Statistical Modeling of Relations Between PET/CT Parameters and CEA in Recurrent and Metastatic Colorectal Cancer

Background: Colorectal cancer (CRC) is a diverse disease with various clinical, pathological and molecular features that affect tumor biological behavior, treatment response and prognosis. Objective: The aim of this study was to evaluate the correlation between metabolic 18F-FDG PET/CT parameters (SUVmax, MTV and TLG) and CEA in recurrent and metastatic CRC and to evaluate prognostic value of metabolic 18F-FDG PET/CT parameters in recurrent and metastatic CRC. Methods: A descriptive study of 100 patients with previously detected and surgically treated CRC referred to PET/CT with a suspicion of recurrent or metastatic CRC. CEA was measured within three months from the imaging. A low-dose PET/CT was performed per institutional protocol. For each hypermetabolic lesion, metabolic PET/CT parameters (SUVmax, MTV, TLG) were calculated semiautomatically. Pathohistology or clinical data from the follow-up were used as the gold standard. Sensitivity, specificity, PPV and NPV for 18F-FDG PET/CT and CEA in detection of recurrent or metastatic CRC were calculated. Correlation between CEA and SUVmax, MTV and TLG was calculated, separately. To assess the prognostic values of metabolic parameters in CRC, survival analysis with 18-month progression-free survival (PFS) as an endpoint was performed. Microsoft Excel sheets, ROC and Kaplan-Meier curves were used to present the data. Logrank and Tarone-Ware test and Cox model of proportional hazards were used to compare the groups. Results: Study included 100 patients, 45 males and 55 females, age range 36-81 years, mean age 61,4 years. Cancer site was colon in 56% and rectum in 44%. Sensitivity, specificity, PPV and NPV of 18F-FDG PET/CT in detection of recurrent or metastatic CRC was 95%, 73%, 70% and 95%, respectively. Sensitivity, specificity, PPV and NPV of CEA in detection of recurrent or metastatic CRC was 58%, 96%, 91% and 78%, respectively. SUVmax, MTV and TLG positively correlated with CEA, but only CEA-TLG correlation was considered significant (r=0,67). The regression model analysis revealed: SUVmax (HR=0,63, 95%CI=0,28-1,41, p=0,214), MTV (0,59, 95%CI=0,28-1,22, p=0,111) and TLG (HR=0,45 95%CI=0,21-0,99, p=0,028), and the prognostic role in CRC was proven for TLG only. Conclusion: Metabolic 18F-FDG PET/CT parameters may have the prognostic value in CRC, but further multicentric prospective studies are required for validation.