Abstract
Gestational choriocarcinoma is a highly malignant form of gestational trophoblastic neoplasia characterized by early vascular invasion and a strong tendency for widespread metastasis. To date, there is no consensus in the FIGO recommendations regarding when chemotherapy should be initiated following diagnosis. This study aimed to evaluate the impact of chemotherapy on survival and develop machine learning (ML) prognostic models for patients with gestational choriocarcinoma. We analyzed data from the SEER database [2000-2020]. Patients with histologically confirmed GC arising from the placenta were included, while those with other malignancies or missing key data were excluded. We conducted a Cox regression analysis for prognostic factors and developed ML models (using 5 algorithms) to predict 5-year survival rates. A validation method incorporating the area under the curve of the receiver operating characteristic curve was used to validate the accuracy and reliability of the ML models. We also investigated the role of multiple therapeutic options using the Kaplan-Meier survival analysis. This study included 732 patients with a median age of 32 years (54.5% ≥30 years); most were White (66.4%), and 44.3% had metastatic disease at diagnosis. Of these, 283 received chemotherapy, 116 underwent surgery alone, and 333 underwent both surgery and chemotherapy. Survival analysis showed no significant differences in survival between the treatment modalities. Multivariate Cox regression analysis identified older age, metastasis, and marital status as significant prognostic factors. Among the ML models, Random Forest Classifier achieved the highest performance. Feature importance analysis identified age, marital status, and metastasis as the most influential survival factors. The study suggests that chemotherapy may not have benefit for survival. Further multicenter prospective studies are needed to evaluate the importance of chemotherapy initiation.