Does diabetes mellitus influence the efficacy of FDG-PET in the diagnosis of cervical cancer?

Abstract

Compared with computed tomography (CT) and magnetic resonance imaging (MRI), positron emission tomography (PET) may have additional value in the assessment of primary and recurrent cervical cancer. However, the degree of tumour uptake of (18)F-2-fluoro-2-deoxy-D: -glucose (FDG) uptake is sometimes influenced by diabetes mellitus (DM). Therefore, we conducted this prospective study to compare the diagnostic ability of FDG-PET in patients with cervical cancer complicated by DM and those without DM. Patients with untreated locally advanced primary or clinically curable recurrent cervical carcinoma were enrolled. Both FDG-PET and MRI/CT scans were performed within 2 weeks. Patients were categorised into the following groups: hyperglycaemic DM (fasting blood sugar >126 mg/dl), euglycaemic DM and non-DM. The lesions were confirmed histologically or by clinical follow-up. The receiver operating characteristic curve method, with calculation of the area under the curve (AUC), was used to evaluate the discriminative power. From February 2001 to January 2003, 219 patients (75 with primary and 144 with recurrent cervical cancer) were eligible for analysis. Sixteen had hyperglycaemic DM, 12 had euglycaemic DM and 191 were in the non-DM group. The diagnostic power of PET in the hyperglycaemic DM, euglycaemic DM and non-DM groups did not differ significantly with regard to the identification of either metastatic lesions (AUC, 0.967/0.947/0.925, P>0.05) or primary tumours/local recurrence (AUC, 0.950/0.938/0.979, P>0.05). Considering all DM patients, PET showed a significantly higher detection power than MRI/CT scans in respect of metastatic lesions (AUC=0.956 vs 0.824, P=0.012). In comparison with its accuracy in non-DM patients, the accuracy of PET in cervical cancer patients with mild to moderate DM was not significantly reduced.