Fully automated treatment planning for head and neck radiotherapy using a voxel-based dose prediction and dose mimicking method

Abstract

Abstract. Recent works in automated radiotherapy treatment planning have used machine learning based on historical treatment plans to infer the spatial dose distribution for a novel patient directly from the planning image. We present an atlas-based approach which learns a dose prediction model for each patient (atlas) in a training database, and then learns to match novel patients to the most relevant atlases. The method creates a spatial dose objective, which specifies the desired dose-pervoxel, and therefore replaces any requirement for specifying dose-volume objectives for conveying the goals of treatment planning. A probabilistic dose distribution is inferred from the most relevant atlases, and is scalarized using a conditional random field to determine the most likely spatial distribution of dose to yield a specific dose prior (histogram) for relevant regions of interest. Voxel-based dose mimicking then converts the predicted dose distribution to a deliverable treatment plan dose distribution. In this study, we investigated automated planning for right-sided oropharaynx head and neck patients treated with IMRT and VMAT. We compare four versions of our dose prediction pipeline using a database of 54 training and 12 independent testing patients. Our preliminary results are promising; automated planning achieved a higher number of dose evaluation criteria in 7 patients and an equal number in 4 patients compared with clinical. Overall, the relative number of criteria achieved was higher for automated planning versus clinical (17 vs 8) and automated planning demonstrated increased sparing for organs at risk (52 vs 44) and better target coverage/uniformity (41 vs 31). The novel dose prediction method with dose mimicking can generate deliverable treatment plans in 12-13 minutes without any user interaction. The method is a promising approach for fully automated treatment planning and can be readily applied to different treatment sites and modalities.

Knowledge‐based automated planning for oropharyngeal cancer

“…This paper adds to the growing literature of KBP‐based automated pipelines, by building on top of the existing clinical planning paradigm that uses an IPP. Our approach combines knowledge‐based planning with inverse optimization to automatically generate treatment plans from patient anatomy.…”

“…One step toward automation is knowledge‐based planning (KBP), which comprises a group of methods that learn from historical treatment plans and predict attributes of desirable plans for new patients . KBP predictions can be input into an automated planning engine to produce a treatment plan, but existing methods introduce a new planning paradigm where planners are typically unable to adjust the final plan using familiar inverse planning techniques (e.g., adjusting objective function weights) . Adjustability of plans is the hallmark of the current clinical planning paradigm, where planners alternate between solving an inverse planning problem (IPP) and tuning model parameters like objective function weights, which quantify the relative importance of various objectives that a planner must optimize.…”

Knowledge‐based dose prediction models for head and neck cancer are strongly affected by interorgan dependency and dataset inconsistency

“…Therefore, achieved OAR doses can be highly institution‐, planner‐ and patient dependent. Proposed solutions are knowledge‐based (KB) treatment plan QA or KB (semi‐)automated treatment planning . With KB treatment plan QA, treatment plans of prior patients are used to predict the achievable doses for a new patient.…”

“…Some are commercially available. The models aim to predict either achievable dose metrics of interest, entire DVHs, or doses to individual voxels …”

Advanced Treatment Planning

“…The dose distribution of a previous plan is used as reference and a new optimization is performed with the objective to, e.g., minimize the doses to healthy structures under the constraint that the dose to each voxel or the DVH of each ROI must be at least as good as in the reference plan . Mimicking can also be used to convert automatically generated dose distributions to deliverable dose distributions and clinically achievable treatment plans with modulated beams or arcs …”