Evaluation of IMRT QA point measurement rocesses using a commercial heterogeneous phantom

Intensity Modulated Radiation Therapy (IMRT) has the potential to deliver a highly conformal dose distribution to the target volume compared to conventional radiotherapy. However, the use of IMRT introduces complexities in dose delivery and verification. Routine IMRT QA is typically performed in a homogeneous solid water phantom and does not verify the accuracy of a treatment planning system's handling of the heterogeneity correction algorithm, which is particularly important in a low density lung medium. The purpose of this work is to evaluate common IMRT QA point measurement processes that take advantage of a commercial heterogeneous phantom [CIRS IMRT thorax phantom (CIRS, Inc., Norfolk, Virginia, USA)]. Dose calculated with Monte Carlo (MC) methods and pencil beam (PB) methods are used. IMRT QA using the CIRS phantom with the MC and PB algorithms was retrospectively analyzed using control charts and a capability index. Fifteen actual IMRT treatment plans of lung cancer patients were used for this study. The dose was measured in the phantom at points located in lung, bone, and tissue with an ion chamber (IC) for 15 cases and thermoluminescent dosimeters (TLDs) for 5 cases. Measurements and calculations in each heterogeneity (e.g., TLD/MC in bone) were considered as separate processes. Control charts and the capability index Cpm were used to evaluate the following processes using the CIRS phantom: IC/MC, PB/MC, TLD/MC for measurements in the lung, tissue and bone. The processes PB/IC and MC/IC using conventional homogeneous water-equivalent slab geometry were also evaluated. In total, 11 IMRT QA processes were considered. Comparison of the data showed that the dose inside the lung calculated with PB was overestimated by 6% on average relative to the MC calculations. On average, MC calculations in bone and tissue agree within 3% with PB calculations and IC measurements. Process capability values (Cpm) greater than 1.33 indicate a well performing process. Using the CIRS phantom, Cpm ranged from 0.25 for the PB/ MC process in lung to 1.41 for the TLD/MC process in tissue. By comparison, the process using the conventional waterequivalent slab phantom showed the PB/IC and MC/IC Cpm values of 1.36 and 1.21, respectively. Nine of the 11 IMRT QA processes studied were not able to meet the clinical specifications of 5%. However, we found the CIRS phantom is versatile to compare both homogeneous and heterogeneous IMRT QA measurements to calculations. Our results indicate that additional refinements of the IMRT QA processes are required. This is especially true for calculations and measurements in lung equivalent media. The capability index is a simple and useful quantitative tool for comparing different approaches to lung IMRT QA.