Purpose: To assist the organ contouring in current treatment planning procedures, we developed a deep learning-based framework to automatize the whole contouring process contributing to the clinical practice.
Methods: This framework was designed for the auto-segmentation of organs at risk (OARs) in head and neck, and pelvis sites, which consisted of a pre-processing module, a deep learning-based segmentation module, and a post-processing module. In the pre-processing module, the original image data was extracted from the clinical DICOM Image file, and then processed for the segmentation network learning. A U-Net like network was designed in the segmentation module, in which three input branches for three consecutive image slices were designed to consider the contour consistency among adjacent image slices, and a deep supervision strategy was integrated to speed up the convergency of the network. The segmentation results would be converted to a DCIOM Structure file in the post-processing module for treatment planning. Moreover, this framework is being developed toward physician-specific to comply with the contour style of each radiation oncologist separately.
Results: The organ-averaged Dice similarity coefficient (DSC) performance of this auto-segmentation framework is 0.71 and 0.74 for 16 head and neck and 11 pelvis OARs, respectively. Once a clinical DICOM Image data is input to the framework, a corresponding DICOM Structure file would be automatically generated containing the OAR segmentation results. The overall operation takes about 1-2 min per case without the need for additional manual interference.
Conclusion: A deep learning-based framework was developed for the auto-segmentation of head and neck, and pelvis OARs with promising contouring accuracy and operation time. This framework is ready to be applied in our clinical practice to improve the efficiency of current treatment planning procedures.
Funding Support, Disclosures, and Conflict of Interest: This research was supported by a grant from Varian Medical Systems (Palo Alto, CA).