Program Information
Towards An Adaptive Radiotherapy Method (ART) Anticipating Morphological Changes Affecting Treatment Quality of Head and Neck Cancer
O Piron*, N Varfalvy , L Archambault , CHUQ Pavillon Hotel-Dieu de Quebec, Quebec, QC
Presentations
SU-E-FS4-5 (Sunday, July 30, 2017) 1:00 PM - 1:55 PM Room: Four Seasons 4
Purpose: To present a new ART method based on relative gamma analysis and automated patient classification for early identification of patients undergoing significant anatomical changes during their radiation therapy, and foresee necessary replanning.
Methods: This study includes 50 patients treated for a head and neck cancer using VMAT. EPID images of every treatment fields were acquired every day and CBCT at least once a week. γ-analysis was performed using the first fraction of the treatment as a reference once validated that it was delivered without error. γ-analysis parameters (<γ>, standard deviation and the top 1%) were used to define categories using either a k-means clustering analysis (KCA) or a hidden Markov model (HMM). From these categories an action threshold is defined and correlated with dosimetric change for 19 out of 50 patients. Performance of this method to correctly detect a dosimetric change was assessed by evaluating both its sensitivity and specificity. Finally, a comparison between the two statistical analysis (KCA and HMM) was performed in order to predict morphological changes before they impact treatment quality.
Results: After performing a clinical retrospective analysis of patient’s categories, an action threshold on the average gamma value was established at 0.42 for head and neck. Of the 19 patients, 11 were above the action threshold at the day of their last fraction and they had a strong dosimetric impact. From those 19 patients, the established threshold had a specificity of 0.9 and a sensitivity of 0.64. The comparison between the two statistical analysis showed great similarities.
Conclusion: This method showed that patients at risk of deviating from their planned treatment can be identified early in the course of their treatment using only EPID images. The possibility of anticipating morphological changes affecting treatment quality before they occur would make necessary treatment replanning more predictable.
Funding Support, Disclosures, and Conflict of Interest: This work was funded in part by Varian Medical Systems
Contact Email: