Abstract
Background
The circumferential resection margin (CRM) is highly prognostic for local recurrence in rectal cancer surgery without neoadjuvant treatment. However, its significance in the setting of long-course neoadjuvant chemoradiotherapy (nCRT) is not well defined.
Methods
Review of a single institution’s prospectively maintained database from 1998 to 2007 identified 563 patients with locally advanced rectal cancer (T3/T4 and/or N1) receiving nCRT, followed after 6 weeks by total mesorectal excision (TME). Kaplan-Meier, Cox regression, and competing risk analysis were performed.
Results
The authors noted that 75 % of all patients had stage III disease as determined by endorectal ultrasound (ERUS) and/or magnetic resonance imaging (MRI). With median follow-up of 39 months after resection, local and distant relapse were noted in 12 (2.1 %) and 98 (17.4 %) patients, respectively. On competing risk analysis, the optimal cutoff point of CRM was 1 mm for local recurrence and 2 mm for distant metastasis. Factors independently associated with local recurrence included CRM ≤1 mm, and high-grade tumor (p = 0.012 and 0.007, respectively). CRM ≤2 mm, as well as pathological, nodal, and overall tumor stage are also significant independent risk factors for distant metastasis (p = 0.025, 0.010, and <0.001, respectively).
Conclusion
In this dataset of locally advanced rectal cancer treated with nCRT followed by TME, CRM ≤1 mm is an independent risk factor for local recurrence and is considered a positive margin. CRM ≤2 mm was associated with distant recurrence, independent of pathological tumor and nodal stage.
At the time of rectal cancer surgery, a positive circumferential radial margin (CRM) indicates incomplete resection and is associated with high rates of local and distant recurrence.1–4 This has been well documented in studies that do not use neoadjuvant therapy. When delivered in 25–28 fractions over 5–6 weeks, chemoradiotherapy induces tumor downsizing and downstaging, which facilitates complete resection.5,6 The optimal radial margin has not been well defined. For instance, in a comprehensive review of the literature, Nagtegaal and Quirke report positive CRM is associated with local recurrence; however, the definition of CRM varied between studies.7 The aim of this study was to further define the relationship between CRM and local and distant recurrence in the setting of long-course neoadjuvant chemoradiotherapy.
MATERIALS AND METHODS
Study Population
Following approval of this study by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center (MSKCC), we performed a review of a prospectively maintained institutional database from 1998 to 2007. We identified 563 patients diagnosed with locally advanced rectal cancer (T3/4 or N1) by endorectal ultrasonography (ERUS) and/or magnetic resonance imaging (MRI). All of these patients received an infusional fluorouracil chemotherapy regimen concurrent with radiotherapy (50.4 Gy in 28 fractions), followed by a 6-week break. The break was followed by resection, performed according to the standard technique of total mesorectal excision. Demographic data, clinical characteristics, operative procedure, and histopathology were examined.
Pathologic Assessment
Specimens were inked and fixed. The entire tumor was sectioned perpendicularly to include the radial margin. The CRM was measured as the distance from the deepest part of the tumor in the inked soft tissue edge.
Pathologists reported the tumor response from neoadjuvant treatment by measuring the proportion of residual viable tumor mass replaced by the amount of fibrosis. Percent response was defined by the percent of lesion composed of fibrous or fibroinflammatory tissues in the primary lesion. The complete pathological response was defined as complete regression of the tumor, by means of no disease left in the specimen.
Statistical Analysis
The Pearson χ2-test for binary parameters was used to compare demographic data between groups. Analysis of CRM as a continuous variable was performed to identify the optimal cutoff point(s) associated with local recurrence and distant metastasis. Univariate analysis of recurrence-free survival was calculated using the Kaplan-Meier procedure-limit method and reported as median survival and estimated 5-year survival. Multivariate analysis was performed using the Cox proportional hazards regression method. A p value ≤0.05 was considered statistically significant.
RESULTS
A total of 563 patients with locally advanced rectal cancer were included in this study. Median age at diagnosis was 60 years (range, 17–88 years). The majority of patients were male (59 %). Also, 477 patients (85 %) underwent low anterior resection (LAR), and 85 (15 %) underwent abdominoperineal resection (APR). Based on ERUS and/or MRI, 75 % were clinical stage III and 25 %were clinical stage II. Of all patients, 20 %were found to have pathological complete response after multimodality treatment. The median follow-up was 39 months with an interquartile range from 27 to 55 months. There were 12 patients (2 %) who developed local recurrence and 98 (17 %) who developed distant metastasis within the follow-up period. The median interval between completion of neoadjuvant treatment and surgery was 48 days.
