Abstract
Objective
To identify predictors of positive circumferential resection margin following rectal cancer resection in the United States.
Background
Positive circumferential resection margin is associated with a high rate of local recurrence and poor morbidity and mortality for rectal cancer patients. Prior study has shown poor compliance with national rectal cancer guidelines, but whether this finding is reflected in patient outcomes has yet to be shown.
Methods
Patients who underwent resection for stage I-III rectal cancer were identified from the 2010-2011 National Cancer Database. The primary outcome was a positive circumferential resection margin. The relationship between patient, hospital, tumor, and treatment-related characteristics was analyzed using bivariate and multivariate analysis.
Findings
A positive circumferential resection margin was noted in 2,859 (17.2%) of the 16,619 patients included. Facility location, clinical T and N stage, histologic type, tumor size, tumor grade, lymphovascular invasion, perineural invasion, type of operation, and operative approach were significant predictors of positive circumferential resection margin on multivariable analysis. Total proctectomy had nearly a 30% increased risk of positive margin compared to partial proctectomy (OR 1.293, 95%CI 1.185-1.411) and a laparoscopic approach had nearly 22% less risk of a positive circumferential resection margin compared to an open approach (OR 0.882, 95%CI 0.790-0.985).
Interpretation
Despite advances in surgical technique and multimodality therapy, rates of positive circumferential resection margin remain high in the United States. Several tumor and treatment characteristics were identified as independent risk factors, and advances in rectal cancer care are necessary to approach the outcomes seen in other countries.
INTRODUCTION
Recent improvements in both survival and quality of life for rectal cancer patients have been attributed to advances in surgical technique, neoadjuvant chemoradiation therapy (nCRT), preoperative staging techniques, and new adjuvant therapies over the past 30 years. Total mesorectal excision (TME), as first described by Heald and Ryall in 1982, has become the standard for surgical care of rectal cancer.1-3 Local recurrence rates decreased from as high as 45% using traditional techniques to less than 10% following TME alone, and to less than 6% following TME in conjunction with nCRT.3-6
At the same time, circumferential tumor spread was recognized as a necessary pathologic evaluation of rectal resection specimens. In 1986 Quirke et al. published their work examining the specimens of 52 rectal cancer excisions. With 14 specimens having positive lateral resection margins and 12 of the 14 going on to have a local recurrence, this early paper clearly emphasized the importance of obtaining a clear CRM.7 In a large meta-analysis comprised of over 17,000 patients, Nagtegaal and Quirke were able to show that a CRM ≤1mm was a strong predictor of local recurrence (HR 2.7, 95%CI 1.7-4.3), distant recurrence (HR 2.8, 95%CI 1.9-4.3), and survival (HR 1.7, 95%CI 1.3-2.3). Furthermore, positive CRM has an even greater association with local recurrence when nCRT is used (HR 6.3, 95%CI 3.7-16.7).8
The integral link between suboptimal surgery, a positive CRM, and poor oncological outcome has been recognized at a national healthcare policy level in many countries, leading to the establishment of quality assurance programs in several northern European countries during the 1990’s. These programs focused on improving outcomes by utilizing evidence based techniques such as a standardized TME technique, appropriate deployment of chemotherapy and radiotherapy and standardized pathological assessment of resected specimens. Many of these goals were achieved through the creation of centers of excellence (CoE), where participating centers had to meet a number of agreed standards for ongoing accreditation. These programs have been associated with an increase in the rate of TME’s performed, a decrease in local recurrence rates, and an increase in 5-year survival.9-12 One of the core quality measures of these programs is CRM status.
Large variation still exists in the quality of treatment received in the United States.13 The majority of rectal cancer patients still receive their care at low volume hospitals with little formal interaction among the providers involved through the medium of multidisciplinary tumor board meetings.14 There is an increasing recognition of the need to measure and track the national quality of cancer care, with the quality of TME and CRM rates previously shown to represent possible metrics in other healthcare settings.14-17 Although CRM status following proctectomy is an accepted quality standard internationally, little is known about the national rates of CRM positivity in current clinical practice in the United States. Therefore the aim of this study was to quantify current rates of positive CRM following rectal cancer resection and to identify predictive factors contributing to an increased risk of a positive margin.18,19
METHODS
Data Source
This study is a review of data from the National Cancer Data Base (NCDB) participant user file (PUF) for rectal cancer only and does not include cases from the rectosigmoid PUF. The NCDB is a hospital based, nationwide cancer registry program sponsored jointly by the American College of Surgeons, the Commission on Cancer (CoC), and the American Cancer Society. This comprehensive database receives more than one million case reports per year from over 1,500 hospitals, representing approximately 70% of all new invasive cancer diagnoses in the United States.20
Study Subjects
Beginning in 2010 the CoC required the reporting of CRM status from all accredited institutions, and at the time of this study the latest data available from the NCDB were cases from 2011. All patients who underwent partial or complete proctectomy (excluding cases of local excision) for histologically confirmed adenocarcinoma, mucinous adenocarcinoma, or signet ring cell carcinoma (identified using international Classification of Disease for Oncology codes) were included. Cases were limited to patients with pathologic stage I-III cancers of the rectum; cases without CRM status were excluded (Fig. 1).
Figure 1.
Flow Chart displaying the breakdown of cases included and excluded from the original 44,448 rectal cancer cases in the National Cancer Database from 2010-2011. Following inclusion and exclusion criteria, 16,619 cases were included in the study.
