Predictive Value of Squamo-Columnar Junction Markers and p16ink4 in Multi-Observer Classification of Cervical Precursor Lesions
Cynthia Jimenez, Brooke E Howitt, Marisa R Nucci, Bradley J Quade, Christopher P Crum, Michael Herfs. Brigham and Women's Hospital, Boston, MA; University of Liege, Liege, Belgium
Background: Currently p16ink4 (p16) staining is recommended for SILs bordering on CIN2, a positive result warranting a diagnosis of HSIL (CIN2). Recent studies have shown that most HSILs express squamo-columnar junction (SCJ) markers (krt7,ARG2) and that two types of LSILs exist; one SCJ marker (-) and associated with a very low risk of HSIL on followup, and one SCJ marker (+), at greater likelihood to be both diagnostically problematic and manifest with an HSIL outcome. This study compared the blind histologic assessment of SILs by three different pathologists against a standard set by an experienced pathologist, SCJ biomarkers and p16 staining pattern. The goal was to address theoretical strengths and weaknesses of these biomarkers.
Design: 158 SILs were classified as LSIL (101) or HSIL (57) and immunostained for Krt7 and p16. Positive p16 staining occupied more than one third of the epithelial thickness and was continuous throughout the lesion. Lesions were subdivided into three categories; SCJ+/p16+ HSIL (57), SCJ+/p16+ LSiL(20), and SCJ- LSIL (81), signifying a descending order of "risk" for HSIL outcome. Each of three observers reviewed the biopsies and scored them as LSIL, HSIL and indeterminate. The results for each observer was compared to the standard. Theoretically, underdiagnosis of SCJ+HSILs increased the risk of recurrence; overdiagnosis of SCJ-, p16+ LSILs increased the risk of unneccesary cone biopsy.
Results: For all observers, an LSIL diagnosis correlated with descending order of risk category (p<.001). Table 1 summarizes the percentage of cases classified as HSIL or indeterminate in each group by each observer.