Predictors of pathologic complete response in patients with residual flat mucosal lesions after neoadjuvant chemoradiotherapy for locally advanced rectal cancer

Changlong Li ,Zhen Guan ,Yi Zhao ,Tingting Sun ,Zhongwu Li ,Weihu Wang ,Zhexuan Li,Lin Wang,Aiwen Wu

1Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing),Gastrointestinal Cancer Center,Peking University Cancer Hospital &Institute,Beijing 100142,China;2 Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing),Department of Radiology,Peking University Cancer Hospital &Institute,Beijing 100142,China;3 Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing),Department of Pathology,Peking University Cancer Hospital &Institute,Beijing 100142,China;4Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing),Department of Radiation Oncology,Peking University Cancer Hospital &Institute,Beijing 100142,China;5Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing),Department of Cancer Epidemiology,Peking University Cancer Hospital &Institute,Beijing 100142,China

Abstract Objective:The accurate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) remains challenging.Few studies have investigated pathologic complete response (ypCR) prediction in patients with residual flat mucosal lesions after treatment.This study aimed to identify variables for predicting ypCR in patients with residual flat mucosal lesions after nCRT for locally advanced rectal cancer (LARC).Methods:Data of patients with residual flat mucosal lesions after nCRT who underwent radical resection between 2009 and 2015 were retrospectively collected from the LARC database at Peking University Cancer Hospital.Univariate and multivariate analyses of the association between clinicopathological factors and ypCR were performed,and a nomogram was constructed by incorporating the significant predictors.Results:Of the 246 patients with residual flat mucosal lesions included in the final analysis,56 (22.8%) had ypCR.Univariate and multivariate analyses showed that pretreatment cT stage (pre-cT) ≤T2 (P=0.016),magnetic resonance tumor regression grade (MR-TRG) 1-3 (P=0.001) and residual mucosal lesion depth=0 mm (P＜0.001)were associated with a higher rate of ypCR.A nomogram was developed with a concordance index (C-index) of 0.759 and the calibration curve showed that the nomogram model had good predictive consistency.The follow-up time ranged from 3.0 to 113.3 months,with a median follow-up time of 63.77 months.The multivariate Cox regression model showed that the four variables in the nomogram model were not risk factors for disease-free survival (DFS) or overall survival (OS).Conclusions:Completely flat mucosa,early cT stage and good MR-TRG were predictive factors for ypCR instead of DFS or OS in patients with LARC with residual flat mucosal lesions after nCRT.Endoscopic mucosal re-evaluation before surgery is important,as it may contribute to decision-making and facilitate nonoperative management or organ preservation.

Keywords: Rectal cancer;preoperative chemoradiotherapy;tumor regression grade;flat mucosal lesions;pathologic complete response

Introduction

Rectal cancer is one of the most common cancers worldwide.Owing to its insidious onset,most patients are diagnosed with locally advanced rectal cancer (LARC).Neoadjuvant chemoradiotherapy (nCRT),followed by total mesorectal excision (TME),has become the standard treatment for LARC.However,radical surgery has many disadvantages,including post-operative complications (such as abdominal infection or anastomotic leakage) and dysfunction (such as urination,sexual dysfunction or anterior resection syndrome).According to the Consensus on the Watch and Wait Policy in Rectal Cancer Patients After Neoadjuvant Treatment (2020 Version) (1),the watch-and-wait (W&W) strategy or transanal local resection in patients with suspected clinical complete response (cCR) or near clinical complete response (nearcCR) after nCRT could alleviate the pain of surgery and avoid a series of problems caused by organ resection.The pathological response is an important prognostic factor for LARC.Studies have shown that approximately 10%-30%of patients with LARC achieve pathologic complete response (ypCR) after nCRT,while up to 70% of patients demonstrate only a partial response to treatment (2-9).The use of existing diagnosis and treatment methods,such as evaluation of clinical characteristics,imaging changes and gross morphology of tumors,to improve the prediction of ypCR is of great significance.Although there have been previous studies on the predictors of ypCR in patients after nCRT,the conclusions have been diverse and inconsistent.However,few studies have been conducted on ypCR prediction in patients with residual flat mucosal lesions after treatment.Despite the fact that these patients exhibit a good tumor response and a large degree of tumor regression,it is difficult to define the status of cCR or ypCR.This study retrospectively analyzed the predictive factors for ypCR in patients with flat mucosal lesions after nCRT for LARC to determine the optimal treatment for such patients.

