Dose to bone marrow using IMRT techniques in prostate cancer patients.
PURPOSE: To investigate the dose distribution in active bone marrow of patients undergoing intensity-modulated radiotherapy (IMRT) for prostate cancer and compare it to the distribution in the same patients, if they had been treated using conformal plans, in order to develop criteria for optimization to minimize the estimated risk of secondary leukemia. PATIENTS AND METHODS: Mean bone marrow doses were calculated for ten patients with localized prostate cancer who underwent whole-pelvis IMRT and compared to three-dimensional conformal (3-D CRT) plans prepared for the same patients. Also for comparison, the IMRT and 3-D CRT plans were produced to simulate the treatment of the prostate gland only. To measure the dose to extrapelvic bone marrow, three thermoluminescent diode (TLD) chips were placed in the middle of the sternum region inside the Rando phantom. RESULTS: For both the pelvic and prostate-only volumes, the IMRT plans were superior to 3-D CRT plans in reducing the high dose volume to the rectum, the bladder and the small bowel while maintaining acceptable coverage of the planning target volume (PTV). For the pelvic treatment group the IMRT plans, compared to 3-D CRT, reduced the high dose volume (> 20 Gy) to os coxae, which is the main contributor of dose to pelvic bone marrow, but increased the middle dose volume (10-20 Gy). No statistically significant differences were observed for lower dose volumes (< 5 Gy). For the prostate-only treatment the IMRT plan increased the high dose volume and slightly decreased the low dose volume of pelvic bone marrow. However, for both treatments the leakage dose to extrapelvic sites was higher by a factor of 2 in IMRT plans. CONCLUSION: There are significant differences in the dose-volume histograms of bone marrow doses from 3-D CRT and from IMRT. Pronounced dose inhomogeneity reduces the risk of leukemia compared to homogeneous radiation exposure of the bone marrow. The mean bone marrow dose is therefore not a useful criterion to judge plan quality, since scattered low doses to distant sites may be more critical than the high dose volumes receiving > 10 Gy. The number of monitor units needed to deliver an IMRT plan affects leakage dose and their incorporation into planning constraints should be considered.