treatment. Recently, the HNC population is considered as two broad groups: tobacco/alcohol-related and HPV-related cancers. HPV + HNC patients have more favorable prognosis and their inclusion in trials may perturb sample size calculations due to dramatically different event rates for many outcomes (See examples later).

For a prospective study, the number of subjects (sample size) needed to address the primary end-point and detect meaningful potential differences requires estimation. The sample should be sufficient to minimize the risk of random errors, unbalanced case inclusion, and bias relating to any intervention (typically addressed by randomization). For a retrospective study with fixed sample size, power analysis can estimate the possibility of identifying statistically significant differences (termed the “power”). A pre-requisite is to specify the H0/H1 and then calculate the sample size to ensure sufficient statistical power to differentiate between these hypotheses, while controlling the probability of incorrectly rejecting the H0.

While mostly applicable to RCTs, the principles of sample size estimation are also important in other studies. There should be a credible judgement about the likely rate for the primary end-point (e.g., OS) in the control group, followed by a similar appreciation of the conceivable medically important impact of the experimental intervention on the end-point. Researchers should avoid overly optimistic effect differences that could result in early trial closure [11] ; alternatively, it may undermine study power as occurred in another study with an ambitious assumption of 15% absolute difference [28] , and may impact ability to detect smaller differences. The likelihood of a false-positive result is normally expressed as the Type I error (or α, typically set at ≤ 0.05), and the false negative rate as the Type II error (or β). By convention “1-β” is referred to as the “statistical power”, e.g., value of 0.8, derived from a β level of ≤ 0.20. The time for trial entry/accrual should be sufficiently short to retain relevance, maintain sensitivity to avoid distracting the research environments from addressing other relevant questions that may emerge over time, and mitigate confounding arising from evolution of treatment/management in such areas as quality or implementation arising during the study accrual period. “Five years” is generally considered an upper limit of desirable accrual duration [29] . Finally, the time period for events to manifest following completion of patient entry influences the design and ultimate trial logistics.

The parameters required for the sample size calculation include significance level (α), statistical power (1-β), and effect size [Δ] [e.g., Cohen’s effect size, odd ratio (OR) or hazard ratio (HR)], and the variation or “spread” of distribution (often using standard deviation) of the study endpoint(s) [ Table 2 ]. Although fixed values of these parameters are often used for sample size determination, they have been criticized for oversimplification by overlooking inherent uncertainties about the assumptions [30] . Different suppositions about parameters are recommended to provide a more comprehensive evaluation of their influence on sample size determination. For early phase clinical trials and pilot observation studies, the significance levels can be less stringent (e.g. α = 0.15 or 0.20 for Phase II trials) [31] , while in some Phase III trials, power is often more stringent (e.g. 0.90) [32] . The estimated effect size is the minimal clinical meaningful difference, ordinarily chosen by interpreting prior research findings. For example, to calculate the impact of CCRT on locally advanced HNC, a strategy might be to choose an effect size based on a robust dataset such as the Meta-Analysis of Chemotherapy in Head and Neck Cancer (MACH-NC) [33] .

Variables Required for Sample Size Calculation.

Abbreviation; PFS: progression free survival.

As an example, the CCTG HN.6 trial ( {"type":"clinical-trial","attrs":{"text":"NCT00820248","term_id":"NCT00820248"}} NCT00820248 ) [34] required 320 patients over 3.2 years to observe a total of 246 events (any relapse or death) assuming the following: alpha 0.05 with 80% power; 2-year PFS of 45% for the control group, and a HR (discussed later) of 0.7 (representing a 30% reduction of the likelihood of an event, corresponding to a 12.2% absolute difference in 2-year PFS); an enrollment of 100 patients/year; and all patients followed for an additional 3 years to ensure the emergence of enough PFS events. If the assumption for any aforementioned parameters changes, the estimated sample size would also change accordingly ( Table 2 ).

Changes in biologic characteristics of disease could also alter the sample size calculation due to changes in assumptions regarding the risk of events. Recent trials in locally advanced HNC [28] , [35] showed dramatically diminished power due to unanticipated emergence of HPV + OPC which changed event rates significantly rendering the original trials, designed before appreciating this phenomenon, virtually obsolete. A lower-than-expected event rate due to unanticipated confounding by the emerging HPV population, e.g., RTOG 0129 [27] , cannot be addressed by longer follow-up. Investigators should be aware of this problem when designing trials to ensure adequate sample size. Planned interim analysis could identify the need to augment sample size. For example, RTOG 1016 ( {"type":"clinical-trial","attrs":{"text":"NCT01302834","term_id":"NCT01302834"}} NCT01302834 ) [15] required sample size expansion from the original 706 to a final accrual of 987 due to a lower-than-estimated event rate.