Materials and methods

Patient selection

Data of 246 patients with residual flat mucosal lesions after intensity-modulated radiation therapy (IMRT) with concurrent capecitabine treatment followed by surgery between 2009 and 2015 were retrospectively collected from the rectal cancer database at Peking University Cancer Hospital (Figure 1).Patients enrolled in our study satisfied the following criteria: 1) age ＞18 years;2) pathologically proven rectal adenocarcinomas;3) rectal tumors located＜15 cm from the anal verge;4) clinical staged as cT3-4 or N+rectal tumors;5) treatment with nCRT and radical surgery;and 6) residual flat mucosal lesions after nCRT.Patients were excluded if they had 1) distant metastasis or recurrent disease;or 2) underwent previous chemotherapy or pelvic radiotherapy.The study was approved by the Medical Ethics Committee of the Peking University Cancer Hospital,and informed consent was waived (No.2021YJZ20).

Figure 1 Flowchart of patient selection.IMRT,intensitymodulated radiation therapy;MR-TRG,magnetic resonance tumor regression grade.

Statistical and clinicopathological data

Data on covariates of interest were collected,including age,sex,height,weight,body mass index (BMI),pre-treatmentcT stage (pre-cT),post-treatment-cT stage (post-cT),pretreatment serum carcinoembryonic antigen (pre-nCRT CEA) levels,post-treatment serum carcinoembryonic antigen (post-nCRT CEA) levels,tumor distance from the anal verge (tumor distance),time interval between chemoradiotherapy and surgery (time interval),magnetic resonance tumor regression grade (MR-TRG),ypCR,extramural vascular invasion (EMVI),circumferential resection margin (CRM),tumor size and residual mucosal lesion depth (lesion depth).Disease-free survival (DFS) and overall survival (OS) were assessed.

Neoadjuvant and surgical treatment

The IMRT regimen consisted of 22 fractions of 2.3 Gy(gross tumor volume,GTV) and 1.9 Gy (clinical target volume,CTV),which has been described in our previous report (10): the total dose of 50.6 Gy (GTV)/41.8 Gy(CTV) was administered 5 times per week over a period of 30 d.IMRT was administered using the Varian RapidArc system.The GTV was defined as the primary tumor,including the mesorectum.CTV was defined as the primary tumor,mesorectal region,presacral region,mesorectal lymph nodes,lateral lymph nodes,internal iliac lymph node chain or pelvic wall area.Capecitabine treatment was administered concurrently with IMRT at a dose of 825 mg/m2orally,twice per day,and surgery based on the principle of TME was recommended 6-8 weeks or more after the initiation of treatment. Adjuvant chemotherapy is routinely recommended.Capecitabine alone,mFOLFOX6 or CapeOx was prescribed at the discretion of the physician.

Image assessment

Preoperative staging was performed using either thoracic or abdominal computed tomography (CT) or abdominal and pelvic magnetic resonance imaging (MRI) and endorectal ultrasonography.Clinical tumor,node and metastasis (TNM) staging was performed according to the American Joint Committee on Cancer (AJCC) 8th staging system.The distance between the lower margin of the tumor and the anal verge on sagittal T2-weighted imaging(T2WI) was defined as the tumor distance.After nCRT,the pelvic MRI was reviewed again before surgery to determine the tumor response to treatment and the MRTRG,which were determined by two radiologists.

Measurement of MR-TRG

The criteria for MR-TRG were similar to those reported in previous literature,and the following were characteristics of each MR-TRG grade (11-13): MR-TRG 1,radiological complete response (absence of tumor signal and barely visible treatment-related scar);MR-TRG 2,good response (dense fibrosis;no obvious residual tumor,signifying minimal residual disease or no tumor);MRTRG 3,moderate response (＞50% fibrosis or mucin and visible intermediate signal);MR-TRG 4,slight response(small areas of fibrosis or mucin but mostly tumor);MRTRG 5,no response (intermediate signal intensity,same appearance as the original tumor or tumor regrowth).MRTRG 1-3 was defined as good responses,and MR-TRG 4-5 as poor responses.

