Can the Sum of Adenoma Diameters (Adenoma Bulk) on Index Examination Predict Risk of Metachronous Advanced Neoplasia?

• Authors: Joseph C. Anderson, Department of Veterans Affairs Medical Center; Carolyn Morris, Dartmouth College; Douglas J. Robertson, Department of Veterans Affairs Medical Center; Elizabeth L. Barry, Dartmouth College; Jane C. Figueiredo, University of Southern California; Marcia Cruz-Correa, University of Puerto Rico; Roberd M Bostick, Emory University; Dennis J. Ahnen, University of Colorado; John A. Baron, University of North Carolina
• Journal Title: Journal of Clinical Gastroenterology
• Volume: Volume 52, Number 7
• Publisher: Lippincott, Williams & Wilkins | 2018-08-01, Pages 628-634
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1097/MCG.0000000000000899
• Subjects: Health Sciences, Medicine and Surgery; Health Sciences, Epidemiology
• Keywords: Science & Technology; Life Sciences & Biomedicine; Gastroenterology & Hepatology; adenoma; colon; colonoscopy; DIMINUTIVE COLORECTAL POLYPS; HIGH-GRADE DYSPLASIA; NATIONAL-SURVEY; WHITE-LIGHT; COLONOSCOPY; CANCER; SURVEILLANCE; GASTROENTEROLOGISTS; MANAGEMENT; DIAGNOSIS

Abstract

Background: Recent data suggest that adenoma size and number are more important predictors of metachronous colorectal neoplasia than advanced histology. Furthermore, there is poor reproducibility in diagnosing advanced histology; high-grade dysplasia and villous histology. Therefore we developed a new metric, adenoma bulk, the sum of diameters of all baseline adenomas, regardless of advanced features. Goal: Compare the predictive value for metachronous advanced neoplasia of adenoma bulk to conventional paradigm. Study: Data were collected prospectively in a multicenter adenomachemoprevention trial (2004 to 2013). For the conventional paradigm, high-risk baseline findings were defined as Z3 adenomas, large adenomas (Z1 cm) or adenomas with villous components or high-grade dysplasia. Adenoma bulk was examined across quartiles and as a continuous variable. Predictive characteristics (sensitivities, specificities, c-statistics) for metachronous advanced neoplasia using conventional criteria and adenoma bulk were calculated. receiver operator characteristic curves were computed using logistic regression. Results: In total, 1948 adults had index and follow-up colonoscopies (mean follow-up, 45.2 mo). Those with an adenoma bulk Z10mm (4th quartile) had a higher metachronous advanced neoplasia risk (14.4% vs. 6.9'8.2% in lower 3 quartiles; P=0.0002). The c-statistics and sensitivities (specificity fixed at 0.73) for the adenoma bulk and conventional models were 0.587 and 0.563 (P=0.17) and 0.396 and 0.390, respectively. Conclusions: Categorizing sporadic adenoma patients as high versus low risk for metachronous advanced neoplasia by adenoma bulk of <versus Z10mm may be comparably predictive as conventional paradigm and simplifies risk stratification by obviating need for additional histology regarding extent of villous component or degree of dysplasia in resected polyps. The adenoma bulk metric and the 10mm cutoff in particular would have to be validated in other populations before it can be used in clinical practice.