Indication-stratified histopathological yield and positive predictive value of sonographic and hysteroscopic impressions in women undergoing hysteroscopy: A 1551-case tertiary center cohort study

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Hysteroscopy is central to evaluating suspected intrauterine pathology, yet the histopathological yield of referral indications and the predictive value of ultrasound and hysteroscopic impressions in tertiary practice remain insufficiently defined. We retrospectively reviewed consecutive women aged ≥18 years who underwent diagnostic or operative hysteroscopy at a tertiary referral center in Ankara, Türkiye, between January and December 2025. Referral indication, transvaginal ultrasound impression, hysteroscopic impression, and final histopathology were recorded. Category-specific positive predictive values (PPVs) were calculated with Wilson 95% confidence intervals (CIs). The sonographic impression was additionally dichotomized against histopathology at both a strict and a broad threshold, and sensitivity, specificity, and likelihood ratios were derived. Reporting followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. Among 1551 procedures, 1250 women were pre/perimenopausal, and 301 were postmenopausal. Final histopathology was available in 1455 cases (93.8%). The most common diagnoses were endometrial polyp (957; 65.8%), benign endometrium (477; 32.8%), and endometrial carcinoma (21; 1.4%). Indications differed by menopausal status ( P < .001): abnormal uterine bleeding predominated before menopause (908; 72.6%), whereas postmenopausal bleeding predominated after menopause (144; 47.8%). Indication-stratified yield also differed significantly ( P < .001). Endometrial carcinoma was rare before menopause and clustered in postmenopausal women with bleeding. Among postmenopausal women, carcinoma was found in 11.1% (16/144) of those referred with bleeding but in only 1.3% (2/157) of those referred for other reasons; among pre/perimenopausal women, the corresponding figures were 0.3% (3/884) and 0% (0/270). Bleeding, therefore, raised carcinoma yield only after menopause. Ultrasound suggesting an endometrial polyp was confirmed in 86.4% (603/698; 95% CI: 83.6–88.7), while ultrasound suspicious for malignancy predicted carcinoma in 85.7% (18/21; 95% CI: 65.4–95.0). A hysteroscopic impression of an endometrial polyp showed a PPV of 90.2% (202/224; 95% CI: 85.6–93.4). For polyp detection, a definite sonographic impression was specific but insensitive (specificity 80.9%, sensitivity 63.0%); admitting suspected polyps raised sensitivity to 95.0%, while specificity fell to 50.2% (negative likelihood ratio, 0.10). Definite focal or malignancy-suspicious ultrasound and hysteroscopic impressions carried a high PPV. Carcinoma risk was governed by the combination of menopausal status and bleeding rather than by bleeding alone, which supports triage based on both factors jointly. Histopathological confirmation remains indispensable.

This publication has 15 references indexed in Scilit: