FDG PET predicts prognosis of patients with resectable metastatic colorectal cancer (CRC) treated with preoperative chemotherapy

Background: Standard evaluation of response to chemotherapy in metastatic CRC is unsatisfactory. Metabolism based imaging using positron emission tomography and fluorodeoxyglucose (FDG-PET) is increasingly used to assess treatment efficacy. Aim: To evaluate the prognostic value of FDG-PET metabolic response to chemotherapy in resectable metastatic CRC. Methods: We reviewed retrospectively 382 patients with metastatic CRC between 7/2002 and 8/2006. Patients curatively operated from CRC metastases treated with preoperative chemotherapy and evaluated by FDG PET at baseline and before surgery were considered eligible for the study. Time to progression (TTP) was defined as the time between surgery and the first proof of recurrence of disease. Total Lesion Glycolysis (TLG) was used as the metabolic parameter. It represents the product of the mean standardised FDG uptake (SUV) of each lesion and its volume. The change of the sum of the TLG of all lesions was calculated. A cut-off of 30% was used to differentiate responders and non- responders. Results: 18 patients (median age 65 yrs) were included. 16 patients were operated from liver mets, and 2 from lung mets. Chemotherapy regimen was FOLFOX 4 in 10 patients, FOLFIRI in 7, and FOLFIRI-bevacizumab in 1 patient. Median number of lesions was 3 (1 to 10) before and 1 after chemotherapy. Median baseline TLG was 115.8 (6.8–841.8). Median post chemo TLG was 12.8 (0–1608.9). The distribution of the metabolic responses was CR 5, PR 9, SD 1, PD 3, with a response rate of 14/18 (78%, 95% CI : 52%-94%). Baseline TLG was not different between responders and non responders (p=0.88). Median follow-up was 422 days: 6 patients died, 15 had recurrent disease. Median TTP was 31 days for non responders, and 197 days for responders (p value not calculated because of the small number of non responding patients). After dichotomization of TLG values using their median as threshold, the post chemo TLG influenced significantly TTP (p=0.001, median TTP 86 days vs 263 days) along with the change of TLG (p=0.002, 107 days vs 263 days). Conclusions: Metabolic response assessment using PET allows a prognostic stratification after preoperative chemotherapy in resectable metastatic CRC.