Statistical Analysis
Analysis was performed using patient, hospital, tumor, and treatment-related characteristics. Patient demographic factors examined included age at diagnosis, gender, race, primary payer, average household income, average education, population density of patient residence, and patient comorbidities. Race was defined as white, black, Hispanic or other. Income and education data were defined by the NCDB using national census data for the zip code of the patient’s residence. Education is reported as a percentage of residents without a high school diploma. Patient comorbidities are categorized using the Deyo classification of the Charlson Comorbidity Score. The NCDB does not provide a breakdown of individual comorbidities, nor does it provide information on body mass index or waist circumference.21
Hospital characteristics provided by the NCDB include hospital type and geographic location. The CoC provides facility accreditation based on facility volume of newly diagnosed cancer cases, availability of diagnostic and treatment services, participation in clinical research including clinical trials, and the training of resident physicians. Community Cancer Programs see between 100 and 500 newly diagnosed cancer cases per year, whereas Comprehensive Community Cancer Programs and Academic Comprehensive Cancer Programs see greater than 500 newly diagnosed cancer cases per year. Resident physician teaching is optional for Comprehensive Community Cancer Programs and Community Cancer Programs, whereas Academic Comprehensive Cancer Programs must provide postgraduate medical education in at least four program areas, including internal medicine and general surgery. Whether or not Comprehensive Community Cancer Programs and/or Community Cancer Programs provided resident physician education was not provided. Hospital location was identified by US census division of reporting facility and categorized by region (Table 1). Hospital volume was defined by the average annual number of rectal cancer resections for each institution over the two participating years. The average number of rectal cancer resections per year at each institution ranged from 1 case per year to 85.5 cases per year. Those centers with 1-10 cases per year were categorized as low volume, 11-30 cases as moderate volume, and 31-85.5 cases per year as high volume centers.
Table 1.
Hospital Location
| Region | States |
|---|---|
| New England | CT, MA, ME, NH, RI, VT |
| Middle Atlantic | NJ, NY, PA |
| South Atlantic | DC, DE, FL, GA, MD, NC, SC, VA, WV |
| East North Central | IL, IN, MI, OH, WI |
| East South Central | AL, KY, MS, TN |
| West North Central | IA, KS, MN, MO, ND, NE, SD |
| West South Central | AR, LA, OK, TX |
| Mountain | AZ, CO, ID, MT, NM, NV, UT, WY |
| Pacific | AK, CA, HI, OR, WA |
| Out of US | All other values |
Tumor-related factors included pathologic stage, clinical tumor and nodal stage, histological type, tumor size, grade, and presence of lymphovascular or perineural invasion. Treatment-related factors included the use of neoadjuvant chemoradiation, type of surgery for the primary site, and operative approach. Tumor size was categorized into 3 groups: <1cm, between 1 and 2cm, and greater than 2cm in concordance with prior papers published on rectal cancer margin status NCDB.22,23 The NCDB defines type of rectal surgery as a partial or total proctectomy. A partial proctectomy includes but is not limited to an anterior resection, Hartmann’s operation, low anterior resection (LAR), or trans sacral rectosigmoidectomy whereas total proctectomy includes abdominoperineal resections (APR), pelvic exenteration, and total proctocolectomy, NOS.
The primary outcome of this study was positive CRM. CRM was reported by the NCDB as a continuous variable by every 0.1mm to the margin. For the purpose of this analysis, a CRM ≤1mm was considered a positive CRM.24 Variations in positive CRM rates were determined by bivariate analysis using Pearson’s Chi-Square test for categorical variables and Student’s t-test for continuous variables. Factors associated with positive CRM with a p-value<0.05 were included in a multivariable logistic regression to identify those variables independently associated with a positive CRM. Out of concern that patient outcomes may not be independent data with clustering of outcomes at the hospital level, we performed a sensitivity analysis using a hierarchical model to account for unequal variance amongst hospitals. Further analysis was performed on the unadjusted relationship of nCRT with CRM status by stratifying for clinical T stage as well as type of operation. The unadjusted relationship between type of operation and CRM positivity was also further evaluated by stratification of overall clinical stage to identify the relationship of operation within these tumor subtypes. A two-sided p-value<0.05 was considered statistically significant. All analyses were performed using SAS/STAT® Software, Version 9.3 of the SAS system for Microsoft Windows (SAS Institute Inc., Cary, NC, USA).
RESULTS
Bivariate Analysis
During the two years of data collection 16,619 patients met study design criteria with 2,859 (17.2%) patients having a positive CRM. Patient and facility characteristics for the cohort can be found in Table 2, and tumor and treatment characteristics in table 3.
Table 2.
