TU-C-17A-11: Progressive Knowledge Modeling for Pelvic IMRT/VMAT Treatment Planning.

PURPOSE: To investigate the feasibility of progressive knowledge modeling for IMRT/VMAT treatment planning for multiple cancer types in the pelvic region. METHODS: The treatment planning knowledge model quantifies correlations between patient anatomical features and the OAR dose sparing features. The model is trained with prior plans using a stepwise regression machine learning technique. The progressive modeling process starts with 20 low risk prostate plans (type 1) which offer simplest PTV-OAR geometry. Cases with more complex PTV-OAR anatomies (prostate with lymph node, type 2 and anal rectal, type 3) are added to the training dataset one by one until the model prediction accuracies reach a plateau. The starting point of the plateau also defines the minimum numbers of type 2 and 3 training cases required for modelling. The DVHs predicted by the knowledge model for bladder, femoral heads and rectum were validated by 20, 9 and 18 cases with type 1, 2, and 3 geometries, respectively (rectum DVHs are omitted for type 3). Mean and standard deviation of differences between the dosimetric parameters sampled from the DVHs and the corresponding actual plan values measure the prediction accuracy of the model. Further, the accuracy was also compared with the single-type models which were trained by single type cases. RESULTS: Prediction accuracy reaches plateau when 6 type 2 and 8 type 3 cases were added to the training dataset. The determination coefficients R2 (should be square, font) for the OARs are: Bladder 0.90, rectum 0.64, and femoral heads 0.82. The prediction accuracies by the multiple-type model and single-type model have no significant differences by F-test (p-value: bladder: 0.58, femoral head: 0.44, rectum: 0.97). CONCLUSION: Progressive knowledge modeling of OAR sparing for multiple cancer types in pelvic region is feasible and has comparable accuracy to single cancer type modeling. Partially supported by NIH/NCI under grant #R21CA161389 and a master research grant by Varian Medical System.

Duke Scholars

Author Qingrong Jackie Wu Radiation Oncology