Prostate Cancer Can Be Distinguished Into Luminal, Basal Subtypes, Potentially Guiding Therapy

Prostate Cancer Can Be Distinguished Into Luminal, Basal Subtypes, Potentially Guiding Therapy

Dr. Felix Feng
 Prostate cancer can be subtyped based on cell of origin, into luminal A, luminal B, and basal cell subtypes, according to a new study (Abstract 3). Luminal B prostate cancer has the poorest prognosis, but also derives the most benefit from androgen deprivation therapy (ADT).

“The cell of origin of prostate cancer is unknown,” said Felix Feng, MD, of the University of California, San Francisco, who presented the study during a prostate cancer Oral Abstract Session at the GU Cancers Symposium. Early research suggested that the malignancy originates from glandular luminal cells, but more recent research has suggested that basal cells may play a role in carcinogenesis.

The researchers hypothesized that, similar to other hormone-associated adenocarcinomas such as breast cancer, prostate cancer could be divided based on those cell subtypes, which could hopefully help guide treatment decisions. To test this, they used the PAM50 classifier test, which measures expression of 50 classifier and five control genes to differentiate basal and luminal cell involvement. They used this assay on 1,567 prostate cancer samples from high-risk patients who underwent radical prostatectomy.

They found that, much like with breast cancer, the test can indeed separate prostate cancer into luminal A, luminal B, and basal subtypes. Heat maps from the test resemble those seen with breast cancer, where the three types are easily distinguishable. “Given the similarity between these heat maps, it suggests that the underlying biology between breast and prostate cancer is somewhat similar,” Dr. Feng said.

This finding was then validated in another 6,300 prostate cancer samples, showing it to be a robust and reproducible classifier.

The more important question is whether these subtypes are clinically meaningful, and further analysis showed that they are, with luminal B disease faring worse than the other two subtypes. The rates of 10-year biochemical recurrence-free survival were 41% and 39% for luminal A and basal subtypes, respectively, compared with 29% for luminal B. Similarly, luminal A and basal disease both had a 10-year freedom from distant metastasis of 73%, compared with 53% with luminal B prostate cancer.

For prostate cancer-specific survival, luminal A had a 10-year rate of 89%, basal cell carcinoma had a rate of 86%, and luminal B had a rate of 78%. For 10-year overall survival, luminal A had a 82% rate, basal had a 80% rate, and luminal B had a 69% rate. On a multivariable analysis, both of the other subtypes had a significantly decreased chance of metastatic progression when compared with luminal B disease. For luminal A, the hazard ratio was 0.55 95% CI [0.43, 0.69]; p = 5.4x10-7); for basal disease, the HR was 0.66 95% CI [0.53, 0.82]; p = 2.0x10-4).

Dr. Angelo M. DeMarzo
The investigators also found that various biological pathways are up- or downregulated in the luminal and basal subtypes. For example, androgen receptor signaling is higher in the luminal subtypes, while proliferation pathways are higher in basal and luminal B disease. Pathways that are involved with epithelial mesenchymal transition, and immune-associated pathways, are upregulated in the basal subtype.

There was also a suggestion that the subtype can predict response to ADT. Patients with luminal B disease seem to derive the most benefit from the therapy, while there was a signal toward actually doing worse with ADT in non-luminal B disease; Dr. Feng noted this could be an artifact of the biases implicit to this sort of retrospective analysis, however.

To assess which patients would most benefit from ADT, Dr. Feng said a new trial, recently approved by a steering committee, will stratify patients according to cellular subtype and then randomly assign them to salvage RT with either placebo or apalutamide.

“This subtyping approach may provide a molecular strategy for treatment selection for patients with aggressive, but potentially curable, cancers,” Dr. Feng said.

Angelo M. DeMarzo, MD, PhD, of Johns Hopkins University School of Medicine, was the Discussant for the abstract, and noted that biomarker development in general remains very challenging. He also pointed out that though the heat maps showing the subtypes seem promising, they do have some differing patterns from the breast cancer heat maps.

“Are these really the right genes, or should we go beyond these?” he asked, meaning the genes tested in the PAM50 assay. He also pointed out that if tissue is collected from prostatectomy, samples could potentially be contaminated with epithelium, where basal cells are found. Dr. Feng noted, though, that high-tech methods such as laser capture microdissection to further isolate tumor cells are unlikely to be available in community practice. “It’s a pragmatic approach,” he said.

–David Levitan