Preoperative staging of patients with liver metastases of colorectal carcinoma. Does PET/CT really add something to multidetector CT?

Abstract

This study was designed to determine prospectively whether the systematic use of PET/CT associated with conventional techniques could improve the accuracy of staging in patients with liver metastases of colorectal carcinoma. We also assessed the impact on the therapeutic strategy. Between 2006 and 2008, 97 patients who were evaluated for resection of LMCRC were prospectively enrolled. Preoperative workup included multidetector-CT (MDCT) and PET/CT. In 11 patients with liver steatosis or iodinated contrast allergy, MR also was performed. Sixty-eight patients underwent laparotomy. Sensitivity, specificity, positive predictive value (PPV), and negative predictive values for hepatic and extrahepatic staging of MDCT and PET-CT were calculated. In a lesion-by-lesion analysis of the hepatic staging, the sensitivity of MDCT/RM was superior to PET/CT (89.2 vs. 55%, p < 0.001). On the extrahepatic staging, PET/CT was superior to MDCT/MR only for the detection of locoregional recurrence (p = 0.03) and recurrence in uncommon sites (p = 0.016). New findings in PET/CT resulted in a change in therapeutic strategy in 17 patients. However, additional information was correct only in eight cases and wrong in nine patients. PET/CT has a limited role in hepatic staging of LMCRC. Although PET-CT has higher sensitivity for the detection of extrahepatic disease in some anatomic locations, its results are hampered by its low PPV. PET/CT provided additional useful information in 8% of the cases but also incorrect and potentially harmful data in 9% of the staging. Our findings support a more selective use of PET/CT, basically in patients with high risk of local recurrence.