Encrypted login | home

Program Information

Quantifying the Performance of Two Different Types of Commercial Software Programs for 3D Patient Dose Reconstruction: Machine Log-File Vs. Machine Log-File with EPID Image

N Kadoya

N Kadoya1*, Y Kon1 , Y Takayama1 , Y Katsuta1 , N Hayashi2 , K Ito1 , S Dobashi1 , K Sato3 , K Takeda1 , K Jingu1 , (1) Tohoku University Graduate School of Medicine, Sendai, Japan ,(2) School of Health Sciences, Fujita Health University, Toyoake, Japan, (3) Tohoku University Hospital, Sendai, Japan,


SU-I-GPD-T-264 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: We clarified the reconstructed 3D dose difference between two different types of commercial software programs.

Methods: Five prostate cancer patients treated with VMAT (74Gy/37fr) were studied. Log-file and cine EPID images (8 frames/image) were acquired for each fraction. Two commercial software programs were used: Mobius version.2.0.1 (Mobius Medical Systems) and PerFRACTION version.1.6.4 (Sun Nuclear Corp). Mobius can reconstruct 3D patient dose using log-file, whereas PerFRACTION can reconstruct patient dose using log-file with EPID image. First, we re-calculated the treatment planning dose on cylindrical phantom and then we acquired log-file and cine EPID images with various acquisition rates (3, 4, 8 and 10 frames/image). We measured the dose at three points using ion chamber and compared the measured point dose with reconstructed point dose in the phantom. Next, we compared dosimetric parameters (mean dose for PTV, rectum and bladder) calculated by Mobius and PerFRACTION for all fractions from five patients.  

Results: For phantom, there was no significant difference in point dose between measurement and reconstructed dose for both software (<2%). For PerFRACTION, reconstructed dose was not changed by frame acquisition rate of EPID (<1%). For patients, differences in dosimetric parameters were within 1% for almost all of fractions. PerFRACTION had wider range of dose difference between first fraction and the other fractions than Mobius (e.g., maximum difference: 0.34% for Mobius vs. 4.76% for PerFRACTION). One reason was that EPID might be detect the subtle error caused by loose or broken t-nut, which could not be detected by log-file. Other possible cause was that EPID data might include the image acquisition failure.

Conclusion: There was small differences in dosimetric parameters between two software, although there were no significant differences between measurement and reconstructed dose for both software in phantom. To clarify the reason, further investigation is needed (e.g., error-induced study).

Contact Email: