Measurement of the interplay effect in lung IMRT treatment using EDR2 films

Abstract

Intra-fraction organ motion during the dynamic delivery of IMRT treatment of lung tumors may cause unexpected hot/cold spots even within the target volume, due to the interplay effect between tumor motion and multi-leaf collimator (MLC) leaf motion. In the past, this has been investigated through theoretical analysis, computer simulation, and experimental measurement using an ionization chamber dosimeter. In the work presented here, the interplay effect was studied experimentally in 2D, using Kodak EDR2 films. A five-field lung IMRT plan was delivered to a solid water phantom with embedded film. The phantom was placed on a motor-driven platform with a sinusoidal motion to simulate the respiration induced tumor motion. The delivery of each field began at one of eight equally spaced initial breathing phases. The dose distribution for each treatment fraction was estimated by combining the dose distributions for all fields with randomly sampled initial breathing phases. The dose variation caused by the interplay effect was estimated by looking at the dose values from 1000 trials of 30 fractions. It was found that, on a day-to-day basis, the standard deviation of the dose to a given pixel in the high dose region could be as high as 2-3% due to the motion interplay effect. After thirty fractions, the standard deviation in the dose to each pixel is reduced to about 0.3-0.5%. However, compared to the static delivery, the dose distribution from a thirty-fraction case in the presence of motion shows some under-dosing in the region of interest. Our results indicate that the interplay effect is of no significant consequence to tumor dose coverage in lung IMRT delivery over the entire course of a thirty-fraction treatment.

Figure 1. The film is aligned using a 3D square and then (b) placed in the NOMOS IMRT QA phantom.

Figure 2. The combination of the five fields delivered to the static platform. For irradiations with the platform in motion, the initial phase was randomly selected. Image (b) is an example of one possible distribution when motion is present. Image (c) is the difference between the example motion distribution and the static distribution. The values of (c) range from −9.8% to +15.4% of the total prescribed dose.

Figure 3. (a) and (b) show TD1{μ} for two different samples of initial motion phases, {μ}, of the platform. Subtracted from each other, (c) shows an example of the possible changes of the daily distribution delivered when motion is present. The values in this example range from −6.5% to +4.4% of the total prescribed dose.

Figure 4. A plot of the value of the standard deviation (1σ) of TD1{μ} at each point in the array after 1000 trials of randomly selected motion phases. The values of the standard deviation range from 1.2% to 5.4%. Within the high‐dose region (90% isodose curve), the variation is 2.0% to 3.8%.

Figure 5. Histogram of the standard deviation of 1000 trials of the single‐ case. Only pixels with a value of at least 90% on the static dose plot were used to calculate this histogram.

Figure 6. The sum of all five fields delivered to a static platform. (b) The average of 30 fractions of five fields delivered to a moving platform with randomly selected initial phases. (c) Some differences in the two distributions may still be seen after the averaging effects of 30 treatments.

Figure 7. The standard deviation (1σ) from 1000 trials of the 30‐fraction case.

Figure 8. The histogram of the standard deviation in each pixel in the region of interest for the 30‐fraction case.

Figure 9. The histograms for 1000 trials of single‐fraction treatments with motion. All the values are calculated within the region of interest (static 90% isodose curve). (a) The maximum (99th percentile) dose is shown along with the static value. (b) The minimum (1st percentile) dose is shown along with the static value. (c) The median dose is shown along with the static value. (d) The mean dose is shown along with the static value.

Figure 10. The histograms for 1000 trials of 30‐fraction treatments with motion. All the values are calculated within the region of interest (static 90% isodose curve). (a) The maximum (99th percentile) dose is shown along with the static value. (b) The minimum (1st percentile) dose is shown along with the static value. (c) The median dose is shown along with the static value. (d) The mean dose is shown along with the static value.