A phase I/II trial of three-dimensionally planned concurrent boost radiotherapy and protracted venous infusion of 5-FU chemotherapy for locally advanced rectal carcinoma

Background: Improving the response to preoperative therapy may increase the likelihood of successful resection of locally advanced rectal cancers. His torically, the pathologic complete response (pCR) rate has been < similar t o 10% with preoperative radiation therapy alone and < similar to 20% with c oncurrent chemotherapy and radiation therapy. Methods and Materials: Thirty-seven patients were enrolled on a prospective Phase I/II protocol conducted jointly at Washington University, St. Louis and the Catholic University of the Sacred Heart, Rome evaluating a three-di mensionally (3D) planned boost as part of the preoperative treatment of pat ients with unresectable or recurrent rectal cancer. Preoperative treatment consisted of 4500 cGy in 25 fractions over 5 weeks to the pelvis, with a 3D planned 90 cGy per fraction boost delivered once or twice a week concurren tly (no time delay) with the pelvic radiation. Thus, on days when the boost was treated, the tumor received a dose of 270 cGy in one fraction while th e remainder of the pelvis received 180 cGy. When indicated, nonaxial beams were used for the boost. The boost treatment was twice a week (total boost dose 900 cGy) if small bowel could be excluded from the boost volume, other wise the boost was delivered once a week (total boost dose 450 cGy). Patien ts also received continuous infusion of S-fluorouracil (1500 mg/m(2)-week) concurrently with the radiation as well as postoperative 5-FU/leucovorin. Results: All 37 patients completed preoperative radiotherapy as planned wit hin 32-39 elapsed days. Twenty-seven underwent proctectomy; reasons for unr esectability included persistant locally advanced disease (6 cases) and pro gressive distant metastatic disease with stable or smaller local disease (4 cases). Actuarial 3-year survival was 82% for the group as a whole. Among resected cases the 3-year local control and freedom from disease relapse we re 86% and 69%, respectively. Twenty-four of the lesions (65%) achieved an objective clinical response by size criteria, including 9 (24%) with pCR at the primary site (documented T0 at surgery). The most important factor for pCR was tumor volume: small lesions with planning target volume (PTV) < 20 0 cc showed a 50% pCR rate (p = 0.02). There were no treatment associated f atalities. Nine of the 37 patients (24%) experienced Grade 3 or 4 toxicitie s (usually proctitis) during preoperative treatment. There were an addition al 7 perioperative and 2 late toxicities. The most important factors for sm all bowel toxicity (acute or late) were small bowel volume (greater than or equal to 150 cc at doses exceeding 4000 cGy) and large tumor (PTV greater than or equal to 800 cc). For rectal toxicity the threshold is PTV greater than or equal to 500 cc. Conclusion: 3D planned boost therapy is feasible. In addition to permitting the use of nonaxial beams for improved dose distributions, 3D planning pro vides tumor and normal tissue dose-volume information that is important in interpreting outcome. Every effort should be made to limit the treated smal l bowel to less than 150 cc. Tumor size is the most important predictor of response, with small lesions of PTV < 200 cc most likely to develop complet e responses. <(c)> 2001 Elsevier Science Inc.