Accuracy of pathologic interpretation of colon polyps by general pathologists

M. Alikhan, D. Rex, O. Cummings, T. Ulbright

Research output: Contribution to journalArticle


Nearly all colorectal cancer (CA) in the West develops in adenomatous polyps (AD). Correct interpretation of polyp (P) histology is important as it often guides subsequent colonoscopy and/or surgery. AIM: To study the accuracy of path interpretation of colon P in clinical practice. METHODS: We identified 20 histologic slides representing different types of P, each chosen because it demonstrated typical features of a type of P. We submitted these slides to blinded review by 20 randomly selected general pathologists(PA) all of whom were practicing general surgical pathology in community hospitals in central Indiana. There were 5 cases of malignant P, 9 AD including 3 with severe dysplasia (SD) and 6 miscellaneous P in the sample. RESULTS: CA was correctly identified in 91% of 100 readings. Of the 91 readings of CA tumor differentiation was described in 50 (55%). 40% of PA identified poor differentiation when present. Only 10 of 20 PA made comments in any of the cases regarding whether the resection margin was free of CA. In the 3 cases where surgery was clearly indicated, resection was recommended in only 15 of 60 readings (25%). In 2 cases where surgery was clearly not indicated resection was recommended in 15 of 40 readings (38%). AD were correctly identified in 169 (94%) of 180 readings. Type of AD (tubular (T), tubulo-villous (TV), villous (V) was noted in 86% (145 of 169 readings). Of TV and V AD, T was incorrectly reported in 2% (2 of 111). SD was correctly identified in 28 (47%) of 60 readings although in 25/28 it was termed "adeno CA-in-situ" or "intramucosal adeno-CA." Invasive CA was incorrectly read in 22% (13 of 60) of AD with SD and in 19 of 60 SD was not called. AD without SD were said to have SD in 12% (14 of 120) and CA in 1%. Of miscellaneous P, hamartomas were read as such by 4 of the 20 PA. The polypoid phase of solitary rectal ulcer syndrome (SRUS) was identified by 2 of 20 PA. Hamartomas and SRUS were most often read incorrectly as AD. Hyperplastic P were correctly identified in 75% (30 of 40), juvenile/retention P in 80% (16 of 20). SUMMARY: 1) Sensitivity for identifying malignant P and AD were both >90%. 2) When a malignant P was identified tumor differentiation, proximity to the resection margin and need for surgical resection were often inadequately or inaccurately reported. 3) Specificity for cancer in neoplastic polyps was high. 4) T AD were often misdiagnosed as TV or V, but TV and V were rarely called T. 5) SD was often not recognized, called CA or designated by out-of-date terminology. 6) Hamartomas and the polypoid phase of SRUS were seldom recognized. CONCLUSION: This study suggests areas to focus continuing education efforts with regard to colon P interpretation. Pathologic interpretation of polyps by community pathologists may be a major source of error in clinical management.

Original languageEnglish (US)
Pages (from-to)AB94
JournalGastrointestinal endoscopy
Issue number4
StatePublished - Dec 1 1998

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Gastroenterology

Fingerprint Dive into the research topics of 'Accuracy of pathologic interpretation of colon polyps by general pathologists'. Together they form a unique fingerprint.

  • Cite this