ESTRO 2022 - Abstract Book

S1450

Abstract book

ESTRO 2022

PO-1652 A new machine learning-based clinical tool for vertebral level verification and treatment planning

T. Netherton 1 , C. Nguyen 1 , C. Cardenas 2 , C. Chung 3 , A. Klopp 3 , L. Colbert 3 , D.J. Rhee 1 , C. Peterson 4 , R. Howell 1 , A. Aggarwal 5 , H. Simonds 6 , L. Court 1 1 University of Texas MD Anderson Cancer Center, Radiation Physics, Houston, USA; 2 University of Alabama Birmingham , Radiation Physics, Birmingham, USA; 3 University of Texas MD Anderson Cancer Center, Radiation Oncology, Houston, USA; 4 University of Texas MD Anderson Cancer Center, Biostatistics, Houston, USA; 5 Guy's and St. Thomas Hospital London, Radiation Oncology, London, United Kingdom; 6 Stellenbosch University, Radiation Oncology, Cape Town, South Africa Purpose or Objective To develop a palliative treatment planning paradigm that verifies all vertebral bodies and creates a treatment plan on diagnostic and simulation CT images. Materials and Methods Vertebrae from any CT and of any length are labeled (C1-L5) using two intendent deep learning models to mirror two different experts labeling the spine. Then, a UNet++ architecture was trained, validation, and tested so that each resulting vertebrae could be contoured using (n=220 CTs). Features from labeling and from contours were input into a random forest classifier to predict whether vertebrae were correctly labeled. This classifier was trained using CBCT, PET-CT, diagnostic, and simulation CT images (n=56 CTs, 752 contours). Autoplans were generated via scripting using auto-verified contours. Each model was combined into a framework to make a fully-automated treatment planning clinical tool. Three radiation oncologist scored plan quality on a new cohort of CT scans (n=60) on a 5-point scale. CTs varied in scan length, presence of surgical implants, imaging protocol, and metastatic burden.

Results Accuracy

for

automatic

labeling

was

94%

with

<

2.2mm

error.

Mean

Dice-Similarity

Coefficient was 85.0%(cervical), 90.3%(thoracic),

and 93.7%(lumbar). The

random forest classifier

predicted

mislabeling across various CT labeling errors (11/11) within treatment regions, including errors from patient plans (6/6) with atypical anatomy (e.g. T13, L6) were detected. Radiation oncologists scored 98% of simulation CT- and 92% of diagnostic CT-based plans as clinically acceptable or needing minor edits for patients with typical anatomy. End-to-end treatment planning time of the clinical tool was less than 8 minutes on average. scan types with precision-recall-AUC=0.82. All contouring and

Conclusion This novel method to automatically verify, contour, and plan palliative spine treatments is efficient, effective, and safe across various CT scan types.

PO-1653 Reproducibility of quantitative PET in radiotherapy setup for multicenter PET/MR in head/neck cancer