Diagnostic yield of the one-year surveillance colonoscopy after curative colorectal cancer resection Authors: Jenna Cross, Garrett Johnson, Harminder Singh, Farhana Shariff, Eric Hyun, Ramzi Helewa Published: 2026 ABSTRACT PurposeGuidelines recommend routine colonoscopy 1 year after curative colorectal cancer resection, despite limited evidence supporting its diagnostic yield. This study evaluated the real-world outcomes of 1-year surveillance colonoscopy following colorectal cancer resection.MethodsThis is a retrospective observational study using pathology, clinic, and endoscopy data patients who underwent segmental oncologic resection for primary colon or rectal. Adenocarcinoma between January 2007 and June 2020 in Winnipeg, Manitoba. Patients were excluded if they had a total proctocolectomy, if no follow-up colonoscopy data were available, or if the date of the first surveillance scope was < 6 months or > 24 months from surgery. Colonoscopy findings were categorized as adenocarcinoma, advanced adenoma, simple adenoma or normal.ResultsA total of 454 patients met the inclusion criteria. The 1-year surveillance colonoscopy detected 2 (0.44%) adenocarcinomas, 3 (0.66%) advanced adenomas, and 87 (19.2%) simple adenomas. The majority (n = 362, 79.7%) had normal colonoscopy results. Preoperative bowel preparation quality was not reported for either recurrent adenocarcinoma. Neither of the recurrent adenocarcinomas was associated with elevated carcinoembryonic antigen (CEA), abnormal imaging, or known hereditary cancer syndromes.ConclusionsOur data demonstrate a low observed rate of clinically significant findings at 1-year surveillance colonoscopy following colorectal cancer resection. Incomplete documentation of index colonoscopy quality limited assessment of whether surveillance findings represented true metachronous lesions or lesions missed at the preoperative examination. Given the low event rate and observational study design, these findings should be interpreted as descriptive real-world data regarding postoperative surveillance outcomes.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00384-026-05187-7. Introduction Colorectal cancer is the third most commonly diagnosed cancer in North America, accounting for approximately 10% of all new cancer cases each year [1]. Several clinical tests and modalities are routinely used to monitor patients for recurrence. With respect to colonoscopy as surveillance, the U.S. Multi-Society Task Force (USMSTF) on Colorectal Cancer [2, 3], which is endorsed by the American Gastroenterological Association (AGA), the American Society for Gastrointestinal Endoscopy (ASGE), and the American College of Gastroenterology (ACG), is widely referenced in North America. These guidelines current support colonoscopy one year after curative colorectal cancer resection [2, 4, 5]. If the 1-year colonoscopy is normal, the next should be done three years later, followed by subsequent colonoscopies at five-year intervals if no abnormalities are found [4, 6]. Despite some randomized controlled trials and systematic reviews evaluating surveillance strategies after curative colorectal cancer resection, there remains no consensus on the optimal surveillance interval. A Cochrane review by Jeffery et al. synthesized six randomized trials and concluded that while intensified follow-up may improve overall survival, there is insufficient evidence to define the ideal timing and frequency of surveillance colonoscopy [7]. Although 1-year surveillance colonoscopy is widely recommended, contemporary real-world data evaluating its diagnostic yield in the setting of modern preoperative colonoscopy remain limited. Consequently, the clinical value and diagnostic yield of this practice remain incompletely understood. In this study, we aimed to characterize findings at the 1-year postoperative colonoscopy and describe rates of local recurrence and metachronous lesions following curative colorectal cancer resection. Methods Study design and setting This was a retrospective cohort study of all patients who underwent elective resection for a primary colon or rectal cancer at St. Boniface Hospital (SBH) in Winnipeg, Manitoba, Canada between January 1, 2007 and June 30, 2020. Patient inclusion was limited to June 30, 2020 to ensure a minimum of 12 months of postoperative follow-up and complete capture of surveillance outcomes within the provincial medical record. This defined study period maintained consistency in colonoscopy documentation and surveillance reporting practices in place at the time, ensuring mature and uniformly recorded data across the cohort. SBH serves as the primary tertiary referral for colorectal surgery in the province of Manitoba, receiving colorectal surgery and endoscopy referrals from healthcare providers across the province, with a population of over 1.4 million people. In Manitoba, postoperative surveillance colonoscopy is typically