Conservative management with progestin is a treatment option for atypical hyperplasia (AH). However, pathologic diagnosis of residual/recurrent lesions is often problematic because of the profound morphologic changes induced by progestin and the lack of established diagnostic criteria for progestin-treated residual AH.
We conducted a longitudinal study of 265 endometrial biopsies from 54 patients with a history of AH on progestin therapy. Patient outcomes were divided into 3 categories after morphologic review and immunohistochemical staining with phosphatase and tensin homolog (PTEN) and paired box 2 (PAX2): (1) persistent or residual disease; (2) recurrent disease; (3) complete response. All specimens were classified into 3 categories based on morphology: (1) persistent/recurrent disease (nonresponse), (2) morphologically uncertain response, (3) optimally treated (complete response). The staining patterns of PTEN/PAX2 were tracked over time in individual patients and correlated with morphologic findings before and after progestin therapy.
Our data showed that aberrant expression patterns of PTEN and/or PAX2 were identified in 48 (88.9%) of the 54 primary biopsies and persisted in persistent/recurrent AH across serial endometrial biopsies (n=99, P<0.00001), while normal PTEN and PAX2 expressions were consistently observed in optimally treated cases (n=84, P<0.00001). More importantly, follow-up biopsies that showed a morphologically uncertain response but a PTEN/PAX2 expression pattern identical to the initial biopsy were significantly correlated with persistent or recurrent disease (n=18, P=0.000182), as evidenced by areas with morphologic features diagnostic of AH on subsequent biopsy.
Biomarker PTEN/PAX2 signatures offer a valuable diagnostic aid to identify residual AH in progestin-treated endometrial samples for which the biomarker status from preprogestin treated AH is known. The findings of this study are promising for a possible future change of diagnostic practice.