Poster Poster Program Therapy Physics

Can Organs at Risk Segmented By AI-Based Auto Segmentation Algorithms be Used In Inverse Planning?

Abstract
Purpose

To investigate whether unedited organs at risk (OARs) automatically segmented by Artificial Intelligence-based algorithms can be used by inverse optimizer to generate plans for lung SBRT cases such that the plans are still clinically acceptable after the contours has been edited by radiation oncologists.

Methods

Planning CT images from twenty-four patients treated with SBRT for lung cancer were selected and automatically segmented by a commercially available AI-based software. The unedited OARs automatically generated by AI-based algorithm were used by an inverse optimizer to manually generate VMAT plans denoted as AIContours_P. The physician-edited OARs were used to inversely generate clinically approved plans denoted as Clinical_P. The plan quality between AIContours_P and Clinical_P were compared. The unedited OARs were fed into an auto planning algorithm to automatically generate VMAT plans denoted as AIContoursAuto_P. The plan quality between AIContoursAuto_P and Clinical_P were compared. Even though unedited OARs were used in inverse planning, only edited contours’ clinically used dosimetric criteria were used in evaluating the plans’ qualities.

Results

All plans manually optimized with AI-segmented OARs met the clinically used scorecards while four out of the twenty-four cases optimized by the auto inverse planner failed at least one dosimetric criteria in the clinically used scorecards. For one case, the Dice Coefficient (DC) between unedited and physician-adjusted heart is 0.92, however, all three plans Clinical_P, AIContours_P and AIContoursAuto_P met the clinical scorecards. For another case, DC between unedited and physician-adjusted great vessel is 0.52, Clinical_P and AIContours_P could meet all dose metrics while AIContoursAuto_P failed at least one dose metric.

Conclusion

It is possible to use unedited OARs segmented by AI-based algorithms to do inverse planning for lung SBRT cases. The plan can still meet all clinically used dose metrics after physician adjusted the contours.

People

Related

Similar sessions

Poster Poster Program
Jul 19 · 07:00
Python-Based Automation Framework for Annual Machine QA Data Archiving In Qatrack+

Annual water-tank measurements help ensure beam characteristics remain consistent with commissioning baselines. However, the lack of a standardized processing workflow and decentralized data storage makes it difficult to analyze...

Syed Bilal Ahmad, PhD
Therapy Physics 0 people interested
Poster Poster Program
Jul 19 · 07:00
User Expectations and Current Availability of HDR Brachytherapy Audits In Europe

The aim of this work was to evaluate the need to implement more dosimetric audits in high‐dose‐rate brachytherapy (HDR-BT) in Europe and to identify which characteristics such audits should meet according to users.

Javier Vijande, PhD Laura Oliver Cañamás
Therapy Physics 0 people interested