Analysis of Carina Position as Surrogate Marker for Delivering Phase-Gated Radiotherapy

Abstract

Respiratory gating can mitigate the effect of tumor mobility in radiotherapy (RT) for lung cancer. Because the tumor is generally not visualized, external surrogates of tumor position are used to trigger respiration-gated RT. We evaluated the suitability of the carina position as a surrogate in respiration-gated RT. A total of 30 four-dimensional (4D) computed tomography (CT) scans from 14 patients with lung cancer were retrospectively analyzed. Both uncoached (free breathing) and audio-coached 4D-CT scans were acquired from 9 patients, and 12 uncoached 4D-CT scans were acquired from 5 other patients during a 2-4-week period of stereotactic RT. The repeat scans were co-registered. The carina position was identified on the coronal cut planes in all 4D-CT phases. The correlation between the carina position and the total lung volume for each phase was determined, and the reproducibility of the carina position was studied in the 5 patients with repeat uncoached 4D-CT scans. The mean extent of carina motion in 21 uncoached scans was 5.3 +/- 1.6 mm in the craniocaudal (CC), 2.3 +/- 1.4 mm in the anteroposterior, and 1.5 +/- 0.7 mm in the mediolateral direction. Audio coaching resulted in a twofold increase in carina mobility in all directions. The CC carina position correlated with changes in the total lung volume (R = 0.89 +/- 0.14), but the correlation was better for the audio-coached than for the uncoached 4D-CT scans (R = 0.93 +/- 0.08 vs. R = 0.85 +/- 0.17; paired t test, p = 0.034). Preliminary data from the 5 patients indicated that the CC carina motion correlated better with tumor motion than did the motion of the diaphragm. The CC position of the carina correlated well with the total lung volume, indicating that the carina is a good surrogate for verifying the total lung volume during respiration-gated RT.