User login
TOPLINE:
A comprehensive model considering patient age, diabetes, colonoscopy indications, and polyp findings can predict colorectal cancer (CRC) risk more accurately than the solely polyp-based model in patients with a first diagnosis of adenoma on colonoscopy.
METHODOLOGY:
- Because colonoscopy surveillance guidelines relying solely on previous polyp findings to assess CRC risk are imprecise, researchers developed and tested a comprehensive risk prediction model from a list of CRC-related predictors that included patient characteristics and clinical factors in addition to polyp findings.
- The comprehensive model included baseline colonoscopy indication, age group, diabetes diagnosis, and polyp findings (adenoma with advanced histology, polyp size ≥ 10 mm, and sessile serrated or traditional serrated adenoma).
- They randomly assigned 95,001 patients (mean age, 61.9 years; 45.5% women) who underwent colonoscopy with polypectomy to remove a conventional adenoma into two cohorts: Model development (66,500) and internal validation (28,501).
- In both cohorts, researchers compared the performance of the polyp findings-only method against the comprehensive model in predicting CRC, defined as an adenocarcinoma of the colon or rectum diagnosed a year after the baseline colonoscopy.
TAKEAWAY:
- During the follow-up period starting 1 year after colonoscopy, 495 patients were diagnosed with CRC; 354 were in the development cohort and 141 were in the validation cohort.
- The comprehensive model demonstrated better predictive performance than the traditional polyp-based model in the development cohort (area under the curve [AUC], 0.71 vs 0.61) and in the validation cohort (AUC, 0.7 vs 0.62).
- The difference in the Akaike Information Criterion values between the comprehensive and polyp models was 45.7, much above the threshold of 10, strongly indicating the superior performance of the comprehensive model.
IN PRACTICE:
“Improving the ability to accurately predict the patients at highest risk for CRC after polypectomy is critically important, given the considerable costs and resources associated with treating CRC and the better prognosis associated with early cancer detection. The current findings provide proof of concept that inclusion of CRC risk factors beyond prior polyp findings has the potential to improve post-colonoscopy risk stratification,” the authors wrote.
SOURCE:
The study, led by Jeffrey K. Lee, MD, MPH, Division of Research, Kaiser Permanente Northern California, Oakland, California, was published online in The American Journal of Gastroenterology.
LIMITATIONS:
External validation of the model’s performance is needed in different practice settings. The generalizability of the findings is limited because the study population did not include individuals without a prior adenoma or those with an isolated serrated polyp. Moreover, the examination of polyp size > 20 mm as a potential predictor of CRC was precluded due to incomplete data.
DISCLOSURES:
The study was conducted within the National Cancer Institute–funded Population-Based Research to Optimize the Screening Process II consortium and funded by a career development grant from the National Cancer Institute to Lee. The authors declared no conflicts of interest.
A version of this article appeared on Medscape.com.
TOPLINE:
A comprehensive model considering patient age, diabetes, colonoscopy indications, and polyp findings can predict colorectal cancer (CRC) risk more accurately than the solely polyp-based model in patients with a first diagnosis of adenoma on colonoscopy.
METHODOLOGY:
- Because colonoscopy surveillance guidelines relying solely on previous polyp findings to assess CRC risk are imprecise, researchers developed and tested a comprehensive risk prediction model from a list of CRC-related predictors that included patient characteristics and clinical factors in addition to polyp findings.
- The comprehensive model included baseline colonoscopy indication, age group, diabetes diagnosis, and polyp findings (adenoma with advanced histology, polyp size ≥ 10 mm, and sessile serrated or traditional serrated adenoma).
- They randomly assigned 95,001 patients (mean age, 61.9 years; 45.5% women) who underwent colonoscopy with polypectomy to remove a conventional adenoma into two cohorts: Model development (66,500) and internal validation (28,501).
- In both cohorts, researchers compared the performance of the polyp findings-only method against the comprehensive model in predicting CRC, defined as an adenocarcinoma of the colon or rectum diagnosed a year after the baseline colonoscopy.
TAKEAWAY:
- During the follow-up period starting 1 year after colonoscopy, 495 patients were diagnosed with CRC; 354 were in the development cohort and 141 were in the validation cohort.
