Machine Learning for Predicting Clinician Evaluation of Treatment Plans for Left-Sided Whole Breast Radiation Therapy.

Purpose: The objective of this work was to investigate the ability of machine learning models to use treatment plan dosimetry for prediction of clinician approval of treatment plans (no further planning needed) for left-sided whole breast radiation therapy with boost. Methods and Materials: Investigated plans were generated to deliver a dose of 40.05 Gy to the whole breast in 15 fractions over 3 weeks, with the tumor bed simultaneously boosted to 48 Gy. In addition to the manually generated clinical plan of each of the 120 patients from a single institution, an automatically generated plan was included for each patient to enhance the number of study plans to 240. In random order, the treating clinician retrospectively scored all 240 plans as (1) approved without further planning to seek improvement or (2) further planning needed, while being blind for type of plan generation (manual or automated). In total, 2 × 5 classifiers were trained and evaluated for ability to correctly predict the clinician's plan evaluations: random forest (RF) and constrained logistic regression (LR) classifiers, each trained for 5 different sets of dosimetric plan parameters (feature sets [FS]). Importances of included features for predictions were investigated to better understand clinicians' choices. Results: Although all 240 plans were in principle clinically acceptable for the clinician, only for 71.5% was no further planning required. For the most extensive FS, accuracy, area under the receiver operating characteristic curve, and Cohen's κ for generated RF/LR models for prediction of approval without further planning were 87.2 ± 2.0/86.7 ± 2.2, 0.80 ± 0.03/0.86 ± 0.02, and 0.63 ± 0.05/0.69 ± 0.04, respectively. In contrast to LR, RF performance was independent of the applied FS. For both RF and LR, whole breast excluding boost PTV (PTV40.05Gy) was the most important structure for predictions, with importance factors of 44.6% and 43%, respectively, dose recieved by 95% volume of PTV40.05 (D95%) as the most important parameter in most cases. Conclusions: The investigated use of machine learning to predict clinician approval of treatment plans is highly promising. Including nondosimetric parameters could further increase classifiers' performances. The tool could become useful for aiding treatment planners in generating plans with a high probability of being directly approved by the treating clinician.

Automation of pencil beam scanning proton treatment planning for intracranial tumours.

PURPOSE: To evaluate the feasibility of comprehensive automation of an intra-cranial proton treatment planning. MATERIALS AND METHODS: Class solution (CS) beam configuration selection allows the user to identify predefined beam configuration based on target localization; automatic CS (aCS) will then explore all the possible CS beam geometries. Ten patients, already used for the evaluation of the automatic selection of the beam configuration, have been also employed to training an algorithm based on the computation of a benchmark dose exploit automatic general planning solution (GPS) optimization with a wish list approach for the planning optimization. An independent cohort of ten patients has been then used for the evaluation step between the clinical and the GPS plan in terms of dosimetric quality of plans and the time needed to generate a plan. RESULTS: The definition of a beam configuration requires on average 22 min (range 9-29 min). The average time for GPS plan generation is 18 min (range 7-26 min). Median dose differences (GPS-Manual) for each OAR constraints are: brainstem -1.60 Gy, left cochlea -1.22 Gy, right cochlea -1.42 Gy, left eye 0.55 Gy, right eye -2.33 Gy, optic chiasm -1.87 Gy, left optic nerve -4.45 Gy, right optic nerve -2.48 Gy and optic tract -0.31 Gy. Dosimetric CS and aCS plan evaluation shows a slightly worsening of the OARs values except for the optic tract and optic chiasm for both CS and aCS, where better results have been observed. CONCLUSION: This study has shown the feasibility and implementation of the automatic planning system for intracranial tumors. The method developed in this work is ready to be implemented in a clinical workflow.

Automatic genetic planning for volumetric modulated arc therapy: A large multi-centre validation for prostate cancer.

PURPOSE: The first clinical genetic autoplanning algorithm (Genetic Planning Solution, GPS) was validated in ten radiotherapy centres for prostate cancer VMAT by comparison with manual planning (Manual). METHODS: Although there were large differences among centres in planning protocol, GPS was tuned with the data of a single centre and then applied everywhere without any centre-specific fine-tuning. For each centre, ten Manual plans were compared with autoGPS plans, considering dosimetric plan parameters and the Clinical Blind Score (CBS) resulting from blind clinician plan comparisons. AutoGPS plans were used as is, i.e. there was no patient-specific fine-tuning. RESULTS: For nine centres, all ten plans were clinically acceptable. In the remaining centre, only one plan was acceptable. For the 91% acceptable plans, differences between Manual and AutoGPS in target coverage were negligible. OAR doses were significantly lower in AutoGPS plans (p < 0.05); rectum D15% and Dmean were reduced by 8.1% and 17.9%, bladder D25% and Dmean by 5.9% and 10.3%. According to clinicians, 69% of the acceptable AutoGPS plans were superior to the corresponding Manual plan. In case of preferred Manual plans (31%), perceived advantages compared to autoGPS were minor. QA measurements demonstrated that autoGPS plans were deliverable. A quick configuration adjustment in the centre with unacceptable plans rendered 100% of plans acceptable. CONCLUSION: A novel, clinically applied genetic autoplanning algorithm was validated in 10 centres for in total 100 prostate cancer patients. High quality plans could be generated at different centres without centre-specific algorithm tuning.