For interstitial applications of high dose rate (HDR) afterloading brachytherapy, generally a single stepping iridium-192 source is used, enabling optimization of the dose distribution by optimization of the relative time (dwell time) that the source remains at a certain position (dwell position). We analysed the effects of geometric optimization in a regular volume implant, with strictly parallel catheters, and in an irregular volume implant, such as an implant for tumours of the base of the tongue characterized by a non-parallel geometry and varying catheter separations. In both examples the reference dose is specified at 85% of the mean central dose (as is done in the Paris system for dose specification) in the non-optimized as well as the optimized plan. The irradiated volume, the dose uniformity, and the choice of the reference dose of optimized and non-optimized dose distributions were compared. This was done by isodose plots for representative planes, volume dose histograms (distributed, contiguous, and natural), and dose non-uniformity ratios (DNRs). For the regular implant, optimization results in a 28% increase in the treated volume with a similar increase in the overdosed volumes. In order to keep the treated volume comparable with the non-optimized dose distribution, 90–95% of the mean central dose should be chosen as a reference dose or the range of active dwell positions should be shortened in case of optimization. In the case of the irregular volume implant at the base of the tongue, the method for dose specification should be kept unchanged after geometric optimization as the volume enclosed by the reference isodose does not increase. It is clear from the volume-dose histograms that there is a reduction of the overdosed volume due to optimization. This is accompanied by an increase in the uniformity index and a decrease of the DNR. In conclusion, geometric optimization appears to be an effective tool to improve the dose distribution of interstitial volume implants. Contiguous and natural volume dose histograms appear, apart from planar dose plots, valuable methods for evaluating the dose distribution of an implant.