Abstract
OBJECTIVE: Acupuncture is acknowledged for its safety and effectiveness in the process of frozen embryo transfer (FET) to improve pregnancy outcomes. The study aimed to develop a clinical prediction model to predict the probability of clinical pregnancy after acupuncture treatment during FET and to identify the most predictive characteristics. METHODS: Two clinical trials on acupuncture treatment during FET containing a total of 390 patients (315 in Trial 1 and 75 in Trial 2) were involved for data training. Eighty baseline clinical characteristics were collected from patients in Trial 1, and the support vector classification (SVC) model was created to predict the improvement of FET clinical pregnancy by acupuncture. Trial 1 was utilized as the internal validation set (divided into internal test and validation sets in a 7:3 ratio), whereas Trial 2 was used as the external validation set to assess the external generalizability of this clinical prediction model. RESULTS: In Trial 1, the prediction model achieved an accuracy of 0.778, a precision of 0.821, a recall score of 0.807, an f1 score of 0.814, and an AUC of 0.772 in predicting the acupuncture response. The in-hospital cycle, vascularized flow index, and transferred embryo number were the essential predictive features identified by the SVC model. For Trial 2, an accuracy of 0.74, a precision of 0.625, a recall score of 0.625, an f1 score of 0.625, an AUC of 0.713 were shown in the LSVC model. CONCLUSION: The clinical prediction model constructed through this study may help physicians determine in advance how patients will respond to acupuncture before FET and provide accurate treatment plans for acupuncture.