Human Papillomavirus mRNA In Situ Hybridization Signal Patterns Are Not Predictive of High-Grade Squamous Intraepithelial Lesion (CIN2) Behavior

Grading cervical intraepithelial neoplasia is important for appropriate patient management. The management of CIN1 and CIN3 is well characterized; however, the management of CIN2 is controversial as the behavior is less predictable. HPV mRNA expression patterns within these lesions have been described; however, the predictive power of these staining patterns has not been studied. The ASCCP recommends that women <24 years with CIN2 are observed due to high likelihood of regression. In this study, we evaluated the staining patterns of CIN2 to determine whether they can be further stratified to predict regression or progression. A retrospective database search was performed from 2013 to 2017, identifying 26 patients <28 years old with CIN2. Exclusion criteria included patients who underwent surgical excision. The mean follow-up time was 10.5 months. HPV E6/E7 mRNA expression patterns were evaluated by RNAscope chromogenic in situ hybridization (CISH). The mRNA expression patterns were graded as pattern 1, superficial diffuse nuclear staining; pattern 2, superficial diffuse nuclear staining and scattered dot-like nuclear and granular cytoplasmic staining; and pattern 3, scattered dot-like nuclear and granular cytoplasmic staining. Thirteen cases had pattern 2 and 13 had pattern 3 staining. Seven of the pattern 2 cases regressed on follow-up Pap smear or biopsy, while six persisted or progressed. Eight of the pattern 3 cases regressed while five persisted or progressed. Therefore, pattern 2 staining was not predictive of regression (P > .9). HPV mRNA expression pattern was hypothesized to be predictive of CIN2 biologic behavior (ie, pattern 2 more likely to regress, pattern 3 likely to persist/progress); however, this small cohort with limited follow-up failed to confirm this hypothesis. This ongoing study will continue to review additional cases with follow-up to gain a larger statistical power and better understand the biologic potential of these lesions.