eFigure. Identification of Patients With Resected Colon Cancer in Ontario During 2002-2008
eTable 1. Patient-Level Factors Associated With Right-Sided Colon Cancer Among Patients With Early-Stage Disease Resected in Ontario During 2002-2008 (n = 6365)
eTable 2. Factors Associated With Right-Sided Colon Cancer Among Patients With Stage III Disease Resected in Ontario During 2002-2008 (n = 2756)
eTable 3. Factors Associated With Survival Among Patients With Resected Colon Cancer Stage I in Ontario During 2002-2008 (n = 1163)
eTable 4. Factors Associated With Survival Among Patients With Resected Colon Cancer Stage II in Ontario During 2002-2008 (n = 2446)
eTable 5. Factors Associated With Survival Among Patients With Resected Colon Cancer Stage III in Ontario During 2002-2008 (n = 2756)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Karim S, Brennan K, Nanji S, Berry SR, Booth CM. Association Between Prognosis and Tumor Laterality in Early-Stage Colon Cancer. JAMA Oncol. 2017;3(10):1386–1392. doi:10.1001/jamaoncol.2017.1016
Is there an association between tumor laterality and prognosis in early-stage colon cancer?
In this population-based cohort study, there was no association between tumor sidedness and overall survival or cancer-specific survival among patients with stage I to III colon cancer.
The observation that laterality is associated with survival in advanced colorectal cancer may not extend to patients with early-stage disease.
Recent data have suggested that disease biology and outcome of colon cancer may differ between right-sided and left-sided tumors. However, the literature on the prognostic value of tumor laterality is conflicting.
To explore differences in laterality based on disease characteristics and outcomes in a population-based cohort of early-stage colon cancer.
Design, Setting, and Participants
This investigation was a population-based retrospective cohort study of patients with early-stage colon cancer from the province of Ontario, Canada. Electronic records of treatment were linked to the Ontario Cancer Registry to identify all patients with colon cancer who underwent resection between January 1, 2002, and December 31, 2008. The date of the final analysis was October 20, 2016. The study population included a 25% random sample of all patients with resected stage I to III disease. Right-sided colon cancer was defined as any tumor arising in the cecum, ascending colon, hepatic flexure, or transverse colon. Left-sided colon cancer was defined as any tumor arising in the splenic flexure, descending colon, sigmoid colon, or rectosigmoid colon.
Main Outcomes and Measures
Overall survival (OS) and cancer-specific survival (CSS) measured from the time of resection.
This study identified 6365 patients with early-stage colon cancer (48.7% [3098 of 6365] female). Their median age was 72 years, and 51.7% (3291 of 6365) had right-sided disease. Stage distribution was 18.3% (1163 of 6365) stage I, 38.4% (2446 of 6365) stage II, and 43.3% (2756 of 6365) stage III. Patients with right-sided colon cancer were more likely to be older (median age, 73 vs 70 years; P < .001) and female (54.4% [1790 of 3291] vs 42.6% [1308 of 3074], P < .001) and have greater comorbidity. Right-sided cancers were more likely to be T4 (19.2% [631 of 3291] vs 15.9% [490 of 3074], P < .001) and poorly differentiated (21.1% [695 of 3291] vs 9.6% [295 of 3074], P < .001) but less likely to be node positive (42.0% [1383 of 3291] vs 44.7% [1373 of 3074], P = .03) compared with left-sided disease. In adjusted analyses, there was no difference in long-term survival for right-sided compared with left-sided colon cancer: the hazard ratios were 1.00 (95% CI, 0.92-1.08) for OS and 1.00 (95% CI, 0.91-1.10) for CSS. These results were consistent when the survival analyses were restricted to stage III disease: the hazard ratios were 1.03 (95% CI, 0.93-1.14) for OS and 1.10 (95% CI, 0.97-1.24) for CSS.
Conclusions and Relevance
In this population-based cohort of early-stage resected colon cancer, disease laterality was not associated with long-term OS or CSS.
