So what about cancer-specific death? In studies from multiple institutions, where they looked at Gleason score 6 cancers and asked what the likelihood was of dying from prostate cancer vs. competing causes, the former had a very low risk of cancer-specific death in patients 60 years and younger.
We performed similar analysis at the Harvard School of Public Health, Boston, a few years ago, looking at the Physicians’ Health Study and the Health Professionals Follow-Up Study and saw that when we re-created the biopsies and radicals, the scores of Gleason 6 and lower went up to Gleason 7 very often. In a blinded analysis, not one of the Gleason score 6 tumors was associated with lethal prostate cancer.
You’re probably thinking about what we know about conservative management in the setting of localized cancer. I would refer you to the work of Dr. Peter C. Albertsen at the Connecticut Tumor Registry, where by looking at similar types of outcomes between cancer-specific deaths and competing causes, it’s shown that there’s very low risk of dying from prostate cancer with Gleason scores 2-6 (statistics are unadjusted for deaths per 1,000 patients).
From the molecular standpoint – we’ve heard a number of times that various cancers have a risk of somatic alterations. Prostate cancer tends to have a very low number of somatic alterations, but copy number alterations are common, going from localized to advanced. We see driver mutation events occurring more in late-stage cancer, and this is provisional data from 333 localized prostate cancer [cases] from many institutions where it’s been organized from the bottom to the top starting with Gleason score 6.
What we can see is an accumulation of molecular alterations, but what I would highlight is that like adenomas, Gleason score 6 has certain characteristic alterations that we often see without driver events.
And going back to the Physicians’ Health study and the Swedish “Watchful Waiting” study, if we look at all lethal prostate cancers, the vast majority were Gleason score 7 and above. The Swedish cohort was pre-PSA, so some of these Gleason score 6 cancers were probably contaminated because of sampling. Even if we include them and look at principle component analysis, we see that the vast majority of the Gleason score 6 cases cluster together; they’re much more homogeneous, molecularly speaking, compared with the Gleason score 8 cancers that are heterogeneous and spread all over.
Another important analysis that will inform next-generation sequencing studies is to look at subclonal alterations. We believe that the subclonal mutations are the ones that are going to merge in as driver mutations, So what you can do is take the Cancer Genome Atlas data, as we’ve done, and go from Gleason score 6 all the way up to the higher-grade tumors. There are more deletions and genomic focus as you go higher, but also you see that the majority of Gleason score 6 cancers are clonal for these alterations. As you move on up to higher-grade tumors, you start seeing some subclonal emergence of driver lesions.
So I would argue that the Gleason pattern 3 or Gleason score 6 cancers have a very, very low risk of disease compared to other cancer grades and is not associated with death following definitive treatment, and is probably dramatically overtreated. It’s genetically homogeneous, and we see little evidence of subclonal driver mutations in Gleason score 6. So my answer to the question “Should Gleason score 6 be called cancer?” is “no.” But I do agree entirely that we need to have categories of risk, and so Gleason score 6 – or whatever we end up calling it – should be recognized as a group of tumors with a very low risk of aggressiveness.
Dr. Mark A. Rubin is the director of the Institute for Precision Medicine at Weill Cornell Medical College, New York. He reported no relevant financial disclosures.
Dr. Epstein and Dr. Rubin made their remarks at the annual meeting of the Society for Urologic Oncology. Reporter Deepak Chitnis covered their presentation.