Published paper AbstractPurpose: To determine the implications of the use of the Anisotropic Analytical Algorithm (AAA) for the production and dosimetric verification of IMRT plans for treatments of the prostate, parotid, nasopharynx and lung. Methods: 72 IMRT treatment plans produced using the Pencil Beam Convolution (PBC) algorithm were recalculated using the AAA and the dose distributions compared. 24 of the plans were delivered to inhomogeneous phantoms and verification measurements made using a pinpoint ionisation chamber. The agreement between the AAA and measurement was determined.Results: Small differences were seen in the prostate plans, with the AAA predicting slightly lower minimum PTV doses. In the parotid plans, there were small increases in the lens and contralateral parotid doses while the nasopharyngeal plans revealed a reduction in the volume of the PTV covered by the 95% isodose (the V 95% ) when the AAA was used. Large changes were seen in the lung plans, the AAA predicting reductions in the minimum PTV dose and large reductions in the V 95% . The AAA also predicted small increases in the mean dose to the normal lung and the V 20 . In the verification measurements, all AAA calculations were within 3% or 3.5mm distance to agreement of the measured doses. Conclusions:The AAA should be used in preference to the PBC algorithm for treatments involving low density tissue but this may necessitate re-evaluation of plan acceptability criteria. Improvements to the Multi-Resolution Dose Calculation algorithm used in the inverse planning are required to reduce the convergence error in the presence of lung tissue. There was excellent agreement between the AAA and verification measurements for all sites.3