coordinated by the operating surgeon. For patients residing in rural or remote communities, surveillance is often arranged locally through community-based endoscopy services rather than at the index surgical center. As a result, postoperative colonoscopies in this study were performed across tertiary, urban community, and rural centers. This retrospective cohort study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement, consistent with the EQUATOR Network recommendations for observational research [8]. University of Manitoba Health Research Ethics Board approval was obtained for this study prior to its commencement (HS21588, H2018:103). Population Patients were identified from our database of elective colorectal surgeries performed at SBH. The details of this database were previously described [9]. In brief, patients were identified from SBH hospital admission records through admission diagnosis, procedure, and admitting hospital service. Hospital, operative, endoscopy and pathology reports were then reviewed and synthesized with clinic chart data. Prophylactic, emergency or palliative surgeries were excluded. From this database, patients who underwent elective segmental resection for adenocarcinoma were included. Patients were excluded if they underwent a total proctocolectomy, if they did not undergo a complete preoperative colonoscopy, if they did not undergo a postoperative follow-up colonoscopy within the provincial medical record, or if the date of the first surveillance colonoscopy was < 6 months or > 24 months from surgery. Accordingly, the definition of ‘1-year surveillance colonoscopy’ included procedures performed between 6 and 24 months to account for true variability in timing of postoperative surveillance encountered in real-world clinical practice. Data sources/variables Data regarding patient characteristics, preoperative colonoscopies, and surgical procedures were obtained from medical records. The findings of the first surveillance colonoscopy after surgery were documented, including timing, location, specialty of the proceduralist, bowel preparation quality, number and characteristics of polyps identified and/or cancer recurrence. Preoperative endoscopy report findings were abstracted (bowel preparation quality, polyp numbers, or incomplete colonoscopy). Additionally, data were collected on whether patients underwent a repeat preoperative colonoscopy and on the type of provider who performed each procedure (e.g., colorectal surgeon, gastroenterologist, or rural practitioner). Patient demographics included age, sex, known genetic cancer predisposition (e.g., Familial Polyposis (FAP), or Lynch Syndrome (LS)), results of other surveillance tests, tumor biology, and operative factors. Data were collected and uploaded digitally on the Research Electronic Data Capture (REDCap) secure online platform [9, 10]. Missing data were assumed to be missing at random. Descriptive analyses used all available documented data. For exploratory comparisons involving incompletely documented variables (e.g., bowel preparation quality), analyses were performed with the limited patient information available. No adjustment for confounders was attempted due to the low event rate. Outcomes Primary outcomes were endoscopy findings on the 1-year surveillance colonoscopy, which were categorized as: normal, simple adenomas, advanced adenomas, and adenocarcinomas. Normal findings were defined as the absence of any adenomatous pathology identified on endoscopy. Advanced adenomas were defined as lesions ≥ 10 mm, with villous histology, or high-grade dysplasia. Adenocarcinomas are reported as identified on the pathology report. Both advanced adenomas and adenocarcinomas were classified as clinically significant neoplasia. Statistical analysis Descriptive statistics were used to summarize patient demographics and surveillance colonoscopy findings. Outcomes were reported as frequencies and proportions of the total study cohort. Univariable and multivariable analyses were planned to identify predictors of positive 1-year colonoscopy findings (i.e., clinically significant neoplasia). However, given the extremely low number of clinically significant findings, formal regression modeling was not pursued, as any estimates would be statistically unstable. Instead, exploratory comparisons were conducted using Fisher’s exact test and descriptive stratification to examine potential associations between surveillance findings and select clinical variables, such as location of procedure (urban vs rural), endoscopist specialty, and documentation of bowel preparation quality. Data were analyzed using Statistical Package for Social Sciences (SPSS) (IBM SPSS version 27, IBM Corp, Armonk, New York). The number needed to scope (NNS) was calculated as the inverse of the proportion of events detected at 1-year surveillance colonoscopy. Results Participant characteristics In total, 962 patients who underwent