The majority of melanoma-related fatalities involve people who had an early-stage diagnosis of the most deadly type of skin cancer before going on to experience a recurrence that was often not noticed until it had spread or metastasized.
An artificial intelligence-based strategy to identify which patients are most likely to experience a recurrence and are therefore anticipated to benefit from intensive therapy has recently been created by a team led by researchers at Massachusetts General Hospital (MGH). A study that was published in NEJM Precision Oncology verified the approach.
Patients with more advanced malignancies frequently get immune checkpoint inhibitors, which successfully boost the immune response against tumor cells but also come with considerable adverse effects. Patients with early-stage melanoma are typically treated with surgery to remove malignant cells.
According to senior author Yevgeniy R. Semenov, MD, an investigator in the Department of Dermatology at MGH, “There is an urgent need to develop predictive tools to assist in the selection of high-risk patients for whom the benefits of immune checkpoint inhibitors would justify the high rate of morbid and potentially fatal immunologic adverse events observed with this therapeutic class.”
“Reliable melanoma recurrence prediction can enable more precise immunotherapy treatment selection, slow the spread of the cancer to distant organs, and increase melanoma survival while minimizing exposure to treatment toxicities.”
To assist in achieving this, Semenov and his colleagues evaluated the efficacy of machine-learning-based algorithms that used information from patient electronic health records to predict melanoma recurrence.
In order to use machine learning algorithms to predict patients’ risk of recurrence, the team gathered 1,720 early-stage melanomas, 1,172 from the Mass General Boston healthcare system (MGB) and 548 from the Dana-Farber Cancer Institute (DFCI). They also extracted 36 clinical and pathologic features of these cancers from electronic health records. Tumor thickness and the rate of cancer cell division were shown to be the most effective predictive characteristics after algorithms were built and tested using multiple MGB and DFCI data sets.
Semenov claims that “our comprehensive risk prediction platform,” which uses cutting-edge machine learning techniques to assess the likelihood of an early-stage melanoma recurrence, “achieved high levels of classification and time to event prediction accuracy.” Our findings imply that machine learning algorithms can extract predictive signals from clinicopathologic parameters for early-stage melanoma recurrence prediction, enabling the identification of patients who may benefit from adjuvant treatment.
Ahmad Rajeh, Michael R. Collier, Min Seok Choi, Munachimso Amadife, Kimberly Tang, Shijia Zhang, Jordan Phillips, Nora A. Alexander, Yining Hua, Wenxin Chen, Diane Ho, Stacey Duey, and Genevieve M. Boland are additional co-authors on the Mass General study
The Melanoma Research Alliance, the National Institutes of Health, the Department of Defense, and the Dermatology Foundation all provided funding for this research.