Patient and Facility Characteristics
| Patients, N=16,619 (%) | CRM Negative, N=13,760 (%) | CRM Positive, N=2,859 (%) | P-value | |
|---|---|---|---|---|
| Mean Patient Age (SD) | 62.5 (±13.0) | 62.5 (±12.9) | 62.8 (±13.4) | 0.158 |
| Sex | 0.939 | |||
| Male | 10005 (60.2) | 8282 (82.8) | 1723 (17.2) | |
| Female | 6614 (39.8) | 5478 (82.8) | 1136 (17.2) | |
| Race | 0.369 | |||
| White | 13492 (81.2) | 11207 (83.1) | 2285 (16.9) | |
| Black | 1338 (8.1) | 1088 (81.3) | 250 (18.7) | |
| Hispanic | 894 (5.4) | 736 (82.3) | 158 (17.7) | |
| Other | 782 (4.7) | 635 (81.2) | 147 (18.8) | |
| Unknown | 113 (0.7) | 94 (83.2) | 19 (16.8) | |
| Insurance | <0.001 | |||
| Uninsured | 695 (4.2) | 551 (79.28) | 144 (20.7) | |
| Government | 8013 (48.2) | 6545 (81.7) | 1468 (18.3) | |
| Private | 7703 (46.4) | 6489 (84.2) | 1214 (15.8) | |
| Unknown | 208 (1.3) | 175 (84.1) | 33 (15.9) | |
| Mean Household Income | 0.410 | |||
| <$30,000 | 2039 (12.3) | 1698 (83.3) | 341 (16.7) | |
| $30,000-$34,999 | 3043 (18.3) | 2485 (81.7) | 558 (18.3) | |
| $35,000-$45,999 | 4388 (26.4) | 3638 (82.9) | 750 (17.1) | |
| >$45,999 | 6165 (37.1) | 5114 (83.0) | 1051 (17.1) | |
| Unknown | 984 (5.9) | 825 (83.8) | 159 (16.2) | |
| Without High School Degree | 0.442 | |||
| >28.9% | 2627 (15.8) | 2158 (82.2) | 469 (17.9) | |
| 20.0-28.9% | 3733 (22.5) | 3080 (82.5) | 653 (17.5) | |
| 14.0-19.9% | 3867 (23.3) | 3232 (83.6) | 635 (16.4) | |
| <14.00% | 5404 (32.5) | 4461 (82.6) | 943 (17.5) | |
| Unknown | 988 (6.0) | 829 (83.9) | 159 (16.1) | |
| Charlson/Deyo Score | 0.930 | |||
| Score 0 | 12567 (75.6) | 10399 (82.8) | 2168 (17.3) | |
| Score 1 | 3175 (19.1) | 2636 (83.0) | 539 (17.0) | |
| Score 2 | 877 (5.3) | 725 (82.7) | 152 (17.3) | |
| Facility Type | 0.223 | |||
| Community Cancer Program | 1703 (10.3) | 1405 (82.5) | 298 (17.5) | |
| Comprehensive Community Program | 9532 (57.4) | 7871 (82.6) | 1661 (17.4) | |
| Academic | 5308 (31.9) | 4415 (83.2) | 893 (16.8) | |
| Other | 76 (0.5) | 69 (90.8) | 7 (9.2) | |
| Facility Location | <0.001 | |||
| New England | 933 (5.6) | 759 (81.4) | 174 (18.7) | |
| Middle Atlantic | 2278 (13.7) | 1971 (86.5) | 307 (13.5) | |
| South Atlantic | 3552 (21.4) | 2942 (82.8) | 610 (17.2) | |
| East North Central | 3076 (18.5) | 2503 (81.4) | 573 (18.6) | |
| East South Central | 989 (6.0) | 813 (82.2) | 176 (17.8) | |
| West North Central | 1542 (9.3) | 1325 (85.9) | 217 (14.1) | |
| West South Central | 1444 (8.7) | 1171 (81.1) | 273 (18.9) | |
| Mountain | 851 (5.1) | 691 (81.2) | 160 (18.8) | |
| Pacific | 1954 (11.8) | 1585 (81.1) | 369 (18.9) | |
| Center Volume | 0.972 | |||
| 1-10 cases | 7721 (46.5) | 6387 (82.7) | 1334 (17.3) | |
| 11-30 cases | 7046 (42.4) | 5839 (82.9) | 1207 (17.1) | |
| 31-85.5 cases | 1852 (11.1) | 1534 (82.8) | 318 (17.2) | |
| Distance Traveled | 0.558 | |||
| ≤30mi | 13166 (79.2) | 10879 (82.6) | 2287 (17.4) | |
| 30-60mi | 1862 (11.2) | 1555 (83.5) | 307 (16.5) | |
| 60-100mi | 830 (5.0) | 685 (82.5) | 145 (17.5) | |
| >100mi | 761 (4.6) | 641 (84.2) | 120 (15.8) | |
| Population Density | 0.244 | |||
| Metro | 12284 (73.9) | 10128 (82.5) | 2156 (17.6) | |
| Urban | 2833 (17.1) | 2374 (83.8) | 459 (16.2) | |
| Rural | 433 (2.6) | 360 (83.1) | 73 (16.9) | |
| Unknown | 1069 (6.4) | 898 (84.0) | 171 (16.0) |
SD, Standard Deviation
Table 3.
