Pretreatment clinical findings predict outcome for patients receiving preoperative radiation for rectal cancer

Background: As a sole modality, preoperative radiation for rectal carcinoma achieves a local control comparable to that of postoperative radiation plu s chemotherapy. Although the addition of chemotherapy to preoperative treat ment improves the pathologic complete response rate, there is also a substa ntial increase in acute and perioperative morbidity. Identification of subs ets of patients who are at low or high risk for recurrence can help to opti mize treatment. Methods: During the period 1977-95, 384 patients received preoperative radi ation therapy for localized adenocarcinoma of the rectum. Ages ranged from 19 to 97 years (mean 64.4), and there were 171 females, Preoperative treatm ent consisted of conventionally fractionated radiation to 3600-5040 cGy (me dian 4500 cGy) 6-8 weeks before surgery in 293 cases or low doses of < 3000 cGy (median 2000 cGy) immediately before surgery in 91 cases. Concurrent p reoperative chemotherapy was given to only 14 cases in this study period. P ostoperative chemotherapy was delivered to 55 cases. Results: Overall 93 patients have experienced recurrence (including 36 loca l failures), Local failures were scored if they occurred at any time, not j ust as first site of failure, For the group as a whole, the actuarial (Kapl an-Meier) freedom from relapse (FFR) and local control (LC) were 74% and 90 % respectively at 5 years. Univariate analysis of clinical characteristics demonstrated a significant (p < 0.05) adverse effect on both LC and FFR for the following four clinical factors: (I) location <5 cm from the verge, (2 ) circumferential lesion, (3) near obstruction, (4) tethered or fixed tumor . Size, grade, age, gender, ultrasound stage, CEA, radiation dose, and the use of chemotherapy were not associated with outcome. Background of the sur geon was significantly associated with outcome, colorectal specialists achi eving better results than nonspecialist surgeons, We assigned a clinical sc ore of 0 to 2 on the basis of how many of the above four adverse clinical f actors were present: 0 for none, I for one or two, 2 for three or four. Thi s sorted outcome highly significantly (p less than or equal to 0,002, Taron e Ware), with 5-year LC/FFR of 98%/85% (score 0), 90%/72% (score 1), and 74 %/58% (score 2), The scoring system sorts the data for both subgroups of su rgeons; however, there are substantial differences in LC on the basis of th e surgeon's experience, For colorectal specialists (251 cases), the 5-year LC is 100%, 94%, and 78% for scores of 0, 1, and 2, respectively (p = 0.004 ). For the more mixed group of nonspecialist surgeons (133 cases), LC is 98 %, 80%, and 65% for scores of 0, 1, and 2 (p = 0,008), In multivariate anal ysis, the clinical score and surgeon's background retained independent pred ictive value, even when pathologic stage was included, Conclusions: For many patients with rectal cancer, adjuvant treatment can b e administered in a well-tolerated sequential fashion-moderate doses of pre operative radiation followed by surgery followed by postoperative chemother apy to address the risk of occult metastatic disease. A clinical scoring sy stem has been presented here that would suggest that the local control is e xcellent for lesions with a score of 0 or (if the surgeon is experienced) 1 , and therefore sequential treatment could be considered. Cases with a clin ical score of 2 should be strongly considered for protocols evaluating more aggressive preoperative treatment, such as combined modality preoperative treatment, (C) 2001 Elsevier Science Inc.