Colorectal cancer (CRC) is the second leading cause of cancer death in the Western world.1 In recent years, a number of prognostic factors have been identified to guide therapeutic decision making. The extent to which the location of the primary tumor influences disease outcome is controversial. The hypothesis that proximal and distal CRCs represent distinct entities originated in the late 1980s.2-4 Anatomically, the right and left colon arise from different embryonic origins; the proximal colon arises from the midgut and receives its main blood supply via the superior mesenteric artery, whereas the distal colon arises from the hindgut and is supplied by the inferior mesenteric artery. A variety of histological5,6 and genetic7-11 differences between the right and left colon have been described. Clinically, studies12-14 have shown that patients with right-sided colon cancer are older and more likely female, are initially seen with more advanced tumor stage, and are more likely to have poorly differentiated tumors.
Several population-based studies15-19 have explored the prognostic relevance of laterality in CRC, with conflicting results. A recent meta-analysis20 of 15 studies reported inferior survival for patients with right-sided colon cancer (hazard ratio [HR], 1.14). Several presentations at the 2016 American Society of Oncology annual meeting and the 2016 European Society of Medical Oncology annual meeting also described inferior survival among patients with right-sided metastatic CRC,18 especially in those with RAS wild-type tumors.21-24
These apparently contradictory findings have renewed our interest in understanding the correlation between primary tumor location and prognosis in both nonmetastatic and metastatic CRC. We undertook a population-based study of patients with resected early-stage colon cancer to further evaluate the prognostic value of disease laterality.
This population-based retrospective cohort study explored the association between laterality and outcome of patients with resected colon cancer in the Canadian province of Ontario. Ontario has a population of approximately 13.5 million people and has a single-payer universal health insurance program. We used the Ontario Cancer Registry (OCR) to identify all incident cases of colon cancer in Ontario diagnosed between 2000 and 2008. To reduce the possibility of locally recurrent disease, we restricted the cohort to patients who underwent primary tumor resection within 6 months of diagnosis. The final study population included patients who underwent resection of stage I, II, or III colon cancer in Ontario between January 1, 2002, and December 31, 2008. The date of the final analysis was October 20, 2016. The OCR does not capture stage of disease for all patients; therefore, we obtained surgical pathology reports for a random sample of 25% of cases. Reports were not available for patients with surgery in 2005; accordingly, the study cohort is restricted to patients who had surgery in 2002 to 2004 and 2006 to 2008. Patients with rectal cancer and nonadenocarcinoma histology were excluded, as were patients without an explicit tumor location or lymph node count recorded in the pathology report. The study was approved by the Research Ethics Board of Queen’s University. The population-level data sets used in this analysis contain anonymized patient-level data. Informed consent was waived by the Research Ethics Board.
The OCR is a passive, population-based cancer registry that captures diagnostic and demographic information on at least 98% of all incident cases of cancer in the province of Ontario.25 The OCR also provides information about vital status and cause of death. Records of hospitalization from the Canadian Institute for Health Information contain information about surgical interventions; these records are known to be complete.26 Provincial physician billing records from the Ontario Health Insurance Plan, treatment records from regional cancer centers, and provincial records of chemotherapy delivery were used to identify chemotherapy use. Surgical pathology reports were obtained from the OCR for a random sample of 25% of cases. A team of trained data abstractors manually reviewed these pathology reports and entered information about laterality, extent of disease, and surgical procedure into an electronic database. Data sets were linked using unique encoded identifiers and analyzed at the Institute for Clinical Evaluative Sciences.
Indicators of the socioeconomic status of the community in which patients resided at diagnosis were linked as described previously.27 Quintiles of the median household income were based on the household income distribution for the entire province of Ontario. Quintile 1 represents the communities where the poorest 20% of the Ontario population resided. Comorbidity was classified using the Charlson Comorbidity Index modified for administrative data.28 Adjuvant chemotherapy was defined as chemotherapy given within 16 weeks after resection of the primary tumor. Right-sided colon cancer was defined as any tumor arising in the cecum, ascending colon, hepatic flexure, or transverse colon. Left-sided colon cancer was defined as any tumor arising in the splenic flexure, descending colon, sigmoid colon, or rectosigmoid colon. Overall survival (OS) and cancer-specific survival (CSS) were measured from the time of resection of the primary CRC. To account for possible cause of death miscoding, CSS included death from any cancer. Complete information about vital status in the OCR was available up to December 31, 2012; cause of death was available up to December 31, 2010.