Tumor and Treatment Characteristics
| Patients, N=16,619 (%) | CRM Negative, N=13,760 (%) | CRM Positive, N=2,859 (%) | P-value | |
|---|---|---|---|---|
| Pathologic Stage | <0.001 | |||
| Stage I | 5403 (32.5) | 4864 (90.0) | 539 (10.0) | |
| Stage II | 5054 (30.4) | 4068 (80.5) | 986 (19.5) | |
| Stage III | 6162 (37.1) | 4828 (78.4) | 1334 (21.7) | |
| Clinical T Stage | <0.001 | |||
| cT0 | 26 (0.2) | 21 (80.8) | 5 (19.2) | |
| cTis | 91 (0.6) | 78 (85.7) | 13 (14.3) | |
| cT1 | 1683 (10.1) | 1517 (90.1) | 166 (9.9) | |
| cT2 | 2262 (13.6) | 1965 (86.9) | 297 (13.1) | |
| cT3 | 8044 (48.4) | 6673 (83.0) | 1371 (17.0) | |
| cT4 | 773 (4.7) | 501 (64.8) | 272 (35.2) | |
| Unknown | 3740 (22.5) | 3005 (80.4) | 735 (19.7) | |
| Clinical N Stage | <0.001 | |||
| N0 | 9717 (58.5) | 8198 (84.4) | 1519 (15.6) | |
| N1 | 4096 (24.7) | 3358 (82.0) | 738 (18.0) | |
| N2 | 759 (4.6) | 552 (72.7) | 207 (27.3 | |
| Unknown | 2047 (12.3) | 1652 (80.7) | 395 (19.3) | |
| Histological Type | <0.001 | |||
| Adenocarcinoma | 15569 (93.7) | 13012 (83.6) | 2557 (16.4) | |
| Mucinous Adenocarcinoma | 945 (5.7) | 686 (72.6) | 259 (27.4) | |
| Signet-Ring Cell Carcinoma | 105 (0.6) | 62 (59.1) | 43 (41.0) | |
| Tumor Size | <0.001 | |||
| ≤10mm | 1012 (6.1) | 892 (88.1) | 120 (11.9) | |
| 11-20mm | 2082 (12.5) | 1784 (85.7) | 298 (14.3) | |
| >20mm | 11336 (68.2) | 9148 (80.7) | 2188 (19.3) | |
| Unknown | 2189 (13.2) | 1936 (88.4) | 253 (11.6) | |
| Tumor Grade | <0.001 | |||
| Well or Moderately differentiated | 13043 (78.5) | 10905 (83.6) | 2138 (16.4) | |
| Poorly or Undifferentiated | 2100 (12.6) | 1594 (75.9) | 506 (24.1) | |
| Unknown | 1476 (8.9) | 1261 (85.4) | 215 (14.6) | |
| Lymphovascular Invasion | <0.001 | |||
| No | 11381 (68.5) | 9575 (84.1) | 1806 (15.9) | |
| Yes | 2791 (16.8) | 2050 (73.5) | 741 (26.6) | |
| Unknown | 2447 (14.7) | 2135 (87.3) | 312 (12.8) | |
| Perineural Invasion | <0.001 | |||
| No | 13099 (78.8) | 11058 (84.4) | 2041 (15.6) | |
| Yes | 1670 (10.1) | 1088 (65.2) | 582 (34.9) | |
| Unknown | 1850 (11.1) | 1614 (87.2) | 236 (12.8) | |
| Neoadjuvant Chemoradiation | 0.190 | |||
| No | 8034 (48.3) | 6663 (82.9) | 1371 (17.1) | |
| Yes | 8550 (51.5) | 7072 (82.7) | 1478 (17.3) | |
| Unknown | 35 (0.2) | 25 (71.4) | 10 (28.6) | |
| Operation | <0.001 | |||
| Partial Proctectomy | 10871 (65.4) | 9202 (84.7) | 1669 (13.4) | |
| Total Proctectomy | 5573 (33.5) | 4407 (79.1) | 1166 (20.9) | |
| Proctectomy, NOS | 175 (1.1) | 151 (86.3) | 24 (13.7) | |
| Surgical Approach | <0.001 | |||
| Open | 10430 (62.8) | 8544 (81.9) | 1886 (18.1) | |
| Laparoscopic | 3464 (20.8) | 2948 (85.1) | 516 (14.9) | |
| Robotic | 865 (5.2) | 718 (83.0) | 147 (17.0) | |
| Unknown | 1860 (11.2) | 1550 (83.3) | 310 (16.7) | |
| Proximal/Distal Margins | <0.001 | |||
| Negative | 15419 (92.8) | 13453 (87.2) | 1966 (12.8) | |
| Positive | 1129 (6.8) | 251 (22.2) | 878 (77.8) | |
| Unknown | 71 (0.4) | 56 (78.9) | 15 (21.1) |
NOS, Not Otherwise Specified
The rate of positive CRM had a statistically significant variation between the geographic regions in the United States from 13.5% in the Middle Atlantic to 18.9% in the West South Central and Pacific regions (p<0.001). In contrast, there was no statistically significant difference in the positive CRM rate among facility types (p=0.222) or the treatment center volume (p=0.972).
All tumor characteristics tested were significantly associated with positive CRM rates (Table 3). While T4 tumors had the highest rate of CRM positivity, they only contributed to 9.5% of positive CRM cases reported. Cases with T3 and T2 tumors still had a high rate of CRM positivity with 17.0% and 13.1% respectively.
The type of operation and the operative approach both had a statistically significant variation in positive CRM rates (p<0.001). The positive CRM rate was 20.9% for total proctectomies versus 13.4% for partial proctectomies. When performing a sub analysis of the relationship between operation type and CRM status by clinical stage, the type of operation did not have a statistically significant difference in CRM status for clinical stage I patients, however, there was a significant difference for clinical stage II and III cases (Fig. 2). The use of nCRT was not significantly related to positive CRM (p=0.190) for the entire cohort. The overall 30-day mortality was 1.4% but higher in those patients who had a positive CRM at 2.0% versus 1.3% for patients with a negative CRM (p=0.002).
Figure 2.
Comparison of positive circumferential resection margin (CRM) rates by clinical stage for each type of operation.
Comparing positive margin rates, 6.8% of the entire study group had a positive proximal/distal margin whereas the positive CRM rate was 17.2% (N=2,859). Furthermore, 77.8% of cases with a positive proximal/distal margin had a positive CRM whereas 68.8% of patients with a positive CRM had a negative proximal/distal margin. In other words, out of the 2,859 patients with a positive CRM, only 31% would have been recognized as a positive margin based on proximal/distal margin status alone.
Multivariable Analysis
Facility location, health insurance status, clinical T stage, clinical N stage, histologic type, tumor size, tumor grade, presence of lymphovascular invasion, perineural invasion, type of operation, and surgical approach were all included in the multivariable analysis for positive CRM (Table 4). Health insurance status was the only variable that did not have an independent association with a positive CRM when adjusting for the other factors in the analysis. Applying these same factors to a hierarchical model accounting for variance in CRM status amongst hospitals did not change the results seen in the multivariable logistic regression model.
