Program Information
Validation and Pitfalls of Strength-Based Medical Event Definition in Prostate Seed Brachytherapy: An Evidence-Based Critical Analysis
D Todor1*, M Hagan2, J Horn3, (1) Virginia Commonwealth University, Richmond, VA, (2) Department of Veterans Affairs, Richmond, VA, (3) Virginia Commonwealth University, Richmond, VA
SU-D-105-5 Sunday 2:05PM - 3:00PM Room: 105Purpose: Previous analysis of 94 prostate implants having good dosimetric quality was used to define the treatment site for strength based reporting of medical events. The current work is a validation of this definition applying the 80%-20% strength-based rule to 20 cases previously reported as failing a D90-based evaluation.
Methods:Pre- and post-implants were analyzed for 20 cases. Structure contours and seed data were extracted from Variseed. Software was developed to identify seed distributions with 20% or greater of the implanted seeds residing in specified zones beyond the prostate ; to determine inter-distances between seeds; and to calculate centers of mass for prostate contours and seed distributions. Similarity between pre- and post-implant plans was evaluated by skewness and kurtosis of the inter-seed distance distributions.
Results:Based on the 94 training case dataset, a treatment site structure created as a 5mm expansion beyond the outer prostate contour was used to assess the 80%-20% rule on the test case dataset. Fifteen test cases (75%) had more than 20% of the seeds implanted outside the treatment site. One metric that singled out dosimetrically poor implants was the shift between center of mass (CM) for the seed distribution and CM of the prostate contour. On preplans, both the training group and the test group had similar CM shifts: 1.93+/-1.02mm vs 2.51+/-1.32mm. On the post-implant plans, however, the CM shifts were very different: 3.28+/-1.41mm vs. 12.42+/-5.66mm. The group of 5 cases that were not flagged by the 80%-20% rule, had an average CM shift of 8.88+/-2.67.
Conclusion:Analysis uncovered scenarios wherein implants with very poor dosimetry (D90 of 40-60%PD) would comply with the strength-based rule. Geometrical descriptors were derived and tested for improved discrimination of implant quality.
Contact Email: