Evaluating the Use of Machine Learning to Predict Expert-Driven Pareto-Navigated Calibrations for Personalised Automated Radiotherapy Planning.

Featured Application: Fully automated and personalised radiotherapy treatment planning. Automated planning (AP) uses common protocols for all patients within a cancer site. This work investigated using machine learning to personalise AP protocols for fully individualised planning. A 'Pareto guided automated planning' (PGAP) solution was used to generate patient-specific AP protocols and gold standard Pareto navigated reference plans (MCO_gs) for 40 prostate cancer patients. Anatomical features related to geometry were extracted and two ML approaches (clustering and regression) that predicted patient-specific planning goal weights were trained on patients 1–20. For validation, three plans were generated for patients 21–40 using a standard site-specific AP protocol based on averaged weights (PGAP_std) and patient-specific AP protocols generated via regression (PGAP-ML_reg) and clustering (PGAP-ML_clus). The three methods were compared to MCO_gs in terms of weighting factors and plan dose metrics. Results demonstrated that at the population level PGAP_std, PGAP-ML_reg and PGAP-ML_clus provided excellent correspondence with MCO_gs. Deviations were either not statistically significant (p ≥ 0.05), or of a small magnitude, with all coverage and hotspot dose metrics within 0.2 Gy of MCO_gs and OAR metrics within 0.7% and 0.4 Gy for volume and dose metrics, respectively. When compared to PGAP_std, patient-specific protocols offered minimal advantage for this cancer site, with both approaches highly congruent with MCO_gs. [ABSTRACT FROM AUTHOR]