Medical device manufacturer, RaySearch Laboratories, has issued a medical device safety alert concerning its treating planning system, RayStation. The affected devices are: RayStation 3.5, RayStation 4.0, RayStation 4.5, RayStation 4.7, RayStation 5, RayStation 6, RayPlan 1 (RayStation 4.9) and RayPlan 2 (RayStation 6).
Electron dose calculation in RayStation uses a Monte Carlo dose engine, where a large number of particles are simulated. Each simulated particle is called a history, adding to the simulated dose. The number of histories can be adjusted for each dose calculation. Generally, a higher number of histories leads to a better dose accuracy due to lower statistical uncertainty, but adds to the calculation time. In beam modeling, using a higher number of histories may also be desired to determine an appropriate normalization value for commissioning the treatment machine.
When starting the dose calculation, the data variable representing the number of histories is erroneously converted to a number format which cannot handle sufficiently large values. This means that when a very large number of histories is requested, the Monte Carlo dose engine may use a lower number of histories than requested. At the end of the dose calculation, the resulting dose is divided by the intended number of histories. This leads to a calculated dose that is lower than the actual dose. The shape of the dose distribution will be correct but the absolute dose level will be wrong.
The error can arise during treatment planning in RayStation and in RayPlan, as well as during beam commissioning in RayPhysics and in RayPlanPhysics. When the number of histories during treatment planning is above the limit, the computed dose is underestimated which leads to overdosage.
Commissioning a machine with a normalization value determined by an erroneous calculation would scale the monitor units so that too few monitor units are prescribed. This results in an underdosage to all patients, regardless of the number of histories used in treatment planning.
If the large number of histories which is required to trigger the error is used, this would result in very long calculation times. This large number of histories is also well above the required number of histories to get a clinically acceptable dose accuracy.
For an individual treatment plan, the error could be detected by common clinical knowledge of the expected dose per monitor unit relationship for electron beams or by secondary calculations. For beam commissioning, the error could be detected by validation of the beam model.
The manufacturer advises affected users to take the following actions:-
According to the manufacturer, the affected products are distributed in Hong Kong.
If you are in possession of the affected products, please contact your supplier for necessary actions.
Posted on 24 November 2017