A new artificial intelligence model can bring much-needed clarity to doctors delivering prognoses and deciding on treatments for patients with colorectal cancer, the second deadliest cancer worldwide. According to a new study published in the journal, ‘Nature Communications’, it was identified that this AI tool accurately predicts how aggressive a colorectal tumor is, how likely the patient is to survive with and without disease recurrence, and what the optimal therapy might be.
The researchers say that the tool is meant to enhance, not replace, human expertise, and is designed to detect and interpret visual patterns on microscopy images that are indiscernible to the human eye. Called MOMA (Multi-omics Multi-cohort Assessment), the model was trained on data from nearly 2,000 patients with colorectal cancer from diverse cohorts including the Health Professionals Follow-up Study, the Nurses’ Health Study, the Cancer Genome Atlas Program, and the NIH’s PLCO (Prostate, Lung, Colorectal and Ovarian) Cancer Screening Trial. During this training period, the model was fed information about the patients’ ages, sexes, cancer stages, outcomes, and the tumors’ genomic, epigenetic, protein, and metabolic profiles.
The test results showed the model’s remarkable capabilities in predicting the patients’ overall survival after diagnosis, as well as how many years the patient will remain cancer-free. Moreover, it accurately predicted how the patient would respond to different therapies, based on whether the tumor harbored particular genetic mutations.
The researchers noted that the model cannot perfectly predict any given patient’s survival and thus recommended that before deploying the model in clinics, it should be tested in a randomized trial to assess its performance in actual patients. The tool is free to researchers and clinicians and the researchers said the model will undergo periodic upgrading as science evolves and new data emerge.
Thanks to the new AI tool, clinicians would be better guided on how to follow up more closely with a patient, consider more aggressive treatments, or recommend clinical trials testing experimental therapies. As such, it could save many lives from the 1 million lives that colorectal cancer claims each year.
The Kun-Hsing Yu, assistant professor of biomedical informatics in the Blavatnik Institute at Harvard Medical School, co-authored the study and led an international team of pathologists, oncologists, biomedical informaticians, and computer scientists. With any AI model, he noted, it is critical to continuously monitor its behavior and performance due to the shifts in disease burden or environmental toxins that can contribute to cancer development.
Aside from providing a feasible solution to the 1 million lives lost to colorectal cancer each year, the study demonstrates the importance of artificial intelligence in medicine. AI-empowered diagnostic platforms can potentially identify patterns and characteristics that would elude the human eye and potentially save more lives. This presents both opportunities and challenges for medical practitioners who, nonetheless, have a responsibility to make an informed decision using AI-assisted models for the benefit of their patients.
