Tumor response to nCRT is considered to be one of the important factors predicting the prognosis and influencing treatment strategy decisions. However, accurate prediction of the tumor response is still difficult to achieve. Consequently, ypCR patients continue to undergo TME surgery because the tumor response could not be accurately determined before tumor removal. For patients with low rectal cancer who need a definitive abdominoperineal (APR) procedure, the subsequent permanent stoma greatly decreases the quality of life. Increasing experience with the W&W strategy for cCR or near-cCR through several international registries or single center reports has stressed the importance of judging the status of tumor shrinkage before deciding the surgical plan. This retrospective study was a large sample study exploring the relationship between mucosal changes after nCRT and ypCR. It investigated the prediction of ypCR in patients with residual flat mucosal lesions after nCRT based on several classical factors, and established a nomogram with a maximum prediction efficiency of 60%.
Mucosal appearance of residual disease and ypCR:
The appearance of post-treatment residual mucosal lesions is important for assessing the treatment response. It is generally believed that the typical endoscopic signs of cCR after chemoradiation therapy include the disappearance of the tumor with healing of the mucosa, decrease in size and complete normalization of the tumor bed, or residual red or white scarring with telangiectasia but without palpable abnormalities [10–12]. However, the reliability of these signs is not very satisfactory, and consistency between cCR and ypCR has not been clarified yet.
Among the 246 patients in our study, 207 patients had incompletely flat lesions (lesion depth > 0 mm), of which 36 (17.4%) were ypCR. However, among the 39 patients with completely flat mucosa (lesion depth = 0 mm), the ypCR rate was as high as 51.3%. The results showed that a lesion depth = 0 mm was associated with ypCR. This suggests the possibility of assessing the post-treatment tumor response by digital rectal exam (DRE) and endoscopy, and if the residual lesions are found to be almost completely flat, these patients could be managed by the W&W strategy. However, there are some other studies that are worth noting. Fraser et al. and Jemma Bhoday et al. showed that most patients with ypT0 or pCR status did not exhibit mucosal complete clinical response[8, 13]. Therefore, we believe that it may be necessary to combine other parameters to improve the accuracy of ypCR prediction. Consequently, in our study, we explored the synergistic effect of lesion depth, MR-TRG, and post-nCRT CEA for predicting ypCR. The results revealed that the ypCR rate could reach 54.5% (in lesions with depth = 0 mm and MR-TRG 1 to 3) and 55.6% (in lesions with depth = 0 mm and post-nCRT CEA ≤ 5 ng/ml) by utilizing the combination of two variables.
MR-TRG and ypCR:
Our results also showed that MR-TRG 1 to 3 could be an effective indicator for ypCR. We merged MR-TRG 1/2/3 together as a good response to make it more practical for clinical application similar to other published data. The MR-TRG stage was chosen as the observed variable rather than MR-T staging because although MRI can provide additional details of the extensive pelvic anatomy compared to ERUS or endoscopy, its accuracy for ypT staging after neoadjuvant therapy is low. The sensitivity of MR for the assessment of post-treatment ypT staging was less than 50%, while the sensitivity for the evaluation of ypT0 was only 19.1%[14, 15].
Similar to our study, Patel et al. reported that good MR-TRG (grade 1 to 3) was significantly associated with a favorable pathology [16]. Battersby et al. also suggested that MR-TRG appears to be the most encouraging method for assessing the tumor response to neoadjuvant therapy, and MR-TRG could be used as a key criterion to assess the complete response on the basis of frequent follow-up [17]. Our results are consistent with the above studies. Thus, in patients with lesion depth = 0 mm, we may use MR-TRG more convincingly to find the candidates suitable for the W&W strategy.
Serum CEA and ypCR:
Previous studies have shown that serum CEA levels might predict the response to treatment and clinical outcome, while the preoperative CEA level was a strong predictor of decreased overall survival and systemic recurrence [18–20]. Whether serum CEA could predict ypCR after nCRT is also worth investigating. In our study, multivariate logistic regression analysis showed that a low level of post-nCRT CEA was associated with a higher incidence of ypCR. Our results were consistent with those of other studies by Kleiman et al. and Peng J et al [21, 22]. We believe that the reason why pre-nCRT CEA had no association with ypCR was that it was obtained before treatment. Hence, it predominantly reflected the tumor’s biologic status rather than its response to nCRT, which is in agreement with Yoon DAE Han's view [23].
However, there is no uniform CEA cut-off value to predict ypCR. Peng J et al. used post-nCRT CEA of ≤ 2 ng/mL as the cut-off value, and they claimed that low post-nCRT CEA levels were associated with ypCR [22]. Moureau–Zabotto et al., Das et al., and Steinhagen et al. reported similar findings with a cut-off of 5 ng/mL or 2.5 ng/mL[24–26]. The CEA cut-off value defined by Peng et al. was 5.33 ng/mL, and that by Sun et al. was 5 ng/mL [27, 28]. In our study, we set the CEA cut-off value as 5 ng/ml because it is the maximum threshold of the normal value, which is more in line with diagnostic protocols.
Time interval between chemoradiotherapy and surgery and ypCR:
Studies showed that prolonging the time interval could increase the ypCR rate [29–31]. However, most of the studies were retrospective studies, consequently, this result should be treated with caution. A multicenter randomized, controlled trial (GRECCAR-6) from France reported that there was no difference in the ypCR rate between the 7-week and 11-week groups[32]. Another prospective study found similar results [33]. In our study, there was no statistical difference in the ypCR rate between the two groups (≤ 8 weeks group and > 8 weeks group). When we changed the cut-off value to 12 weeks, the ypCR rate of patients in the > 12 weeks group was higher, but there was no statistical difference. To date, the optimum interval between nCRT and surgery for LARC, in fact, remains controversial. Prospective, randomized, multicenter trials with larger samples are required to draw definitive conclusions.