A feasibility study of automated inverse treatment planning for cancer of the prostate

Purpose: The development of automated "inverse planning," utilizing intensity-modulated radiation therapy (IMRT) raises the question of whether this new technique can provide a practical and efficient means of dose escalation in conformal treatment of cancer of the prostate. The purpose of this feasibility study was to determine a single set of inverse-planning parameters that can be used for a variety of different prostate patient geometries to automatically generate escalated dose (greater than or equal to 81 Gy) IMRT plans that satisfy normal tissue constraints for rectal and bladder walls. Methods: We studied a subset of the 46 patients who were previously treated at Memorial Sloan Kettering Cancer Center (MSKCC) to a total dose of 81 Gy using a 3D conformal approach, Six patients were selected for our study and replanned using an analytical inverse-planning algorithm (referred to as OPT3D) applied to 8 intensity modulated, co-axial radiation beams. A set of more than a dozen inverse planning parameters were adjusted by trial and error until the resulting dose distributions satisfied the critical organ dose-volume constraints imposed by our study rules (D30 less than or equal to 75.6 Gy and D10 less than or equal to 80 Gy for the rectal wall; D15 less than or equal to 80 Gy for the bladder wall) for the sample of patients selected. The OPT3D-generated plans were compared to hand-generated BEV plans using cumulative DVH analysis. Results: A single set of inverse-planning parameters was found that was able to automatically generate IMRT plans meeting all critical organ dose-volume constraints for all but one of the patients in our study. [The exception failed to meet bladder dose constraints for both IMRT and BEV methods, due to extensive overlap between the planning target volume (PTV) and bladder contours]. Based upon analysis of the cumulative dose-volume histogram (DVH) for the prostate PTV, the D95 (DX is defined such that x% of the volume receives a dose greater than or equal to DX), averaged over all patients, was approximately 81 Gy. The average D90 and mean dose values were 85 Gy and 93 Gy, respectively. Although a similar D95 was achieved using the BEV-generated plans, the D90 and mean dose values were substantially higher for the inverse planning (OPT3D) method. Conclusion: This limited "paper study" shows IMRT with inverse planning to be a promising technique for the treatment of prostate cancer to high doses. We determined a small set of inverse-planning parameter values that was able to automatically design intensity-modulated radiotherapy (IMRT) plans for a subset of 6 patients previously treated at MSKCC to 81 Gy using BEV planning techniques. With one minor exception, the resulting plans succeeded in meeting predetermined dose-volume constraints while at the same time allowing an increase in the mean dose and D90 to the prostate PTV. These 8 field plans also resulted in reduced dosage to the femoral heads. This automated technique is efficient in terms of planning effort and, with proper software for computer-controlled MLC, may be appropriate for clinical use. The clinical feasibility of this approach for a larger group of patients is currently under study. (C) 1998 Elsevier Science Inc.