MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Automated Detection of Missed and Failed Breath-Holds with SDX Motion Management

T Nano*, M Feng, T Solberg, A Witztum, University of California, San Francisco, San Francisco, California

Presentations

(Saturday, 4/4/2020)   [Mountain Time (GMT-6)]

Purpose: To analyze spirometry-based breathing waveforms in patients undergoing voluntary breath-hold (BH) radiotherapy to detect errors that may be missed clinically.

Methods: An automated process was developed to analyze SDX (Dyn’r Medical Systems Aix-en-Provence, France) BHs post-treatment. A total of 596 waveforms in 89 fractions (47 inspiration and 42 expiration) from 27 patients were analyzed. For each fraction, multiple BH waveforms were recorded and quantitative metrics extracted to assess treatment, including BH time range, number of missed holds (BHs shorter than 2 seconds), and number of failed holds (BHs longer than 2 seconds but shorter than mean BH minus 5 seconds). Following each fraction delivery, a summary report is generated to be added to the patient chart and alert the clinical team.

Results: For all treatments, total BH time was greater than beam-on time. There were 14 misses and 28 fails in total (7% of BHs), with 7 patients that had at least one missed BH and 13 patients had at least one fail. Multiple fractions were recorded for 15 patients, and 9 patients showed errors (both missed and failed) at some time-point. In 3 of these patients, all errors were in the 1st and 2nd fraction. During inspiration holds, there were 2 misses and 11 fails whereas expiration holds had a greater number of errors with 12 misses and 17 fails. For all treatments, mean BH time and range (max-min) was 36.7s and 1.3s for fractions without errors, and 28.0s and 26.3s for those with at least 1 failed BH respectively.

Conclusion: A method to automatically detect and report missed and failed breath-holds during treatments with SDX motion management was developed. Manual intervention during treatments is a potential source of error, and detection of prior breath-hold inconsistencies may help identify patients that could benefit from additional monitoring or training.

ePosters

Contact Email