Logo image
A Technique to Enable Efficient Adaptive Radiation Therapy: Automated Contouring of Prostate and Adjacent Organs
Journal article   Open access   Peer reviewed

A Technique to Enable Efficient Adaptive Radiation Therapy: Automated Contouring of Prostate and Adjacent Organs

Daniel E. Hyer, Joseph Caster, Blake Smith, Joel St-Aubin, Jeffrey Snyder, Andrew Shepard, Honghai Zhang, Sean Mullan, Theodore Geoghegan, Benjamin George, …
Advances in radiation oncology, Vol.9(1), 101336
01/2024
DOI: 10.1016/j.adro.2023.101336
PMCID: PMC10801646
PMID: 38260219
url
https://doi.org/10.1016/j.adro.2023.101336View
Published (Version of record) Open Access

Abstract

Purpose The purpose of this work was to investigate the use of a segmentation approach that could potentially improve the speed and reproducibility of contouring during MR-guided adaptive radiotherapy. Methods and Materials The segmentation algorithm was based on a hybrid deep neural network and graph optimization approach that also allows rapid user intervention (Deep LOGISMOS+JEI). A total of 115 MR-datasets were used for training and quantitative assessment. Expert segmentations were used as the independent standard for the prostate, seminal vesicles, bladder, rectum, and femoral heads for all 115 datasets. In addition, three independent radiation oncologists contoured the prostate, seminal vesicles, and rectum for a subset of patients such the inter-observer variability could be quantified. Consensus contours were then generated from these independent contours using a STAPLE approach and the deviation of Deep LOGISMOS+JEI contours to the consensus contours was evaluated and compared to the inter-observer variability. Results The absolute accuracy of Deep LOGISMOS+JEI generated contours was evaluated using median absolute surface-to-surface distance (ASSD) which ranged from a minimum of 0.20 mm for the bladder to a maximum of 0.93 mm for the prostate when compared to the independent standard across all datasets. The median relative surface-to-surface distance was less than 0.17 mm for all organs, indicating that the Deep LOGISMOS+JEI algorithm did not exhibit a systematic under- or over-segmentation. Inter-observer variability testing yielded a mean ASSD of 0.93, 1.04, and 0.81 mm for the prostate, seminal vesicles, and rectum, respectively. In comparison, the deviation of Deep LOGISMOS+JEI from consensus STAPLE contours was 0.57, 0.64, and 0.55 mm for the same organs. On average, the Deep LOGISMOS algorithm took less than 26 seconds for contour segmentation. Conclusions Deep LOGISMOS+JEI segmentation efficiently generated clinically acceptable prostate and normal tissue contours, potentially limiting the need for time intensive manual contouring with each fraction.

Details

Metrics

52 Record Views
Logo image