Measurement of gross tumor specimens and pathologic outcomes

After radical surgery,gross tumor specimens and pathologic outcomes were evaluated by experienced pathologists.The maximum diameter of the gross tumor specimen was defined as the tumor size.The deepest portion of the lesion was cut perpendicular to the intestinal wall.The distance between the bottom of the lesion and the horizontal plane of the mucosa was measured and recorded as the lesion depth.The flat mucosal lesions were defined as tumors that regressed after treatment and remained as superficial ulcers or scars with a depth of less than or equal to 5 mm (Figure 2).ypCR was determined using paraffin pathology.ypCR was defined as postoperative pathology that confirmed no residual tumor cells in the primary tumor,mesangial adipose tissue,or regional lymph node tissue.The CRM positive was defined as the circumferential resection margin ≤ 1 mm under the microscope.The EMVI positive was defined by the presence of a tumor within the mesorectal vasculature beyond the muscularis propria.

Figure 2 Typical findings of residual flat mucosal lesions after neoadjuvant chemoradiotherapy on examination of gross resected specimens.Lesions with depths of 0 mm (A),1 mm (B),2 mm (C),3 mm (D),4 mm (E) and 5 mm (F) are shown.

Follow-up

Patients were regularly followed up every 3 months for the first 2 years and every 6 months thereafter for 3 years.After 5 years,follow-up visits were performed once a year until death or loss to follow-up. Follow-up examinations included serum tumor markers,thoracoabdominal/pelvic CT or MRI,and enteroscopy.DFS was defined as the time from radical surgery to the first recurrence (local or distant),final follow-up,or death (without recurrence or metastasis).OS was defined as the time from radical surgery to death from any cause or final follow-up.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics (Version 25.0;IBM Corp.,New York,USA) and R software (Version 4.0.4;R Foundation for Statistical Computing,Vienna,Austria).Categorical variables were analyzed using the Chi-square test (2×2) or Fisher’s exact test (2×C),where applicable.Multivariate analysis was performed using a binary logistic regression model(Forward: LR).The “rms” package were used to construct the nomogram prediction model.The bootstrap method was used for internal validation of the nomogram predictive model.The number of repeated samplings was 1,000.A concordance index (C-index) was used to evaluate the predictive performance of the model.Calibration of the nomogram for predicting ypCR rates was performed by comparing the predicted probability and actual status after bias correction.The “survival” and “survminer” packages were used for survival analysis,and “ggplot2” package was used for plotting.Survival was estimated using the Kaplan-Meier method,and the differences between curves were verified by the log-rank test.A multivariate Cox regression model was used to identify the primary predictors of survival.Statistical significance was set at P＜0.05.

Results

A total of 246 patients with residual flat mucosal lesions were included in the final analysis,and 56 (22.8%) had ypCR.The mean lesion depth was 0.31±0.19 cm.

Univariable analysis

Univariate analysis showed that age,sex,BMI,tumor distance,time interval,post-cT,pre-nCRT CEA level,tumor size,EMVI and CRM were not related to ypCR.There were differences in the rate of ypCR between the pre-cT groups (44.8%vs.19.8%,P=0.003),post-nCRT CEA groups (25.2%vs.3.6%,P=0.010),MR-TRG groups(29.7%vs.5.6%,P＜0.001) and lesion depth groups (51.3%vs.17.4%,P＜0.001) (Table 1).

Multivariate logistic regression analysis

Variables with P-values of ≤0.20 by univariate analysis were incorporated into a logistic regression model.The results showed that pre-cT ≤T2 [P=0.016,odds ratio (OR)=2.944,95% confidence interval (95% CI): 1.226-7.065],MRTRG 1-3 (P=0.001,OR=6.524,95% CI: 2.146-19.830)and residual mucosal lesion depth of 0 mm (P＜0.001,OR=4.708,95% CI: 2.159-10.266) were predictive factors for ypCR in LARC patients with residual flat mucosal lesions after nCRT.The multivariate logistic regression analysis results showed that patients with post-nCRT CEA≤5 ng/mL had an increased ypCR risk by 7.885 times,and the P-value (0.052) was close to the value of 0.05 (Table 2).

