Establishment of a Predictive Model for the Efficacy of High-Intensity Focused Ultrasound in the Treatment of Uterine Fibroids.

OBJECTIVES

High-intensity focused ultrasound (HIFU) has demonstrated efficacy as a non-invasive treatment for uterine fibroids, though individual variability exists. This study aims to develop a risk scoring model using clinical and biochemical features to predict HIFU treatment outcomes.

METHODS

This study collected clinical data from patients receiving HIFU treatment, including demographic characteristics, clinical symptoms, treatment information, and biochemical indicators. A risk scoring model was constructed using the random forest analysis method, and its performance was evaluated. Meanwhile, the impact of risk models and other factors on the efficacy of HIFU was evaluated. Furthermore, the interrelationships between the risk model and other factors were explored through interaction analysis. Finally, a nomogram was developed to evaluate its clinical utility.

RESULTS

The risk model, 4 or more treatments, age, and tumor necrosis factor levels were identified as independent influencing factors, with the risk model demonstrating the best performance (area under the curve (AUC) = 0.693). Interaction analysis revealed a significant synergistic effect between the risk model and receiving 4 or more treatments. The nomogram analysis indicated that lower risk scores and fewer treatment sessions were associated with better HIFU treatment outcomes. The receiver operating characteristic curves and calibration curves in both the training and validation sets demonstrated good performance of the nomogram.

CONCLUSIONS

This study successfully constructed a risk scoring model based on clinical features and biochemical indicators, which can effectively predict the efficacy of HIFU treatment for uterine fibroids. There is a significant interaction between the risk model and 4 or more treatments. The constructed nomogram provides strong support for individualized treatment.