Comparison of human papillomavirus DNA prevalence in atypical squamous cells of undetermined significance subcategories as defined by the original Bethesda 1991 and the new Bethesda 2001 Systems.

CONTEXT: The new Bethesda System 2001 (TBS 2001) minimized the subclassification of atypical squamous cells of undetermined significance (ASCUS). OBJECTIVE: The primary goal of this study was to determine the impact of the new subclassification on the accuracy of Papanicolaou (Pap) test diagnosis by examining the prevalence of human papillomavirus (HPV) DNA in different ASCUS subcategories, as defined by the new TBS 2001 versus the original TBS 1991. The second goal was to identify specific morphologic features of atypical squamous cells that are more frequently associated with HPV detection. DESIGN: Consecutive cases of ThinPrep Pap tests were retrospectively reviewed by a panel of pathologists to obtain consensus diagnoses. The study group consisted of ASCUS cases; the positive control group consisted of low- and high-grade squamous intraepithelial lesions (LSILs and HSILs, respectively); and the negative control group consisted of cases "negative for intraepithelial lesion or malignancy." All ASCUS cases were subclassified according to TBS 1991 into the following categories: favor reactive (ASCUS-R), favor LSIL (ASCUS-L), favor HSIL (ASCUS-H), and not otherwise specified (ASCUS-NOS). In a separate review, ASCUS cases were subclassified according to TBS 2001 into the following categories: atypical squamous cells of undetermined significance (ASC-US) and atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion (ASC-H). Furthermore, morphologic ASCUS subtypes were recorded (atypical mature, immature, parakeratotic, and atrophic cells); in addition, individual morphologic features of atypical cells were recorded. Broad- spectrum HPV DNA amplification and genotyping was performed using short PCR fragment (SPF 10) polymerase chain reaction/Line Probe assays. RESULTS: In cases classified according to TBS 1991, HPV was detected in 32% of negative, 49% of ASCUS, and 93% of LSIL/HSIL cases. On the second review, using the diagnostic categories of TBS 2001, which eliminated the ASCUS-;R category, the number of ASCUS cases decreased by 45%. The prevalence of HPV DNA in ASCUS cases downgraded to the negative category was 38%, which was not significantly different from HPV prevalence in negative cases as diagnosed under TBS 1991. Furthermore, HPV was detected in 56% of ASC-US and 71% of ASC-H cases. The prevalence of HPV in different morphologic subtypes of ASCUS was not significantly different, and none of the 8 individual morphologic features of atypical cells were more frequently associated with HPV detection. CONCLUSION: Elimination of the ASCUS-R category in TBS 2001 resulted in a significant decrease in the number of ASCUS diagnoses. Downgraded cases had a relatively low prevalence of HPV DNA. It is expected that TBS 2001 will increase specificity of the Pap test without compromising its sensitivity.