Molecular markers predict outcomes for patients with liver metastases of colorectal cancer
Researchers led by Ludwig Chicago’sSean Pitroda reported in JAMA Oncology in July a method to accurately predict which patients with limited metastases—or oligometastases—of colorectal cancer are likely to have a favorable outcome following surgical removal of tumors that spread to the liver. The findings could help improve the personalization of colorectal cancer therapy. The study built on previous research led by Sean and Ralph Weichselbaum, co-director of Ludwig Chicago, who in 2018 reported unique molecular patterns that identified patients with a subtype of colorectal liver metastases that was associated with robust 10-year survival rates after surgery. Both that and the most recent study were broadly based on a hypothesis proposed in 1995 by Ralph and Samuel Hellman—a former board member of the Ludwig Institute for Cancer Research—that cancer metastases exist on a continuum and that those that are relatively limited could be cured with localized treatment. Sean and his colleagues devised an artificial intelligence neural network classifier that predicted the molecular subtype of the disease with 96% accuracy and validated this classifier on a cohort of 147 patients with limited liver metastases who were treated with chemotherapy and surgery in a randomized clinical trial in the U.K. Their classification, combined with clinical features like tumor size, predicted treatment outcomes in patients with high fidelity.