Comparisons of proportions between study groups were made using the χ2 test. The OS and CSS were determined using the Kaplan-Meier method. Factors associated with laterality were evaluated using log binomial regression. Factors associated with OS or CSS were evaluated using the Cox proportional hazards regression model. Results were considered statistically significant at P < .05. All analyses were performed using statistical software (SAS, version 9.4; SAS Institute Inc).
Linked administrative databases identified 25 613 potentially eligible patients who underwent surgery between 2002 and 2008. Surgical pathology reports were reviewed for 7519 cases; 270 (3.6%) of these cases were excluded due to nonadenocarcinoma histology (n = 56), rectal cancer (n = 134), inconsistent dates between diagnosis and surgery (n = 36), and delivery of preoperative chemotherapy (n = 44); also excluded were patients with stage IV (n = 443) or stage 0 (n = 26) disease (eFigure in the Supplement). Among the remaining 6780 patients with stage I, II, or III colon cancer, 83 (1.2%) were excluded due to unknown lymph node count and 332 (4.9%) due to missing or bilateral laterality classification. The final study population consisted of 6365 patients.
Characteristics of the study population are listed in Table 1. For the entire cohort, the median age was 72 years, and 51.7% (3291 of 6365) had right-sided disease. Stage distribution was 18.3% (1163 of 6365) stage I, 38.4% (2446 of 6365) stage II, and 43.3% (2756 of 6365) stage III. Patients with right-sided cancer were older (median age, 73 vs 70 years; P < .001), more likely to be female (54.4% [1790 of 3291] vs 42.6% [1308 of 3074], P < .001), and had greater comorbidity. These patient-level differences persisted in adjusted analyses (eTable 1 in the Supplement) and were consistent when analyses were restricted to stage III disease (eTable 2 in the Supplement). Right-sided cancers were more likely to be T4 tumors (19.2% [631 of 329] vs 15.9% [490 of 3074], P < .001) and were more likely to have poorly differentiated histology (21.1% [695 of 3291] vs 9.6% [295 of 3074], P < .001) compared with left-sided cancers. Right-sided disease was less likely to be node positive (42.0% [1383 of 3291] vs 44.7% [1373 of 3074], P = .03).
Factors associated with long-term survival are listed in Table 2. In adjusted analyses, laterality was not associated with differences in long-term survival: the HRs were 1.00 (95% CI, 0.92-1.08) for OS and 1.00 (95% CI, 0.91-1.10) for CSS. Stage-specific Cox proportional hazards regression model analyses (Table 3 and eTables 3, 4, and 5 in the Supplement) did not show any association between laterality and OS or CSS in any stage of disease. For stage I, the HRs were 1.10 (95% CI, 0.88-1.39) for OS and 0.73 (95% CI, 0.48-1.09) for CSS. For stage II, the HRs were 0.89 (95% CI, 0.78-1.02) for OS and 0.83 (95% CI, 0.69-1.00) for CSS. For stage III, the HRs were 1.03 (95% CI, 0.93-1.14) for OS and 1.10 (95% CI, 0.97-1.24) for CSS.
We undertook exploratory analyses within the stage III subset to identify prognostic factors that may differ between right-sided and left-sided disease. There was no evidence of interaction between laterality and benefit from adjuvant chemotherapy by test for interaction: the HRs for OS were 0.43 (95% CI, 0.37-0.50) for left-sided tumors and 0.46 (95% CI, 0.40-0.54) for right-sided tumors (P = .52), while the HRs for CSS were 0.41 (95% CI, 0.34-0.49) for left-sided tumors and 0.45 (95% CI, 0.38-0.54) for right-sided tumors (P = .41). There was some evidence of interaction between laterality and histological grade (P = .09 for OS and P = .02 for CSS by test for interaction). Stage-specific Cox proportional hazards regression model analyses revealed that poorly differentiated histology was associated with a larger effect size in left-sided colon cancer (HRs of 1.52 [95% CI, 1.25-1.85] for OS and 1.75 [95% CI, 1.41-2.17] for CSS) than in right-sided colon cancer (HRs of 1.15 [95% CI, 0.99-1.34] for OS and 1.18 [95% CI, 1.00-1.40] for CSS). There was no strong evidence of interaction for other disease-related factors.