Optimal CRM Cutoff Point
In an attempt to identify the most prognostic cutoff point, CRM was analyzed as a continuous variable (Fig. 1). Unfortunately, no clear inflection point was noted for either local or distant recurrence. Competing risk analysis was then used to determine the association of CRM with local and distant recurrence. For local recurrence, creation of cohorts based on CRM ≤1 and > 1 mm (1-mm cutoff point) resulted in the greatest separation of recurrence curves, compared with 2 mm and 4 mm (Fig. 2a). A total of 20 patients had CRM ≤ 1 mm (Table 1). Patients undergoing APR showed a higher incidence of CRM ≤ 1 mm compared with those undergoing low anterior resection (LAR) (10.3 vs 3.6 %, p = 0.016). In contrast, when analyzing distant recurrence, a cutoff point of 2 mm created the greatest separation of survival curves (Fig. 2b).
FIG. 1.
Analysis of correlation between CRM as continuous variable and 5-year recurrence-free survival. a Local recurrence, b distant recurrence
FIG. 2.
Competing risk analysis of a local recurrence and 1-, 2-, and 4-mm CRM cutoff points and b of distant recurrence and 1-, 2-, and 4-mm CRM cutoff points
TABLE 1.
Demographic data of 563 patients with locally advanced rectal cancer
| CRM ≤1 mm (N = 20) |
CRM >1 mm (N = 409) |
P valuea | |
|---|---|---|---|
| Age (years) | 60.2 | 59.2 | 0.993 |
| Gender | 0.649 | ||
| Female | 7 | 164 | |
| Male | 13 | 245 | |
| Distance from anal verge | 0.667 | ||
| >5 cm | 6 | 145 | |
| ≤5 cm | 13 | 253 | |
| Operation | 0.016 | ||
| LAR | 13 | 348 | |
| APR | 7 | 61 | |
| Preoperative staging | 0.995 | ||
| Stage II | 4 | 80 | |
| Stage III | 13 | 261 | |
CRM circumferential resection margin, LAR low anterior resection, APR abdominoperineal resection
Pearson χ2 test for binary parameters
Local Recurrence
Univariate analyses demonstrated that low rectal cancer (≤5 cm from the anal verge; p < 0.001), advanced pathological staging (ypT3–4; p < 0.02), high-grade tumor (poorly differentiated to undifferentiated; p < 0.005), and CRM ≤ 1 mm (p <0.001) were associated with local recurrence (Table 2). The 5-year local recurrence-free survivals were 66 and 98 % in CRM ≤ 1 and >1 mm, respectively. Multivariate analysis confirmed that CRM and tumor grade are independent prognostic factors for local recurrence (Table 3).
TABLE 2.
Univariate analysis of clinicopathological prognostic factors associated with local and distant recurrence-free survival
| Variable | Local recurrence-free survival | Distant recurrence-free survival | |||||
|---|---|---|---|---|---|---|---|
| No.a | 5-Year survival (%) | P valueb | No.a | Median (months) | 5-Year survival (%) | P valueb | |
| Gender | 0.609 | 0.416 | |||||
| Female | 230 | 95 | 230 | 95 | 68 | ||
| Male | 333 | 94 | 333 | 113 | 77 | ||
| DTAV | <0.001 | 0.002 | |||||
| ≤5 cm | 153 | 90 | 153 | 70 | 59 | ||
| >5 cm | 398 | 97 | 398 | 110 | 79 | ||
| Tumor response | 0.289 | <0.001 | |||||
| Partial | 452 | 95 | 452 | 95 | 68 | ||
| Complete | 108 | 98 | 108 | NR | 96 | ||
| ypT | 0.020 | <0.001 | |||||
| ypT 0–2 | 304 | 97 | 304 | 113 | 92 | ||
| ypT 3–4 | 257 | 92 | 257 | 70 | 54 | ||
| ypN | 0.063 | <0.001 | |||||
| ypN 0 | 435 | 97 | 435 | NR | 81 | ||
| ypN 1–2 | 128 | 89 | 128 | 60 | 49 | ||
| Tumor grade | 0.005 | 0.003 | |||||
| 1–2 | 499 | 97 | 499 | 110 | 75 | ||
| 3–4 | 59 | 85 | 59 | NR | 54 | ||
| CRM | |||||||
| ≤1 mm | 20 | 66 | <0.001 | – | – | – | – |
| >1 mm | 409 | 98 | – | – | – | ||
| ≤2 mm | – | – | – | 45 | 59 | 41 | 0.001 |
| >2 mm | – | – | 384 | 110 | 78 | ||
| Distal margin | 0.153 | 0.417 | |||||
| ≤2 cm | 238 | 96 | 238 | 110 | 72 | ||
| >2 cm | 259 | 94 | 259 | 98 | 71 | ||
DTAV distance from anal verge, ypT pathological tumor staging, ypN pathological nodal staging, CRM circumferential resection margin, NR median recurrence-free survival not reached
Number of patients available for analysis
Survival estimates calculated by Kaplan–Meier product limit method; differences analyzed by log-rank test
TABLE 3.
