Effect of understaging on local recurrence of rectal cancer

Abstract Background and Objectives Magnetic resonance imaging of the pelvis has a limited accuracy to detect positive lymph nodes but does dictate neoadjuvant treatment in rectal cancer. This study aimed to investigate preoperative lymph node understaging and its effects on postoperative local recurrence rate. Methods Patients were selected from a retrospective cross‐sectional snapshot study. Patients with emergency surgery, cM1 disease, or unknown cN‐ or (y)pN category were excluded. Clinical and pathologic N‐categories were compared and the impact on local recurrence was determined by multivariable analysis. Results Out of 1548 included patients, 233 had preoperatively underestimated lymph node staging based on (y)pN category. Out of the 695 patients staged cN0, 168 (24%) had positive lymph nodes at pathology, and out of the 594 patients staged cN1, 65 (11%) were (y)pN2. Overall 3‐year local recurrence rate was 5%. Clinical N‐category was not associated with local recurrence when corrected for pT‐category, neoadjuvant therapy, and resection margin, neither in patients with (y)pN1 (hazard ratio [HR]: 1.67 (95% confidence interval [CI]: 0.68‐4.12) P = .263) nor (y)pN2‐category (HR: 1.91 95% CI: [0.75‐4.84], P = .175). Conclusion Preoperative understaging of nodal status in rectal cancer is not uncommon. No significant effect on local recurrence or overall survival rates were found in the present study.

determined by digital rectal examination. 10 MRI has made substantial contributions to optimize patient selection and to tailor treatment. 11 However, preoperative clinical staging using MRI has some restrictions, especially accuracy for suspected positive lymph nodes is limited. 9,12,13 Furthermore, since organ preserving surgery and nonsurgical management for rectal cancer are gaining prominence, adequate staging might become even more important in the future.
This study aimed to investigate the number of patients who underwent surgical resection for rectal cancer with a preoperative underestimated lymph node category and to analyze the effects of nodal understaging on local recurrence rates.

| METHODS
All data were obtained from the Dutch Snapshot Research Group which was described in detail previously. 14

| Staging
Staging included an abdominal computed tomography (CT)-scan, thoracic X-ray, and in most patients a pelvic MRI. The fifth edition TNM staging was applied, because this version was used during the study period in Dutch hospitals. 15 Based on this edition, patients were classified as N0 in the absence of lymph node metastasis, as N1 in case of up to three regional tumor positive lymph nodes, and N2 for four of more positive nodes. The national guideline-recommended considering lymph nodes positive when their size was equal to or greater than 5 mm on MRI images. Tumor nodules in the perirectal adipose tissue without lymphatic tissue were regarded as lymph node metastases when the diameter exceeded 3 mm at pathology. Positive non-regional lymph nodes were staged as M1 disease. The national guideline of 2008, which was still valid in 2011, recommended neoadjuvant radiotherapy for all cT2-4 tumors, with the exception of proximal cT2N0 tumors. Chemoradiotherapy could be considered for cT3/4 and cN2 categories, while short-course radiotherapy was recommended for the remaining patients.

| Variables
Low anterior resection was defined as a TME with the formation of an anastomosis, with or without diverting stoma. Abdominoperineal resection was defined as a rectal resection according to TME principles including the anal sphincter complex with a permanent colostomy. Low Hartmann's procedure was defined as a (low) anterior resection with closure of the rectal stump and the formation of an end colostomy. Any disease recurrence in the pelvis, at the anastomosis, or in the perineal wound was defined as local recurrence. Recurrence at other locations not present at the time of rectal resection was defined as distant recurrence and termed metastasis-free survival within the current study. Metastasis-free survival and local recurrence-free survival were defined as the time from surgery to recurrence or last follow-up. Overall survival was defined as the time from surgery to death or last follow-up.

