Personalizing Prostate Cancer Treatment with Math
Imaging Software for Individualized Treatment with Radiotherapy
In the United States alone, prostate cancer is the third leading cause of cancer deaths in men. In 2017, the American Cancer Society estimates approximately 161,350 new cases of prostate cancer and 26,730 deaths attributed to prostate cancer. Treatments targeting prostate cancer often involve different forms of radiation therapy.
Prostate cancer patients are often treated with multi-fractionated intensity modulated radiotherapy (IMRT), an external beam therapy that aims high doses of radiation at the prostate tumor. Image-guided radiation therapy (IGRT) allows physicians to further utilize imaging principles to take pictures of the prostate and more accurately target it, since it moves during treatment.
Prostate motion is unique from person-to-person and mediated by factors such as size; a person with a smaller prostate, for instance, might exhibit increased motion compared to a patient with a larger prostate. However, while individuals experience distinct motion movements during radiation, current methods typically utilize 5-10-mm uniform radiation treatment margins around the prostate for all patients. Radiation aimed using these uniform margins may not adequately target the moving tumor, and might instead expose surrounding organs—such as the bladder and the rectum—to increased toxicity, especially if very large margins are used.
Mohammad K. Khan, MD, PhD, associate professor in radiation oncology in the School of Medicine, is uniquely suited to address this problem. Khan, who holds doctorate degrees in both medicine and nuclear engineering, is accustomed to analyzing large data sets using engineering and statistical principles. He even saw motion management in action during his residency as part of his research and prospective trial. “Different patients have different organ sizes, and different organ motions… and yet, we treat everyone the same,” notes Khan. “We now have new technology and innovations in place to be able to take this variability into account, and potentially integrate them into future radiation treatments.”
Khan says researchers can better take into account differences among patients regarding how their prostates and normal organ moves, and individualize their radiation treatments into a “more personalized approach,” using engineering principles. Kahn, in collaboration with colleagues from University of Tennessee and Parkridge Medical Center, developed a mathematical model that utilizes Bayesian statistics to predict prostate movement based on motion collected, either before radiation treatment begins or within the first few radiation treatments. Such a model allows physicians to collect data and adjust radiation treatment margins, and therefore minimize harmful radiation exposure to the patient. For instance, within the first few treatments, Khan can analyze a patient’s prostate motion and movement variability; he can then perform a predictive analysis on how that motion is going to project itself in the future.
“We can then adapt his treatment margins based on his individualized motions,” Khan says, noting that these individualized margins might differ from a second patient who comes in for treatment.
Khan sees this algorithm being easily integrated into any treatment planning system. He also notes that this concept can apply to other types of cancers in the body that may involve non-random tumor motion and be utilized to individualize treatments on a larger scale. “For example, pancreatic cancer and lung cancer, could be other potential targets,” says Khan. “We can take non-random motion and do predictive analysis on it.”
Hyeon (Sean) Lee, the case manager who oversees this technology for the Office of Technology Transfer adds, “I am certain Dr. Khan’s Individualized Treatment technology will maximize the effectiveness of radiation therapy and greatly benefit the patients undergoing treatment.”
Read our technology brief