An AI-powered method for detecting tumor DNA in blood has shown unprecedented sensitivity in predicting cancer recurrence, in a study led by researchers at Weill Cornell Medicine, NewYork-Presbyterian, the New York Genome Center (NYGC ) and Memorial Sloan Kettering Cancer Center. (MSK). The new technology has the potential to improve cancer care with very early detection of recurrence and close monitoring of tumor response during therapy.
In the study, which appears June 14 in Nature medicine, researchers demonstrated that they could train a machine learning model, a type of artificial intelligence platform, to detect circulating tumor DNA (ctDNA) based on DNA sequencing data from patient blood tests, with very high sensitivity and accuracy . They successfully demonstrated the technology in patients with lung cancer, melanoma, breast cancer, colorectal cancer, and precancerous colorectal polyps.
“We were able to achieve a remarkable signal-to-noise ratio improvement and this allowed us, for example, to detect cancer recurrence months or even years earlier than standard clinical methods,” said study co-corresponding author Dr. Dan Landau, professor of medicine in the division of hematology and medical oncology at Weill Cornell Medicine and senior faculty member at the New York Genome Center.
The study’s co-first author and co-corresponding author was Dr. Adam Widman, a postdoctoral fellow in the Landau Laboratory and a breast cancer oncologist at MSK. The other co-authors were Minita Shah of NYGC, Dr. Amanda Frydendahl of Aarhus University, and Daniel Halmos of NYGC and Weill Cornell Medicine.
Liquid biopsy technology has been slow to realize its great promise. Most approaches to date have targeted relatively small sets of cancer-associated mutations, which are often too sparsely present in the blood to be reliably detected, resulting in cancer recurrences that go undetected.
Several years ago, Dr. Landau and his colleagues developed an alternative approach based on whole-genome sequencing of DNA in blood samples. They showed that they could collect much more “signal” this way, allowing for more sensitive (and logistically easier) detection of tumor DNA. Since then, liquid biopsy developers have increasingly adopted this approach.
In the new study, the researchers took a step forward again, using an advanced machine learning strategy (similar to that of ChatGPT and other popular AI applications) to detect subtle patterns in the sequencing data, in particular, to distinguish patterns that suggest cancer from those that suggest cancer. of sequencing errors and other “noises”.
In one test, the researchers trained their system, which they called MRD-EDGE, to recognize patient-specific tumor mutations in 15 colorectal cancer patients. After the patients’ surgery and chemotherapy, the system predicted from blood data that nine had residual cancer. Five of these patients were found (months later, with less sensitive methods) to have cancer recurrence. But there were no false negatives: None of the patients deemed free of tumor DNA by MRD-EDGE experienced recurrence during the study window.
MRD-EDGE showed similar sensitivity in studies of patients with early-stage lung cancer and triple-negative breast cancer, with early detection of all but one recurrence and monitoring of tumor status during treatment.
The researchers showed that MRD-EDGE can detect even mutant DNA from precancerous colorectal adenomas, the polyps from which colorectal tumors develop.
“It was not clear that these polyps removed detectable ctDNA, so this is a significant advance that could guide future strategies aimed at detecting premalignant lesions,” said Dr. Landau, who is also a member of the Sandra and Edward Meyer Cancer Center at Weill. Cornell Medicine and hematologist/oncologist at NewYork-Presbyterian/Weill Cornell Medical Center.
Finally, the researchers demonstrated that even without prior training on sequencing data from patients’ tumors, MRD-EDGE could detect responses to immunotherapy in patients with melanoma and lung cancer, weeks before detection with standard imaging-based in x-rays.
“Overall, MRD-EDGE addresses a huge need and we are excited about its potential and working with industry partners to try to deliver it to patients,” Dr. Landau said.
Research for this story was funded in part by the National Cancer Institute, part of the National Institutes of Health, through grant number R01 CA266619.