Table 4.
Multivariable Analysis of Factors Associated With Positive Circumferential Resection Margin
| Variable | P-Value | Odds Ratio (95% Confidence Interval) |
|---|---|---|
| Facility Location | ||
| New England | Reference | |
| East North Central | 0.740 | 1.034 (0.850, 1.257) |
| East South Central | 0.891 | 0.983 (0.773, 1.251) |
| Middle Atlantic | <0.001 | 0.677 (0.548, 0.837) |
| Mountain | 0.557 | 1.077 (0.841, 1.379) |
| Pacific | 0.191 | 1.149 (0.933, 1.414) |
| South Atlantic | 0.588 | 0.948 (0.781, 1.151) |
| West North Central | 0.009 | 0.738 (0.589, 0.926) |
| West South Central | 0.742 | 1.038 (0.832, 1.294) |
| Insurance | ||
| Uninsured | Reference | |
| Government | 0.999 | 1.000 (0.817, 1.224) |
| Private | 0.168 | 0.867 (0.707, 1.062) |
| Unknown | 0.342 | 0.812 (0.528, 1.248) |
| Clinical T Stage | ||
| cT1 | Reference | |
| cT0 | 0.297 | 1.741 (0.614, 4.934) |
| cTis | 0.198 | 1.499 (0.809, 2.777) |
| cT2 | 0.066 | 1.213 (0.987, 1.491) |
| cT3 | <0.001 | 1.538 (1.285, 1.841) |
| cT4 | <0.001 | 3.510 (2.787, 4.420) |
| Unknown | <0.001 | 1.834 (1.509, 2.228) |
| Clinical N Stage | ||
| cN0 | Reference | |
| cN1 | 0.223 | 1.068 (0.961, 1.188) |
| cN2 | <0.001 | 1.445 (1.203, 1.734) |
| Unknown | 0.807 | 0.982 (0.846, 1.139) |
| Operation | ||
| Partial Proctectomy | Reference | |
| Total Proctectomy | <0.001 | 1.293 (1.185, 1.411) |
| Proctectomy, NOS | 0.566 | 0.878 (0.562, 1.370) |
| Surgical Approach | ||
| Open | Reference | |
| Laparoscopic | 0.026 | 0.882 (0.790, 0.985) |
| Robotic | 0.944 | 0.993 (0.821, 1.202) |
| Unknown | 0.452 | 0.948 (0.826, 1.089) |
| Histologic Type | ||
| Adenocarcinoma | Reference | |
| Mucinous Adenocarcinoma | <0.001 | 1.705 (1.459, 1.991) |
| Signet-Ring Cell Carcinoma | <0.001 | 2.254 (1.473, 3.451) |
| Tumor Size | ||
| <10mm | Reference | |
| 11-20mm | 0.343 | 1.119 (0.887, 1.410) |
| >20mm | 0.004 | 1.348 (1.100, 1.650) |
| Unknown | 0.330 | 0.888 (0.700, 1.127) |
| Tumor Grade | ||
| Well or Moderately Differentiated | Reference | |
| Poorly or Undifferentiated | <0.001 | 1.260 (1.119, 1.420) |
| Unknown Differentiation | 0.678 | 0.967 (0.824, 1.134) |
| Lymphovascular Invasion | ||
| No | Reference | |
| Yes | <0.001 | 1.375 (1.232, 1.534) |
| Unknown | 0.001 | 0.790 (0.688, 0.908) |
| Perineural Invasion | ||
| No | Reference | |
| Yes | <0.001 | 2.227 (1.971, 2.517) |
| Unknown | 0.050 | 0.857 (0.735, 1.000) |
NOS, Not Otherwise Specified
After adjusting for the patient, facility, tumor, and treatment factors, the type of operation and surgical approach remained statistically significantly associated with a positive CRM. Total proctectomy had nearly a 30% increased risk of a positive CRM compared to partial proctectomy (OR 1.293, 95%CI 1.185-1.411) and a laparoscopic approach had nearly 22% less risk of a positive CRM compared to an open approach (OR 0.882, 95%CI 0.790-0.985). When the use of nCRT was included in the multivariable model it had no independent association with a positive CRM (p=0.875) even after adjusting for clinical T stage, type of operation, and all other factors seen in table 4 (results not shown).
DISCUSSION
A large body of literature has established that CRM status plays a significant role in the long-term outcomes of patients with rectal cancer.8 TME is widely accepted as the standard surgical technique for achieving a clear CRM and for providing patients with the best opportunity for recurrence free survival. Nonetheless, using recent data from the NCDB, this study has shown that current rates of CRM positive resections within the United States are excessive with one in every six rectal resections having a positive margin. In comparison to the rate of 17% seen in this study, other large population based studies following the adoption of TME have seen rates of positive CRM between 8 and 13%.8,18,25,26 Multiple tumor-specific characteristics as well as treatment factors were independently predictive of a positive CRM in this large cohort of patients. These factors not only provide prognostic value but may also be recognized as high-risk features that warrant preoperative recognition and potential for alteration to the individual patient’s treatment plan.