segmental oncologic resection for a primary colon or rectal adenocarcinoma between January 2007 and June 2020 at SBH in Winnipeg, Manitoba, were identified in this study. Details from each patient’s case are described in Table 1. Patients were excluded if they had a total proctocolectomy (n = 10), if no follow-up colonoscopy data was available (n = 458), or if the date of the scope was < 6 months (n = 15) or > 24 months from surgery (n = 25), resulting in a total of 454 patients analyzed, as shown in Fig. 1. The majority of excluded patients had no record of postoperative colonoscopy within the provincial medical record. Table 1Baseline characteristics of patients undergoing segmental resection for colorectal adenocarcinoma (n = 454). Patient demographics, tumor characteristics, and surgical approach. Values are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variablesFig. 1Flow diagram of patient inclusion and exclusion for analysis of 1-year surveillance colonoscopy. Flow of patients through the study, including total number of resections performed, exclusions (total proctocolectomy, missing data, surveillance colonoscopy outside 6–24 months), and final study population (n = 454) Table 1Baseline characteristics of patients undergoing segmental resection for colorectal adenocarcinoma (n = 454). Patient demographics, tumor characteristics, and surgical approach. Values are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables Baseline characteristics of patients undergoing segmental resection for colorectal adenocarcinoma (n = 454). Patient demographics, tumor characteristics, and surgical approach. Values are presented as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables Flow diagram of patient inclusion and exclusion for analysis of 1-year surveillance colonoscopy. Flow of patients through the study, including total number of resections performed, exclusions (total proctocolectomy, missing data, surveillance colonoscopy outside 6–24 months), and final study population (n = 454) Colonoscopy yield Of the total cohort, the mean time interval from surgery to postoperative surveillance colonoscopy was 13.5 ± 3.8 months, reflecting the real-world variability in follow-up timing. 81.7% of surveillance colonoscopies were performed within Winnipeg, at urban community hospitals or tertiary centers (Table 2). Table 2Timing and setting of 1-year surveillance colonoscopy (n = 454). Interval between surgery and surveillance colonoscopy expressed as mean ± SD. Location of procedure categorized as a tertiary center, an urban community hospital, or a rural site Table 2Timing and setting of 1-year surveillance colonoscopy (n = 454). Interval between surgery and surveillance colonoscopy expressed as mean ± SD. Location of procedure categorized as a tertiary center, an urban community hospital, or a rural site Timing and setting of 1-year surveillance colonoscopy (n = 454). Interval between surgery and surveillance colonoscopy expressed as mean ± SD. Location of procedure categorized as a tertiary center, an urban community hospital, or a rural site Among all the surveillance colonoscopies examined, 79.7% (n = 362) were normal, 19.2% (n = 87) detected a simple adenoma, 0.7% (n = 3) detected an advanced adenoma, and 0.4% (n = 2) detected an adenocarcinoma, as shown in Fig. 2. The diagnostic yield for clinically significant neoplasia (patients having either an advanced adenoma or recurrent adenocarcinoma) was 1.1% (n = 5). The number needed to scope (NNS) to detect one recurrent adenocarcinoma was 227 procedures, and 91 procedures to detect one clinically significant neoplasia (advanced adenoma or adenocarcinoma).Fig. 2Distribution of findings at 1-year surveillance colonoscopy (n = 454). Values presented as absolute numbers and percentages of the study cohort Distribution of findings at 1-year surveillance colonoscopy (n = 454). Values presented as absolute numbers and percentages of the study cohort The two patients with a detected recurrent adenocarcinoma on surveillance colonoscopy are described in Table 3. Both underwent right hemicolectomy for proximal colon cancers and had normal CT imaging at the time of surveillance. Recurrent lesions were identified at 13 and 16 months during surveillance colonoscopy performed in community settings. Neither patient had documented hereditary cancer syndromes. Table 3Adenocarcinoma patient characteristicsCharacteristicPatient 1Patient 2SexFemaleMaleAge at time of surgery (years)6970Surveillance interval (months)1613Index to surveillance scope interval (months)1712Index scope locationCommunityTertiarySurveillance scope locationCommunityCommunityIndex cancer location in colonAscending colonTransverse colonType of resectionRight HemicolectomyRight HemicolectomyNumber of polyps (surveillance scope)11Location of polyps (surveillance scope)Transverse colonSigmoid colonCT findingsNormalNormalDemographic and surgical data of patients