Intraprostatic fiducials for localization of the prostate gland: Monitoring intermarker distances during radiation therapy to test for marker stability

Purpose: The use of intraprostatic fiducials as surrogates for prostate gland position assumes that the markers are rigidly positioned within the prostate. To test this assumption, the intermarker distances (IMD) of implanted markers was monitored during the full course of radiation therapy to determine marker stability within the prostate gland. Methods and Materials: The analysis is performed on 56 patients treated with intensity-modulated radiotherapy. A total of 168 markers (3 markers per patient) were implanted. Two high-resolution X-rays were acquired before treatment delivery to visualize the position of the implanted markers. A total of 2,037 daily alignments were performed on the 56 cases (average: 36 alignments per patient). Each pair of X-ray images allows the computation of the 3 IMDs. A total of 6,111 IMDs were available for analysis. To study variations in marker position, daily IMDs were compared with the IMD that was observed during the first alignment. We defined the variation in the IMD as the important measure of intrinsic marker position variation. The standard deviation (SD) of IMD variations was studied as a measure of the extent of marker position variation. Particular attention was given to cases in which significant intermarker variations were observed. Results: The average directional variation of all IMDs (+/- SD) was -0.31 (+/- 1.41) mm. The average absolute variation of all IMDs (+/- SD) was 1.01 (+/- 1.03) mm. The largest observed variation in IMD was 10.2 mm. Among the individual 56 patients, the SDs of the IMD variations were computed and found to range from 0.4 to 4.2 mm. In 54 of the 56 patients (96%), the variations of all 3 IMDs had SD of 4.0 mm or less, which indicates little variation in the relative position of the markers. Only in 2 patients did any of the IMDs vary, with SD that exceeded 4.0 mm, which indicated noticeable and consistent marker-position variation. The maximum observed SD in the IMD variation was 4.2 mm. In each of the 2 cases, 2 IMDs were found to fluctuate, while the third IMD remained fairly constant. This finding means that 1 of 3 markers varied frequently in its relative position throughout the treatment. Therefore, only 2 of the 168 markers (1%) showed frequent changes in their relative positions. A review of these 2 cases revealed that the observed marker mobility was likely not caused by migration of the marker itself but caused by prostate deformation, secondary to rectal filling. To investigate the frequency of extreme situations, the maximum observed IMD variation was determined for each patient. In 47 of the 56 patients (84%), the maximum difference in IMDs was at least 2 mm. The corresponding numbers for 3, 4, and 5 mm were 23 (41%), 10 (18%), and 5 (9%) patients, respectively. Conclusion: This study is the largest reported series of localized prostate cancer patients with implanted intraprostatic markers used for daily target localization in which individual marker positions were registered and IMDs were computed to test for marker position variation. Only 2 of 168 implanted markers showed a relatively significant and consistent change in their relative position throughout a course of treatment. However, these variations in position were most likely not caused by marker migration but caused by prostate deformation. Typically, the IMDs varied minimally, which indicated relatively little deformation of the gland as well as the absence of significant marker migration. However, during a typical course of treatment, the IMD is likely to vary by several millimeters in some instances, which indicates infrequent but significant deformation. In these instances, an alignment based on the 3 markers' center of mass will still provide a meaningful alignment of the prostate within the radiation field. Intraprostatic implanted fiducials in the prostate allow a reliable and simple localization of the prostate gland, even in the presence of organ deformation. (c) 2005 Elsevier Inc.