Analyzing human decisions in IGRT of head-and-neck cancer patients to teach image registration algorithms what experts know

Abstract

BackgroundIn IGRT of deformable head-and-neck anatomy, patient setup corrections are derived by rigid registration methods. In practice, experienced radiation therapists often correct the resulting vectors, thus indicating a different prioritization of alignment of local structures. Purpose of this study is to transfer the knowledge experts apply when correcting the automatically generated result (pre-match) to automated registration.MethodsDatasets of 25 head-and-neck-cancer patients with daily CBCTs and corresponding approved setup correction vectors were analyzed. Local similarity measures were evaluated to identify the criteria for human corrections with regard to alignment quality, analogous to the radiomics approach. Clustering of similarity improvement patterns is applied to reveal priorities in the alignment quality.ResultsThe radiation therapists prioritized to align the spinal cord closest to the high-dose area. Both target volumes followed with second and third highest priority. The bony pre-match influenced the human correction along the crania-caudal axis. Based on the extracted priorities, a new rigid registration procedure is constructed which is capable of reproducing the corrections of experts.ConclusionsThe proposed approach extracts knowledge of experts performing IGRT corrections to enable new rigid registration methods that are capable of mimicking human decisions. In the future, the deduction of knowledge-based corrections for different cohorts can be established automating such supervised learning approaches.

Evaluation of intrafraction couch shifts for proton treatment delivery in head‐and‐neck cancer patients: Toward optimal imaging frequency

“…6,7 Because of the high sensitivity of proton dose distribution to the patient setup, IGRT setup becomes the ultimate choice to treat each fraction of the treatment plan. 4,[8][9][10] In-room and onboard imaging methods obtain information on target position and movement in intra-and interfractions, compare these images with reference imaging taken initially, and give feedback to optimize patient setup and target localization. Radiation oncology incident learning systems have demonstrated that incorrect or omitted patient shifts during treatment are common near misses or incidents in radiation therapy.…”

“…Several studies investigated the setup accuracy comparison based on cone-beam computed tomography (CBCT) and kilovoltage (kV) images. 10,12,13,[15][16][17] Kraan et al 7 found increased target dose deterioration with increasing setup errors by introducing a systematic shift of isocenter by 2 mm. They also found that the treatment intent of 90% of population to have D98% > 95% of the prescription dose was no longer satisfied.…”

“…However, the physical properties of protons make IMPT more sensitive than photons to planning uncertainties 6,7 . Because of the high sensitivity of proton dose distribution to the patient setup, IGRT setup becomes the ultimate choice to treat each fraction of the treatment plan 4,8–10 . In‐room and onboard imaging methods obtain information on target position and movement in intra‐ and interfractions, compare these images with reference imaging taken initially, and give feedback to optimize patient setup and target localization.…”

Head and neck radiotherapy on the MR linac: a multicenter planning challenge amongst MRIdian platform users

“…In recent years various advances in image-guided radiotherapy (IGRT) have led to improved precision and accuracy in radiation treatment delivery for head and neck cancer (HNC) [1][2][3]. Three-dimensional image guidance for patient setup allows corrections and could also reduce fail-Fig.…”

The implementation of MOSAIQ‐based image‐guided radiation therapy image matching within radiation therapy education

“…An important task was to undertake a review of the literature to learn from several studies that reported and discussed the rationale for image matching competency training in clinical environments. Several articles reviewed highlighted it is imperative RTs are trained in IGRT, particularly image analysis and decision‐making 7‐12 . An enquiry into the clinical use of IGRT technology and MOSAIQ OIS was important as we needed to understand how we would embed MOSAIQ and actual clinical practice through a range of teaching and learning strategies, whilst understanding the tools which would effectively support teaching.…”