Watch your steps, especially when changing treatment for locally advanced rectal cancer

Treatment for rectal cancer has evolved over the last decades. After the introduction of low-dose radiotherapy (RT) prior to surgery, local control rates improved, although the effect on survival was not convincing [1–3]. From Sweden a series of studies have investigated the effects from short-course preoperative RT (SCRT) [4]. With the seminal publication on the results from total mesorectal excision (TME) trial, the first data indicated that surgery could yield 3.7% local recurrence rate without RT [5]. Norway was one of the first countries to accept the new TME surgery and arranged TME training courses for rectal cancer surgeons. Motivated by the lack of survival effect from preoperative RT alone, TME was initially implemented without additional treatments. Local recurrence rates improved from 25-30% to 14.5%, but remained higher than the target level around 5% [6].

Soon, it became evident that the distance from tumour to the mesorectal fascia (CRM or MRF) evaluated in the operation specimen was a strong prognostic factor [7,8]. With the introduction of magnetic resonance imaging (MRI) it became possible to measure this distance preoperatively [9]. A minimal MRF or CRM distance of 1–2 mm by MRI is now accepted as an indication for preoperative RT. The changing treatment guidelines and patterns have translated results to better outcomes, which is demonstrated in new data from The Norwegian Cancer Registry that now show local recurrence rates of 4%–5% compared to 25%–30% before the introduction of TME [6]. Distant spread remains the predominant clinical challenge with distant metastases reduced from 25% to 20% for all operable rectal cancer patients with or without preoperative RT in Norway. Treatment of rectal cancer is constantly evolving. There is consensus that organ preservation represents a paradigm shift in the management of early rectal cancer [10–13]. In this issue of Acta Oncologica, two previously published studies using different preoperative RT scheduled have been re-analysed [14]. Three schedules were used: SCRT alone before surgery or SCRT followed by three courses of de Gramont chemotherapy (CT) with delayed surgery, and conventional preoperative chemo RT (CRT) with 50 Gy given by 2 Gy per fraction over 5 weeks, combined with 5-fluorouracil and leukovorin (Nordic schedule). The authors motivate the investigation with concern for a potentially inferior treatment using SCRT with respect to complete clinical response. Heterogeneity in treatment with the different regimens and potential bias in patient selection may be related to treatment choice at the discretion of the participating institutions in the observational studies presented. Especially the SCRT alone group included many early stages. The study revealed no significant differences in complete clinical response in relation to RT schedule. Although the authors conclude that the data did not demonstrate superiority of CRT over delayed SCRT or SCRT + CT and underline that this is hypothesis generating, the study is underpowered to demonstrate superiority.

Demonstration of potential superiority would require randomised data with comparable groups. Hence, the data presented document that SCRT can lead to a complete clinical and/or pathological complete response and lend further support to the watch and wait strategy [15,16]. The authors presented complete response data separately for the earliest tumours, defined as less than 4 cm length and less than 50% of the rectal circumference. Among 152 patients in this group complete clinical response rates for SCRT, SCRT + CT, and CRT were 26%, 38% and 45%, respectively. It is not obvious that these three regimens are equal based on this small sample. In a recent large register study 764 patients treated with SCRT were compared with 5070 patients treated with CRT. Pathological complete response rates were 9.3 vs. 17.5% [17]. Whether clinical or pathological complete response represent ideal surrogate endpoints remains unclear [10,18]. These endpoints are linked to favourable outcomes using disease-free survival or cancer-specific survival. However, in a study 17% of the patients with pathological complete response in the primary tumour had viable cells in harvested lymph nodes indicating increased risk of pelvic recurrence if the nodes were not removed [19]. In the study by Pietrzak L et al. [14] the primary endpoint is local regrowth with death of any cause handled as a competing risk. The authors apply a Gray test to avoid the inherited overestimation using Kaplan–Maier calculations. Also, the definition of time to surgery from day one of radiation start to date of decision of clinical complete response may influence relevant endpoints.

In Norway the Norwait study (https://clinicaltrials.gov/NCT03402477) was designed to diagnose complete clinical response after standard preoperative CRT with capecitabine or 5-fluorouracil or SCRT followed by CT. In Norway, the national treatment guidelines call for preoperative RT for patients with locally advanced rectal adenocarcinomas, i.e. T3 tumours with <2 mm MFR (CRM) or 1 mm for an involved lymph node, and T4 tumours or tumours with extra-mesorectal lymph nodes. This contrasts most published watch and wait studies where most included tumours are T2 and T3. After including 93 patients, the study has been temporarily stopped due to a higher-than-expected rate of distant metastases at one institution. In the study distant metastases have developed in 12.9% of all included patients. This can be compared to 20% for all rectal cancers treated by surgery alone in Norway, 26.8% in the Rapido trial CRT arm and 20.0% in the experimental SCRT + CAPOX or FOLFOX 4 [20]. The results can be regarded as expected in the Norwait study, given that all locally advanced rectal cancer patients were eligible.

There is more information to be drawn from MRI before the treatment goal is decided: A reasonable way to move forward is to add to the inclusion criteria the well-documented prognostic factors published for colorectal cancer. For T3 tumours it is possible by MRI to identify extramural venous invasion (EMVI+) [21,22], extramural extension (EMD) [23,24], or satellite tumour deposits in mesorectum (N1c) in addition to T4 and mucinous tumours [15]. These factors are associated with poor prognosis and can be a warning for inclusion in watch and wait studies [11,25]. Locally advanced rectal cancer with poor prognostic features should rather be selected for more intense preoperative treatment regimens [20].

In conclusion, the jury is still out related to optimal selection for watch and wait strategies. This concept may represent an important strategy for early-stage low rectal cancers with good prognostic factors. For higher tumours where sphincter-saving procedures can be used, and for all patients with high risk of distant spread, more intense chemoradiation strategies aimed at reduction of distant spread should be considered and discussed with the patient. The data from the Polish study and the early findings from the Norwait study serve as reminders of the need for careful selection of rectal cancers suitable for a watch and wait strategy [11,25].

Disclosure statement

No potential conflict of interest was reported by the author(s).