with recurrent adenocarcinoma detected at surveillance colonoscopy Table 3Adenocarcinoma patient characteristicsCharacteristicPatient 1Patient 2SexFemaleMaleAge at time of surgery (years)6970Surveillance interval (months)1613Index to surveillance scope interval (months)1712Index scope locationCommunityTertiarySurveillance scope locationCommunityCommunityIndex cancer location in colonAscending colonTransverse colonType of resectionRight HemicolectomyRight HemicolectomyNumber of polyps (surveillance scope)11Location of polyps (surveillance scope)Transverse colonSigmoid colonCT findingsNormalNormalDemographic and surgical data of patients with recurrent adenocarcinoma detected at surveillance colonoscopy Adenocarcinoma patient characteristics Demographic and surgical data of patients with recurrent adenocarcinoma detected at surveillance colonoscopy Predictors of positive scopes As part of the predictor analysis, potential factors associated with positive findings at the 1-year surveillance colonoscopy were evaluated. Notably, both cases of recurrent adenocarcinoma occurred in patients without biochemical, radiographic, or symptomatic indicators of recurrence. There were no identifiable clinical predictors such as tumor stage, location, or perioperative carcinoembryonic antigen (CEA), that differentiated these patients from the broader cohort. Index preoperative colonoscopy quality was reviewed as a potential contributor to pathology on surveillance colonoscopies. Documentation of bowel preparation quality and Boston Bowel Preparation Scores (BBPS) was absent in approximately 70–80% of index procedures, limiting the ability to assess preparation as a predictive variable. Exact bowel preparation quality is described in Table 4. Among the minority of cases with documented BBPS, there was no statistically significant association between recorded bowel preparation quality and the presence of advanced adenoma or recurrent carcinoma at the 1-year scope. Table 4Documentation of bowel preparation quality and Boston Bowel Preparation Scores (BBPS) at index and surveillance colonoscopy. Data are expressed as n (%). Missing values reflect absent documentation in procedure reports. BBPS, Boston Bowel Preparation Score Table 4Documentation of bowel preparation quality and Boston Bowel Preparation Scores (BBPS) at index and surveillance colonoscopy. Data are expressed as n (%). Missing values reflect absent documentation in procedure reports. BBPS, Boston Bowel Preparation Score Documentation of bowel preparation quality and Boston Bowel Preparation Scores (BBPS) at index and surveillance colonoscopy. Data are expressed as n (%). Missing values reflect absent documentation in procedure reports. BBPS, Boston Bowel Preparation Score Discussion This study provides real-world provincial data regarding findings at 1-year surveillance colonoscopy following curative colorectal cancer resection. In this cohort, clinically significant pathology at surveillance colonoscopy was uncommon, with only five patients identified as having advanced adenoma or recurrent adenocarcinoma. These findings should be interpreted within the context of substantial missing data, broad surveillance intervals (6–24 months post-resection rather than a fixed 12-month examination), and a very low event rate, which limit the ability to draw conclusions regarding surveillance effectiveness or optimal follow-up strategies. In addition to the infrequent findings, the number needed to scope was 227 to detect one recurrent adenocarcinoma and 91 to detect one clinically significant neoplasia. These values provide context regarding the frequency of clinically significant findings within this cohort but should be interpreted cautiously given the low event rate. It is important to note that colonoscopy cannot be evaluated solely on procedural yield. Even a small number of asymptomatic recurrences detected at an earlier and potentially treatable stage may carry important oncologic value. The present study was not designed to assess salvageability, recurrence-specific survival, or downstream oncologic outcomes, and therefore does not determine the clinical effectiveness of 1-year surveillance colonoscopy. Despite this low yield, from a clinical perspective, the 1-year colonoscopy serves functions beyond the detection of metachronous neoplasia. It may identify anastomotic or local recurrence and provides confirmation that perioperative colonic evaluation was complete. This is particularly relevant in patients with obstructing tumors or suboptimal preoperative examinations, where full visualization of the colon may not have been achieved prior to resection. As such, the role of the 1-year examination should be interpreted within this broader context. Importantly, clinically significant findings identified on surveillance colonoscopy may represent biologically distinct entities, including metachronous neoplasia, missed synchronous