Treatment verification by portal films in chest and mantle field irradiation

The results of a portal film investigation regarding patients treated for lung tumors and Hodgkin's lymphoma with 6 and 18 MV X-Rays (from 6/100 and 1800 Clinac Varian) are reported. 150 portal films and 68 simulation films of 68 patients (anterior-posterior (AP) fields) have been collected and analyzed evaluating cranio-caudal and lateral shifts, comparing the patient positioning during therapy with respect to the set-up during simulation. Fields have been divided in 'complex' and 'non-complex' ones: we considered as 'complex' ones the mantle fields for the treatment of Hodgkin's disease and the lung-mediastinum fields which included in the treatment volume the supraclavicular nodes, having a complex 'quasi-mantle' shape. 'Non complex' ones were the fields irradiating the mediastinum and portions of lungs, without including supraclavicular nodes. Mean discrepancies and standard deviations have been respectively -0.1 and 3.6 mm for cranio-caudal shifts and -0.8 and 3.1 mm for lateral shifts. Non complex fields showed a standard deviation not significantly different from that of complex fields. Among the considered patients treated for lung diseases, 8 had CT-based treatment planning. For these patients, all treated with 18 MV X-Rays (except one treated with 6 MV X-Rays), exit dose distributions on the central axial slice have been measured by portal film exit dosimetry (Kodak X-Omat V in localization cassette) and compared with those calculated by our treatment planning system (Cadplan Dosetek Varian), with and without applying inhomogeneity correction algorithms (Batho and Equivalent TAR (ETAR)). Off-axis exit doses in the region corresponding to the lungs have been considered in order to assess an index of accuracy in the calculation of the treatment planning. Results show a mean deviation of -11.5% (σ = 7.4%) between calculated and measured relative exit doses (% (calculated-measured)/calculated) when calculating dose distributions without inhomogeneity correction; of -2.5% (σ = 4.4%) when using the Batho algorithm; of +0.3% (= 2.9%) when using Equivalent TAR algorithm. The study confirms the wide dosimetric possibilities of portal film in-vivo dosimetry (especially in the area of verification of the treatment and of treatment planning systems).