Knowledge-based planning for multi-isocenter VMAT total marrow irradiation

Abstract

PurposeTotal marrow irradiation (TMI) involves optimization of extremely large target volumes and requires extensive clinical experience and time for both treatment planning and delivery. Although volumetric modulated arc therapy (VMAT) achieves substantial reduction in treatment delivery time, planning process still presents a challenge due to use of multiple isocenters and multiple overlapping arcs. We developed and evaluated a knowledge-based planning (KBP) model for VMAT-TMI to address these clinical challenges.MethodsFifty-one patients previously treated in our clinic were selected for the model training, while 22 patients from another clinic were used as a test set. All plans used a 3-isocenter to cover sub-target volumes of head and neck (HN), chest, and pelvis. Chest plan was performed first and then used as the base dose for both the HN and pelvis plans to reduce hot spots around the field junctions. This resulted in a wide range of dose-volume histograms (DVH). To address this, plans without the base-dose plan were optimized and added to the library to train the model.ResultsKBP achieved our clinical goals (95% of PTV receives 100% of Rx) in a single day, which used to take 4-6 days of effort without KBP. Statistically significant reductions with KBP were observed in the mean dose values to brain, lungs, oral cavity and lenses. KBP substantially improved 105% dose spillage (14.1% ± 2.4% vs 31.8% ± 3.8%), conformity index (1.51 ± 0.06 vs 1.81 ± 0.12) and homogeneity index (1.25 ± 0.02 vs 1.33 ± 0.03).ConclusionsKBP improved dosimetric performance with uniform quality. It reduced dependence on planner experience and achieved a factor of 5 reduction in planning time to produce quality plans to allow its wide-spread clinical implementation.

A comparison of knowledge-based dose prediction approaches to assessing head and neck radiotherapy plan quality

“…This notion has emphasized the importance of continued clinical efforts to improve the consistency and quality of radiotherapy planning using a variety of planning tools. [2][3][4][5][6][7][8][9] These include traditional knowledge-based planning tools that predict dosevolume histograms (DVHs), and deep learning-based tools, that predict dose distributions on computed tomography images for later DVH calculation. The more traditional knowledge-based planning tools typically use a library of plans from previously treated patients and develop models associating geometric features with their corresponding dosimetry to predict possibly achievable dosimetry for a new patient.…”

Effects of model size and composition on quality of head‐and‐neck knowledge‐based plans

“…Several clinics that utilize KBP demonstrated that KBP plans were non-inferior to manual planning and that planning time could be decreased. [13][14][15][16] To create high-quality plans using a KBP model, a database of "good" plans is required to train the model. Quality filtering is suggested to improve trained models.…”

“…These predictions can then be used to automatically generate optimization objectives, decreasing the amount of trial and error needed in manual treatment planning. Several clinics that utilize KBP demonstrated that KBP plans were non‐inferior to manual planning and that planning time could be decreased 13–16 …”

“…Several clinics that utilize KBP demonstrated that KBP plans were non‐inferior to manual planning and that planning time could be decreased. 13 , 14 , 15 , 16 …”

Considerations for intensity modulated total body or total marrow and lymphoid irradiation

“…Over the last two decades, multiple centers have developed and implemented intensity-modulated total body irradiation (IM-TBI), total marrow irradiation (IM-TMI) and total marrow and lymphoid irradiation (IM-TMLI) using modern linear accelerators (linac) and image-guidance [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] . Common to these techniques are the treatment of the patient in the supine position at the machine isocenter, volumetric imaging for image guidance, and inverse optimization within a treatment planning system (TPS).…”