P-0296 Could Pre-Treatment Cea be an Effective Predictor of Relapse in High Risk Rectal Cancer?

Abstract

ABSTRACT Introduction Over 14,000 cases of rectal cancer occur per year in the UK. Radical surgery with total mesorectal excision (TME) is the main curative treatment for disease localised to the pelvis. Some cases are judged on pelvic MRI to be at high risk of positive margins from surgery alone. They are offered pre-operative chemo-radiotherapy (CRT) before surgery with TME. This hopes to downsize and downstage tumors to provide clear circumferential resection margins and reduce the likelihood of relapse. Good pathological responses to CRT are associated with improved cancer outcomes. We investigated whether carcinoembryonic antigen (CEA) levels at presentation could act as a simple predictor of relapse when using CRT followed by TME for patients with high-risk rectal cancers. Methods Retrospective data was collected on high-risk rectal cancer patients treated radically with pre-op CRT and TME surgery at Ipswich Hospital between March 2005 and Jan 2012. All received at least 45 Gy in 25 fractions of radiotherapy using a three-field CT plan and 5 weeks of oral capecitabine chemotherapy. CEA levels were obtained prior to radiotherapy. Following surgery, rectal specimens were assessed by our pathologists to evaluate response to treatment and whether complete pathological response (pCR) was achieved. Follow-up was by our oncology department and colorectal surgeons with annual CT scanning, 1-3 yearly colonoscopies and clinical surveillance for 5 years. Results 56 patients met inclusion criteria. Male to female ratio: 1.8:1. 96% (n=54) patients received pre-op chemotherapy alongside radiotherapy: 2 patients were considered too high-risk for this component of treatment. Mean follow-up from commencing radiotherapy was 33 months (range 7-69). Mean time from last dose of radiotherapy to surgery was 11 weeks (range 5-18). 93% (n=52) had (R0) resection margins following surgery. 7.1% (n=4) had (R1) incomplete resection margins following surgery; 2 of these 4 suffered local relapse in the pelvis. 30% (n=17) achieved a pCR following CRT, none of the patients who achieved a pCR have relapsed. 77% (n=43) presented with a normal CEA (≤5, range 5, range 6-54). Normal CEA group: 35%, (n=15) achieved pCR following CRT and 14% (n=6) relapsed. High CEA group: 15% (n=2) achieved pCR following CRT and 46%, (n=6) relapsed. There was a significant between group difference in relapse (Fisher's exact test p=0.02). Further survival analysis for relapse revealed a statistically significant hazard ratio of 4.5 (95% CI 1.04 to 19.4). Conclusion 1. High CEA on presentation was associated with an estimated relative risk of relapse 4.5 times greater than presenting with a normal CEA. 2. Therefore high CEA at presentation could be a useful biomarker to identify patients with the highest risk of relapse who may benefit from more intensive treatment. 3. Our high rates of R0 resection margin (93%) and pCR (30%) could be due to a relatively prolonged time from radiotherapy to surgery in our cohort.