Reconstitution of internal target volumes by combining four-dimensional computed tomography and a modified slow computed tomography scan in stereotactic body radiotherapy planning for lung cancer

BACKGROUND: To evaluate the volumetric and geometric differences in the ITVs generated by four-dimensional (4D) computed tomography (CT), a modified slow CT scan, and a combination of these CT methods in lung cancer patients treated with stereotactic body radiotherapy (SBRT). METHODS: Both 4D CT and modified slow CT using a multi-slice CT scanner were performed for SBRT planning in 14 patients with 15 pulmonary targets. Volumetric and geometric analyses were performed for (1) ITVall, generated by combining the gross tumor volumes (GTVs) from all 8 phases of the 4D CT; (2) ITV2, generated by combining the GTVs from 2 extreme phases of the 4D CT; (3) ITVslow, derived from the GTV on the modified slow CT scan; (4) ITVall+slow, generated by combining ITVall and ITVslow; and (5) ITV2+slow, generated by combining ITV2 and ITVslow. Three SBRT plans were performed using 3 ITVs to assess the dosimetric effects on normal lung caused by the various target volumes. RESULTS: ITVall (11.8 ± 8.3 cm(3)) was significantly smaller than ITVall+slow (12.5 ± 8.9 cm(3)), with mean values of 5.8% for the percentage volume difference, and a mean of 7.5% of ITVslow was not encompassed in ITVall. The geometric coverages of ITV2 and ITVslow for ITVall were 84.7 ± 6.6% and 76.2 ± 9.3%, respectively, but the coverage for ITVall increased to 90.9 ± 5.9% by using the composite of these two ITVs. There were statistically significant increases in the lung-dose parameters of the plans based on ITVall+slow compared to the plans based on ITVall or ITV2+slow. However, the magnitudes of these differences were relatively small, with a value of less than 3% in all dosimetric parameters. CONCLUSIONS: Due to its ability to provides additional motion information, the combination of 4D CT and a modified slow CT scan in SBRT planning for lung cancer can be used to reduce possible errors in true target delineation caused by breathing pattern variations.