Purpose: This study tested the beam path length (BPL) of daily cone beam CT (CBCT) images as an adaptive planning indicator of head-and-neck (HN) patients treated with proton beam therapy (PBT).
Methods: In this IRB approved study, we retrospectively analyzed 70 HN patients who were treated with PBT. Out of 70 patients, 30 of them underwent adaptive planning on their quality assurance CT scans (QACTs) due to different reasons, e.g. weight loss (10 patients), anatomical changes (11 patients), tumor shrinkage (8 patients) and loss of target coverage (1 patient). Each beam was ray traced from the isocenter to the skin surface along the central axis using the daily CBCT images. The patterns of change in BPL of CBCT images have been compared to those of the adaptive planning using weekly QACT.
Results: In our study, we used a threshold of BPL change of 3 mm, beyond which an adaptive plan must be needed. Out of 119 beams for all 30 the adaptively planned patients, 58 (48.7 %) of them had BPL changes more than 3 mm. Similarly, out of 128 beams for the 40 non-adaptive planned patients, only 3 (2.3 %) beams had BPL changes of more than 3 mm. The maximum change in BPL for 29 out of 30 adaptive planned patients were larger than 3 mm, this led to a sensitivity of 96.7 % for BPL change as an indicator of adaptive planning. 37 out of 40 non-adaptive planned patients had BPL changes less than 3 mm, led to a specificity of 92.5 %.
Conclusion: The beam path length (BPL) change from isocenter to skin surface along the beam central axis on CBCT images can be used as an adaptive plan indicator for proton beam therapy (PBT) head-and-neck patients with 96.7 % sensitivity and 92.5 % specificity.