Quality of interstitial PDR-brachytherapy-implants of head-and-neck-cancers: predictive factors for local control and late toxicity?

BACKGROUND AND PURPOSE Parameters and indices related to the implant geometry in use for describing the quality of volume implants in interstitial brachytherapy were developed on the basis of LDR-brachytherapy. The aim of our study was to evaluate their usefulness for predicting late toxicity and local control in the PDR-brachytherapy of head-and-neck-tumors. PATIENTS AND METHODS Between January 2000 and October 2004, 210 patients were treated with PDR-brachytherapy which was administered either postoperatively or as definitive treatment. Brachytherapy was used as sole treatment in some cases while in others a combination with EBRT was used. For assessment of quality of implants we analyzed the following indices and parameters using the univariate chi2 test and multivariate logistic regression analysis: V85, V120 and V150 (volume enclosed by the surface of the 85%-, 120%- and 150%-isodose), UI (uniformity index), QI (quality index), HI (homogeneity index), VGR (volume gradient ratio), DNR (dose non-uniformity ratio), LD (low dose), HD (high dose), PD (peak dose) and the intersource spacing. RESULTS After a median follow-up of 24 months (4-50) the rate of - usually transient - soft tissue necrosis (STN) was 11%, osteoradionecrosis (ORN) was seen in 7.6% of cases and local relapse occurred in 7% of cases. Univariate analysis shows a significant influence on the development of soft tissue necrosis for V85, and on osteoradionecrosis for HD and PD. In the multivariate analysis a correlation between soft tissue necrosis and QI was found. For local control a correlation with QI, VGR and minimal tube distance was found using univariate analysis. CONCLUSIONS Using interstitial PDR-brachytherapy in head-and-neck-tumors the probability of local control and of the development of soft tissue necrosis or osteoradionecrosis is dependent on dose and volume parameter like the volume of the reference isodose, the high and peak dose values, on the homogeneity of the dose distribution, quantified by the quality index or the volume gradient ratio as well on the minimal tube distance.