Analysis of Dynamic MLC Files

Visualizing a dynamic sliding MLC window in a single image can be tricky. As the sliding window changes its outline constantly, the width of the field opening defined by each leaf pair is a function of time. In this composite image the colour plot shows the sliding window width for each leaf pair and shape as a function of time, encoded in colors. Red means that the actual field opening is less than 1 cm, yellow means 1-2 cm and so on. The largest gap is between 5 and 6 cm wide (purple).

The abscissa could also be labelled with shape index, dose fraction or accumulated MUs - they all mean the same thing. If the treatment is performed by beaming 250 MUs @ 600 MU/min and the DMLC file has 200 shapes, a dose fraction of 0.65 is reached after 163 MUs or 16 seconds. After 16 seconds shape number 130 is reached. So far the theory - if dose rate is not stable, things look different. But after all, this plot shows how things should look like.

At the dose fraction of 0.65 (or after 16 seconds), a snapshot of the actual MLC shape 130 is also shown standing perpendicular to the plot. The width of each leaf pair in the shape is encoded in different colours in the Contour Plot at the position of the vertical white line.

Calculation of Leaf Velocities

So far, there was not much dynamics in the example. Further analysis of the DMLC file can be done by calculating the leaf velocities of all leaves during the treatment. This should be done to check the Leaf Motion Calculator's output. When the LMC performs its optimization, it is aware of the physical speed limit of the leaves (here: 2.5 cm/s). When analyzing the LMC output, no leaf should exceed this limit during the treatment. For the velocity calculation, the simple equation s = v * t is taken, and a constant velocity within a segment is assumed. This is also what happens on the machine, since leaf speeds within a segment are constant. If the number of MUs (250) and the dose rate (600MU/min) is taken from above, velocities in cm/s can be calculated simply by taking the difference of leaf positions for subsequent shapes and dividing this distance by the time it take to "beam" from dose fraction i to dose fraction i+1.

Plots can be drawn for the leaf velocities of leaf bank A and B (vA, vB), or the maximum of both (Max[vA, vB]).

Dose rate modulations can occur as a result of wrong user settings or bad mechanical parts (e.g., if a leaf motor draws too much current as a result of increased friction). For instance, if CadPlan is configured to perform dose dynamic treatments at 300 MU/min and the user selects 600 MU/min on the machine, such modulations may occur. Or, if the dynamic leaf tolerance is reduced to a very small value, modulations of output are very likely to be the result. But this does not mean that the dose applied to the patient will be different! The resulting dose distribution will be the same, independent of dose rate variations. This is guaranteed by the control mechanisms and was already thoroughly investigated on our machines.

back to Medical Physics home