Introduction To create and clinically validate knowledge‐based planning (KBP) models for gynaecologic (GYN) and rectal cancer patients. Assessment of ecologic generalisability and predictive validity of conventional planning versus single calculation KBP was reviewed against practical metrics of planning time (PT) and radiation oncologist plan preference. Method Study cohorts were 34 and 42 consecutively treated GYN and rectal cancer patients dosimetrically archived within the centre’s research databank. For model training, structures and dose distributions from 22 and 32 GYN and rectal volumetric‐modulated arc therapy (VMAT) plans were used in RapidPlan?. Prescription doses ranged from 45 to 60Gy in 25 fractions using a simultaneous integrated boost to 2–4?targets and up to 9 organ‐at‐risk volumes. For model validation, 12 GYN and 10?rectal were independent of the archive and a single pass KBP VMAT plan was created. Each plan was evaluated against the archived treated plan under blinded conditions for radiation oncologist preference using standard dosimetric quality parameters. Results All 22 plans generated in the KBP validation cohort met pre‐set GYN and rectal cancer dosimetric quality metrics. Fifty per cent of GYN plans and eighty per cent of rectal plans were judged superior to the manually optimised plans. KBP reduced PT considerably for both tumour sites. Conclusion Single pass KBP for GYN and rectal cancer patients produced clinically acceptable treatment plans which were non‐inferior to conventionally optimised plans in 14 of 22 cases. Efficiencies captured by KBP will have predictable impacts on institutional workflows and resource allocation to facilitate adaptive planning.
展开▼
