EPID based in vivo dosimetry system: clinical experience and results

Abstract

Mandatory in several countries, in vivo dosimetry has been recognized as one of the next milestones in radiation oncology. Our department has implemented clinically an EPID based in vivo dosimetry system, EPIgray, by DOSISOFT S.A., since 2006. An analysis of the measurements per linac and energy over a two-year period was performed, which included a more detailed examination per technique and treat-ment site over a six-month period. A comparison of the treatment planning system doses and the doses estimated by EPIgray shows a mean of the differences of 1.9% (± 5.2%) for the two-year period. The 3D conformal treatment plans had a mean dose difference of 2.0% (± 4.9%), while for intensity-modulated radiotherapy and volumetric-modulated arc therapy treatments the mean dose difference was -3.0 (± 5.3%) and -2.5 (± 5.2%), respectively. In addition, root cause analyses were conducted on the in vivo dosimetry measurements of two breast cancer treatment techniques, as well as prostate treatments with intensity-modulated radiotherapy and volumetric-modulated arc therapy. During the breast study, the dose differences of breast treatments in supine position were correlated to patient setup and EPID positioning errors. Based on these observations, an automatic image shift correc-tion algorithm is developed by DOSIsoft S.A. The prostate study revealed that beams and arcs with out-of-tolerance in vivo dosimetry results tend to have more complex modulation and a lower exposure of the points of interest. The statistical studies indicate that in vivo dosimetry with EPIgray has been successfully imple-mented for classical and complex techniques in clinical routine at our institution. The additional breast and prostate studies exhibit the prospects of EPIgray as an easy supplementary quality assurance tool. The validation, the automatization, and the reduction of false-positive results represent an important step toward adaptive radiotherapy with EPIgray.

Figure 1

Chart of the workflow of the IVD controls, on the example of IMRT plans.

Figure 2

Example of IVD images and points of interest of breast treatment plans in (a) supine and (b) lateral positions in EPIgray.

Figure 3

Example of IVD images of a prostate treatment obtained for (a) an IMRT beam and (b) a VMAT arc, in EPIgray.

Figure 4

Dose difference distribution of the IVD results for the period of July 2013 to June 2015.

Figure 5

Dose difference distribution of the IVD results for the period of July 2013 to January 2014.

Figure 6

Illustration of the effect of positioning errors on IVD results: (a) in presence of a setup error, the point of interest receives the correct dose but is associated with the wrong CT parameters, such as patient thickness; (b) the signal at the reference point is different than expected, thus a large dose difference is calculated; (c) by realigning the image with the TPS contours, the correct signal is reassigned to the reference point.

Figure 7

Distribution of the differences between recorded EPID shifts and the Y component of the manual translation performed on IVD images.

Figure 8

Distribution of the transit image translations in the imager's lateral (X) direction, for internal and external mammary fields.