AI and CT combine to predict efficacy of immunotherapy for melanoma

by Thomas Dworetzky, Contributing Reporter | March 04, 2022
Artificial Intelligence Business Affairs
Artificial intelligence has helped more accurately predict immunotherapy treatment outcomes for melanoma, according to Columbia University researchers in a paper in JAMA Oncology.

The researchers created a machine learning algorithm that looked at patient CT scans and made a biomarker, called a radiomic signature, that correlated with “high accuracy” evaluating how well the melanoma would respond to immunotherapy.

“We hope to take a patient early on who looks like they are not doing well on a given therapy because of their signature and enhance, change, or add another drug to the therapy,” author Dr. Lawrence H. Schwartz of the Department of Radiology at Columbia University Vagelos College of Physicians and Surgeons said in a statement.

The effort is another step toward optimizing individualized cancer care in real time for patients, he noted.

Plans now call for the project to broaden to other types of tumors — including lung, colon, prostate and renal, and to treatments other than immunotherapy.

At present, tumor size is the main way to determine therapy benefit. “Most of the current response criteria were developed several decades ago to assess the response to systemic treatments like chemotherapy,” noted first author Dr. Laurent Dercle of the Department of Radiology at the college. Since immunotherapy can lead to a transitory enlargement of tumors before a response, “we needed to create new tools in order to predict treatment success,” he added.

The program's algorithm looks at tumor volume, and also “spatial heterogeneity, or the non-uniform distribution of cancer cells across disease sites; and texture, which looks at the variation of pixel intensities across the tumor CT image,” noted the authors.

The algorithm-derived radiomic signature used baseline and 3-month-follow-up CT images to predict 6-month survival with a high degree of accuracy — outperforming standard methods using tumor diameter.

“The field of radiology and imaging in general has never been more exciting [than] with this artificial intelligence revolution,” Schwartz said in the Columbia statement. “We’ve always looked at advances in terms of new machines, new tracers, and things like this. But this gives us an opportunity to optimize the information that we have from all of our imaging modalities to speed diagnosis, to become more accurate and precise, and give patients more effective treatments.”

AI was also in the news in late January, when findings from a new algorithm developed at Duke University, published in Nature Machine Intelligence, showed that it could analyze cancerous lesions in mammograms to not only help indicate if a cancer patient should undergo an invasive biopsy but also could show radiologists how it reached its conclusions.

“By suggesting an interpretable diagnosis, the algorithm can help radiologists to improve their performance and consistency. Long term, we hope that this approach can avoid many benign biopsies, which would benefit not only the patients but also expedite the time-consuming processes of diagnostic workup and biopsy scheduling,” Joseph Lo, professor of radiology at Duke University School of Medicine, told HCB News.

You Must Be Logged In To Post A Comment