Triggered GTV-Based Offline Adaptation for LINAC Head and Neck Radiotherapy

Abstract

<h3>Purpose/Objective(s)</h3> Adaptive head and neck radiotherapy has shown great promise in ensuring delivered dose matches physician intent. Optimal points for adaptation are still unknown and replanning is always a balance between clinical benefit and resources. In this work we explore a gross tumor volume (GTV) based trigger, enabled by in-house automation software, that alerts the clinical team of significant changes in the GTV to aid in their decisions regarding adaptive replanning. <h3>Materials/Methods</h3> Twenty-eight patients with mainly cancer of the oropharynx receiving definitive radiotherapy with no resection were included. Patients with visibly bulky primary or nodal disease were selected and prescription dose to the high-risk planning target volume (PTV) was 70Gy in 33/35 fractions. All patients were scheduled on an in-house software program called AWARE (Automated Watchdog for Adaptive Radiotherapy Environment) that automatically tracks preselected structures for volume changes on weekly CBCTs. GTVs were tracked and a change in GTV greater than 20% triggered a resimulation and a discussion with the physician regarding the need for adaptation. Absent 20% changes, all patients were resimulated 3 weeks after the start of treatment. If deemed clinically necessary following resimulation, the patient was replanned. The treatment timeline was recorded for each patient including the AWARE trigger point which was compared between adapted and non-adapted patients. For adapted patients the dose to the organs at risk and PTVs were compared between what would have been delivered with the old plan on the resimulation scan and the dose that was delivered with the replan on the resimulation. Finally, the correlation between the AWARE-tracked GTV change and the PTV change in the adapted plan was determined. <h3>Results</h3> Of the 28 patients that were tracked, 16 received an adaptive plan. On average, the resimulation occurred 23 days [range: 14-30] after treatment start and the adaptive plan was treated for 11.5 fractions [range: 9-20]. Of the 16 patients that were treated with an adaptive plan, 13 (81%) reached the AWARE trigger by end of week 4. By comparison only 2 out of 12 (17%) patients that were not adapted reached the trigger. Adaptation resulted in a statistically significant decrease in the mean dose to the submandibular glands and the oral cavity (p-value < 0.002). In 3 patients the PTV coverage increased by more than 10% with the adapted plan. The correlation between AWARE tracked GTV and the PTV change was 0.68 (p-value=0.004). <h3>Conclusion</h3> Our initial experience with automated tracking shows that it can provide triggers that agree well with the decision to adapt and changes in the resimulated target volumes. Once adapted, the dosimetric benefit for minimizing dose delivered to the normal tissue can be highly significant. This can translate into meaningful toxicity reduction, while an adaptive plan can also improve coverage to the target when changes occur.