For example, advanced clinical T and N stage, tumors >2cm, mucinous adenocarcinomas and signet ring cell carcinomas, high grade tumors, and lymphovascular and perineural invasion were identified as features independently associated with a positive CRM. This study is not the first to identify these factors as being associated with a positive margin but rather confirms these results in a large national database across thousands of resections and hundreds of hospitals.19,22,27,28 These features also remained significantly associated with positive CRM even after adjusting for the use of nCRT and type of surgical technique. These high-risk features, although not necessarily modifiable, are preoperatively identifiable. Location and depth of tumor invasion, nodal involvement, and tumor size should all be recognized with routine preoperative high-resolution pelvic MRI. In the MERCURY Study Group experience, high-resolution pelvic MRI has been proven to be a very successful staging tool for rectal cancer and for identifying patients at high risk for positive CRM.27 Additionally, preoperative MRI assessment of the CRM has been shown to predict short-term and long-term outcomes.29 Our findings suggest that current strategies for treatment are inadequate. Wider adoption of dedicated high-resolution rectal MRI may be one approach to help aid in the decision for adjunct therapies like chemoradiation, as well as planning for the most appropriate surgical approach to provide the greatest opportunity for a negative CRM.
While the type of surgery and surgical approach used were found to be independently associated with CRM status on multivariable analysis in our study, these results are subject to potential selection bias and confounders that are not controllable with the variables provided by NCDB. Furthermore, surgical specialty training is also a variable not available and therefor no analysis could be performed on the relationship between specialists and sphincter sparing operations or CRM status. Nonetheless, these results do emphasize the significance of proper surgical technique in achieving the ideal TME specimen and highlight the potential benefits from increasing rates of sphincter preserving operations and minimally invasive surgery as seen with colorectal surgeons in centers of excellence.9,12 A growing body of evidence suggests that there is a strong association between the quality of TME surgery and outcomes. Quirke et al. undertook a detailed pathological analysis of the mesorectal specimens in 1,156 patients enrolled in the CR07 randomized control trial. The overall CRM positivity rate was 11% and the plane of surgery was defined as good (mesorectal plane) in 52%, intermediate in 34% (intramesorectal), and poor (muscularis propria) in 13%. The CRM positivity rate was significantly associated with plane of surgery (mesorectal plane=9% vs. muscularis propria=19%, p<0.001). The quality of surgical plane correlated with subsequent 3-year local recurrence rates which were only 4% for patients with a good plane of dissection compared to 13% for patients with a poor plane of dissection (p=0.0039), in accordance with previous studies.30-32
In a study using NCDB data from 1998-2007, Russell et al. studied predictors of a positive margin following surgery for rectal and rectosigmoid tumors and developed a nomogram for a risk-adjusted pathologic margin positivity rate. In this original work they found that low outliers for a positive resection margin were more commonly high volume and academic centers.22 The group later went on to publish a paper in which they showed that the same low outliers for positive margin also had improved nodal evaluation, greater nCRT use, performed more sphincter preserving operations, had a lower 30-day mortality, and ultimately a greater 5-year survival compared to high outliers.23 This would suggest that auditing of outcome measures such as margin status could help identify high and low performing institutions. Circumferential margin status represents an excellent outcome measure as it can be accurately assessed pre-operatively on MRI, is associated with oncological outcomes, and accurately measured following surgical resection. MRI results can be incorporated into the deliberations on the need for neoadjuvant therapy and subsequent surgical decision-making. Subsequent pathological assessment of the CRM provides a valuable feedback loop to both the radiologist and surgeon. These steps are predicated on a standardized patient assessment pathway, which is facilitated by a multidisciplinary tumor board and audit, and forms the basis of the program of Centers of Excellence being proposed by the OSTRiCh Clinical Consortium.9
Since our study found no hospital level factors associated with CRM status and largely poor CRM rates across the nation, these results further support the role for surgical improvement and teaching programs proposed and encouraged by the OSTRiCh Clinical Consortium through the CoC.9 Multidisciplinary programs aimed at improving rectal cancer outcomes in large populations have been proven successful in the past. With the goal of reducing local recurrence rates and improving compliance with evidence-based treatment guidelines, in 2002 British Columbia held educational programs that were attended by approximately 80% of surgeons performing rectal cancer resections in that region. Following these sessions, the Province recognized a significant improvement in patient outcomes. The rate of negative CRM increased from 41.6% in 1996 to 86.9% (p<0.001) in 2003/04. Compliance with recommendations for preoperative radiation significantly improved and they experienced a significant decrease in 2-year pelvic recurrence among patients with stage III cancer from 18.2% to 9.2% (p=0.020).26,33
While our study does provide significant power through the large number of patients provided by the NCDB, we do acknowledge several limitations. Several important variables were not available for adjustment in this analysis, including tumor location, distance from the anal verge, surgeon specialty and operative volume, pathologic response to nCRT, diagnostic modality utilized, and information on the plane of surgery. Furthermore, the NCDB has limited data on operative details including a description as to what procedural codes comprise the two major categories for resection, partial and total proctectomy. Similar to other studies using NCDB, this study only includes data from CoC accredited hospitals and may misrepresent the true rates of positive CRM in this country and the relationship of hospital characteristics to non-CoC accredited institutions. However, it is reasonable to suggest that institutions achieving successful CoC accreditation are unlikely to have worse outcomes than non-accredited hospitals. Additionally, racial minorities appear to be under-represented in this cohort of patients as has been seen in other NCDB studies and warrant additional consideration.22 Lastly, the definition of “center” in the NCDB may include a heterogeneous group of institutions ranging from single site to multiple site facilities.
CONCLUSION
National rates of positive CRM following rectal cancer resection remain high, with one in six patients having an inadequate oncologic resection resulting in high risk for poor long-term outcomes. These rates are higher than those rates reported in contemporaneous International studies. Seen alongside data from the same source relating to apparent failure of guideline adherence, it strongly suggests that the United States is lagging significantly behind other countries in delivery of optimal care for patients with rectal cancer. Whereas many patient- and hospital-level characteristics were not significant predictors of positive CRM rates, multiple tumor-related and treatment-related factors were identified as high-risk factors for a positive CRM. By greater adherence to basic surgical tenets of rectal cancer care such as TME, improved pathologic assessment (grading of TME specimens, notation of CRM status), advanced imaging techniques, devotion to appropriate neoadjuvant therapy, and commitment to a multidisciplinary team approach to rectal cancer care a substantial reduction in CRM positivity and its attendant negative oncological outcomes can be delivered.
