Abstract
Adaptive Radiotherapy (ART) is a tool with the potential to improve head and neck (H&N) cancer patient outcomes by creating new treatment plan(s) throughout the treatment regimen, reducing dose to nearby organs at risk (OAR) whenever possible. While online ART makes use of specific and costly machinery to create daily treatment plans for patients, a lack of standardization prevents large-scale offline ART research and/or clinical trials. Therefore, a study to develop an offline ART workflow may be beneficial in developing techniques which most benefit patients with as few resources as possible. This would be most beneficial for low-resource clinics unable to practice existing methods of ART.
A previous study performed by Andy Fanning through the University of Nebraska Medical Center and Creighton University demonstrated that the middle of treatment (midpoint) was the most time-efficient point of assessment for patient plan adaptation using dose-volume histogram (DVH) metrics and their Velocity-based workflow. In this study, we demonstrated a proof-of-concept workflow with the same midpoint assessment scheme, but instead utilizing normal tissue complication probability (NTCP) metrics.
NTCP values were retrospectively calculated and compared between the planned dose and true delivered dose. Adaptive replanning was deemed warranted for a patient plan if the change in NTCP (∆NTCP) for an OAR was 5% or greater. Furthermore, an estimated midpoint assessment NTCP value was calculated and compared to the true value to determine its prediction capabilities. These results were also tested and compared against the same study using DVH metrics.
We found that methods utilizing NTCP metrics have the potential to be 12.5% more accurate at predicting the need for adaptation than DVH metrics. While the use of NTCP metrics requires an additional 0.5-1 hour of labor, it has the potentialto save 1-2 days of clinical time needed for a false positive assessment as it was shown in this study. Additionally, there is a large amount of research required for a more robust characterization of the NTCP midpoint assessment scheme that would allow for this proof of concept to be used clinically.