Significance of subclassifying high-grade squamous intraepithelial lesions into moderate dysplasia/CIN II versus severe dysplasia/CIN III/CIS in the Bethesda System terminology.

Recent revisions of the Bethesda System (TBS III) did not modify the original two-tiered low-grade and high-grade squamous intraepithelial lesion (LSIL/HSIL) terminology and continued to offer the option of subclassifying HSIL into moderate dysplasia/cervical intraepithelial neoplasia (CIN II) and severe dysplasia/CIN III. The purpose of this study is to evaluate the significance of the HSIL moderate dysplasia/CIN II subclassification and the causes of Pap test-biopsy discordance for this subclassification. HSIL/moderate dysplasia/CIN II Pap tests were identified from the University of California Davis Medical Center laboratory file for a 5-yr period (1997-2001) and correlated with follow-up cervical biopsies. Cervical biopsies with CIN II diagnoses were identified from the laboratory file for the same time period, and correlated with their preceding Pap test. Discordant Pap test-biopsy pairs from either group were reviewed by two pathologists for the following causes for discordance: cytologic overcall or undercall, biopsy overcall or undercall, and sampling error or possible lesion regression. During the review period, 161 of 378 total HSIL Pap tests had a HSIL/CIN II result (42%), and 81/161 (50%) had a follow-up cervical biopsy. 38/81 (47%) were concordant, and 43 (53%) were discordant; 39/43 discordant cases had both Pap test and biopsy slides available for review. CIN I was the most common discordant biopsy diagnosis. Sampling error was the most common cause of discordance and was three times more common than the other causes; 536 CIN cervical biopsies with CIN were identified in the laboratory files during this review period, and 108/536 (20%) received a diagnosis of CIN II. Sixty-seven of 108 (62%) had a preceding Pap test result available; 32 of the 67 (48%) had a concordant preceding Pap test with a HSIL/CIN II result, and 35 (52%) were discordant; 32/35 discordant cases had both Pap test and biopsy slides available for review. LSIL was the most common discordant Pap test diagnosis. Sampling error was also the most common cause of discordance, and was also three time more frequent than any other cause for discordance. HSIL/CIN II is a meaningful subclassification in our laboratory. We support the TBS III recommendation for optional subclassification of HSIL and suggest that individual laboratories may wish to monitor use of this subclassification through internal correlation studies. Copyright 2004 Wiley-Liss, Inc.