80 Chapter 5 Our results thus indicate that ymrT-staging mainly has merit to re-define the local tumour stage prior to surgery in patients who show no or limited tumour regression, but is of little to no value in well-responding patients who show predominant fibrosis. In these patients, radiologists tend to ‘err on the safe side’ and regard any areas with persistent abnormalities, including fibrosis, at risk to harbour (small volumes of) residual tumour. Our results confirm those of previous literature that – though MRI is relatively accurate in identifying patients with residual tumour – it has a limited sensitivity detect patients with a complete tumour response (ypT0 / pTRG1) [18]. When aiming to select candidates for organ-preservation, we should thus never rely solely on ymrT-staging but combine the findings of MRI – including useful morphologic and DWI response patterns [19-21] – with those of digital rectal examination and endoscopy to more accurately assess the local tumour response [9]. As also demonstrated in a previous study by Haak et al. [22], MRI mainly has a role to discern patients with gross residual disease (who will definitely require surgical resection) from patients with a (very) good or potential complete response who benefit from further clinical and endoscopic evaluation to establish their eligibility to go for organ-preservation. A near-complete response is often used clinically to describe small tumour remnants that may be treated with local excision (or local radiotherapy) instead of complete surgical resection. How to define a near-complete response in terms of TRG and yT-stage, is a topic of debate. Fragmentation of tumour cells within fibrosis may result in low TRG scores but persistent high T-stage (e.g., 9 out of 17 pTRG2 cases in our study were still ypT3) [23]. Conversely, patients with a low ypT-stage will sometimes show a poor tumour regression at histology (e.g., 3 out of 17 ypT2 cases in our study were still pTRG4-5). Better established clinical definitions of how to diagnose a “near-complete” tumour response are therefore urgently needed. When applying ymrT-staging more selectively for patients with a poor tumour regression, it is important that MRI is able to recognize these patients pre-operatively. On average 80% of the pTRG4-5 patients in our studywere classified as mrTRG3-5 on MRI (versus only 34% as mrTRG4-5), suggesting that an mrTRG score of 3-5 may be the most appropriate threshold to select poor responders. These results are in line with those of previous reports showing that the tumour regression grade as assessed on MRI does not match exactly to the same TRG subgroups at histopathology [24,25]. Despite the relatively good performance for ymrT-staging in the subgroup of poor responders, the performance of MRI to diagnose yT4 disease was low. From a clinical point of view, this is an important subgroup as persistent invasion of organs or structures outside the standard total mesorectal excision (TME) plane would involve expanding the surgical field. Though the total number of ypT4 tumours in our cohort was small, these cases were consistently understaged by more than half of the study readers, including the experts. This is in line with a previously published meta-analysis in which 49% of ypT4 tumours were under-staged on MRI [26]. At the same time, a significant proportion
RkJQdWJsaXNoZXIy MjY0ODMw