785P - Online prognostic calculator for gestational trophoblastic neoplasia (GTN): A tool for personalized treatment planning

Background

GTN involves malignant entities, such as choriocarcinoma (GCC), epithelioid trophoblastic tumor (ETT), and placental site trophoblastic tumor (PSTT), all of which can metastasize and be fatal if not treated in a timely and effective manner. We aimed to develop machine-learning (ML)-based prognostic models for patients with GTN to improve treatment decisions and personalized medicine.

Methods

The SEER database provided the data used for this study’s analysis (2000–2019). Patients were excluded if they did not meet certain criteria, such as a histology-confirmed diagnosis, previous history of cancer or other concurrent malignancies, non-placental origins, or unknown data. Cox regression and ML algorithms were used to construct prognostic models for predicting the 5-year survival. Patient records were randomly divided into a training set (70%) and a validation set (30%). A validation method incorporating the area under the curve (AUC) of the receiver operating characteristic curve was used to validate the accuracy and reliability of ML models.

Results

This study included 725 patients, of whom 139 had ETT, 107 had PSTT, and 479 had GCC. The median age was 37 years, and the most common stage was distant metastasis (n=330). The average tumor size was 6.13 cm, and 75.9% of patients underwent chemotherapy, 57.9% underwent surgery, and 4.3% underwent radiotherapy. Multivariate Cox regression revealed that older age, marital status, metastasis, and radiotherapy were poor prognostic factors, whereas surgery was a good prognostic factor. ML models revealed that the most important factors that influenced the model prediction performance were tumor size, years of diagnosis, and race. The online calculator, which is based on ML models, allows for real-time personalized survival predictions, and can be accessed through the following link: https://sakhrshwayyat.shinyapps.io/Gestational_Trophoblastic_Neoplasia/.

Conclusions

Our study revealed the critical components of clinical decision-making. The designed models and online prediction resources can significantly facilitate personalized treatment planning and enhance the clinical outcomes of patients with GTN.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.