Encrypted login | home

Program Information

Pitfalls of the Radiosurgery Conformity Index Reported by the BrainLab Treatment Planning System


N Mensah-Brown

N Mensah-Brown*, H Saleh , The University of Kansas Cancer Center, Lee's Summit, MO

Presentations

SU-F-T-595 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall


Purpose: To propose a new method to more accurately compute the conformity index (CI) for stereotactic radiosurgery patients with multiple lesions planned with the BrainLab iPlan treatment planning system.

Methods: The dose files of the patients being studied were exported from the BrainLab iPlan system to VelocityAI, an imaging informatics software tool that provides an independent method of assessing the quality of treatment plans. An independent method of calculating CI is necessary as BrainLab incorrectly reports CI for multiple lesions. BrainLab recommends deleting the other treatment isocenters to account for this inaccuracy.

In VelocityAI, a dose object of the isodose line in question was obtained. The CI was then calculated using the volume of the dose object, the PTV, and the percentage dose coverage. The method was tested for patients with single lesions and then applied to patients with multiple lesions. For multiple lesions, a Boolean operator was used to separate the isodose line in question from the isodose lines of the other lesions. The results of the new method were compared to those obtained following BrainLab’s recommendations.

Results: The absolute relative CI discrepancy between Brainlab and VelocityAI for single lesions averaged 1.63±0.9% for lesions >1cc. This discrepancy should be the same for multiple lesions. The results obtained using BrainLab’s recommendations differed from our method by 8.27±8% for multiple lesions >1cc.

Conclusion: BrainLab’s recommendation for obtaining CI for multiple lesions is not accurate. Exporting isodose lines from Brainlab into VelocityAI is one method of obtaining conformity indices for multiple targets. However, the accuracy of the result depends strongly on the size of the lesion. Although Brainlab does not directly report a correct CI for multiple lesions, it does rely on user feedback. The reader is encouraged to request CI calculation capabilities in Brainlab for multiple targets lesions.




Contact Email: