Molecular Subtypes Predict Response to Radiation in Prostate Cancer

October 2016, Vol 7, No 9

Boston, MA—Three different molecular subtypes of prostate cancer have been identified in the largest genomic analysis to date of men with primary prostate cancer. The good news is that these subtypes are prognostic and predict response to radiation, according to Daniel E. Spratt, MD, Chief, Genitourinary Radiotherapy Program, University of Michigan, Ann Arbor, who presented the study results at the 2016 American Society for Radiation Oncology annual meeting. The 3 subtypes are dubbed subtypes A, B, and C, with subtype A having the best prognosis.

“This is very exciting work and exactly what we need in prostate cancer, understanding the genes and how they interact. I can’t wait to see this data incorporated in clinical trials,” said press conference moderator Colleen Lawton, MD, FASTRO, Vice-Chair, Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee. “Once we have these data we will treat differently. This is tapping into something we didn’t have before.”

“We currently use imperfect tools to diagnose prostate cancer, such as digital rectal exam, biopsy, Gleason score, and PSA [prostate-specific antigen]. These measures do not give us information on disease biology, risk of recurrence, or whether a patient will benefit from a specific therapy. Tumors that look clinically similar can be genetically different. For example, a tumor that appears to be at high-risk clinically may be genetically low-risk, and vice versa. We found 3 clusters of tumors that predict risk and response to radiotherapy,” said lead investigator Dr Spratt.

Dr Spratt and colleagues studied 9 large cohorts of patients, with 4236 samples drawn from men with primary prostate cancer who had undergone radical prostatectomy. Of the cohorts, 7 were from institutions, and were studied retrospectively to validate the clusters. Two cohorts were studied prospectively to validate the clusters.

First, they identified the 100 most common genes, which fell into 3 distinct clusters (ie, subtypes) based on decreased and increased gene expression. The subtypes were validated in all cohorts, and were highly reproducible, Dr Spratt said.

The subtypes separated patients according to prognosis for distant metastasis-free survival, with patients in subtype A having the most favorable prognosis. In a multivariate analysis correcting for other clinical and pathological features, subtype was an independent predictor of prognosis.

Next, the researchers looked at response to radiotherapy according to subtype in 350 men with radical prostatectomy followed by radiation. The investigators found that subtypes B and C had a more than 2-fold response to radiation compared with subtype A, suggesting that subtype A patients do not benefit from radiation and may not need it.

“This may be one of the first biomarkers to tell us which patients will benefit from radiation. This gives us a method of personalizing care,” Dr Spratt told attendees.

Several companies are interested in developing genomic platforms to identify these subtypes, he said. The next step is to incorporate testing for the 3 subtypes upfront in clinical trials to determine genomic risk and response to other therapies. For now, these are preliminary data, but the goal is to use the subtypes to select patients for therapy.

Related Articles