Choosing the best treatment for a patient with cancer involves gathering as much information as possible about the disease. Recent advances in prostate cancer imaging now allow for much more detailed information than previously. More information about the disease does not, however, automatically lead to better treatment outcomes. The evidence behind our current clinical guidelines is based on studies relying on urologists’ index fingers (DRE) for local staging and CT in combination with bone scan for metastasis staging. It will be a long time before high-level evidence from studies using modern imaging for staging can guide treatment choices. Until then, observational studies linking findings on modern imaging with oncological outcomes can help us to better predict what an individual patient may expect from a certain treatment.

One such study was recently published in the BJU International: Dutch researchers investigated the association between preoperative staging with MRI plus prostate-specific membrane antigen (PSMA) positron emission tomography (PET)/CT and short-term oncological outcomes after robot-assisted radical prostatectomy with extended lymph node dissection in 493 men with unfavourable intermediate- to very high-risk prostate cancer [1]. In multivariable analysis, seminal vesicle invasion on MRI (radiological [r]T3b, 18% of the patients) and of lymph node metastasis on PSMA PET/CT (molecular imaging [mi]N1, 13% of the patients) were the only variables significantly associated with a postoperative PSA level of ≥0.1 ng/mL (16% of the patients). In addition to rT3b and miN1, biopsy Gleason Grade Group 4–5 and preoperative PSA were associated with biochemical recurrence (including persistent PSA) after a median follow-up of 13 months. All odds ratios for rT3b and miN1 were as high as 3–5 (95% CIs 2–10). Approximately 80% of the patients with miN1 were estimated to experience recurrence within 2 years. This proportion was similar for those who had PSMA uptake in one and two or more nodes.

What does this study tell us? It is not surprising that surgery rarely cures patients with evidence of lymph node metastasis on PET/CT but the results from this study are valuable for preoperative counselling, basically preparing miN1 patients for the fact that further treatment will likely be needed within a couple of years if they opt for surgery. The 20% recurrence-free rate at 2 years is probably a best-case scenario, as patients with multiple lymph node metastases or locally advanced tumours were probably recommended radiotherapy rather than surgery. The poor oncological outcome for miN1 patients does not necessarily mean that surgery is a bad choice as a first step in multimodal treatment. However, given the high-level evidence for primary radiotherapy plus hormonal treatment for oligometastatic prostate cancer, surgery should be done within prospective trials or be reserved for carefully selected miN1 patients (e.g. those with urinary retention, bowel problems, or a strong wish to postpone hormonal treatment).

More interesting is the strong association between rT3b and recurrence, with odds ratios as high as those for miN1. Even more so, as the negative prognostic impact of a preoperative rT3b status might, just as for miN1, have been underestimated as massive seminal vesicle invasion on MRI probably often led to patients being treated with radiotherapy rather than surgery.

The study by Meijer et al. [1] adds to a growing body of evidence that local staging of prostate cancer with MRI is clinically important, particularly when seminal vesicle invasion is found [2-4]. The current TNM classification is clearly at odds with this evidence as it states that MRI findings should not be used for local (T) staging [5]. According to the TNM classification, a impalpable tumour with massive seminal vesicle invasion on MRI should be categorised as T1c, which in turn means that the cancer should be classified as intermediate risk if it is also a Gleason Grade Group ≤3 cancer and the PSA level is <20 ng/mL. This is not reasonable! Using MRI for local staging has its limitations but even more so has a urologist’s finger. Most likely, the presence or absence of obvious extra-prostatic extension or seminal vesicle invasion on MRI is more strongly associated with oncological treatment outcomes than local stage assessed by DRE alone. The next time the TNM classification is revised, the committee will have to consider results from studies like the one by Meijer et al. [1]. My guess is that they will then conclude that MRI results should be included in the clinical T-stage categorisation in the next version of the TNM classification.