Table 1 Univariable analysis results of ypCR

Table 2 Multivariate logistic regression analysis results of pathologic complete response

Establishment and verification of the nomogram

A nomogram was constructed by incorporating the significant predictors identified using multivariate logistic regression (Figure 3).Each subtype within these variables was assigned a score on a point scale,which was then added to obtain the total score and probability of ypCR.The original concordance index (orig_c-index) was 0.769 and the bias-corrected concordance index (bias_corrected_c_index) was 0.759. The results showed that the discrimination ability of the nomogram model was satisfactory.In addition,the calibration curve demonstrated that the nomogram model had good predictive consistency,with a mean absolute error (MAE) of 0.008 and mean squared error (MSE) of 0.00011 (Figure 4).

Figure 3 Nomogram to predict the probability of achieving ypCR. Post-CEA,post-treatment serum carcinoembryonic antigen;pre-cT,pretreatment-cT stage;deep,lesion depth;MRTRG,magnetic resonance tumor regression grade; ypCR,pathologic complete response.

Figure 4 Calibration curve for predicting ypCR.The x-axis represents nomogram-predicted probability and the y-axis represents actual probability of ypCR.A perfect prediction would correspond to the 45° blue dashed line.Red dotted line represents the entire cohort (n=246),and black solid line is biascorrected by bootstrapping (B=1,000 repetitions),indicating observed performance.Mean absolute error=0.008.ypCR,pathologic complete response.

Survival analysis

The follow-up time ranged from 3.0 to 113.3 months,with a median follow-up of 63.77 months.Figures 5,6show DFS and OS analyses by Kaplan-Meier curves for the predictors of ypCR in the nomogram model,respectively.The median DFS and OS times of the pre-cT,lesion depth and MR-TRG groups were not reached,and there were no significant differences between groups with respect to either DFS or OS in the Kaplan-Meier analysis.However,there were significant differences between the post-nCRT CEA groups with respect to DFS (P＜0.001) and OS(P=0.028).The median DFS time for patients with postnCRT CEA＞5 ng/mL patients was 29.9 months,and the median DFS time for patients with post-nCRT CEA≤5 ng/mL patients was not reached.The median OS of the post-nCRT CEA group was not reached.The multivariate Cox regression model showed that the four predictors of ypCR in the nomogram model were not risk factors for DFS or OS. The results of the multivariate Cox proportional hazards regression models are shown inTables 3,4.

Table 3 Univariate and multivariate cox regression analysis of disease-free survival

Table 4 Univariate and multivariate cox regression analysis of overall survival

Figure 5 DFS analysis by Kaplan-Meier curves for pre-cT (P=0.129) (A),post-nCRT CEA (P＜0.001) (B),MR-TRG (P=0.258) (C) and lesion depth groups (P=0.800) (D).DFS,disease-free survival;pre-cT,pre-treatment-cT stage;post-nCRT CEA,post-treatment serum carcinoembryonic antigen;MR-TRG,magnetic resonance tumor regression grade.

Figure 6 OS analysis by Kaplan-Meier curves for pre-cT (P=0.164) (A),post-nCRT CEA (P=0.028) (B),MR-TRG (P=0.053) (C) and the lesion depth groups (P=0.528) (D).OS,overall survival;pre-cT,pre-treatment-cT stage;post-nCRT CEA,post-treatment serum carcinoembryonic antigen;MR-TRG,magnetic resonance tumor regression grade.

Discussion

Tumor response to nCRT is considered one of the important factors predicting prognosis and influencing treatment strategy decisions.This study was a large-sample study exploring the relationship between mucosal changes after nCRT and the occurrence of 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 80%.In the W&W era,our data may further support routine endoscopic re-evaluation before surgical planning to facilitate nonoperative management or organ preservation and share decisionmaking following patients’ preferences.

