Intensity modulated radiotherapy of non-small-cell lung cancer incorporating SPECT ventilation imaging

Methods: For ten consecutive stage III non-small-cell lung cancer patients, the authors obtained both ventilation/perfusion SPECT scans and four-dimensional CT scans for treatment planning purposes. Each ventilation scan was registered with the corresponding planning CT and ventilation volumes corresponding to either >= 50% (vv50) or >= 70% (vv70) of the maximum SPECT count were automatically segmented. For each patient, three IMRT plans were generated: One using nine equally spaced beams optimized according to nonfunctional lung based mean lung dose and lung v20; a second using nine equally spaced beams optimized to avoid vv50 and vv70; and a third plan using only three beams with gantry angles chosen based on minimum mean ventilated lung dose calculated for each conformal beam at every 10 degrees gantry angle avoiding vv50 and vv70. Resultant dose volume histogram indices were calculated for each plan and were compared with respect to calculated SPECT-based ventilation parameters in order to quantify the potential utility of ventilation SPECT in this setting. Results: Two patient groups were identified based on (i) the overlap volume between PTV and vv50 and (ii) the average angular mean ventilated lung dose (AAMvLD). The first parameter quantifies the proximity of the PTV to well ventilated lung and the second parameter quantifies the degree of ventilation that surrounds the PTV. For group 1 patients, < 5% of the vv50 overlapped with the PTV. For group 2 patients, >5% of the vv50 overlapped the PTV. Group 1 was further classified into subgroups 1A and 1B: For subgroup 1A, AAMvLD is >18 Gy, implying that the functional lung surrounds the PTV; for subgroup 1B, AAMvLD is < 18 Gy, implying that the well ventilated lung does not completely surround PTV. For subgroup 1A, the plans generated using ventilated lung avoidance reduced dose to vv50 and vv70, with below tolerance dose to normal lung and acceptable coverage of the PTV. For subgroup 1B, the dose to the total lung and well ventilated lung are reduced with the beam direction optimization for the three-beam plan. For group 2, there was no significant dosimetric advantage of using SPECT-based ventilation information in IMRT plan optimization. Conclusions: In conclusion, it is feasible to use SPECT ventilation scans to optimize IMRT beam direction and, subsequently, to reduce dose to ventilated lung when overlap of the PTV and the ventilated lung is minimal and that the PTV is not surrounded by the ventilated lung. The potential benefit of ventilation SPECT scanning can be determined by preplanning assessment of overlap volumes and the AAMvLD.