Growing up in Ukraine in the wake of the Chernobyl nuclear disaster, Eugene Shenderov was diagnosed with leukemia at an early age. He spent much of his childhood in hospitals, developing a "lifelong fascination" with the disease that threatened his life.
Shenderov survived the cancer. Now, as an associate professor and researcher at Johns Hopkins University, he helps others do the same.
Shenderov and his team are fighting cancer on several different fronts, including promising new trials to treat prostate cancer with immunotherapy. But perhaps most promising of all is Shenderov's work with artificial intelligence, which he pursues with colleagues Mark Levis, Alex Baras, and Ivo Francischetti, as well as a team of student researchers.
Scientists across Hopkins have spent the past few years fascinated with how AI can improve medical diagnoses. The technology is being used to flag cases of sepsis, Lyme disease, strep throat, degenerative eye conditions, COVID-19— and, of course, different types of cancer. These technologies can help physicians reach faster diagnoses, improving patient care and outcomes.
For Shenderov, the disease in the spotlight is acute promyelocytic leukemia (APL), a rare subform of leukemia that affects the white blood cells. Only 800 Americans are diagnosed with it each year.
APL is extremely aggressive and can cause sudden, life-threatening brain bleeds. But there is one silver lining: Once diagnosed, APL goes from being the most lethal leukemia to the most curable one. Thanks to an extremely effective treatment that involves giving patients essentially high doses of vitamin A, APL's five-year survival rate has reached more than 90%.
"In this case, diagnosis directly leads to saving lives, period," Shenderov explains. "If you can catch it three hours into a patient's stay rather than three days, you potentially have saved them the morbidity of brain bleeds and even death."
The numbers Shenderov mentions aren't hypothetical. Right now, only specialized hospitals can make the diagnosis, with multi-day delays being the norm. But with the help of Shenderov's artificial intelligence program, which he calls the "Leukemia Smart Physician Aid," doctors can reach the same conclusion in just a few hours, allowing them to start crucial treatment as soon as possible.
"That's a true paradigm shift," Shenderov says. "This algorithm has the potential to add to existing human capacity and not merely substitute for it."
The Leukemia Smart Physician Aid uses peripheral blood smears to flag potential cases of APL. These smears, which involve examining a patient's blood under a microscope to look for abnormal cells, are relatively inexpensive and done all over the world. This has huge implications for patients in rural and low-resource areas, where APL is less likely to be properly diagnosed. Right now, patients usually need genomic testing to be diagnosed with the disease. This means traveling to a major academic medical center, sometimes across international borders, which can have major time and cost implications.
Each delay makes the disease more dangerous. Sometimes, even when a doctor finally recognizes the APL, the information comes too late.
"If we had diagnosed them sooner, we could have saved them," Shenderov says. "The hours are the difference."
To Shenderov, APL is just the start. He believes that his team's algorithm could have applications far beyond what it's currently designed for. With some tweaks and testing, the technology could detect a wide variety of blood diseases, like malaria or other leukemias. Shenderov is even discussing with NASA-affiliated scientists whether this technology can be applied to blood testing for space travel. But as federal cuts to research funding threaten cancer research around the country, Shenderov is facing uncertainty with his current projects. For example, the Department of Defense, which funds his work with prostate cancer, recently cut its budget by 57%, eliminating whole programs aimed at fighting pancreatic, lung, kidney and other cancers.
Right now, the Leukemia Smart Algorithm needs thorough testing to make sure it's ready for clinical rollout. Typically, that would involve government funding. Now, Shenderov isn't so sure. The problem worries him. Less government research funding means fewer chances at finding cancer cures.
"You never know what you need in order to save a life until you've looked into it," he says. "I'm a prostate cancer doctor and a GU immune-oncologist, so I see mostly prostate cancer, kidney cancer, bladder cancer, etc. Well, why am I even working on a leukemia project? Because our research can take unexpected detours, and saving lives is saving lives. ... If you start to eliminate branches of research because you don't believe that they're going to be useful, you may never have a researcher who's working in that field come up with a discovery that has implications for a vastly different field."
Shenderov will be in Washington, D.C. on April 7 and 8 with the American Society of Clinical Oncology as an alumnus of its Government Relations Committee. There, he hopes to make a difference in the minds of legislators.
"We're going to tell our elected officials from both sides of the aisle that cancer is nonpartisan," he says. "Cancer doesn't care if you're a Republican, a Democrat, an Independent. It doesn't care if you're old or young, unfortunately. ... Research literally translates to new potential cures in cancer. Research over the years has led directly to the therapies we use today. Research dollars are there to saves lives."
