Ensuring that the desired dose is delivered to the prescribed location for each patient requires a well thought out QA program. Use of pre-treatment QA often requires a disproportionate amount of physics time and resources relative to the number of failures found, and traditional array-based pre-treatment QA methods have been shown to provide low sensitivity to error detection (ref 1). Additional methods must be considered if we are truly to provide complete QA for all patients. Independent in vivo verification of the plan during delivery is a significant step toward ensuring dose is delivered as prescribed. How does your clinic determine whether plans are delivered accurately during each and every fraction?

Independent automated methods using integrated EPID panels are one solution. The latest release of Adaptivo™ provides improved workflow for daily treatment evaluation. Adaptivo compares exit dosimetry measurements to predicted images to evaluate both accuracy and consistency of delivery. The predicted image is generated based on the plan, planning CT, and a LINAC specific beam model. This provides an independent assessment both that the correct plan is being delivered and that the correct patient setup and alignment occurred each day. Incorrect patient setup on day 1 which is corrected in subsequent fractions is shown in Figure 1.

Adaptivo was designed by clinical physicists with clinical evaluation in mind. Traditional gamma analysis provides no information concerning whether delivery is hotter or cooler than planned. Adaptivo’s hot/cold gamma analysis (Fig 1) is the default view because of the added clinical information it provides. The traditional gamma is given a sign based on if the measured value is hotter (positive/red) or cooler (negative/blue) for ease in evaluation and troubleshooting. Hot/cold gamma is calculated via post-processing of the gamma index map using the following the equation:

Γ_signed=Γ*[2*H(D_meas-D_pred)-1]

where H is the heaviside function.

In addition to confirmation of the patient plan delivery and alignment, weekly average alerts are a mechanism to trigger replanning conversations between the radiation oncologist and medical physicist. Daily differences either due to setup or variable anatomy such as bladder or bowel changes need to be differentiated from longer term changes such as tumor shrinkage or weight loss. Adaptivo facilitates review using weekly average data alerts and treatment delivery trends of the gamma pass rates. In the example shown in Figure 1, replanning triggered by the automatic alerts on the Adaptivo dashboard (Figure 2a) and backed up with trend data (Figure 2b) resulted in improved treatment. In this example, Adaptivo alerts helped facilitate significant reduction in the treatment volume, an increase in tumor dose, and a reduction in dose to normal tissue (Data courtesy of Brian Pomije, MS, DABR Radiation Therapy Physicist, GenesisCare – August 2020).

Figure 1: Incorrect setup of a breast plan during fraction 1 which is corrected on subsequent treatment days due to alert from Adaptivo. Analysis shown is 5%/3mm hot/cold gamma analysis with an alert threshold at 90% and failure threshold at 80%.

Figure 2: (a) Adaptivo’s dashboard provides a daily pass/alert/fail indication and weekly averages highlighted in grey for 38 delivered fractions. (b) Gamma analysis trends for each beam/arc support timing of replanning.

We are “Very satisfied at CHR Metz-Thionville with the results of our in-vivo transit dosimetry solution Adaptivo … Predicted images are automatically compared with EPID acquisitions showing that even our highly modulated VMAT beams are delivered with accuracy” – Paul Retif, Ph.D, Chief Medical Physicist (Figure 3)

Figure 3: Paul Retif using Adaptivo to review beam specific gamma and profile comparisons to ensure accuracy.

About the Author

Mary Napolitano, PhD is Product Manager for patient QA solutions at Standard Imaging. She obtained her graduate degrees at Georgia Institute of Technology and her bachelor’s degree in Nuclear Engineering from the University of Cincinnati. Prior to Standard Imaging, Mary has worked for multiple vendors in the radiation oncology domain.

Reference

1 - Kry SF, Molineu A, Kerns JR, et al. Institutional patient-specific IMRT QA does not predict unacceptable plan delivery. Int J Radiat Oncol Biol Phys. 2014; 90: 1195– 1201.