Planned sample size is also critical in studies on precision/molecular radiotherapy research. Studies with limited numbers of patients can be used for exploratory or pilot analysis and hypothesis generation. Multicenter collaborations and integrative analysis on such trials are encouraged for further confirmation/validation.

2.3. Randomization, stratification and intention-to-treat

Randomization is a fundamental pillar of prospective trials because it provides the opportunity to balance the distribution of all baseline covariates (observed and unobserved) across treatment groups. The date of randomization also provides a useful initiation date for cohort analysis to minimize potential lead-time bias due to potential differences in duration of treatments under comparison (e.g., surgery vs non-surgical treatment).

Stratification should improve the efficiency of a RCT by reducing the variation of the treatment effect. Stratified randomization can be conducted by assigning patients with certain characteristics equally to each treatment arm. The study randomization list should be generated by an independent biostatistician, and distributed/monitored by an independent administration center.

An intention-to-treat analysis is an additional important principle to reduce confounding by analyzing patients according to their original randomization assignment, regardless of the treatment they actually received.

3. Data analysis and reporting

3.1. understanding study endpoints.

The most commonly used oncological endpoints in studies include: OS, PFS/disease-free survival (DFS), and cause-specific survival (CSS) [36] ( Table 3 ). The advantage of OS is its objective definition (alive or dead) and consequent less susceptibility to misreporting. However, it does not distinguish index-cancer death from competing mortality. Alternatively, CSS restricts events to index-cancer death and therefore addresses ablative or tumoricidal effects of a treatment, but the accuracy of cause of death attribution remains a concern. PFS/DFS has become more popular in clinical trials recently because both treatment failure and death are considered “events”, which garners more incidents resulting in greater power and reduced study sample size. However, the terms “disease-free” or “progression-free” can both be misleading because death from other cause, such as cardiac event/suicide/car accident, are also counted as “events” although unrelated to “the cancer-of-interest”. OS and PFS/DFS all suffer from other consequences such as the detrimental effect of smoking on cancer survival. While the effect on OS is consistent and rational in HPV + OPC patients, the effect on disease control is not consistent [37] , [38] . It is possible that the lower DFS or OS in heavy smokers results from death due to competing risk, and does not necessarily indicate that smoking has induced a more aggressive tumor phenotype. In turn, it does not indicate that smokers would uniformly benefit from intensified treatment. Furthermore, in a landmark initial study addressing this hypothesis, the occurrence of a second primary cancer was included as an event, together with survival and disease recurrence, in attempting to unravel the impact of smoking on outcome of these patients [39] . A subsequent publication from the same group did not observe worse cancer specific outcomes in smokers [38] .

Definition of Commonly Used Oncologic Outcome Endpoint and Analytic Procedure.

Abbreviation: N/A: not applicable; OS: overall survival; CSS: cause-specific survival; RFS: recurrence-free survival; PFS: progression-free survival; DFS: disease-free survival; LC: local control; RC: regional control; DC: distant control.

Oncologic outcomes are often time-to-event endpoints and their analysis differs from simple calculations of frequencies. The event may not be observable for all subjects due to attrition of cases from the sample or termination of follow-up, which are considered as censored data. For time-to-event endpoints, the uniformly agreed analysis is the Kaplan-Meier method with log-rank test for comparison [40] . It provides an estimate of event-free probability at any time point during the follow-up period, and permits censoring and varying lengths of follow-up. However, it does not take into account death due to competing-risk and could overestimate the event-of-interest when calculating a disease-specific endpoint (e.g. local/regional/distant failure), since a competing-risk event can preclude the event-of-interest from occurring [41] , exemplified in Fig. 1 . For these endpoints, the competing-risk model is more appropriate. This is especially important for vulnerable populations, including the elderly, susceptible to competing mortality. While many HNC patients are heavy tobacco users, additional alcohol use contributes further to their inherent risk of non-cancer mortality. Table 3 summarises commonly accepted terms and analytic procedures (“censoring”, and “competing risk calculations”) for various oncologic endpoints.

Actuarial Rate of Locoregional Failure Estimated by Kaplan-Meier Method vs Competing Risk Method in HPV-negative OPC Patients Treated at Princess Margaret Cancer Centre, Toronto, Canada.