Multivariate analysis of clinicopathological factors associated with local and distant recurrence
| Variable | HR | 95 % CI | P valuea |
|---|---|---|---|
| Local recurrence | |||
| CRM ≤1 mm | 7.39 | 1.55–35.24 | 0.012 |
| Grade 3–4 | 5.82 | 1.62–20.90 | 0.007 |
| ypT 3–4 | 2.05 | 0.37–11.25 | 0.407 |
| ypN 1–2 | 1.44 | 0.39–2.67 | 0.581 |
| Distant recurrence | |||
| CRM ≤2 mm | 1.80 | 1.08–3.03 | 0.025 |
| Grade 3–4 | 0.93 | 0.42–2.04 | 0.858 |
| ypT 3–4 | 1.91 | 1.17–3.11 | 0.010 |
| ypN 1–2 | 5.11 | 2.49–10.48 | <0.001 |
HR hazard ratio, 95 % CI 95 % confidence interval, CRM circumferential resection margin, ypT pathological tumor stage, ypN pathological nodal stage
Multivariate analysis performed by Cox proportional hazards regression method
Distant Metastasis
Univariate analysis demonstrated that low rectal cancer (p < 0.003), advanced pathological staging (p < 0.001), high-grade tumor biology (p < 0.004), partial tumor response (p < 0.001), pathological nodal (ypN) disease (p < 0.001), and CRM ≤2 mm (p < 0.001) were significantly associated with distant recurrence (Table 2). Multivariable analysis confirmed that CRM, as well as pathological tumor (ypT) and nodal (ypN) stage, were independent prognostic factors for distant metastasis (Table 3).
Correlation Between CRM and Other Clinicopathologic Factors
CRM was associated with distance of the lesion from the anal verge; lower tumors were more likely to have CRM ≤1 mm, consistent with the association between CRM and APR (Fig. 3a). We identified no association between CRM and primary tumor response to chemoradiotherapy (Fig. 3b).
FIG. 3.
Correlation analysis between CRM and pathological percentage of tumor response to neoadjuvant therapy
DISCUSSION
The strong association between CRM and local recurrence was clearly demonstrated in 1986 by Quirke et al. in a series of patients undergoing rectal cancer surgery without neoadjuvant therapy. Local recurrence developed in 85 % of all patients (11 of 13) with a positive CRM (≤1 mm), compared with 3 % of those with a clear margin (1 of 38).1 These results were reproduced by several investigators, again in the setting of no preoperative therapy.2,3,8,9 The optimal cutoff for defining positive CRM is still debated, however, and some investigators propose 2 mm rather than 1 mm.2,8 Nagtegaal et al. reported on a series of 656 rectal cancer patients undergoing rectal resection without neoadjuvant treatment. They found that patients with CRM ≤2 mm had a 16 % risk of local recurrence at 2 years, compared with 5.8 % risk for those patients with CRM > 2 mm (p < 0.001). The authors also noted that CRM ≤1 mm was associated with distant metastasis.2 However, most of the literature defines a positive CRM as tumor within 1 mm of the cut surface, and this definition is currently endorsed by the NCCN Guidelines.1–4,10,11
The prognostic significance of CRM status in the setting of neoadjuvant treatment is not as well defined, because the majority of studies excluded patients receiving neoadjuvant treatment. One might expect different results in patients treated neoadjuvantly, as long-course neoadjuvant chemoradiotherapy results in tumor downsizing and downstaging in a majority of cases.5,6 In a comprehensive review of the literature, Nagtegaal and Quirke7 noted that CRM was a more powerful predictor of local recurrence in patients treated with neoadjuvant therapy (hazard ratio [HR] = 6.3; 95 % confidence interval [95 % CI], 3.7–16.7 in studies using neoadjuvant therapy vs HR = 2.0; 95 % CI, 1.4–2.9 in studies not using neoadjuvant therapy).
