Dosimetric evaluation of a commercial inverse treatment planning system: UAB experience

A commercial inverse treatment planning system was evaluated for its dose computational accuracy. Three types of verification studies were performed: I. phantom plans versus measurements, where phantom plans were optimized to simulated targets in the phantoms, II. patient-phantom hybrid plans versus measurements, where hybrid plans were generated by combining the phantom geometry with beam configurations of patient plans; and III. inverse plans versus forward plans, in which non-modulated plans were calculated by the inverse planning system and compared to those by a forward treatment planning system. Three different phantoms were used with two ionization chambers and a film dosimetry system for measurements of absolute dose and relative dose distribution. After applying a MU correction factor (1.024 and 1.008 for the 1 and the 2 cm leaf mode, respectively) determined in the type I study, the absolute doses of the 20 hybrid plans agreed with measurements with a mean difference /spl plusmn/ SD of -0.3%/spl plusmn/1.3%. Planned isodose lines also agreed with the measured ones within 2 mm in high dose areas (>50% dose lines) and generally within 3 mm in low dose areas (<50% dose lines), except for some star-shaped low isodose lines where mismatches as high as 5 mm were observed. Comparison between the inverse and forward plans showed an agreement better than 3% in MU calculation, and an isodose line match within 2 mm or 2%. It is concluded that the dosimetric accuracy of the inverse treatment planning system is clinically acceptable.