Accuracy of ultrasound in detecting retained products of gestation after second-trimester medical termination of pregnancy

Objectives: To determine the diagnostic accuracy of ultrasound parameters in detecting retained products of gestation (RPOG) after second-trimester medical termination of pregnancy (MTOP) and to identify predictors of RPOG. Methods: Medical records of women who underwent second-trimester MTOP, followed by ultrasound evaluation and surgical evacuation for suspected RPOG, at Queen Elizabeth Hospital, Hong Kong, between 1 January 2019 and 31 December 2024 were reviewed. Clinical features were documented, as were ultrasound features, including endometrial thickness, the presence of endometrial masses and heterogeneous echogenicity, and endometrial vascularity. The diagnostic accuracy of endometrial thickness in detecting RPOG was assessed using the area under the curve. Diagnostic performances of other ultrasound variables were also determined. Multivariate logistic regression was performed to determine predictors of RPOG. Results: In total, 237 women were included in the analysis. Of these, 196 (82.7%) had histopathology-confirmed RPOG and 41 (17.3%) had no histological evidence of RPOG. The RPOG-positive group had a lower median gestational age at termination (16 vs 19 weeks, p=0.008), a higher percentage of women with prior surgical evacuation (27.6% vs 7.3%, p=0.006), and a greater median endometrial thickness (22.5 vs 21.0 mm, p=0.018). The area under the curve for endometrial thickness in diagnosing RPOG was 0.613. The highest Youden index (0.205) was observed at an endometrial thickness of 27.9 mm, yielding a sensitivity of 23% and specificity of 97.6%. The presence of intrauterine mass and heterogeneous echogenicity demonstrated high sensitivity (72.4% and 81.1%, respectively) but low specificity (22.0% and 7.3%). Among 11 women who underwent colour Doppler ultrasound, vascularity yielded a sensitivity of 90% and specificity of 100%. Multivariate logistic regression identified lower gestational age at termination (adjusted odds ratio [OR]=0.854, p=0.004), prior abortion requiring surgical evacuation (adjusted OR=5.993, p=0.016), and increased endometrial thickness (adjusted OR=1.104 per mm, p=0.003) as independent predictors of RPOG. Conclusion: Increased endometrial thickness is an ultrasound predictor of RPOG after second-trimester MTOP, with high specificity but low sensitivity at a high cut-off value. The favourable diagnostic performance of vascularity supports routine use of colour Doppler ultrasound. Clinicians should adopt a comprehensive, individualised approach that integrates clinical and ultrasound parameters.