Incremental diagnostic value of serum progesterone combined with transvaginal ultrasound in differentiating early ectopic pregnancy: a single-center real-world study

ObjectiveTo evaluate serum progesterone combined with transvaginal ultrasound for distinguishing ectopic pregnancy (EP) in patients with pregnancy of unknown location (PUL), and to assess diagnostic efficacy across risk stratifications.MethodsInternally unvalidated analysis included 148 PUL patients (January 2022–December 2025). Final diagnosis by surgical pathology or follow-up divided patients into EP (49) and non-EP (99; intrauterine pregnancy or natural regression). All underwent initial transvaginal ultrasound to confirm PUL. The prediction point was the 48-hour reassessment. Predictors were abnormal adnexal findings on follow-up ultrasound at 48 h, serum progesterone, and 48-hour β-hCG results available at that visit. The intended application is risk re-stratification and follow-up planning at 48-hour reassessment, not replacement of surgical/histopathological diagnosis or immediate treatment. Progesterone and β-hCG were measured; groups were compared. Logistic regression screened independent predictors. ROC curves evaluated single and combined indicators. We assessed if progesterone improved the ultrasound + M6NP model, with M6NP defined as M6 excluding initial progesterone.ResultsThe EP group had lower serum progesterone, initial β-hCG, 48-hour β-hCG change rate, and endometrial thickness, and higher abnormal adnexal ultrasound detection (P < 0.05). Multivariate analysis showed log progesterone, abnormal adnexal ultrasound, previous EP, and M6NP high-risk classification (≥5%) were independent predictors (P < 0.05). All AUC values are apparent/in-sample estimates. Ultrasound + M6NP had an apparent AUC of 0.913; adding progesterone increased it to 0.947. In the M6NP high-risk subgroup, progesterone plus ultrasound had an apparent AUC of 0.945 and specificity of 91.67%. Because the model was developed and evaluated in the same dataset without bootstrap or internal validation, estimates are preliminary and require external validation.ConclusionInternally unvalidated analysis, serum progesterone combined with transvaginal ultrasound was associated with improved apparent diagnostic accuracy for EP in PUL, especially within M6NP risk stratification. These estimates are apparent/in-sample, preliminary, and hypothesis-generating.