It was reported to adac that the mlc shapes (and presumably block shapes) for the two control points set up for the open and wedge portions of the motor wedge beam are not linked leading to the potential for an error in dose distribution for minimally wedged fields. No injuries were reported as a result of this event.

Treatment beams that contain motorized wedges are composed of two distinct beam components (the wedged control point and the open control point). These two control points are intended to have identical treatment parameters. If the positions of any of the leaves of the multi-leaf collimator are manually modified in the mlc editor, the changes are intended to be applied to both control points. However, due to an issue in pinnacle3, these manual leaf position modifications are only applied to the wedged control point, and not to the open one. The computed dose distribution is calculated from the edited mlc positions of the wedged control point and the unedited mlc positions of the open control point. Furthermore, when exported via dicom rt, only the mlc changes to the wedged control point are transferred. As a result, the parameters used to compute the dose are not the parameters that are exported to the record and verify system. The possibility exists that a treatment could be delivered in which the computed dose does not match the delivered dose. All sites with this software were notified of this issue by phone, which was followed up with a customer letter.

Add'l info attached as follows: during the course of investigating this issue, other related info was discovered. The issue described in the initial medwatch report can happen when the dose has been fully or partially calculated (i. E. , user stops calculation after the wedged dose is calculated, but there the un-wedged dose is calculated). Edits to the jaw position (using the graphical jaw editing tool) were only appied to the wedged control point. Also, edits to beam modifiers may not invalidate dose appropriately.