- The comprehensive model demonstrated better predictive performance than the traditional polyp-based model in the development cohort (area under the curve [AUC], 0.71 vs 0.61) and in the validation cohort (AUC, 0.7 vs 0.62).
- The difference in the Akaike Information Criterion values between the comprehensive and polyp models was 45.7, much above the threshold of 10, strongly indicating the superior performance of the comprehensive model.
IN PRACTICE:
“Improving the ability to accurately predict the patients at highest risk for CRC after polypectomy is critically important, given the considerable costs and resources associated with treating CRC and the better prognosis associated with early cancer detection. The current findings provide proof of concept that inclusion of CRC risk factors beyond prior polyp findings has the potential to improve post-colonoscopy risk stratification,” the authors wrote.
SOURCE:
The study, led by Jeffrey K. Lee, MD, MPH, Division of Research, Kaiser Permanente Northern California, Oakland, California, was published online in The American Journal of Gastroenterology.
LIMITATIONS:
External validation of the model’s performance is needed in different practice settings. The generalizability of the findings is limited because the study population did not include individuals without a prior adenoma or those with an isolated serrated polyp. Moreover, the examination of polyp size > 20 mm as a potential predictor of CRC was precluded due to incomplete data.
DISCLOSURES:
The study was conducted within the National Cancer Institute–funded Population-Based Research to Optimize the Screening Process II consortium and funded by a career development grant from the National Cancer Institute to Lee. The authors declared no conflicts of interest.
A version of this article appeared on Medscape.com.
TOPLINE:
A comprehensive model considering patient age, diabetes, colonoscopy indications, and polyp findings can predict colorectal cancer (CRC) risk more accurately than the solely polyp-based model in patients with a first diagnosis of adenoma on colonoscopy.
METHODOLOGY:
- Because colonoscopy surveillance guidelines relying solely on previous polyp findings to assess CRC risk are imprecise, researchers developed and tested a comprehensive risk prediction model from a list of CRC-related predictors that included patient characteristics and clinical factors in addition to polyp findings.
- The comprehensive model included baseline colonoscopy indication, age group, diabetes diagnosis, and polyp findings (adenoma with advanced histology, polyp size ≥ 10 mm, and sessile serrated or traditional serrated adenoma).
- They randomly assigned 95,001 patients (mean age, 61.9 years; 45.5% women) who underwent colonoscopy with polypectomy to remove a conventional adenoma into two cohorts: Model development (66,500) and internal validation (28,501).
- In both cohorts, researchers compared the performance of the polyp findings-only method against the comprehensive model in predicting CRC, defined as an adenocarcinoma of the colon or rectum diagnosed a year after the baseline colonoscopy.
TAKEAWAY:
- During the follow-up period starting 1 year after colonoscopy, 495 patients were diagnosed with CRC; 354 were in the development cohort and 141 were in the validation cohort.
- The comprehensive model demonstrated better predictive performance than the traditional polyp-based model in the development cohort (area under the curve [AUC], 0.71 vs 0.61) and in the validation cohort (AUC, 0.7 vs 0.62).
- The difference in the Akaike Information Criterion values between the comprehensive and polyp models was 45.7, much above the threshold of 10, strongly indicating the superior performance of the comprehensive model.
IN PRACTICE:
“Improving the ability to accurately predict the patients at highest risk for CRC after polypectomy is critically important, given the considerable costs and resources associated with treating CRC and the better prognosis associated with early cancer detection. The current findings provide proof of concept that inclusion of CRC risk factors beyond prior polyp findings has the potential to improve post-colonoscopy risk stratification,” the authors wrote.
SOURCE:
The study, led by Jeffrey K. Lee, MD, MPH, Division of Research, Kaiser Permanente Northern California, Oakland, California, was published online in The American Journal of Gastroenterology.
LIMITATIONS:
External validation of the model’s performance is needed in different practice settings. The generalizability of the findings is limited because the study population did not include individuals without a prior adenoma or those with an isolated serrated polyp. Moreover, the examination of polyp size > 20 mm as a potential predictor of CRC was precluded due to incomplete data.
DISCLOSURES:
The study was conducted within the National Cancer Institute–funded Population-Based Research to Optimize the Screening Process II consortium and funded by a career development grant from the National Cancer Institute to Lee. The authors declared no conflicts of interest.
A version of this article appeared on Medscape.com.