In this population-based cohort study, we explored the extent to which disease laterality is associated with survival among patients with early-stage colon cancer. Several important findings have emerged. First, older, female patients, with greater comorbidity, are more likely to have right-sided colon cancer. Second, we found that right-sided cancers are more likely to have a higher T stage, to be poorly differentiated, and to be node negative. Third, our results do not demonstrate any significant difference in long-term survival between right-sided and left-sided colon cancers.
Our findings are in keeping with those by Weiss et al,16 who showed that, in a cohort of 53 801 patients who underwent curative-intent surgery for colon cancer in the United States during 1992 to 2005, disease laterality was not associated with survival. The primary analysis did not show any association across all stages; however, in subgroup analysis, the authors found that right-sided cancers in stage II disease were associated with decreased mortality (HR, 0.92; 95% CI, 0.87-0.97; P = .001). Although not statistically significant, our results also demonstrate a trend toward improved outcome among patients with right-sided stage II disease. Weiss et al16 found greater all-cause mortality with right-sided disease in stage III (HR, 1.12; 95% CI, 1.06-1.18; P < .001); our results suggest a trend toward inferior CSS among patients with right-sided stage III disease. It is notable that the study by Weiss et al16 included only patients 65 years and older, and their study was conducted over an earlier and longer period than our own (1992-2005 vs 2002-2008).
In a study using the Surveillance, Epidemiology, and End Results (SEER) program database, Meguid et al15 identified 77 978 patients who underwent resection of colon cancer between 1988 and 2003. In adjusted analyses, right-sided cancers were associated with increased overall mortality (HR, 1.04; 95% CI, 1.02-1.07). Stratified analysis demonstrated no difference in outcome for stage I disease and improved survival for stage II right-sided cancers. Accordingly, most of the association between right-sided disease and inferior survival was driven by patients with stage III and IV disease. However, a major limitation of their study was the inability to control for adjuvant chemotherapy use. Warschkow and colleagues17 reported another analysis using SEER data in which they identified 91 416 patients with stage I to III colon cancer resected between 2004 and 2012. In a propensity score analysis, they found improved OS (HR, 0.92; 95% CI, 0.89-0.94) and CSS (HR, 0.89; 95% CI, 0.84-0.87) for patients with right-sided cancers. This result was driven by a strong association in stage I and II disease because laterality was not associated with survival among patients with stage III disease. A meta-analysis20 of 15 studies published up to May 2014 in patients with stage I to IV colon cancer demonstrated inferior OS for patients with right-sided disease (HR, 1.14; 95% CI, 1.06-1.22). Among the 3 studies that reported CSS, there was no statistically significant association between laterality and outcome (HR, 1.14; 95% CI, 0.86-1.51). Finally, a more recent meta-analysis29 of 66 studies published from 1995 to 2016 showed that left-sided colon cancers were associated with improved survival (HR, 0.82; 95% CI, 0.79-0.84). Studies that included patients with only stage IV disease compared with those that included stage I to III disease demonstrated a greater effect on mortality (HR, 0.73; 95% CI, 0.69-0.78 vs HR, 0.84; 95% CI, 0.79-0.89).
Three studies presented at the 2016 annual meeting of the American Society of Clinical Oncology further explored the association between disease laterality and outcome. Using SEER data for patients with colon and rectal cancer from 2007 to 2011, Schrag and colleagues18 observed inferior OS for patients with right-sided disease compared with left-sided disease and rectal cancer. This association was strongest for patients with stage III and IV disease. Although these results were adjusted for age, sex, race/ethnicity, and year, they were not adjusted for comorbidity. Cancer-specific survival was not reported. Kennecke and colleagues19 described relapse-free survival and survival after relapse in patients with stage I to III colon cancer. They found that 5-year relapse-free survival was superior for left-sided stage II tumors, but there was no difference in outcome for stage III disease. However, in patients who relapsed, those who initially had right-sided tumors had worse 5-year survival after relapse. Finally, Venook et al24 explored disease laterality in stage IV CRC in the context of the Intergroup 80405 clinical trial. They reported inferior OS among KRAS wild-type patients with right-sided primary tumors (HR, 1.56; 95% CI, 1.32-1.84).