The current study evaluates the significance of CRM in a homogeneous cohort of locally advanced rectal cancer patients receiving standard long-course neoadjuvant chemoradiotherapy, followed by total mesorectal excision. The results suggest that CRM remains an independent prognostic factor for local recurrence. The 5-year local recurrence-free survival is significantly different in patients with CRM ≤ 1 mm than in patients with CRM > 1 mm (66 vs 98 %, respectively, p < 0.001). This result supports the current NCCN Guidelines.
Furthermore, CRM is also an independent prognostic factor for distant recurrence. Of note, while a wider CRM was associated with fewer recurrences, 2 mm appears to create the greatest separation of survival curves, indicating the optimal cutoff point compared with 4- and 1-mm margins. The 5-year distant recurrence-free survival is 41 % in CRM ≤2 mm and 78 % in CRM >2 mm (p = 0.001). One could hypothesize that those tumors that do threaten the mesorectal margin and do not regress with neoadjuvant therapy represent aggressive tumor biology. Although one may expect that tumors with greater response to chemoradiation will demonstrate more regression from the mesorectal fascia, we did not identify a relationship between CRM and response to chemoradiation (percent tumor regression or tumor regression grade). This may result from the fact that not all stage II and III tumors encroach upon the mesorectal fascia at presentation. Multivariate analysis further demonstrated that CRM and pathological staging are independent prognostic factors, whereas the treatment response is not a significant factor. This result is consistent with previously published data from our institution concluding that postoperative pathological tumor and nodal staging are more important prognostic factors than treatment response.5 As demonstrated in our study, the association of CRM with distant recurrence is identified not only in the setting of neoadjuvant treatment but in the setting of no neoadjuvant treatment as well.2,9,12,13
The higher rate of positive CRM in patients undergoing APR versus LAR is again demonstrated in this study. This finding is consistent with the literature, which reports positive CRM in 15–30 %.2,8,9,14 Even with optimal cylindrical/extralevator resection, CRM positivity rates approach 15 %.10,15 The Polish study comparing short-course radiotherapy to long-course chemoradiotherapy noted a significant reduction in positive CRM in the long-course chemoradiation cohort.16 In this study, the rate of positive CRM was demonstrated in seven of 68 patients undergoing APR (10.3 %).
This study shares the limitations associated with any retrospective review. Although treatment protocols were standardized, there was some variability in the time from end of chemoradiation to surgery, which could affect CRM. Furthermore, the low rate of local recurrence can limit statistical analysis.
In conclusion, in the setting of neoadjuvant chemoradiotherapy, CRM ≤ 1 mm is considered to be a positive margin and remains a poor prognostic factor for local recurrence. CRM ≤ 2 mm is a prognostic factor for distant metastasis, independent of pathological tumor (T) and nodal (N) staging.
ACKNOWLEDGMENT
Dr. Trakarnsanga received a Traveling Fellowship from the Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.
Footnotes
This paper was presented at the 97th Annual Clinical Congress of the American College of Surgeons, October 23–27, 2011, San Francisco, CA.
