Despite a growing number of therapeutic regimens for the disease, the 5-year survival remains low at less than 10%. Given the variety of treatment options, improving clinicians ability to predict patients responses to various treatments would allow them to alter the therapeutic course for patients with unresponsive tumors. This new study suggests an approach that overcomes some of the limitations of the current gold standards for disease assessment, which are either limited to treatments that ultimately affect tumor size or those that cannot be applied, in many cases, until three months have elapsed after treatment. A total of 26 mCC patients from a single hospital completed the study, which involved pretreatment PET/CT and treatment with repeating cycles of capecitabine/oxaliplatin (CAPOX) and bevacizumab. Early and late PET/CT evaluations were performed, as well as peripheral blood collection at 20 days after the first treatment cycle and 20 days after the fourth cycle. In addition to functional imaging using FDG PET/CT, the group assessed the levels of the glycoprotein carcinoembryonic antigen (CEA), tissue inhibitor of metalloproteinases-1 (TIMP-1), liberated domain 1 of urokinase plasminogen activator receptor [uPAR(I)] and the mutation status of KRAS and BRAF oncogenes, all of which are known predictors of treatment response and/or prognosis. When patients were divided into responders and nonresponders, the early metabolic response category assessed by PET could predict the late morphological response category determined by CT in 73% of cases, with 80% sensitivity and 69% specificity. Early determination of metabolic nonresponse was significantly predictive of disease progression, with nonresponders having a greater than 3 times higher risk of progression compared to responders. A nonsignificant reduction in overall survival was also observed among metabolic nonresponders compared with responders, according to early PET evaluation. A 10-mm diameter increase in the largest lesion observed at pretreatment PET/CT was significantly predictive of poorer survival, as was the number of metastatic sites observed. High levels of TIMP-1 at both pretreatment and early post-treatment evaluations were significantly associated with an increased risk of death. Additionally, treatment significantly decreased TIMP-1 levels at the early evaluation. While CEA levels were characteristically high in almost all the mCC cases, there was no association between their levels and patient survival, although CEA did decrease following treatment in a borderline statistically significant manner. Finally, high uPAR(I) levels at both pretreatment and early evaluation were predictive of poorer survival, and the levels of this biomarker also decreased after treatment.