Squamous intraepithelial lesions in cervical tissue samples of limited adequacy and insufficient for grading as low or high grade: outcome, clinico-pathological correlates, and predictive role of p16INK4a and Ki67 biomarker staining.

ABSTRACT

OBJECTIVE: Cervical tissue samples of limited adequacy but with pathological features of squamous intraepithelial lesions (SIL) may not be gradable and result in a diagnosis of ungraded SIL (SILQ). SILQ outcome, clinico-pathological correlates, and the predictive role of biomarker staining are unknown. MATERIALS AND METHODS: Among 17,551 colposcopy attendees, 478 (2.7%) had SILQ. Glass slides of 472 were reviewed. Positive [high SIL (HSIL), adenocarcinoma in situ (AIS), or carcinoma] and negative [negative for intraepithelial lesion or malignancy (NILM) or low SIL (LSIL)] outcomes were based on the worst pathology in 24 months of follow-up. p16 and Ki67 immunohistochemistry of 80 random SILQ and 149 controls (44 NILM, 15 LSIL, 75 HSIL, and 15 AIS) was scored as unsatisfactory, positive, or negative. Biomarker and outcome status were correlated, and sensitivity, specificity positive predictive value (PPV), and negative predictive value (NPV) were calculated. RESULTS: Of the total cases, 332 (1.9%) were reviewed as SILQ, and follow-up for 329 was positive in 134 (41%). Atypical glandular cells, AIS, atypical squamous cells (cannot exclude HSIL), HSIL referral Pap test (70% vs. 47%, p < .001), and HSIL colposcopic impression (33% vs. 19%, p < .001) were more frequent among positive compared with negative outcomes. Best SILQ sensitivity (89%) and NPV (77%) occurred with combined biomarkers, and best specificity (52%) and PPV (58%) occurred with Ki67. All 4 performance metrics among the controls were high. CONCLUSIONS: The 2% frequency and 41% positive outcome highlight the clinical importance of SILQ. The referral Pap test and colposcopic impression could prioritize follow-up colposcopy for some SILQ, and negative staining with both biomarkers could eliminate further colposcopy in others.