CONFLICT OF INTEREST The author(s) declare that they have no competing interests
REFERENCE
- 1.Quirke P, Durdey P, Dixon MF, Williams NS. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection: Histopathological study of lateral tumor spread and surgical excision. Lancet. 1986;2:996–999. doi: 10.1016/s0140-6736(86)92612-7. [DOI] [PubMed] [Google Scholar]
- 2.Nagtegaal ID, Marijnen AM, Kranenbarg EK, van de Velde CJ, van Krieken JH. Circumferential margin involvement is still an important predictor of local recurrence in rectal carcinoma: not one millimeter but two millimeters is the limit. Am J Surg Pathol. 2002;26:350–357. doi: 10.1097/00000478-200203000-00009. [DOI] [PubMed] [Google Scholar]
- 3.Adam IJ, Mohamdee MO, Martin IG, Scott N, Finan PJ, Johnston D, et al. Role of circumferential margin involvement in the local recurrence of rectal cancer. Lancet. 1994;344:707–711. doi: 10.1016/s0140-6736(94)92206-3. [DOI] [PubMed] [Google Scholar]
- 4.de Haas-Kock DF, Baeten CG, Jager JJ, Langendijk JA, Schouten LJ, Volovics A, Arends JW. Prognostic significance of radial margins of clearance in rectal cancer. Br J Surg. 1996;83:781–785. doi: 10.1002/bjs.1800830617. [DOI] [PubMed] [Google Scholar]
- 5.Quah HM, Chou JF, Gonen M, Shia J, Schrag D, Saltz LB, et al. Pathological stage is most prognostic of disease-free survival in locally advanced rectal cancer patients after preoperative chemoradiation. Cancer. 2008;113:57–64. doi: 10.1002/cncr.23516. [DOI] [PubMed] [Google Scholar]
- 6.Weiser MR, Quah HM, Shia J, Guillem JG, Paty PB, Temple LK, et al. Sphincter preservation in low anterior cancer is facilitated by preoperative chemoradiation and intersphincteric dissection. Ann Surg. 2009;249:236–242. doi: 10.1097/SLA.0b013e318195e17c. [DOI] [PubMed] [Google Scholar]
- 7.Nagtegaal ID, Quirke P. What is the role of circumferential margin in the modern treatment of rectal cancer? J Clin Oncol. 2008;26:303–312. doi: 10.1200/JCO.2007.12.7027. [DOI] [PubMed] [Google Scholar]
- 8.Tilney HS, Rasheed S, Northover JM, Tekkis PP. Influence of circumferential resection margin on long-term outcomes following rectal cancer surgery. Dis Colon Rectum. 2009;52:1723–1729. doi: 10.1007/DCR.0b013e3181b54fbd. [DOI] [PubMed] [Google Scholar]
- 9.Shirouzu K, Akagi Y, Fujita S, Ueno H, Takii Y, Komori K, et al. The study group of the Japanese Society for Cancer of the Colon and Rectum (JSCCR) on clinical significance of the mesorectal excision of rectal cancer. Clinical significance of the mesorectal excision of rectal cancer: a Japanese multi-institutional study. Ann Surg. 2011;253:704–710. doi: 10.1097/SLA.0b013e3182119331. [DOI] [PubMed] [Google Scholar]
- 10.Martijnse IS, Dudink RL, West NP, Wasowicz D, Nieuwenhuijzen GA, van Lijnschoten I, et al. Focus on extralevator perineal dissection in supine position for low rectal cancer has led to better quality of surgery and oncological outcome. Ann Surg Oncol. 2011;19:786–793. doi: 10.1245/s10434-011-2004-9. [DOI] [PubMed] [Google Scholar]
- 11.National Comprehensive Cancer Network. The NCCN guideline TM version 1.2012 for rectal cancer. [Accessed 1 Feb 2012];2012 Available: http://www.nccn.org/professionals/physician_gls/pdf/rectal. [Google Scholar]
- 12.Bernstein TE, Endreseth BH, Romundstad P, Wibe A. Norwegian Colorectal Cancer Group. Circumferential resection margin as a prognostic factor in rectal cancer. Br J Surg. 2009;96:1348–1357. doi: 10.1002/bjs.6739. [DOI] [PubMed] [Google Scholar]
- 13.Hall NR, Finan PJ, Al-Jaberi T, Tsang CS, Brown SR, Dixon MF, et al. Circumferential margin involvement after mesorectal excision of rectal cancer with curative intent: Predictor of survival but not local recurrence? Dis Colon Rectum. 1998;41:979–983. doi: 10.1007/BF02237384. [DOI] [PubMed] [Google Scholar]
- 14.Birbeck KF, Macklin CP, Tiffin NJ, Parsons W, Dixon MF, Mapstone NP, et al. Rates of circumferential resection margin involvement vary between surgeons and predict outcomes in rectal cancer surgery. Ann Surg. 2002;235:449–457. doi: 10.1097/00000658-200204000-00001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.West NP, Finan PJ, Anderin C, Lindholm J, Holm T, Quirke P. Evidence of the oncologic superiority of cylindrical abdominal excision for low rectal cancer. J Clin Oncol. 2008;26:3517–3522. doi: 10.1200/JCO.2007.14.5961. [DOI] [PubMed] [Google Scholar]
- 16.Bujiko K, Nowacki MP, Nasierowska-Guttmejer A, Michalski W, Bebenek M, Pudełko M, et al. Sphincter preservation following preoperative radiotherapy for rectal cancer: report of a randomized trial comparing short-term radiotherapy vs. conventionally fractionated radiochemotherapy. Radiother Oncol. 2004;72:15–24. doi: 10.1016/j.radonc.2003.12.006. [DOI] [PubMed] [Google Scholar]