It is important to highlight that most of these studies evaluated OS (rather than CSS) and did not adjust for comorbidity. Given the well-known association between more advanced age and right-sided disease, it is possible that results of some studies may be biased by confounding due to unmeasured comorbidity. This conjecture is supported by the observation that, among those studies that evaluated CSS, outcomes were not inferior for right-sided colon cancer.
Our study should be interpreted in the context of certain methodological limitations. The study population was identified using linked administrative databases that do not include information regarding various known prognostic factors in colon cancer, such as smoking status, diet, or race/ethnicity. In addition, information on performance status was not collected. Therefore, residual confounding is a major limitation in this study. In addition, because the correlation between medical record–derived and International Classification of Diseases, Ninth Revision, code–derived Charlson Comorbidity Index is moderate,30 our adjustment for the Charlson Comorbidity Index may not be optimal. As in most population-based studies, we were unable to determine the molecular subtype of the tumor, which likely has a role in prognosis. Specifically, tumors that exhibit microsatellite instability (MSI) have been shown to have better prognosis, and it is estimated that 20% to 25% of stage II and 15% of stage III right-sided cancers are MSI positive.31 In an attempt to understand why right-sided cancers may have worse cancer-specific mortality, despite being more likely to be MSI positive, Phipps et al32 correlated MSI status with oncologic outcomes of more than 3000 patients with CRC. Patients who had higher cancer-specific mortality were limited to those with microsatellite-stable tumors. Some of those patients also had other mutations in their tumor, including CIMP and BRAF, which are associated with poor prognosis.
The major strengths of this study include the large sample size and the fact that we captured all cases of resected early-stage colon cancer using the OCR. By reviewing all surgical pathology reports, we were also able to control for important disease-related factors (eg, grade and lymphovascular invasion). Finally, we describe OS and CSS; CSS may be less vulnerable than OS to confounding by age and comorbidity, which is relevant given the known differences in patient case mix between right-sided and left-sided colon cancer. To our knowledge, this investigation is the largest study that has evaluated both OS and CSS.
In summary, we found no strong association between disease laterality and outcome of early-stage resected colon cancer. Results of previously published studies are not consistent; therefore, the extent to which disease laterality is prognostic remains uncertain. Studies in this field should control for comorbidity and explore CSS in addition to OS. Future research that can incorporate molecular and genetic elements of disease will provide important insights into how tumor location is associated with patient outcome.
Accepted for Publication: March 8, 2017.
Corresponding Author: Christopher M. Booth, MD, Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, 10 Stuart St, Kingston, ON K7L 3N6, Canada (email@example.com).
Published Online: June 8, 2017. doi:10.1001/jamaoncol.2017.1016
Author Contributions: Dr Booth had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Karim, Nanji, Booth.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Karim, Booth.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Karim, Brennan.
Obtained funding: Nanji, Booth.
Administrative, technical, or material support: Nanji, Berry.
Study supervision: Nanji, Booth.
Conflict of Interest Disclosures: Dr Berry reported serving on advisory boards for Amgen and Lilly. No other disclosures were reported.
Funding/Support: This work was supported by the Canada Foundation for Innovation and the Canadian Institutes of Health Research. This study was also supported by the Institute for Clinical Evaluative Sciences, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information. Dr Booth is supported as a Canada Research Chair in Population Cancer Care.
Role of the Funder/Sponsor: Neither the Institute for Clinical Evaluative Sciences nor the Ontario Ministry of Health and Long-Term Care had any role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Disclaimer: The opinions, results, and conclusions reported in this article are those of the authors and are independent from the funding sources and the Canadian Institute for Health Information. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred. Parts of this material are based on data and information provided by Cancer Care Ontario; however, the analysis, conclusions, opinions, and statements expressed herein are those of the authors and not necessarily those of Cancer Care Ontario.