Cyberknife lung SBRT: from semi-homogeneous to monte carlo dose calculation.

Purpose: To re-evaluate planning target volume (PTV) and organ at risk (OAR) dosimetry using Monte Carlo recalculation of treatment plans initially calculated using the Ray-Trace algorithm (effective path length method, EPL). Any differences would be correlated with known clinical outcome in terms of local failure (LF) and toxicity. Methods: Lung CK SBRT treatment plans initially calculated using an EPL algorithm (without correction for tissue inhomogeneity) were re-calculated with MC algorithm. PTV coverage and doses to OARs were compared. We reviewed PTV coverage of patients with LF, volume of the chest wall receiving ≥ 30 Gy (V30) of patients with costal tenderness or rib fracture and percentage total lung volume receiving 5 Gy (V5) and 20 Gy (V20) for patients with grade ≥ 3 pneumonitis. Results: Twenty four plans were recalculated. Median clinical follow-up on these patients was 19 months. Median prescription dose was 60 Gy (range, 48-60 Gy) in 3 fractions (range, 3-5). With EPL and MC respectively, median PTV coverage was 95 % and 91%. Thirteen patients (54%) had a 5% or greater drop in PTV coverage after recalculation (median 9%, range (+6 to -64%). However, in 2 of the 3 cases of LF PTV coverage was actually better than expected (95% PTV coverage for EPL vs. 99% by MC). The chest wall V30 increased by a median of 1.4 cc following re-calculation. Unsurprisingly, the median V30 was greater in the 12 cases of chest wall toxicity (27cc by EPL vs. 28cc by MC) that for the other patients (10 cc using either EPL or MC). Of the patients with chest wall toxicity, 3 had ≥ 5cc (range, 8.4-16cc) V30 increase on MC re-calculation. EPL and MC lung V5 and V20 differences were not significant. Two patients with pulmonary fibrosis developed grade 5 pneumonitis. Whereas median V20 was 5% (EPL and MC) (range, 2-10%) for the other patients, V20 for these 2 patients was19%(MC). Conclusion: MC recalculation predicted worse PTV coverage than EPL, however EPL inaccuracies do not appear to be implicated in LF. No significant differences were found in either chest wall V30, lung V5 or lung V20. Our work confirms that chest wall toxicity correlates with higher V30 and suggested that more severe lung constraints should be used for patients with pulmonary fibrosis. Despite improving the accuracy of patient dosimetry, the clinical benefits of MC dose calculation remain to be demonstrated. Disclosure: No significant relationships. [ABSTRACT FROM AUTHOR]