Abstract
OBJECTIVE: To establish a predictive model for distinguishing testicular benign or yolk sac tumors in children. METHODS: We retrospectively analyzed data for 119 consecutive patients with unilateral testicular tumors treated at a single institution from June 2014 to July 2020. The patients were divided into the benign (n = 90) and yolk sac (n = 29) tumor groups based on the pathological diagnosis. We recorded patient age, serum markers [serum alpha-fetoprotein (AFP), human chorionic gonadotropin], and tumor ultrasonic findings (maximum diameter, ultrasonic echo, blood flow signal). Predictive factors were identified using descriptive statistical methods. A nomogram was established for preoperative prediction. An additional 46 patients were used as a validation cohort to verify the model. RESULTS: Patients with testicular yolk sac tumors were younger (median age: 14.0 vs. 34.0 months, P = 0.001) and had a higher incidence of elevated AFP levels (93.1% vs. 2.2%, P < 0.001). Ultrasonography indicated that testicular yolk sac tumors tended to have larger maximum diameters (26.5 ± 11.3 vs. 16.6 ± 9.2 cm, P < 0.001), a higher proportion of hypoechoic masses (44.8% vs. 8.9%, P < 0.001), and a higher incidence of masses with strong blood flow signals (93.1% vs. 5.6%, P < 0.001). A nomogram based on age, AFP levels, and ultrasound blood flow signals effectively predicted the probability of yolk sac tumor in children, with an accuracy of 0.98 (95% confidence interval: 0.984-1.003). The Brier score of the nomogram was 0.0002. CONCLUSION: A nomogram based on age, AFP levels, and ultrasound blood flow signals can effectively predict the probability of testicular yolk sac tumor preoperatively, aiding in clinical decision-making and patient counseling.