Go to navigation (press enter key)


A Unique Vassar-Mayo Clinic Collaboration

When Ming-Wen An, associate professor of mathematics and statistics, attends oncology conferences, her name tag nearly always raises eyebrows. “I just went to one, and people would see my name tag and say, `Vassar College?’”  

Professor An and URSI fellows Siqi Fan ’17 and Melanie Lai Wai ‘16

The people who typically attend these oncology conferences work at cancer research centers or medical institutions, not small liberal arts colleges.  But An has a unique research collaboration with a colleague at the Mayo Clinic in Rochester, MN.  She and Sumithra Mandrekar, professor of biostatistics and oncology and faculty statistician for the clinic’s lung cancer clinical trials research program, are just finishing up a three-year National Institutes of Health (NIH) grant-funded project studying end points in Phase II cancer clinical trials.

“There are generally three phases of clinical trials,” An explains. “Phase I  mainly looks at the safety profile to make sure the drug is safe for humans and at what dose level.  Phase II looks for signs of efficacy and continues to monitor safety.  And Phase III looks at the effect of a drug on long-term outcome, typically overall survival, which is ultimately what we hope the drug can improve.  Each of these phases has what we call an endpoint where you look at the data and determine whether the trial is a success or not.  Our question is whether the endpoint commonly used in Phase II trials—tumor response--is the most appropriate endpoint.”

One problem is that oftentimes drugs deemed a success in Phase II based on short-term endpoints don’t ultimately improve long-term survival rates, i.e. they fail in Phase III.  “That could be due to a number of reasons,” says An, “but one possibility is that we need robust Phase II endpoints--for example, a continuous endpoint such as changes in tumor sizes over time, rather than just: did the tumor respond, yes or no?”

The seeds for this collaboration were planted long ago.  An grew up in Rochester, and her father worked at the Mayo Clinic as a researcher in orthopedics.  When she was a student at Carleton College majoring in math, her father told her about the field of biostatistics.  Intrigued to learn more, she completed a summer internship with the biostatistics group at Mayo.  “I had a very generous mentor who invited me to all of his meetings with medical doctors, and that experience really piqued my interest in the field.  It blended math and statistical thinking with medicine and science, and it was truly collaborative.”

After graduating from Carleton, An worked for two years with the Cancer Center Statistics group at Mayo before pursuing her PhD in biostatistics at the Johns Hopkins School of Public Health.  In 2008, she joined the Mathematics Department (now called the Mathematics and Statistics Department) at Vassar.  “My doctoral work wasn’t related to cancer clinical trials, but after I arrived at Vassar, my mom was diagnosed with pancreatic cancer, and that renewed my interest in cancer research,” says An. “So for my first sabbatical, I contacted Mayo, and we decided that we would use that sabbatical to apply for a grant.” 

An and Mandrekar did some preliminary work and wrote an article together the fall before her sabbatical. The following spring, they coauthored the NIH grant.  “And we got it!  At first it may seem a challenge for our work that she is in Minnesota and I am in New York, but we have developed a synergistic collaboration and rhythm.  We have bi-weekly conference calls and I occasionally visit Mayo for short-term visits.”

Selfie of Siqi Fan ’17 (front) and Melanie Lai Wai ’16 with the Mayo brothers at the Mayo Clinic

This arrangement works well for them.  “Sumithra is in the center of the action of clinical trials.  On a daily basis, she works with clinical investigators to design and conduct trials and often is needed to address urgent needs of on-going trials.  This leaves relatively little time for her own research.  And I, on the other hand, have more time to do research of my own, but I don’t have access to the kind of clinical trials data we need to develop new methods.  Working together, both our time and resources seem to multiply.”

In their project on Phase II endpoints, they had access to a large database of clinical trials and tumor measurements—but only tumor measurements, no other patient data.   What they’ve found so far is that the alternative tumor measurement-based endpoints they’ve considered are not necessarily more predictive of overall survival than the binary endpoint of tumor response, and they have published these findings in high-impact oncology journals.  “But that’s not to say that there aren’t other better endpoints,” says An. “For example, we may need to include data from other technologies like PET scans or functional MRIs or something beyond just tumor measurements.”

Under the auspices of Vassar’s Undergraduate Research Summer Institute (URSI), An and Mandrekar are now laying the groundwork for a new grant proposal on a promising new technology.   In collaboration with An, URSI fellows Melanie Lai Wai (’16, Mathematics) and Siqi Fan (’17, Neuroscience and Math) are developing statistical methods to assess the predictive ability of a mouse model.  “Avatar mouse models have recently been introduced as a promising approach to personalized medicine,” says Lai Wai. “In this technology, a patient’s tumor is engrafted into several mice that each receive different types of treatment.  By observing the outcomes in the mice and in the corresponding patient, researchers at Mayo are trying to see whether a connection can be drawn between the two.”

Lai Wai and Fan are in the very early stages of this project, developing a statistical model that can be used to run simulations.  “Can the avatar serve as a marker that predicts whether some patients will benefit more from a given treatment than other patients, or whether some treatments work better for a given patient?  That’s the question we’re trying to answer,” says Lai Wai. “Then we can assess the predictive behavior of these markers via statistical models.  We expect different trajectories for different profiles of predictive behavior.”

One of the highlights of their URSI experience was a trip to the Mayo Clinic with Professor An. “It was very impressive,” says Fan. “We got to sit in on meetings with Professor An’s colleagues, including Dr. Mandrekar.  It was great just to listen to them talk and to see how they work.  And we also had a meeting with an MD who will collaborate with us on the project and advise us on its clinical aspects.”

Lai Wai, who is considering a career in statistics, says that the trip was inspiring. “The atmosphere there seemed very friendly and welcoming, and the biostatisticians we met all seemed to love their jobs and to love working at Mayo.  The only drawback is that it’s in Rochester!  They have six-month winters there!”

--Julia Van Develder

Posted by Office of Communications Wednesday, August 26, 2015