3.2. Median follow-up and actuarial estimation

The purpose of reporting median follow-up in survival analysis studies is to describe the maturity of data. It is generally calculated on surviving patients only, which ideally should be specifically stated since it is important to appreciate if sufficient time has elapsed to permit most events to occur. Survival estimates become less accurate when they extend beyond the median follow-up time due to insufficient numbers at risk. Thus, it is unrealistic to estimate 5-year OS in a study with only 3 years of median follow-up time.

Survival rates are often derived by Kaplan-Meier analysis which uses median time-to-event as an estimation. However, median time can also be unstable and susceptible to outliers, such as patients who die shortly after treatment or those with long survival. This is relevant when comparing long-term outcomes beyond the traditional 5-year period when two arms could exhibit significantly different median follow-up. Restricted mean survival time (RMST) calculates mean survival time over a pre-specified, clinically important time point. It is equivalent to the area-under-the-Kaplan-Meier-curve from the beginning of a study through that pre-specified time points (e.g., 2-year or 3-year) [42] , [43] . It is complimentary to Kaplan-Meier analysis, and may augment time-dependent data analyses in clinical trials and meta -analyses [44] . A case study of individual patient data (IPD) network meta -analysis (NMA) on nasopharyngeal cancer has shown different results using both methods [45] . RMST difference is valid and interpretable even if the proportional hazards assumption is violated [45] .

Ideally, clinical trials should also have sufficient follow-up to appreciate late toxicity, which might alter the conclusion of the trial [46] , [47] . For example, the RTOG 91–11 trial initially reported superior 5-year laryngeal preservation and locoregional control with similar OS using CCRT compared to induction chemotherapy, while radiotherapy-alone fared the worst [47] . However, long-term results [46] showed a trend for better OS with induction chemotherapy compared to CCRT, leading to speculation that unexplained death might be attributed to greater long -term toxicity (e.g., silent aspiration) with the latter approach.

3.3. Interim analysis

Interim analysis is important and should preferably be pre-planned and undertaken in a controlled manner, generally under the auspices of a data monitoring committee that includes experts who are not investigators on the trial. The focus is often directed at the safety of patients (a principal reason) in the event that a trial needs to be paused or terminated for several reasons: 1). Excessive toxicity mandating immediate closure, as occurred in an altered fractionation radiotherapy trial in locally advanced HNC where only 82 of 226 planned cases were eventually accrued [48] , 2). Clear superiority of one treatment compared to another may be grounds for closure for ethical reasons, especially when the primary question may have been addressed and there is no further rationale to continue expending resources, and further patients could continue receiving a proven inferior approach: this was seen with the experimental treatment in the highly influential trial of chemoradiotherapy in nasopharyngeal carcinoma that changed practice globally [49] , 3). Unexpected significantly worse performance of an experimental arm also warrants immediate closure as was evident in the DAHANCA 10 trial using darbepoetin alfa to improve anemia in HNC (HR for OS: 1.30) or the ARTSCAN III trial ( {"type":"clinical-trial","attrs":{"text":"NCT01969877","term_id":"NCT01969877"}} NCT01969877 ) [11] comparing cetuximab versus cisplatin-chemoradiotherapy (HR for OS: 1.63), 4). Other reasons for premature closure include futility, relating to inadequate power consequent on slow accrual. Examples include the rare trial that compared chemo-radiotherapy versus definitive surgery in HNC [50] , [51] and the PARADIGM induction chemotherapy trial [52] . Unplanned interim analysis may occasionally be useful if the investigators respond to new evidence from other studies during the course of the trial [11] , and may result in amendments or premature closure.

Alternatively, multiple interim analyses may inflate false positive findings. This multiplicity problem dictates the need for methodologies developed for statistical adjustment on stopping rules. Group sequential design is a commonly used procedure which defines p-values for considering trial stoppage at an interim analysis while preserving the overall type I error [53] , [54] .

Rather than focussing only on trial closure, an important alternative consideration for the data safety monitoring committee, may be the observation during the trial that borderline differences exist justifying the addition of either more patients or an extended duration [55] .

Finally, an important factor for the broader research landscape concerns the impact of stopping comparative effectiveness trials which may still contribute useful information by enhancing the power of subsequent meta -analyses addressing important questions or may identify value to treatments in later follow-up if they are less invasive, or less expensive/inconvenient [56] .

4. Addressing confounding variables

4.1. observational studies and propensity score matching.

