Researchers Develop Gene Classifier to Aid Treatment Decisions

San Diego—With the increasing use of precision medicine, there has not yet been a broad impact of genomic classifications of prostate cancer on patient care, according to researchers led by Sungyong You, PhD. The research team presented results of an integrative analysis of 14 disease-related pathways in more than 4600 clinical specimens and 25 prostate cancer preclinical models during a poster session at the AUA 2016 Annual Meeting. The poster was titled Three Intrinsic Subtypes of Prostate Cancer with Distinct Pathway Activation Profiles Differ in Prognosis and Treatment Response.

The analyses were validated in ten independent patients series. Kaplan-Meier analysis and Cox proportional hazard regression were performed to correlate subgroup assignment to time to metastatic progression and prostate cancer specific mortality. Drug sensitivity of patient subgroups was examined using a comparative Gene Set Enrichment Analysis approach. The researchers also developed a 37-gene classifier using a random forest machine learning algorithm that allows individual tumors to be subtyped with high accuracy.

The novel prostate cancer classification scheme developed by the researchers included three subtypes (PCS1, PCS 2, and PCS3). The classification scheme predicts disease progression and drug resistance.

PCS1 tumors, including those with low Gleason score, progress most rapidly to metastatic disease. PCS1 and PCS3 subtypes are over-represented in treatment-resistant tumors. PCS1 and PCS2 have characteristics of luminal cells; PCS3 exhibits basal features. Common prostate cancer cell lines could be categorized using this approach; however, the researchers said that none of the mouse models they analyzed resembles PCS3.

In conclusion, the researchers said, “This new subtyping method applies to primary as well as metastatic and castration-resistant tumors and provides novel opportunities for patient stratification and therapeutic decisions.”

Source: You S, Knudsen B, Erho N, et al. Three intrinsic subtypes of prostate cancer with distinct pathway activation profiles differ in prognosis and treatment response. Abstract of a poster presented at the American Urological Association 2016 Annual Meeting, May 7, 2016, San Diego, California.