Background and objective: To prospectively analyze the diagnostic accuracy of PET/CT in non-Hodgkin's lymphoma (NHL) and to evaluate the most appropriate study protocol of this technique.
Patients and method: Seventy-six biopsy proven NHL patients were enrolled in this prospective study for 3 years. Patients initially underwent a low-dose CT without intravenous contrast, then a PET emission scan and finally a full-dose CT with intravenous contrast. For every patient, two modalities of PET/CT images were reconstructed: a low-dose unenhanced PET/CT and a full-dose enhanced PET/CT. Each modality was evaluated by either of two pairs of readers, different for each modality. Enhanced CT and PET images were evaluated by an independent radiologist and nuclear medicine physician respectively.
Results: Agreement between reference standard and techniques was as follows: 52.2% of patients with enhanced CT (κ=0.458), 46% with PET (κ=0.335), 75% with low-dose unenhanced PET/CT (κ=0.664) and 76.8% with full-dose enhanced PET/CT (κ=0.679), with p<0.001. Although all techniques underestimated the stage in comparison to gold standard, the lowest percentage was for full-dose enhanced PET/CT (20.3%).
Conclusions: PET/CT improved staging accuracy of NHL, being full-dose enhanced PET/CT the most accurate technique in our study.
Copyright © 2010 Elsevier España, S.L. All rights reserved.