Observational, often retrospective, studies are often considered less impactful than prospective trials because of compromised ability to address case eligibility and biases, the temptation to apply risks and assessments from post treatment outcomes to the baseline prognostic framework, and generally have less rigor to evaluate endpoints that may not be predefined, and a higher likelihood of imbalanced baseline characteristics compared to clinical trials. Propensity score matching may help to address this [57] , [58] by creating matched groups of untreated and treated cases with the same likelihood of clinical behaviour or treatment response for a given a set of observed covariates. Ideally propensity score matching requires large samples with a reasonable spread of baseline variables across the population and substantial overlap between the comparison groups. The process generally includes: (1) choosing variables to be included in the propensity score, (2) choosing matching and weighting strategies to balance covariates across treatment groups, (3) balancing covariates after matching or weighting the sample, and (4) interpreting treatment effect estimates [59] . The covariates used in propensity score matching are identified from variables predictive of the outcomes-of-interest [60] .

Two types of propensity score matching designs predominate: the most common identifies propensity score matched samples [61] ; the other creates propensity scores, and conducts outcome analysis using all samples adjusting for the subsequent propensity scores [62] . One-to-one or one-to-two matching are commonly used. Since propensity score matching can only control for observable covariates, hidden bias may remain due to unobserved variables after matching [63] .

4.2. Univariable vs multivariable analysis

Univariable analysis (UVA) is commonly used to assess association between a single predictor or risk factor and the study endpoint. However, biased inference may be derived from UVA due to pre-existing confounding effects [64] . Multivariable analysis (MVA) is a statistical method to adjust for observed confounding factors to correct for and enable accurate inference.

For MVA model construction, four selection procedures are typical: forward , backward , stepwise , and best subset selection. All choose candidate variables for inclusion in the MVA, usually identified from significant variables in UVA, or important risk factors related to the study endpoint, or frequent confounders such as age and treatment. The forward selection algorithm starts by adding candidate variables sequentially; attributes with the lowest p-value below the selection criteria (e.g., 0.05), are chosen iteratively until no new variables can be added. Backward selection starts by including all candidate variables followed by sequential iterative removal according to highest p-value exceeding the selection criteria, until no variables can be removed. Stepwise algorithm uses a combination of backward and forward selection. Best subset selection assesses combinations of variables (“subset of variables”) and identifies the most optimal model using model evaluation criteria, such as the Akaike Information Criterion (AIC), the Bayes Information Criterion (BIC) [65] , [66] , adjusted R-square, residual sum of squares, Mallow’s Cp Statistic, and concordance index (C-index). C-Index (ranges from 0 to 1) is the standard performance measure for survival model assessment, and a higher value indicates a higher predictability in a survival model.

To construct a reliable and robust multivariable model, the minimal number of “samples” (referring to “events” in time-to-event outcome) per variable is important for model performance and estimation. Generally, the minimum number of “samples”/“event” per variable lies between 5 and 20 [67] , [68] . In survival analysis, ten “events” per variable is often the minimum required sample size for linear regression models to ensure accurate prediction in subsequent subjects [69] , [70] , [71] .

4.3. Difference between multivariate analysis and multivariable analysis

Although often used interchangeably, the terms “ multivariable analysis ” and “ multivariate analysis ” are distinct. A multivariable model is an analysis with a single endpoint but multiple independent variables, whereas a multivariate analysis describes multiple study outcome endpoints, i.e., different adverse events with multiple independent variables [72] , [73] , [74] , using a single model and provides unbiased and precise parameter estimation and potentially increased statistical power.

5. Statistical modeling

5.1. risk classifications and prediction.

Evaluation of clinical factors includes both association studies and predictive studies . Association studies identify prognostic factors associated with study outcomes. Predictive studies address multiple predictors with combined effects on response to treatment and outcome prediction.

The development of a clinical prediction model involves three components: model building, validation, and implementation. Both model building and validation are guided by model prediction performance evaluations. The first step is to select candidate risk factors, including clinical factors or biomarkers with strong preliminary data suggesting prognostic impact, and previously established clinical factors or biomarkers that could be confounders or effect modifiers [75] .

The second step is model construction where the decision about the most important variables to predict outcome is usually conducted through multivariable regression modeling based on model selection algorithms such as stepwise or backward selection . Another aspect of model specification is the interactive effects of risk factors. After predictive model construction, patients can be classified into high-risk vs low-risk groups.

