Worse treatment response to neoadjuvant chemoradiotherapy in young patients with locally advanced rectal cancer

A retrospective study based on our prospectively maintained database was performed. LARC patients who underwent NCRT and radical resection between 2011 and 2018 were identified. The inclusion criteria were as follows: 1) clinical stage II or III (cT3/4 or cN1/2) disease; 2) pathologically proven rectal adenocarcinomas; and 3) tumors located within 12 cm from the anal verge. Exclusion criteria included: 1) previous of malignancies or concurrent with malignancies; 2) patients who underwent emergent surgery, palliative resection, or local excision. Finally, a total of 901 LARC patients were included. Since the majority of patients were pathologically confirmed by endoscopic biopsy when admitted to our hospital (a tertiary referral hospital), colonoscopy samples from only 169 patients were available for the immunohistochemical analysis. This study was approved by the Institutional Review Board (IRB) of Fujian Medical University Union Hospital. A patient flow diagram was shown in Fig 1.

Comprehensive assessments for tumor staging of patients were performed by a digital rectal examination, colonoscopy, chest radiography, abdominopelvic magnetic resonance imaging (MRI), and/or transrectal ultrasound (ERUS). Preoperative long-course radiotherapy consisted of a total dose of 45 Gy to the pelvis, delivered in 25 fractions for 5 consecutive weeks (180 cGy per fraction, 5 days a week), followed by a boost of 5.4 Gy to the primary tumor. Preoperative concurrent chemotherapy was initiated on the first day of radiotherapy by using 5-FU plus oxaliplatin (FOLFOX) or capecitabine plus oxaliplatin (CapeOX).

Surgical operation was performed at an interval of 6–8 weeks after the completion of radiation. Surgical resection for rectal cancer was performed according to the principle of total mesorectal excision (TME) and high ligation of the inferior mesenteric artery. The surgical procedure consisted of low anterior resection (LAR), abdominoperineal resection (APR), or Hartmann’s procedure. About 3–4 weeks after surgery, postoperative adjuvant chemotherapy was administered to patients (using FOLFOX or CapeOX) for 6 months.

Tumor response to NCRT was graded according to the tumor regression grade (TRG) system [17]. pCR was defined as the absence of viable tumor cells in either the primary tumor site or the resected lymph nodes. Postoperative morbidity was classified according to the Clavien-Dindo classification, grades I-II was considered as minor complications, and grades III-V as major complications. Tumor size was divided according to the quartile intervals to make it more applicable in clinical practice (≤1.8 cm, 1.9–3.1 cm, ≥3.2 cm). Perioperative mortality was defined as any death within 30 days of surgery or occurring in the hospital.

CD133 (Affinity Biosciences, AF5120, Polyclonal, 1:100) protein expression in specimens obtained before NCRT in 169 LARC patients was measured using the immunohistochemical streptavidin-biotin complex method (Fig. 3e, f, g, h) [18]. Phosphate-buffered saline (PBS) was used as the negative control and the image of the positive control from GE Healthcare Life Sciences. Immunoreactivity was scored by semi-quantitative analysis, and the fields were randomly selected in five directions (up, center, down, left, and right) under high magnification (× 400). The color was determined based on the intensity score as follows: 0 (no staining), 1 (light yellow), 2 (brown), and 3 (deep brown). The percentage of positive cells was scored as 0 (< 5%), 1 (5–25%), 2 (25–50%), 3 (50–75%), and 4 (> 75%). The mean value was calculated for each case with the aforementioned scoring methods and the final score was obtained by multiplying these two scores. All analyses were performed in a double-blinded manner.

Statistical analysis was performed using SPSS version 23.0 (SPSS INC., Chicago, USA) and R software packages, version 3.5.1 (The R Foundation for Statistical Computing, http://​www.​r-project.​org/​). The categorical variables were presented as frequencies and percentages and assessed using the Chi-squared (χ²) test or Fisher’s exact test. The continuous variables were described as means ± standard deviation and assessed using Student’s t-test. All significant variables in the univariate analysis were entered into a Logistic regression model to identify predictors of pCR. Based on the multivariable analysis, a predictive nomogram was developed using the R project. The performance of the nomogram was evaluated by calculating the Harrell’s concordance index (c-index). Decision curve analysis (DCA) was performed to evaluate the clinical utility of the model for pCR. DCA is a method for evaluation and comparison of the predictive value between different prediction models [19, 20]; therefore, this method was used to evaluate the clinical utility of the model for pCR. The x-axis of the DCA represents the percentage of threshold probability, and the y-axis represents the net benefit of the predictive model. The net benefit was calculated according to the following formula: Net benefit = (true positives/n) − (false positives/n) * (pt/(1 − pt). “Number high risk” indicated the number of patients classified as positive (high risk) by a model including age group according to various threshold probabilities. “Number high risk with the event” was the true positive patient number according to various threshold probabilities. The optimal cut-off values for CD133 expression were calculated and determined by using the X-tile program (http://​www.​tissuearray.​org/​rimmlab/​), a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization, which identified the cut-off with the minimum p values from log-rank χ2 statistics in terms of OS [21]. The optimal cut-off score was identified as 11 for CD133. Thus, we defined CD133 expression score ≤ 11 as low expression, and > 11 as high expression. Survival outcomes were assessed using the Kaplan–Meier method and log-rank test. P < 0.05 was considered to indicate statistical significance.