| Statistical analysis
All continuous variables were displayed as median with interquartile range, with the exception of hazard ratios for which the 95% confidence interval was reported. Categorical variables were reported as numbers with percentages and differences were tested using χ 2 tests. Survival curves were generated using the Kaplan-Meier method, and differences between groups were tested using logrank tests. Multivariable analyses were performed using the Cox-proportional hazard method. For local recurrence and metastasis-free survival, the effect of clinical nodal staging was corrected for pT-category (dichromate variable (y)pT0-2 vs

| RESULTS
In total, 2095 underwent resection for rectal cancer in 2011 in The Netherlands and were included in the snapshot study cohort.
Eighty-seven patients underwent emergency surgery and were excluded. A further 134 patients were excluded due to M1 disease.
Finally, 326 patients were excluded due to either undetermined or missing cN category (n = 293) and/or pN category (n = 43). Baseline characteristics of the 1548 included patients are shown in Table 1.
Multivariable analysis revealed that for patients with (y)pN1 disease, short-course radiotherapy before surgery was associated with lower local recurrence rates (Table 3). Underestimated cN category was not predictive for local recurrence. For patients with a (y)pN2-category, multivariable analysis identified only advanced T-category to be associated with local recurrence, and short-course radiotherapy before surgery approached statistical significance (Table 3).

| DISCUSSION
In a cohort of 1548 patients who underwent surgical resection for rectal cancer in 2011 in The Netherlands, 233 patients (15%) were clinically understaged considering the actual postoperative lymph node category. Clinical lymph node understaging had no significant impact on local recurrence rates, did not impact metastasis-free survival, and was not associated with lower overall survival rates.
Several randomized trials have shown the prognostic significance of local lymph node metastases on local recurrence rates which was confirmed in the nationwide data. 16 The other with 75 articles found a slight but nonsignificant advantage for endoscopic ultrasound. 20 The reported sensitivity for the detection of nodal disease ranged from 57% to 85%, meaning positive nodes were overlooked in one out over every two to seven patients.
Size alone was insufficient to discriminate tumor positive nodes on MRI, contour and signal intensity have been shown to increase discrimination. [21][22][23] Up to 58% of positive nodes were less than 5 mm, negative nodes more than 10 mm were not uncommon and in patients with positive nodes concurrent reactive nodes were often of similar or even greater size. 21,22 Combining size with the morphological criteria increased the accuracy but to the limited extent mentioned above.
In The Netherlands, nodal status is one of the criteria for the selection of patients for the most appropriate neoadjuvant treatment regimen. Almost all randomized rectal cancer trials on neoadjuvant treatment did not incorporate routine MRI staging. Several studies have reported that node positive disease was a risk factor for local recurrence. 10,24 The Dutch TME trial showed that short-course radiotherapy reduced local recurrence at 2 years by over threefold at univariable analysis. 10 A similar Swedish trial also revealed a twofold reduction in local recurrence by preoperative short-course radiotherapy in node positive disease. 25 These results indicated that patients understaged as cN1 at preoperative MRI and who did not undergo short-course radiotherapy were at increased risk of local recurrence. Whether MRI staging can improve risk stratification beyond the clinical staging as used in these previous trial remains to be investigated in future trials. Furthermore, quality of TME surgery has significantly improved over time, which has diminished the absolute risk reduction that can be obtained by adding neoadjuvant radiotherapy. The overall rate of positive resection margins in the TME trial was 15%, while the current rate of incomplete resection is below 10% overall, and below 5% for early to intermediate risk rectal cancer without neoadjuvant radiotherapy. [26][27][28] Patients with a preoperative underestimated lymph node stage Unexpectedly positive lymph nodes suggests the use of adjuvant therapy. Adjuvant therapy after resection with positive circumferential margin seems to have no benefit. 34 Some studies do suggest a reduction in local recurrence rates with adjuvant therapy for high-risk patients who did not receive neoadjuvant treatment. [35][36][37] However this subject is controversial, and considering the low overall local recurrence rates would result in substantial overtreatment. Due to the probable high number of patients needed to treat, the benefits are unlikely to outweigh the negative side effects of adjuvant therapy.
As the criteria for positive lymph nodes on MRI images are not unequivocal, not only understaging, but also overstaging is a problem.
Although analyzing overstaging was not the aim of the analysis, Table 2 shows it does occur. Although the numbers are small, due to the neoadjuvant therapies that hamper analyses on overstaging, it is a problem discussed in literature. 38,39 Overstaging can lead to unnecessary neoadjuvant treatment that might not offer a benefit, but does do harm. Short-course radiotherapy is associated with more frequent fecal incontinence, sexual problems, and delayed wound healing. 40 were not available in the current dataset, while these might be able to explain some differential outcomes. 45 In conclusion, understaging of nodal status is not uncommon but not associated with higher rates of (local) recurrence or inferior overall survival. Although the overall outcomes of the cohort were comparable to large randomized trials, future studies should assess the effect of preoperative staging accuracy on outcomes with the current more restricted neoadjuvant therapy indications.