The Use Of Electronic Portal Imaging In Continuous Mode To Correlate External Marker Motion With Internal Tumor Motion During Respiratory-Monitored Treatment

Purpose/Objective(s)To investigate the use of electronic portal imaging acquired during treatment as a method to correlate internal tumor motion with external fiducial motion and to determine intra-fraction variability of non-gated, respiratory-gated, and respiratory-gated IMRT delivery without adding radiation dose to the patient.Materials/MethodsA Varian AS1000 Electronic Portal Imaging Device (EPID) was used in continuous acquisition mode to acquire high quality images at a rate as high as 13 images/sec during multiple treatment fractions to 5 free breathing lung cancer patients and a motion phantom. One patient received non-gated 3D-CRT, 1 patient received respiratory-gated 3D-CRT, and 3 patients received respiratory-gated step-and-shoot IMRT. As a proof of concept, a motion phantom composed of a water equivalent sphere embedded in a cedar cylinder that moved 2 cm in the S/I direction with a 4s period received each delivery technique. The A/P motion of the phantom where a marker block was placed was examined both in phase with the moving cylinder and out-of-phase by 1s. All patients and the phantom were monitored using the Varian RPM Respiratory Gating System. The log file generated by the RPM software during delivery was used to determine the “beam-on” time and assign an external marker position to each acquired EPI. For gated IMRT delivery, a dynalog file, which records the actual dose fraction vs. MLC leaf position every 50 ms, was used to determine “beam-off” time due to leaf transition while within the gating window. The internal tumor position was determined for each EPI. The external marker position was then correlated with the internal tumor motion using the Pearson Correlation Coefficient (r2). The time shift, t, was calculated by shifting one curve until r2 was maximized. The intra-fraction variability in tumor position for each treatment fraction was defined as standard deviation of the mean position of the tumor generated on each EPI acquired during delivery.ResultsThe phantom study showed r2=0.98 when the marker block was moving in phase with the moving cylinder and, r2=0.14 when out-of-phase by 1s. However, r2=0.98 when the RPM curve was shifted 1s. Similar results were obtained for both gated 3D-CRT and gated IMRT. For the patient studies, the correlation between external motion and internal motion ranged between 0.954 ≥ r2≥0.886 with a time shift as high as 0.15s. Intra-fraction variation was (2.1+/−0.57) mm for non-gated delivery and (0.94+/−0.44) mm for gated delivery.ConclusionsElectronic Portal Imaging in continuous acquisition mode can be used effectively to correlate the external motion with internal motion during treatment and to determine intra-fraction variability without adding radiation dose to the patient. Purpose/Objective(s)To investigate the use of electronic portal imaging acquired during treatment as a method to correlate internal tumor motion with external fiducial motion and to determine intra-fraction variability of non-gated, respiratory-gated, and respiratory-gated IMRT delivery without adding radiation dose to the patient. To investigate the use of electronic portal imaging acquired during treatment as a method to correlate internal tumor motion with external fiducial motion and to determine intra-fraction variability of non-gated, respiratory-gated, and respiratory-gated IMRT delivery without adding radiation dose to the patient. Materials/MethodsA Varian AS1000 Electronic Portal Imaging Device (EPID) was used in continuous acquisition mode to acquire high quality images at a rate as high as 13 images/sec during multiple treatment fractions to 5 free breathing lung cancer patients and a motion phantom. One patient received non-gated 3D-CRT, 1 patient received respiratory-gated 3D-CRT, and 3 patients received respiratory-gated step-and-shoot IMRT. As a proof of concept, a motion phantom composed of a water equivalent sphere embedded in a cedar cylinder that moved 2 cm in the S/I direction with a 4s period received each delivery technique. The A/P motion of the phantom where a marker block was placed was examined both in phase with the moving cylinder and out-of-phase by 1s. All patients and the phantom were monitored using the Varian RPM Respiratory Gating System. The log file generated by the RPM software during delivery was used to determine the “beam-on” time and assign an external marker position to each acquired EPI. For gated IMRT delivery, a dynalog file, which records the actual dose fraction vs. MLC leaf position every 50 ms, was used to determine “beam-off” time due to leaf transition while within the gating window. The internal tumor position was determined for each EPI. The external marker position was then correlated with the internal tumor motion using the Pearson Correlation Coefficient (r2). The time shift, t, was calculated by shifting one curve until r2 was maximized. The intra-fraction variability in tumor position for each treatment fraction was defined as standard deviation of the mean position of the tumor generated on each EPI acquired during delivery. A Varian AS1000 Electronic Portal Imaging Device (EPID) was used in continuous acquisition mode to acquire high quality images at a rate as high as 13 images/sec during multiple treatment fractions to 5 free breathing lung cancer patients and a motion phantom. One patient received non-gated 3D-CRT, 1 patient received respiratory-gated 3D-CRT, and 3 patients received respiratory-gated step-and-shoot IMRT. As a proof of concept, a motion phantom composed of a water equivalent sphere embedded in a cedar cylinder that moved 2 cm in the S/I direction with a 4s period received each delivery technique. The A/P motion of the phantom where a marker block was placed was examined both in phase with the moving cylinder and out-of-phase by 1s. All patients and the phantom were monitored using the Varian RPM Respiratory Gating System. The log file generated by the RPM software during delivery was used to determine the “beam-on” time and assign an external marker position to each acquired EPI. For gated IMRT delivery, a dynalog file, which records the actual dose fraction vs. MLC leaf position every 50 ms, was used to determine “beam-off” time due to leaf transition while within the gating window. The internal tumor position was determined for each EPI. The external marker position was then correlated with the internal tumor motion using the Pearson Correlation Coefficient (r2). The time shift, t, was calculated by shifting one curve until r2 was maximized. The intra-fraction variability in tumor position for each treatment fraction was defined as standard deviation of the mean position of the tumor generated on each EPI acquired during delivery. ResultsThe phantom study showed r2=0.98 when the marker block was moving in phase with the moving cylinder and, r2=0.14 when out-of-phase by 1s. However, r2=0.98 when the RPM curve was shifted 1s. Similar results were obtained for both gated 3D-CRT and gated IMRT. For the patient studies, the correlation between external motion and internal motion ranged between 0.954 ≥ r2≥0.886 with a time shift as high as 0.15s. Intra-fraction variation was (2.1+/−0.57) mm for non-gated delivery and (0.94+/−0.44) mm for gated delivery. The phantom study showed r2=0.98 when the marker block was moving in phase with the moving cylinder and, r2=0.14 when out-of-phase by 1s. However, r2=0.98 when the RPM curve was shifted 1s. Similar results were obtained for both gated 3D-CRT and gated IMRT. For the patient studies, the correlation between external motion and internal motion ranged between 0.954 ≥ r2≥0.886 with a time shift as high as 0.15s. Intra-fraction variation was (2.1+/−0.57) mm for non-gated delivery and (0.94+/−0.44) mm for gated delivery. ConclusionsElectronic Portal Imaging in continuous acquisition mode can be used effectively to correlate the external motion with internal motion during treatment and to determine intra-fraction variability without adding radiation dose to the patient. Electronic Portal Imaging in continuous acquisition mode can be used effectively to correlate the external motion with internal motion during treatment and to determine intra-fraction variability without adding radiation dose to the patient.