Click here to


Are you sure ?

Yes, do it No, cancel

Dose Error in Using Point Source Approximation for Permanent Brain Seed Implants

S Sorensen*, D Pinnaduwage , E Youssef , S Srivastava , X Yan , S Jani , D Brachman , Barrow Neurological Institute, Phoenix, AZ


(Sunday, 7/29/2018) 3:00 PM - 6:00 PM

Room: Exhibit Hall

Purpose: To quantify dose differences in using a point source (PS) vs a more accurate line source (LS) model for several permanent implant radioisotopes, and to investigate corresponding differences with an independent dose calculation.

Methods: Ten patients who underwent intraoperative brain brachytherapy with Cs-131 seeds imbedded in a collagen carrier/spacer on an IRB approved trial were retrospectively identified. For each patient the PS model was replaced by a LS model in the treatment planning system (TPS) (BrachyVision, Varian, Palo Alto, CA). Dose differences were compared at 15, 30, 60, and 100 Gy. The LS model required identification of both the location and orientation of each seed. The same comparison was made for I-125 and Pd-103. Substituting these isotopes, we used the exact same seed locations and orientations and modified the activities to match the previously achieved 60Gy Cs-131 PS dose volumes. Reference points placed at the aforementioned dose levels in the Cs-131 plans were sent to an independent dose calculation software (RadCalc, LifeLine Software, Tyler, Tx). Differences in dose to these points provided a comparison of the PS and LS models in RadCalc with those from the TPS.

Results: For Cs-131 the average percent error was 4.3, 4.5, 4.7, and 4.6 at 15, 30, 60, and 100 Gy. At the same dose levels the percent error was 5.6, 6.3, 6.6, 6.8, and 10.2, 11.8, 13.6, 15.4 for I-125 and Pd-103, respectively. The independent PS and LS check using RadCalc showed an average difference of 1% and 2% respectively.

Conclusion: Depending on the radioisotope, using the PS model for permanent seed implants overestimates the dose by ~5% or more for this particular TPS. The implementation of the PS and LS models differed in the TPS and RadCalc. Default seed orientations in RadCalc resulted in larger LS differences.


Brachytherapy, Brain


TH- Brachytherapy: General (most aspects)

Contact Email