Radiotherapy of prostate cancer with or without intensity modulated beams:A planning comparison

Purpose: To evaluate whether intensity modulated radiotherapy (IMRT) by sta tic segmented beams allows the dose to the main portion of the prostate tar get to escalate while keeping the maximal dose at the anterior rectal wall at 72 Gy. The value of such IMRT plans was analyzed by comparison with non- IMRT plans using the same beam incidences. Methods and Materials: We performed a planning study on the CT data of 32 c onsecutive patients with localized adenocarcinoma of the prostate. Three fi elds in the transverse plane with gantry angles of 0 degrees, 116 degrees, and 244 degrees were isocentered at the center of gravity of the target vol ume (prostate and seminal vesicles). The geometry of the beams was determin ed by beam's eye view autocontouring of the target volume with a margin of 1.5 cm. In study 1, the beam weights were determined by a human planner (3D -man) or by computer optimization using a biological objective function wit h (3D-optim-lim) or without (3D-optim-unlim) a physical term to limit targe t dose inhomogeneity. In study 2, the 3 beam incidences mentioned above were used and in-field un iform segments were added to allow IMRT. Plans with (IMRT-lim) or without ( IMRT-unlim) constraints on target dose inhomogeneity were compared. In the IMRT-lim plan, target dose inhomogeneity was constrained between 15% and 20 %. After optimization, plans in both studies were normalized to a maximal r ectal dose of 72 Gy. Biological (tumor control probability [TCP], normal ti ssue complication probability [NTCP]) and physical indices for tumor contro l and normal tissue complication probabilities were computed, as well as th e probability of the uncomplicated local control (P+). Results: The IMRT-lim plan was superior to all other plans concerning TCP ( p < 0.0001). The IMRT-unlim plan had the worst TCP. Within the 3D plans, th e 3D-optim-unlim had the best TCP, which was significantly different from t he 3D-optim-lim plan (p = 0.0003). For rectal NTCP, both IMRT plans were su perior to all other plans (p < 0.0001). The IMRT-unlim plan was significant ly better than the IMRT-lim plan (p < 0.0001). Again, 3D-optim-unlim was su perior to the other 3D plans (p < 0.0007). Physical endpoints for target showed the mean minimal target dose to be the lowest in the IMRT-unlim plan, caused by a large target dose inhomogeneity (TDI). Medial target dose, 90th percentile, and maximal target dose were s ignificantly higher in both IMRT plans. Physical endpoints for the rectum showed the IMRT-unlim plan to be superior compared to all other plans. There was a strong correlation between the 65 th percentile (Rp6S) and rectal NTCP (correlation coefficient greater than or equal to 89%). For bladder, maximal bladder dose was significantly highe r in the IMRT-unlim plan compared to all other plans (p less than or equal to 0.0001). P+ was significantly higher in both IMRT-plans than in all other plans. The 3D-optim-unlim plan was significantly better than the two other 3D plans ( p < 0.0001). Conclusion: IMRT significantly increases the ratio of TCP over NTCP of the rectum in the treatment of prostate cancer. However, constraints for TDI ar e needed, because a high degree of TDI reduced minimal target dose. IMRT im proved uncomplicated local control probability. In our department, IMRT by static segmented beams is planned and delivered in a cost-effective way. IM RT-lim has replaced non-modulated conformal radiotherapy as the standard tr eatment for prostate cancer. (C) 2000 Elsevier Science Inc.