lesions from the index examination, or local/anastomotic recurrence. These categories have different implications for surveillance effectiveness and interpretation of diagnostic yield. In the present cohort, recurrence patterns appeared heterogeneous; however, limited clinical detail and incomplete documentation restricted the ability to reliably distinguish between local/anastomotic recurrence, metachronous neoplasia, and lesions potentially missed at the index examination. Given incomplete documentation of index colonoscopy quality and limited recurrence-specific detail, the present study was not designed to distinguish among these mechanisms reliably. Another observation in our study was the poor documentation of perioperative colonoscopy quality. Boston Bowel Preparation Scores (BBPS) or clear assessments of bowel prep adequacy were missing in over 70% of cases, limiting confidence in the completeness of index exams. This lack of standardized documentation limited the assessment of index colonoscopy quality and restricted the ability to determine whether findings at surveillance colonoscopy represented true metachronous lesions or lesions potentially present but undocumented at the index examination. As a result, causation cannot be established from the available data alone. In addition to this missing data, variability in documentation highlights a broader limitation in real-world endoscopy reporting. This is supported by prior regional findings from Johnson et al. [11], who identified a high rate of repeat preoperative endoscopies even after the implementation of electronic synoptic reporting systems. Their study highlights variability in initial scope quality and documentation, contributing to downstream repeat procedures [11]. Another key implication of our study is the limited utility of pre-scope clinical risk factors in predicting recurrence. Neither of the two patients with adenocarcinoma at follow-up had elevated CEA, abnormal CT imaging, or documented hereditary cancer syndromes. This underscores the challenge of relying solely on pre-scope risk stratification to triage surveillance needs and suggests that some recurrence may only be detectable through direct visualization. When contextualized within the existing literature, our observed diagnostic yield at the 1-year surveillance colonoscopy of 0.4% for adenocarcinoma and 0.7% for advanced adenomas was notably lower than previously reported. A 2019 systematic review and meta-analysis of observational data on this topic, including 27 studies and 15,589 patients, reported rates of combined anastomotic and non-anastomotic colorectal cancer recurrences detected on surveillance colonoscopy 6 to 24 years after surgery of 4.3% [12]. Notably, the yield of surveillance colonoscopy in this systematic review was likely overestimated, as most included studies occurred before the advent of high-quality colonoscopy, and only 3 of 27 studies clearly specified or assessed the quality of colonoscopy at all. This systematic review also could not account for patient dropouts in the study, and as patients with recurrent cancer are more likely to become symptomatic and therefore present for their surveillance testing, likely overestimates the incidence of positive findings. More recently, a 2019 Dutch retrospective cohort study of 572 patients reported adenocarcinoma and advanced adenoma rates at a mean of 13.7 months of 1.7% and 9.7%, respectively [13]. This rate of significant findings was much higher than the present study. This discrepancy may also be attributable to differences in preoperative colonoscopy quality. Our study included only patients with a preoperative complete colonoscopy to the cecum, whereas the Dutch study reported a cecal intubation rate of only 86.2%, with only two-thirds of the unexamined segments being removed during surgery. Therefore, differences in completeness of preoperative colonoscopy may partially explain variation in surveillance findings between studies. Cecal intubation is a key quality metric, with current guidelines recommending a rate ≥ 95% for screening and diagnostic colonoscopies [14]. Suboptimal completion rates raise concerns about the quality of the initial exams and may explain the substantially higher postoperative scope yield in the Dutch study. Our results are more similar to other studies where complete preoperative colonoscopy was documented. For example, a 2015 Dutch study of 5156 patients reported a metachronous colorectal cancer rate detected between 6 and 24 months from segmental resection of 0.52% [15]. Although this study was designed to identify the rate of metachronous colorectal cancers after resection, rather than to test the yield of the first diagnostic scope specifically. Similarly, in a 2017 study from Portugal, out of 215 patients who had undergone a complete baseline colonoscopy prior to surgery, only two (0.9%) developed a metachronous cancer