Click here to


Are you sure ?

Yes, do it No, cancel

Assessing the Performance of Automated Breast Treatment Planning Software

I Dragojevic1*, J Hoisak1 , R Manger2 , (1) UC San Diego, Chula Vista, CA, (2) UC San Diego, Encinitas, California


(Sunday, 7/14/2019)  

Room: ePoster Forums

Purpose: To assess the dosimetric and auto-contouring performance of automated breast planning software.

Methods: Fifteen breast tangent electronic compensator plans that were manually planned in Eclipse (Varian Medical Systems) according to RTOG 1005 were re-planned using EZFluence (EZF) automated breast planning software (Radformation, Inc.). Using the same field borders, EZF was used to automatically derive fluences for the electronic compensator. EZF plans were normalized such that the maximum point dose was less than or equal to the protocol plan. The following dosimetric parameters were compared: D95%, D90%, V105%, V95%, CI95% (the quotient of V95 and PTV volume), and total monitor units (MU). EZF auto-contours a structure called PTV_Eval_EZ by subtracting 5 mm from the field edges, skin and specified OARs (e.g. lung, heart, etc). This PTV_Eval_EZ was compared against the manually contoured RTOG 1005 whole breast PTV_eval structure according to volume differences and the Dice similarity coefficient.

Results: The average number of MU for the EZF plans, 452, was significantly less than the manual protocol plans, 538 (p=0.0003). V105% (cc) was, on average, 64 cc less in the EZF breast plans. Compared to the manual protocol plans, the EZF average D95% was significantly greater (96.2% vs. 93.5% p=0.018). The CI95% was less (1.15 vs. 1.52, p=0.05). The average difference in volume between the protocol and EZF PTV volumes was 312 cc (relative difference, 21.4%, range 0.4%-87.6%); only one EZF PTV was larger. The average Dice similarity was 0.79 (range 0.46 – 0.92).

Conclusion: EZFluence produces electronic compensator plans with improved dosimetric homogeneity, coverage, and MU efficiency than manually-edited plans. The EZF auto-contoured PTV provides a consistent breast target to create dosimetrically similar or superior plans but is not equivalent to the RTOG 1005 PTV_eval and should not be used for RTOG 1005 dosimetric evaluation.


Not Applicable / None Entered.


TH- External beam- photons: treatment planning/virtual clinical studies

Contact Email