Immediate histologic correlation in women with atypical squamous cells of undetermined significance cytology and positive high‐risk HPV: A retrospective review of 6000 cases in a large academic women's hospital

Abstract

Background Previous studies of the histologic correlation in women who have atypical squamous cells of undetermined significance cytology (ASC‐US) and are positive for high‐risk human papillomavirus (hrHPV+) have predominantly utilized the Hybrid Capture 2 (HC2) test, whereas the use of other US Food and Drug Administration (FDA)‐approved methods is relatively limited. Methods Cases of ASC‐US/hrHPV+ that were tested using HC2, Cervista, or Aptima were retrieved, and the immediate histologic correlations were analyzed. Results Overall, 53.9% (n = 3238) of women with ASC‐US/hrHPV+ had immediate histologic correlation. The detection rates for grade 1 cervical intraepithelial neoplasia (CIN1) were similar among the 3 methods (37.1%‐41.8%), while the overall detection rates for grade 2 CIN or higher (CIN2+) were much lower (5.1%‐9.0%). CIN2+ detection rates were similar among different age groups for HC2 testing, whereas for Cervista and Aptima testing, detection rates were highest in the youngest group (age <25 years) and decreased with age. Statistical analyses revealed that the CIN2+ detection rate was significantly higher in younger women who were tested with Cervista. Conclusion These hrHPV testing methods revealed low CIN2+ detection rates in women with ASC‐US/hrHPV+ but demonstrated different patterns when using age‐stratified analyses. The high rate of CIN2+ detection in women aged <25 years by the Cervista and Aptima testing platforms, but not the HC2 platform, suggest that continual investigation of FDA‐approved clinical hrHPV testing—especially large‐scale, long‐term, prospective studies—is needed to evaluate the best options for management of ASC‐US/hrHPV+ in this age group.

Evaluation of High-Risk Human Papillomavirus Testing in Women with Atypical Squamous Cells of Undetermined Significance to Detect Cervical Intraepithelial Neoplasia

“…We understand that there are some serious limitations in this study. Although the prevalence of abnormal findings in this study seemed to be reasonable compared with those in some previous studies, 10 the sample size of this study may be small. Moreover, in this retrospective study, we cannot verify the presence of diagnosis difference or reliability among physicians.…”

Prevalence of genotype-specific human papillomavirus in cytology specimens and cervical biopsies, and its implication in cervical cancer risk stratification: a retrospective study of 10647 cases

“…In addition, HPV-16 and 18/45 accounted for 47.1% of all CIN2+, and 79.8% of all SCC with HPV-16 alone accounting for 69.8% of all SCC. Thus, we are providing further evidence that persistent hrHPV infection is the primary cause of high-grade cervical disease, and persistent infection with HPV-16 and 18/45 pose higher risk for CIN2+ and SCC compared to other high-risk genotypes 31 , 35 - 37 . Our findings are consistent with those findings in the POBASCAN and ATHENA trials 25 , 26 , both of which demonstrated higher risk for precancer and cancer in HPV-16/18+ women than women positive for other high-risk genotypes.…”

“…A pooled hrHPV assay has been routinely used worldwide to stratify risks in women with ASC-US 31 . The ATHENA study demonstrated that HPV genotyping may be a promising risk stratification tool in aiding management decision-making for women with ASC-US 36 .…”

“…Immunohistochemical staining with p16 and Ki-67 was also liberally used to support the CIN2/3 diagnoses. In patients with more than one tissue sample, the highest grade diagnosis was recorded 31 .…”

Scrutinizing high‐risk patients from ASC‐US cytology via a deep learning model

“…Additionally, suspected glandular epithelial lesions are commonly diagnosed as atypical glandular cells (AGC), and they rarely occur in cases diagnosed as ASC‐US. The adenocarcinoma ratio among women with ASC‐US cytology is 0.18% and 0.1% according to statistics from our center and an American medical center, respectively 17,35 . Therefore, only squamous cells were used in training our model because training data contained annotated cells from cases diagnosed as LSIL, HSIL, and SCC.…”