Multi-center Dose Prediction Using Attention-aware Deep learning Algorithm Based on Transformers for Cervical Cancer Radiotherapy

Abstract

Aims

Accurate dose delivery is crucial for cervical cancer volumetric modulated arc therapy (VMAT). We aimed to develop a robust deep learning (DL) algorithm for fast and accurate dose prediction of cervical cancer VMAT in multi-center datasets and then explore the feasibility of the DL algorithm to endometrial cancer VMAT with different prescriptions.

Materials and methods

We proposed the AtTranNet algorithm for 3D dose prediction. A total of 367 cervical patients were enrolled in this study. 322 cervical patients from 3 centers were randomly divided into 70%, 10%, 20% as training, validation, testing sets. 45 cervical patients from another center were used as external testing. Moreover, 70 patients of endometrial cancer with different prescriptions were further used to test the model. Prediction precision was evaluated by dosimetric difference, dose map and dose volume histogram metrics.

Results

The prediction results were all clinically acceptable. The mean absolute error within the body in internal testing were 0.66±0.63%. The maximum |δD| for PTV was observed in D98, which is 1.24 ± 2.73 Gy. The maximum |δD| for OARs was observed in Dmean of bladder, which is 4.79 ±3.14 Gy. The maximum |δV| were observed in V40 of pelvic bones, which is 4.77±4.48 %.

Conclusion

AtTranNet showed the feasibility and reasonable accuracy in the dose prediction for cervical cancer in multi-center. The model can also be generalized for endometrial cancer with different prescriptions without any transfer learning.