An Overdose of Overdiagnosis
Low grade prostate cancer might look like cancer down a microscope, but it doesn't behave like cancer.
The overview effect is what astronauts get when they look at Earth from space. Not a borderline to be seen. The petty details of humanity and its too-often sordid history erased by the view from their distant orbital perspective. A sense of awe and an overwhelming desire to protect.
As surgeons, we must pay attention to detail. Our patients’ lives can depend on it. But we can also get stuck in the weeds, don’t see the wood for the trees, or get distracted by decoys.
We’re dazzled, for example, by the flurry of new robotic surgery technologies (1). And we get wowed by the survival benefit of adding even more hormone-based therapy, with all its morbid side effects, for metastatic disease (2). We can be too close, sometimes, to see the bigger picture.
But if we zoom out, what might appear in our field of vision that makes more sense for improving the lives of our patients? The long haul back to Australia at 35,000 feet is not quite space travel, but it’s not a bad opportunity to get some distance.
The EAU25 conference in Madrid was a month ago now, but a pattern emerged there which has stuck with me since. In every prostate cancer session, it seemed, was the unrelenting concern about overdiagnosis in early detection. Overdiagnosis in ERSPC (3), overdiagnosis in ProBase (4), overdiagnosis in ProScreen (5). You might say we were overdosing on overdiagnosis.
But what exactly is this overdiagnosis we are all so worried about? It’s the detection of ISUP Grade Group 1, or Gleason 6 cancer in the old language. We call it overdiagnosis because we know that this form of prostate cancer is itself harmless (6), apart from the anxiety it causes patients (7), and because still too many patients with it are unnecessarily treated (8). We say that we would prefer to not detect it at all because of its indolent nature. (Yet, illogically, when we do detect it, we follow these patients closely on active surveillance – work that out!)
We know from the most rigorous studies in PSA testing that screening of men aged 50-70 with PSA on a repeated basis can reduce death from prostate cancer (9). Too often ignored is also screening’s ability to reduce prostate cancer metastasis (and its morbid systemic treatment), at a rate around double that of its mortality reduction. ERSPC showed that this occurs even prior to MRI being introduced to the diagnostic algorithm.
MRI has now been standard of care in the diagnostic process for prostate cancer for several years, as recommended in the EAU Guidelines (10). By using it to triage for biopsy, it can avoid up to half of all biopsies – the main source of detection of GG1 disease, i.e. of overdiagnosis (11). But still we worry about overdiagnosis! And as a result, the argument remains against prostate cancer screening due to harms apparently outweighing benefits.
And let’s not kid ourselves, the benefits of treatment when we do detect prostate cancer early are nowhere near as good as we’d like to think. For example, according to Vickers’ paper last year using statistical modelling to contemporise raw data from the SPCG-4 and PIVOT trials, we would have to operate on 30-50 patients with GG2 prostate cancer to prevent a single death from the disease (12). The vast majority of GG2 patients we treat with standard radical therapy (surgery or radiation), therefore, do not enjoy a mortality benefit. But this is an argument for surveillance, or in the near future, focal therapy, for such patients. It is not an argument against screening.
If the true argument against screening is overdiagnosis of GG1 disease as the harm that tips the scales away from the demonstrated benefit, then the simplest and most far-reaching way of avoiding overdiagnosis is to stop calling GG1 cancer. Overdiagnosis problem: solved!
Eggener and others have been urging this for years now but have been stonewalled by our pathologist colleagues who have as yet been unable to see past their microscopes to the human beings beyond (13). We can duly acknowledge their protests that GG1 does indeed have features consistent with a cancer, and at the same time argue philosophically for the greater good of still not labelling it as cancer. This is the greater good not only for the patient whose biopsy is on the slide, but for all future patients who stand to benefit from the rollout of risk-adapted screening programs.
Remember how we used to get hung up on ASAP or HG-PIN? Nowadays we ignore these and rightly so. With MRI now enabling us to target lesions, what may have been diagnosed as GG1 in the past era of systematic-only biopsies, is now more likely to be detected as GG2. MRI-targeted biopsy is therefore causing a grade shift (14). For this reason, we are now rightly offering active surveillance not only to GG1 patients, but to many with GG2 disease also.
From a clinical perspective, the GG1 of today is the HG-PIN of the past. It is time, therefore, for the sake of realising the benefits of screening without the blight of overdiagnosis, for GG1 to be relabelled as non-cancer. We must continue to engage with and respectfully lobby our pathologist colleagues to consider making this positive change together.
As the EU and Australia move away from current opportunistic testing for prostate cancer with its inherent health inequities that disadvantage the less health-literate in our communities, and we edge towards planning risk-adapted screening, removal of this remaining impediment will only accelerate screening’s implementation.
Prostate cancer screening can then finally and rightfully sit alongside the well-established screening programs for the other common cancers such as cervix, breast and bowel. One day, as we sit back on our commercial spaceflight in contemplation, we may wonder how we could ever have practised without it.