Experience of micromultileaf collimator linear accelerator based single fraction stereotactic radiosurgery: tumor dose inhomogeneity, conformity, and dose fall off.

Purpose: Sharp dose fall off outside a tumor is essential for high dose single fraction stereotactic radiosurgery (SRS) plans. This study explores the relationship among tumor dose inhomogeneity, conformity, and dose fall off in normal tissues for micromultileaf collimator (mMLC) linear accelerator (LINAC) based cranial SRS plans.

Methods: Between January 2007 and July 2009, 65 patients with single cranial lesions were treated with LINAC-based SRS. Among them, tumors had maximum diameters < or = 20 mm: 31; between 20 and 30 mm: 21; and > 30 mm: 13. All patients were treated with 6 MV photons on a Trilogy linear accelerator (Varian Medical Systems, Palo Alto, CA) with a tertiary m3 high-resolution mMLC (Brainlab, Feldkirchen, Germany), using either noncoplanar conformal fixed fields or dynamic conformal arcs. The authors also created retrospective study plans with identical beam arrangement as the treated plan but with different tumor dose inhomogeneity by varying the beam margins around the planning target volume (PTV). All retrospective study plans were normalized so that the minimum PTV dose was the prescription dose (PD). Isocenter dose, mean PTV dose, RTOG conformity index (CI), RTOG homogeneity index (HI), dose gradient index R50-R100 (defined as the difference between equivalent sphere radius of 50% isodose volume and prescription isodose volume), and normal tissue volume (as a ratio to PTV volume) receiving 50% prescription dose (NTV50) were calculated.

Results: HI was inversely related to the beam margins around the PTV. CI had a "V" shaped relationship with HI, reaching a minimum when HI was approximately 1.3. Isocenter dose and mean PTV dose (as percentage of PD) increased linearly with HI. R50-R100 and NTV50 initially declined with HI and then reached a plateau when HI was approximately 1.3. These trends also held when tumors were grouped according to their maximum diameters. The smallest tumor group (maximum diameters < or = 20 mm) had the most HI dependence for dose fall off. For treated plans, CI averaged 2.55 +/- 0.79 with HI 1.23 +/- 0.06; the average R50-R100 was 0.41 +/- 0.08, 0.55 +/- 0.10, and 0.65 +/- 0.09 cm, respectively, for tumors < or = 20 mm, between 20 and 30 mm, and > 30 mm.

Conclusions: Tumor dose inhomogeneity can be used as an important and convenient parameter to evaluate mMLC LINAC-based SRS plans. Sharp dose fall off in the normal tissue is achieved with sufficiently high tumor dose inhomogeneity. By adjusting beam margins, a homogeneity index of approximately 1.3 would provide best conformity for the authors' SRS system.