Acknowledgments
Grant Support: None
Consortium for Optimizing Surgical Treatment of Rectal Cancer (OSTRiCh) members
John R. T. Monson, University of Rochester Medical Center; D. W. Dietz, Cleveland Clinic; J. W. Fleshman, Baylor University Medical Center; F. L. Greene, University of North Carolina School of Medicine; S. D. Wexner, Cleveland Clinic Florida; F. H. Remzi, Cleveland Clinic; M. A. Abbas, Kaiser Permanente; G. J. Chang, MD Anderson Cancer Center; T. E. Read, Lahey Clinic; D. A. Rothenberger, University of Minnesota; J. Garcia-Aguilar, Memorial Sloan-Kettering Cancer Center; W. Peters, Columbia Surgical; R. D. Madoff, University of Minnesota; N. Baxter, St. Michael’s Hospital, Toronto, Ontario, Canada; D. Beck, Ochsner Clinic; A. Bennett, Cleveland Clinic; M. Berho, Cleveland Clinic Florida; G. Chang, MD Anderson Cancer Center; J. Efron, Johns Hopkins University; A. Fichera, University of Chicago Medical Center; J. Goldblum, Cleveland Clinic; J. G. Guillem, Memorial Sloan Kettering Cancer Center; M. Kalady, Cleveland Clinic; E. Kennedy, Mt. Sinai, Toronto, Ontario, Canada; L. Kosinski, Medical College of Wisconsin; D. Larson, Mayo Clinic; A. Lowry, University of Minnesota; K. A. Ludwig, Medical College of Wisconsin; H. M. MacRae, Mt. Sinai, Toronto, Ontario, Canada; J. Merlino, Cleveland Clinic; P. Roberts, Lahey Clinic; L. Rosen, Cleveland Clinic Florida; B. Safar, St. Louis, L. Stocchi, Cleveland Clinic; and M. Whiteford, Oregon Clinic.
Footnotes
Financial Disclosures:
Funding: No financial disclosures
References
- 1.Heald RJ, Husband EM, Ryall RD. The mesorectum in rectal cancer surgery--the clue to pelvic recurrence? Br J Surg. 1982;69:613–616. doi: 10.1002/bjs.1800691019. [DOI] [PubMed] [Google Scholar]
- 2.Heald RJ, Ryall RD. Recurrence and survival after total mesorectal excision for rectal cancer. Lancet. 1986;1:1479–1482. doi: 10.1016/s0140-6736(86)91510-2. [DOI] [PubMed] [Google Scholar]
- 3.MacFarlane JK, Ryall RD, Heald RJ. Mesorectal excision for rectal cancer. Lancet. 1993;341:457–460. doi: 10.1016/0140-6736(93)90207-w. [DOI] [PubMed] [Google Scholar]
- 4.Peeters KC, Marijnen CA, Nagtegaal ID, et al. The TME trial after a median follow-up of 6 years: increased local control but no survival benefit in irradiated patients with resectable rectal carcinoma. Ann Surg. 2007;246:693–701. doi: 10.1097/01.sla.0000257358.56863.ce. [DOI] [PubMed] [Google Scholar]
- 5.Martling AL, Holm T, Rutqvist LE, et al. Effect of a surgical training programme on outcome of rectal cancer in the County of Stockholm. Stockholm Colorectal Cancer Study Group, Basingstoke Bowel Cancer Research Project. Lancet. 2000;356:93–96. doi: 10.1016/s0140-6736(00)02469-7. [DOI] [PubMed] [Google Scholar]
- 6.Sebag-Montefiore D, Stephens RJ, Steele R, et al. Preoperative radiotherapy versus selective postoperative chemoradiotherapy in patients with rectal cancer (MRC CR07 and NCIC-CTG C016): a multicentre, randomised trial. Lancet. 2009;373:811–820. doi: 10.1016/S0140-6736(09)60484-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Quirke P, Durdey P, Dixon MF, et al. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour spread and surgical excision. Lancet. 1986;2:996–999. doi: 10.1016/s0140-6736(86)92612-7. [DOI] [PubMed] [Google Scholar]
- 8.Nagtegaal ID, Quirke P. What is the role for the 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]
- 9.Dietz DW Consortium for Optimizing Surgical Treatment of Rectal C. Multidisciplinary management of rectal cancer: the OSTRICH. J Gastrointest Surg. 2013;17:1863–1868. doi: 10.1007/s11605-013-2276-4. [DOI] [PubMed] [Google Scholar]
- 10.Archampong D, Borowski D, Wille-Jorgensen P, et al. Workload and surgeon’s specialty for outcome after colorectal cancer surgery. Cochrane Database Syst Rev. 2012;3 doi: 10.1002/14651858.CD005391.pub3. CD005391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kreiter E, Yasui Y, de Gara C, et al. Referral rate to oncologists and its variation by hospital for colorectal cancer patients. Ann Surg Oncol. 2012;19:714–721. doi: 10.1245/s10434-011-2063-y. [DOI] [PubMed] [Google Scholar]
- 12.Wexner SD, Rotholtz NA. Surgeon influenced variables in resectional rectal cancer surgery. Dis Colon Rectum. 2000;43:1606–1627. doi: 10.1007/BF02236751. [DOI] [PubMed] [Google Scholar]
- 13.Monson JR, Probst CP, Wexner SD, et al. Failure of evidence-based cancer care in the United States: the association between rectal cancer treatment, cancer center volume, and geography. Ann Surg. 2014;260:625–632. doi: 10.1097/SLA.0000000000000928. [DOI] [PubMed] [Google Scholar]