detected on surveillance colonoscopy over 5 years of follow-up. Limitations A major limitation of this study is the exclusion of nearly half of the initial cohort due to lack of documented follow-up colonoscopy within the provincial record. This introduces potential selection bias, as included patients may differ systematically from those without recorded follow-up. The omitted patients without documented follow-up may have differed systematically with respect to recurrence risk, mortality, geographic access to endoscopy services, or surveillance adherence. Accordingly, the external generalizability of these findings may be limited. Missing data also limited assessment of index colonoscopy quality, with bowel preparation quality or standardized scores absent in approximately 70–80% of cases. This restricts interpretation of whether findings at surveillance colonoscopy represent true metachronous lesions or lesions missed at the index examination. Consistent with other published studies, this raises the possibility that a portion of the observed yield may be attributable to missed lesions related to suboptimal preparation or incomplete reporting. Importantly, incomplete documentation does not necessarily indicate poor procedural quality, but it does limit the ability to draw firm conclusions regarding the relationship between index colonoscopy quality and subsequent findings. Similarly, clinical variables such as genetic risk factors (e.g., LS, and FAP) and interval symptoms were underreported or inconsistently available, which may have limited their inclusion in risk stratification. For analyses involving variables that were not consistently documented (such as bowel preparation quality), only patients with available information were included. Because some data were missing, these findings should be interpreted cautiously, as missing information may have affected the results of the study. If the assumption of missing at random was violated, our estimates of colonoscopy yield and related outcomes could be susceptible to bias. Future prospective multicenter studies with standardized documentation and broader population sampling are needed to strengthen these results and guide refinements in surveillance guidelines. The study definition of ‘1-year surveillance colonoscopy’ included procedures performed between 6 and 24 months postoperatively in order to reflect real-world surveillance variability. However, this broad interval could have introduced clinical heterogeneity, as examinations performed closer to 6 months versus 24 months may reflect different surveillance indications and recurrence risks. Accordingly, the findings should be interpreted as descriptive of postoperative surveillance practice rather than a strict evaluation of a 12-month colonoscopy exactly. Lastly, the low incidence of recurrent adenocarcinoma and advanced adenomas limited the ability to conduct meaningful analyses for predictors of positive findings. Because of the low event rate, comparisons are to be interpreted as descriptive only, as the study was not statistically powered to detect meaningful associations between clinical variables and surveillance findings. Conclusions This study provides real-world provincial data describing findings at 1-year surveillance colonoscopy following colorectal cancer resection. Clinically significant pathology was infrequently observed in this cohort, with recurrent cancers identified in a small number of asymptomatic patients without clinical or biochemical indicators of recurrence. Given the observational design, substantial exclusions, and limited documentation of index colonoscopy quality, these findings should be interpreted cautiously and viewed as descriptive rather than practice-changing. Importantly, 1-year surveillance colonoscopy may serve roles beyond detection of metachronous neoplasia, including assessment for local or anastomotic recurrence and confirmation of adequate perioperative colonic evaluation. Further prospective multicenter studies are needed to better define surveillance outcomes and identify patients at greatest risk of clinically significant findings. Incomplete documentation of index colonoscopy quality in this cohort limits interpretation, but highlights an important gap in real-world practice that may contribute to uncertainty in surveillance decision-making. Improving the standardization and reporting of colonoscopy quality metrics may help optimize follow-up strategies and, thereby, improve patient care. Further prospective, multicenter studies with standardized quality metrics and clearer risk stratification are needed to determine whether surveillance strategies can be safely individualized and whether select patients may benefit from tailored follow-up intervals. Supplementary Information Below is the link to the electronic supplementary material.ESM 1Supplementary Material 1 (PDF 178 KB) Supplementary Material 1 (PDF 178 KB)