Michael Tallhamer, MSC, DABR
We typically don’t see this type of variation in the rotational deltas with the variability in output typically sub 0.1 degree.
A few things to look into…
It may be due to the phantom surface but it would be very difficult to say without seeing the raw sensor data. You can look at the sensor data in the phantom patient folder and see if there are large over exposed (i.e. hot white spots) in the image which may indicate you need to have the apertures readjusted and the histograms rebalanced.
You can’t test the system with a real patient but you or a volunteer may lay on the table in a mask and evaluate the RTD exports for monitoring that person for say 5 minutes and compare the stats of that person to the phantom to make sure the phantom is a proper representation of what you will see on the patients you will treat (please use a reference capture and not a CT of the volunteer and please don’t image them during the process…I only say something because people have asked me). If the RTD stats differ between the volunteer and the phantom you can try adjusting some of your exposure settings for the surface settings in the software, change the phantom complexion (using masking tape works well) or call service to look into the camera settings.
If you place isocenters all over the cranial vault and monitor you should see some spread of the noise in the signal which I have always attributed to the decrease in focus of the speckle patter away from some optimal point but again it is always sub 0.1 degree for my systems and others I have commissioned in the past. I have seen it around 0.1 on a one or two of them.
Procedurally, you want to be careful to stand off the head of the table with you back to the gantry while adjusting the head adjuster so that you aren’t shadowing the ROI with your body or arm while adjusting the head plate. The surface for an SRS in high resolution and a larger ROI combined with the higher resolution surface will produce some “jumpier” deltas if you are obstructing large sections of the ROI while adjusting it.
Hope some of that gets you down the road to finding the solution