- 14.Delivering High-Quality Cancer Care: Charting a New Course For a System in Crisis. Insitute of Medicine; 2013. [PubMed] [Google Scholar]
- 15.Wibe A, Moller B, Norstein J, et al. A national strategic change in treatment policy for rectal cancer--implementation of total mesorectal excision as routine treatment in Norway. A national audit. Dis Colon Rectum. 2002;45:857–866. doi: 10.1007/s10350-004-6317-7. [DOI] [PubMed] [Google Scholar]
- 16.Dahlberg M, Glimelius B, Pahlman L. Changing strategy for rectal cancer is associated with improved outcome. Br J Surg. 1999;86:379–384. doi: 10.1046/j.1365-2168.1999.01040.x. [DOI] [PubMed] [Google Scholar]
- 17.Khani MH, Smedh K. Centralization of rectal cancer surgery improves long-term survival. Colorectal Dis. 2010;12:874–879. doi: 10.1111/j.1463-1318.2009.02098.x. [DOI] [PubMed] [Google Scholar]
- 18.Tekkis PP, Heriot AG, Smith J, et al. Comparison of circumferential margin involvement between restorative and nonrestorative resections for rectal cancer. Colorectal Dis. 2005;7:369–374. doi: 10.1111/j.1463-1318.2005.00767.x. [DOI] [PubMed] [Google Scholar]
- 19.Rullier A, Gourgou-Bourgade S, Jarlier M, et al. Predictive factors of positive circumferential resection margin after radiochemotherapy for rectal cancer: the French randomised trial ACCORD12/0405 PRODIGE 2. Eur J Cancer. 2013;49:82–89. doi: 10.1016/j.ejca.2012.06.028. [DOI] [PubMed] [Google Scholar]
- 20.Bilimoria KY, Stewart AK, Winchester DP, et al. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol. 2008;15:683–690. doi: 10.1245/s10434-007-9747-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 22.Russell MC, You YN, Hu CY, et al. A novel risk-adjusted nomogram for rectal cancer surgery outcomes. JAMA Surg. 2013;148:769–777. doi: 10.1001/jamasurg.2013.2136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Massarweh NN, Hu CY, You YN, et al. Risk-Adjusted Pathologic Margin Positivity Rate As a Quality Indicator in Rectal Cancer Surgery. J Clin Oncol. 2014 doi: 10.1200/JCO.2014.55.5334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Washington K, Tang LH, Berlin J, et al. Protocol for the examination of specimens from patients with primary carcinoma of the colon and rectum. College of American Pathologists. 2011 [Google Scholar]
- 25.Wibe A, Syse A, Andersen E, et al. Oncological outcomes after total mesorectal excision for cure for cancer of the lower rectum: anterior vs. abdominoperineal resection. Dis Colon Rectum. 2004;47:48–58. doi: 10.1007/s10350-003-0012-y. [DOI] [PubMed] [Google Scholar]
- 26.Phang PT, McGahan CE, McGregor G, et al. Effects of change in rectal cancer management on outcomes in British Columbia. Can J Surg. 2010;53:225–231. [PMC free article] [PubMed] [Google Scholar]
- 27.Oberholzer K, Menig M, Kreft A, et al. Rectal cancer: mucinous carcinoma on magnetic resonance imaging indicates poor response to neoadjuvant chemoradiation. Int J Radiat Oncol Biol Phys. 2012;82:842–848. doi: 10.1016/j.ijrobp.2010.08.057. [DOI] [PubMed] [Google Scholar]
- 28.Hiranyakas A, da Silva G, Wexner SD, et al. Factors influencing circumferential resection margin in rectal cancer. Colorectal Dis. 2013;15:298–303. doi: 10.1111/j.1463-1318.2012.03179.x. [DOI] [PubMed] [Google Scholar]
- 29.Taylor FG, Quirke P, Heald RJ, et al. Preoperative magnetic resonance imaging assessment of circumferential resection margin predicts disease-free survival and local recurrence: 5-year follow-up results of the MERCURY study. J Clin Oncol. 2014;32:34–43. doi: 10.1200/JCO.2012.45.3258. [DOI] [PubMed] [Google Scholar]
- 30.Quirke P, Steele R, Monson J, et al. Effect of the plane of surgery achieved on local recurrence in patients with operable rectal cancer: a prospective study using data from the MRC CR07 and NCIC-CTG CO16 randomised clinical trial. Lancet. 2009;373:821–828. doi: 10.1016/S0140-6736(09)60485-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Maslekar S, Sharma A, Macdonald A, et al. Mesorectal grades predict recurrences after curative resection for rectal cancer. Dis Colon Rectum. 2007;50:168–175. doi: 10.1007/s10350-006-0756-2. [DOI] [PubMed] [Google Scholar]
- 32.Strassburg J, Ruppert R, Ptok H, et al. MRI-based indications for neoadjuvant radiochemotherapy in rectal carcinoma: interim results of a prospective multicenter observational study. Ann Surg Oncol. 2011;18:2790–2799. doi: 10.1245/s10434-011-1704-5. [DOI] [PubMed] [Google Scholar]
- 33.Phang PT. Evolving rectal cancer management in British Columbia. Can J Surg. 2010;53:222–224. [PMC free article] [PubMed] [Google Scholar]