ABSTRACT
PURPOSE: The study aimed to analyze the Pro-Qura database in terms of patient implant sequence number for each institution to determine evidence for a dosimetric learning curve. METHODS AND MATERIALS: In the Pro-Qura database, 2833 of a total of 4614 postplans from 57 brachytherapists were analyzed for evidence of a dosimetric learning curve. The median time between implant and postimplant CT scan was 30 days. I-125 was used in 2123 patients (1687 monotherapy and 536 boost) and Pd-103 in 710 patients (367 monotherapy and 343 boost). Preimplant prostate volume was 35.3 and 32.9 cm3 in the I-125 and Pd-103 cohorts, respectively. The mean I-125 seed activity was 0.32 and 0.26 mCi for monotherapy and boost, whereas for Pd-103 the mean seed activity was 1.59 and 1.27 mCi, respectively. Postimplant dosimetry was performed in a standardized fashion by overlaying the preimplant ultrasound and the postimplant CT scan. Criteria for implant adequacy included a D90 >90% and a V100 >80% for both isotopes. An adequate V150 was defined as <60% for I-125 and <75% for Pd-103. RESULTS: The mean V100 and D90 were 88.9% and 101.9% of prescription dose, respectively. When analyzed in terms of patient sequence number for each institution, the mean V100 for the first 10 patients was 87.4% and increased to 88.6% for patients 11-20 (p = 0.036). Similarly, the mean D90 for the first 10 patients was 98.9%, whereas for the second cohort of 10 patients the mean D90 increased to 102.2% (p = 0.001). In terms of mean V100 and D90, there was minimal further change for subsequent 10 patient institutional groupings of patient sequence numbers. For the first 10 cases, 27.2% were deemed "too cool" (V100 <80% and/or D90 <90%). Approximately 16% of all implants were deemed "too hot" (D90 >140% or V150 >60% for I-125 or >75% for Pd-103). CONCLUSIONS: Although a learning curve exists for prostate brachytherapy, high-quality brachytherapy is achievable in approximately 75-80% of patients treated at community centers.