A total of 901 LARC patients were eligible for our analysis. Among them, 75 (8.3%) were assigned to the young group and 826 (91.7%) patients were assigned to the old groups. The median ages in the two groups were 34.3 and 58.1 years, respectively. Additionally, the old group was associated with a higher American Society of Anaesthesiology (ASA) grade (P < 0.05). As shown in Table 1, there were no significant differences between the two groups in terms of sex, the interval between NCRT and surgery, distance from the anal verge, clinical T stage, clinical N stage, pre-NCRT CA19–9 level, pre-NCRT CEA level, post-NCRT CA19–9 level, and post-NCRT CEA level (all P > 0.05),

No significant differences were observed between the two groups in terms of estimated blood loss, operation time, surgical approach, peri-NCRT complication rates, and sphincter-saving procedure (all P > 0.05, Table 2). There were no significant differences between the two groups in terms of postoperative hospital stay and postoperative complications (P = 0.124, P = 0.736, respectively). Similarly, the chemotherapy regimen did not differ between the two groups (P = 0.461). Compared to the young group, mucinous or signet ring cell carcinoma, (17.3% vs. 7.6%, P = 0.008) or poorly differentiated tumors (24.3% vs. 9.0%, P < 0.001) were more common in the old group. Moreover, the young group was associated with a higher TRG (P = 0.003), as well as a higher rate of perineural invasion (17.3% vs. 6.5%, P = 0.002). Pathological TNM stage and pathological type were similar in both groups (P = 0.957, P = 0.936). Similarly, there were no differences between the two groups in terms of vascular invasion and tumor size did (P = 0.346, P = 0.069). A positive circumferential resection margin (CRM) was observed in one patient (1.3%) in the young group, 10 patients (1.2%) in the old group, with no significant difference between the two groups (P = 1.000). Moreover, Kaplan-Meier curve analysis demonstrated that young patients were associated with poorer prognosis in LARC patients following NCRT. The 3-year OS rate in the old group (≥40 years) was significantly higher than that in the young group (< 40 years) (88.3% vs. 71.6%; P = 0.01, Fig. 3m). Moreover, the 3-year DFS rate for the old group (≥40 years) was higher than that in the young group (< 40 years) (83.8% vs. 68.6%; P = 0.204, Fig. 3l).

To explore the association between young age and treatment response to NCRT, we identified predictive factors for pCR by Logistic regression analysis. In the univariate analysis, tumor size (≤1.8 cm vs. 1.9–3.1 cm OR = 6.764, P < 0.001; vs. ≥3.2 cm OR = 2.022, P = 0.007), age (≥40 years vs. < 40 years, OR = 2.087, P = 0.044), pre-NCRT clinical T stage (OR = 0.731, P = 0.031), pre-NCRT clinical N stage (OR = 0.550, P = 0.016), distance from the anal verge (OR = 0.919, P = 0.018), and post-NCRT CEA (OR = 0.381, P < 0.001) were significantly associated with pCR in LARC patients. In the multivariate analysis, tumor size (≤1.8 cm vs. 1.9–3.1 cm OR = 6.764, P < 0.001; vs. ≥3.2 cm OR = 2.022, P = 0.007), age (≥40 years vs. < 40 years, OR = 2.382, P = 0.027), pre-NCRT clinical N stage (OR = 0.560, P = 0.029), and post-NCRT CEA (OR = 0.873, P = 0.001) were independent predictive factors for pCR in LARC patients (Table 3).

By incorporating the significant determinants in the Logistic regression analysis, we developed a predictive nomogram for pCR in LARC patients after NCRT as shown in Fig. 2a. The c-index of the nomogram was 0.70 (95% CI 0.70–0.71). The calibration curve (Fig. 2b) presented good statistical performance upon internal validation between predicted and actual pCR rates. DCA was used to evaluate the performance of the nomogram. As shown in Fig. 2c, the model including age group provided more predictive power than either the pCR scheme or the non-pCR scheme. The clinical impact curve (Fig. 2d) shows the prediction of risk stratification of 1000 patients using a resampling bootstrap method.

We further investigated the reason for the lower pCR rates in young LARC patients by examining CSC frequency (based on CD133 expression, Fig. 3e, f, g and h). A total of 169 LARC patients were eligible for immunohistochemical analysis. Since the CD133 expression scores were continuous variables, the X-tile program was utilized to identify the optimal cut-off points for determining the greatest actuarial survival difference. Using this method 11 was identified as the cut-off value for CD133 expression (Fig. 3a and b). Based on these cut-off points for CD133 expression, we divided the cohort into low (n = 127) and high (n = 42) subgroups in terms of OS and DFS.

Higher CD133 scores were associated with poorer prognosis in LARC patients following NCRT. The 3-year OS rate for the low CD133 group was significantly higher than that in the high CD133 group (95.1% vs. 62.9%; P < 0.01, Fig. 3d). Moreover, lower CD133 scores were correlated with improved DFS (Fig. 3c). The 3-year DFS rate for the low CD133 group was significantly higher than that in the high CD133 group (92.8% vs. 50.7%; P < 0.01). As shown in Fig. 3i, young patients had a higher CD133 expression score compared to that of old patients (P < 0.01). Moreover, correlation analysis demonstrated that CD133 expression was associated with patient age (P < 0.01, Fig. 3j). Moreover, we examined the CD133 expression in the pCR and non-pCR groups, the results demonstrated that the CD133 expression value in the pCR group was significantly lower than in the non-pCR group (P < 0.01, Fig. 3k). Taken together, these findings indicated that young LARC patients were associated with a higher CD133+ CSC burden, which might contribute to the lower pCR rates.