| Popis: |
PURPOSE: To reduce workload and inconsistencies in organ segmentation for radiation treatment planning, we developed and evaluated general and custom auto-segmentation models on CT for three major tumor sites using a well-established deep convolutional neural network (DCNN). METHODS AND MATERIALS: Five CT-based auto-segmentation models for 42 organs at risk (OARs) in head and neck (HN), abdomen (ABD) and male pelvis (MP) were developed using a full 3D DCNN architecture. Two types of DL models were separately trained using either general diversified multi-institutional datasets or custom well controlled single institution datasets. To improve segmentation accuracy, an adaptive spatial resolution approach for small and/or narrow OARs and a pseudo scan extension approach, when CT scan length is too short to cover entire organs, were implemented. The performance of the obtained models was evaluated based on accuracy and clinical applicability of the auto-segmented contours using qualitative visual inspection and quantitative calculation of dice similarity coefficient (DSC), mean distance to agreement (MDA), and time efficiency. RESULTS: The five DL auto-segmentation models developed for the three anatomical sites were found to have high accuracy (DSC ranging from 0.8 – 0.98) for 74% OARs and marginally acceptable for 26% OARs. The custom models performed slightly better than the general models, even with smaller custom datasets used for the custom model training. The organ-based approaches improved auto-segmentation accuracy for small or complex organs (e.g., eye lens, optic nerves, inner ears, and bowels). Compared with traditional manual contouring times, the auto-segmentation times, including subsequent manual editing, if necessary, were substantially reduced by 88% for MP, 80% for HN, and 65% for ABD models. CONCLUSIONS: The obtained auto-segmentation models, incorporating organ-based approaches were found to be effective and accurate for most OARs in the male pelvis, head and neck and abdomen. We have demonstrated that our multi-anatomical deep learning auto-segmentation models are clinically useful for radiation treatment planning. |
Plný text