Finally, either external validation or internal validation should be conducted to verify the developed predictive model. Cross-validation is one of the common techniques for internal validation [76] , [77] . More stringent validation is achieved by addressing external validity, using a different, independent dataset from a similar patient population. Nomograms or web applications are commonly used implementation tools underpinned by outputs derived from prediction models [74] .

5.2. Estimates of comparative risk association

When comparing two treatments, both the magnitude of the treatment effect between both treatments and its direction (i.e., an improved or a detrimental result) are important. Several measures of comparative risk association, including relative risk (RR) and odds ratio (OR), can be used to assess differential effects according to the interventions at static time points [78] using binary measures (e.g., toxicity vs no toxicity, response vs no response). However, the most frequently used method for contemporary clinical studies is the HR which applies to time-to-event outcomes. HRs are estimated for an event (e.g., death) over the entire trial duration between two treatments and are a convenient measure of the treatment effect in efficacy studies, although the number of events in either arm is not shown directly. Simplistically, a HR is calculated by the ratio of hazard rates of experimental divided by that of control arm.

Using the CCTG HN.6 trial [34] as an example again, the 2-year PFS was assumed to be 45% with the corresponding hazard rate of 0.40 [−log(0.45)/2 years] for the control arm. With 12.2% absolute difference, the 2-year PFS would be 57.2% for the experimental arm, corresponding to a hazard rate 0.28 [−log(0.572)/2 years], and a HR of 0.7 [0.28/0.40]. When the results are analyzed, if the HR is 1.0, the treatments are considered equivalent, while values < 1.0 indicate superiority and values > 1.0 indicate that the experimental arm is worse. In the example shown, a HR of 0.7 means that the experimental arm has a 30% decrease in hazard of death compared to the control. Correspondingly, if the HR is 1.3, the experimental treatment would have a 30% higher hazard of death compared to the control. It is also usual to indicate 95% confidence intervals (CI) of the HR. It should not overlap unity (1.0) if the effect between the two arms is statistically significant at the level of p < 0.05. This is important for the reader, since it is possible to see comparative survival curves, including when significant differences exist, displayed with only HRs and CIs, but without the p-values. Finally, HRs can be adjusted for covariates within the multivariable Cox regression model that generated the hazard rates.

6. Addressing data heterogeneity

6.1. sensitivity analysis and subgroup analysis.

The goal for personalized medicine is often to identify best treatment for subsets of patients based on demographic, clinical and genetic characteristics. Understanding heterogeneity of treatment response is complex due to the intricate oncology environment. In clinical trials, subgroup analysis should be pre-planned and specified in the trial protocol and readers should be extremely wary when attempting to implement management derived from results of unplanned analyses. However, subset analysis is often useful to understand results of a trial and for hypothesis generation when designing future trials.

Besides subgroup analysis, sensitivity analysis is also important to assess the robustness of a statistical model to its assumptions. It is often used to evaluate consistency in results and conclusions given different parameters of a particular model, including comparison of models using differing clinical covariates, with and without interactive effects. Various statistical models can be applied to the same study to evaluate the estimation of association and outcome prediction. The same analysis methods can be applied to different sample cohorts such as intention-to-treat, per-protocol cohorts, and safety cohorts (randomized patients who received at least a component of the treatment) to evaluate the robustness of parameter estimation and statistical inference.

6.2. Multiple comparison adjustment

When multiple models or statistical tests are conducted on a single study, especially in biomarker research, one of the important issues is multiple comparison adjustments or multiplicity. Due to the large number of potential hypotheses and the discovery-based nature of such studies, investigators may be overwhelmed by the large number of potential analyses possible or become distracted by signals that may inflate false-positives. The multiplicity issues arising within cancer studies are classic problems in drug evaluation and have been heavily studied by regulatory agencies, pharmaceutical/biotech industries, and research institutes [79] . Statistical algorithms, such as the Bonferroni correction, and Hochberg procedure [80] , referred to as multiplicity adjustment procedures (MAPs), have been developed based on the logic that multiplicity can be adjusted by applying more stringent criteria on type I error control.

7. Meta-analysis

Meta-analysis studies are a synthesis of pooled information from existing studies to draw statistics inference. Several types of meta -analyses exist: literature-based, IPD-based, and NMA. Many meta -analyses are derived from published literature, but these are vulnerable to publication bias, “file drawer” effect (i.e., never see the light of day), and variation in quality of separate studies related to methodology (including eligibility) and outcome assessment. In contrast, IPD is considered the gold standard which contains the data of each individual patient, but may not always be available due to confidential policy or data transfer issues, or logistical/operational costs. Finally, NMA summarizes relative treatment effects from independent trials which infers indirect treatment comparisons. However, indirect evidence should be interpreted with caution since it may be more susceptible to imbalanced stratification [81] . Notably, an important caveat when interpreting results for any meta -analyses is that historical migration (demographics, staging, and treatment techniques/systemic agents, etc.) may occur if trials are conducted over different eras.