The appearance of residual mucosal lesions is important for assessing treatment response.Typical endoscopic signs of cCR after chemoradiation therapy include the disappearance of the tumor with the 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 (14-17).However,the reliability of these signs is not satisfactory,and the consistency between cCR and ypCR has not yet been clarified.In our study,the ypCR rate of patients with completely flat mucosa (lesion depth=0 mm) was as high as 51.3%,which was much higher than previous reports(3,8,18-20).The results showed that completely flat mucosa was associated with ypCR.This suggests the possibility of assessing the post-treatment tumor response by digital rectal exam and endoscopy,and if the residual lesions are found to be almost completely flat,these patients could be managed using the W&W strategy.However,other studies are worth noting.Fraseret al.(21)and Bhodayet al.(12) reported that most patients with ypT0 or ypCR status did not exhibit a complete mucosal clinical response.Therefore,we believe that it may be necessary to combine other parameters to improve the accuracy of the ypCR prediction.Consequently,we explored the synergistic effect of lesion depth,MR-TRG,pre-cT stage,and post-nCRT CEA level for predicting ypCR.The nomogram model revealed that the total number of points for the four variables was 318,and the ypCR risk reached 80%.

MR is one of the most accurate imaging modality to stage LARC and MR T-stage is independently associated with ypCR (22,23).In our study,pre-cT stage ≤T2 was an independent significant factor for predicting ypCR and may be considered a predictor for achieving cCR in future practice.The results also showed that post-cT stage was not related to ypCR because of its low sensitivity of MR for the assessment of ypT staging (24,25).Our results also showed that MR-TRG 1-3 could be effective indicators of ypCR,consolidating previous reports by Patelet al.(26)and Battersbyet al.(27).Previous studies have shown that serum CEA levels might predict treatment response and clinical outcome (28-34).In our study,the multivariate logistic regression analysis results showed that patients withpost-nCRT CEA ≤5 ng/mL had an increased ypCR risk by 7.885 times (P=0.052).However,there is no uniform CEA cut-off value to predict ypCR.Studies have reported similar findings with cut-off values of 2,2.5,5 ng/mL or 5.33 ng/mL (29,33,35-41).In our study,we set a CEA cut-off value of 5 ng/mL,which is more in line with routine diagnostic protocols.

Although the ypCR rate was as high as 51.3% in 39 patients with completely flat mucosa,the survival outcome remained unchanged.We believe that these results are due to patient selection bias and the low predictive accuracy of a single variable.Therefore,we built a nomogram ypCR prediction model to improve the accuracy of ypCR prediction,and the C-index suggested that the model had an excellent predictive ability.

Our study had some limitations.First,it was a retrospective study.Since all included patients had flat lesions,this selection bias might have excluded some patients with a poor tumor response,which might have affected the comparison of ypCR rates.If all patients were included in the analysis,the results would be more comparable.Second,lymph node metastasis was not considered.However,studies have shown that the cN stage is not associated with ypCR.In addition,clinical lymph node staging cannot confirm pathological lymph node staging.Last,evaluation and measurement of the lesions using fixed specimens may not be completely consistent with fresh specimens and without visual details when compared with endoscopic findings.

The present study consolidated the importance of mucosal changes after nCRT and reported the ypCR predictive value of the combination of pre-cT stage,residual mucosal lesion depth,MR-TRG and post-nCRT CEA levels.The nomogram we built reflected the significant factors influencing the ypCR rate for reference.In the W&W era,our findings are useful for highly selecting radiosensitive cases and supporting the diagnostic criteria for cCR,such as the MSKCC criteria (17).

Conclusions

Patients with residual flat mucosal lesions after nCRT for LARC should be treated cautiously when developing treatment strategies.Completely flat mucosa,early cT stage and good MR-TRG are predictive factors of ypCR in patients with LARC with residual flat mucosal lesions after nCRT.Our nomogram ypCR prediction models may aid in evaluating the treatment response of patients with residual flat mucosal lesions after nCRT,and guide the selection of clinical treatment measures. Endoscopic mucosal reevaluation before surgery is important and may contribute to decision making and facilitate nonoperative management or organ preservation.

Acknowledgements

This study was supported by grants from the National Natural Science Foundation of China (No.82173156) and Beijing Hospitals Authority Clinical Medicine Development of Special Funding Support (No.ZYLX202116).

Footnote

Conflicts of Interest: The authors have no conflicts of interest to declare.