Diagnostic accuracy of whole-body MRI/DWI image fusion for detection of malignant tumours: a comparison with PET/CT.

Abstract

To prospectively evaluate the diagnostic accuracy of whole-body T2-weighted (wbT2), whole-body diffusion-weighted imaging (wbDWI) and wbT2/wbDWI image fusion for malignant tumour detection compared with PET/CT. Sixty-eight patients (44 men; 60 ± 14 years) underwent PET/CT for staging of malignancy and were consecutively examined by 1.5-Tesla MRI including wbT2 and wbDWI. Two radiologists independently assessed wbDWI, wbT2, wbT2 + wbDWI (side-by-side) and wbT2 + wbDWI + wbT2/wbDWI image fusion for the presence of malignancy. PET/CT served as a reference standard. PET/CT revealed 374 malignant lesions in 48/64 (75%) patients. Detection rates and positive predictive value (PPV) of wbT2 and wbDWI alone were 64% and 84%, and 57% and 93%, respectively. Detection rates and PPV of wbT2 and wbDWI for side-by-side analysis without and with fused images were 72% and 89%, and 74% and 91%, respectively. The detection rate was significantly higher with side-by-side analysis and fused image analysis compared with wbT2 and wbDWI alone (p = .0159; p < .0001). There was no significant difference between fused image interpretation and side-by-side analysis. WbDWI allows detection of malignant lesions with a similar detection rate to wbT2. Side-by-side analysis of wbT2 and wbDWI significantly improves the overall detection rate and fused image data provides no added value.