8. Common statistical pitfalls

Common pitfalls are seen in the oncology literature including incomplete/inappropriate study design, mis-specified statistical models and tests, incomprehensible scientific reports, and tables and figures using incorrect formats. Additional drawbacks include unadjusted analysis of treatment effects without multivariable analysis, insufficient adjustment for baseline measurements, the use of covariates measured after the start of treatment, and composite response measures ( Table 4 ). For longitudinal studies with repeated measurement over time, researchers should take into account all measurements instead of limiting analyses to baseline measures [82] .

Common pitfalls in study design, analysis, and report.

9. Conclusions

This paper provides an overview of statistical principles for clinical and translational research studies and demonstrates how proper study design and correctly specified statistical models are important for the success of cancer research studies. We emphasize the practical aspects of study design, and assumptions underpinning power calculations and sample size estimation. The differences between RR, OR, and HR, understanding outcome endpoints, and statistical modeling to minimize confounding effects and bias are also discussed. Finally, we describe commonly encountered statistical pitfalls that can be avoided by following correct statistical principles and guidance to improve the quality of research studies.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

• International edition
• Australia edition
• Europe edition

Tasmanian devil facial tumour research challenged: disease may not be declining after all

Cambridge scientists critique study that concluded the cancer was no longer a threat to species’ survival

• Follow our Australia news live blog for latest updates
• Get our morning and afternoon news emails , free app or daily news podcast

Cambridge researchers have challenged a previous study which had concluded that a facial cancer that devastated the Tasmanian devil population was on the decline.

Devil facial tumour disease, a fatal cancer spread through biting and sharing of food, emerged in the 1980s. The spread of DFTD led to the species being listed as endangered by the International Union for Conservation of Nature in 2008.

The original study, published in the journal Science in 2020, found that the rate of transmission had slowed, so that an affected animal would only infect one other animal – previously an infected devil would affect another 3.5.

The researchers in 2020 were “cautiously optimistic” that the devils had developed a natural immune response to the cancer and concluded the disease was no longer a threat to the species’ survival.

Sign up for Guardian Australia’s free morning and afternoon email newsletters for your daily news roundup

Cambridge scientists replicated the study, and concluded the key findings of the original study could not be reproduced, leaving the future of Tasmanian devils uncertain, in a critique published in Royal Society Open Science.

Elizabeth Murchison, a professor of comparative oncology and genetics at the University of Cambridge and one of the critique’s senior authors, said the original researchers sequenced DNA half the recommended number of times.

She said it was recommended that scientists sequence DNA at least 30 times when analysing tumours to have confidence that a variant is actually a mutation. Her reanalysis found the researchers in the original study sequenced DNA an average of 15 times.

Murchinson said the mutation rate recorded by the original researchers was “implausibly high” and suggested that the mutations recorded were likely non-meaningful.

The authors of the initial study disagreed, and said they stood by their research. They said they had published papers in the years since that “support the basic conclusion that continued survival of Tasmanian devils in the wild is likely and that there has been rapid evolution of devils in response to the disease”.

In a joint statement, they said the Cambridge researchers had previously approached the journal Science to publish a critique of the first study, but the publication rejected it. They said Murchison and her fellow authors had now published “an almost identical” critique in the Royal Society Open Science and not afforded them the right to respond before publication, contrary to “usual procedure”.

The Cambridge researchers said they reproduced the initial study after noticing that the tree mapping out how the tumour evolved over time generated by the original researchers “looked nothing” like their own tree.

after newsletter promotion

Carolyn Hogg, a population biologist at the University of Sydney who was not involved in either study, commended the depth of sequencing analysis the Cambridge researchers took. She said they were “by far the global experts” in the field.

She disagreed with the conclusions the authors drew in their initial paper and did not see how their conclusions were supported by their data.

“I don’t know if the [initial] researchers did anything wrong,” she said. “They probably weren’t aware [of sequencing depths] … because they’re not cancer researchers.

“It’s a cautionary tale for scientists to be cautious of the conclusions they draw if they’re not an expert.”

Hogg said the “best bet” for Tasmanian devils was a vaccine being developed by the Menzies Institute for Medical Research.

Most viewed