OBJECTIVETo evaluate the value of whole-body diffusion weighted imaging (WB-DWI) on detection of malignant metastasis.

METHODSForty-six patients with malignant tumors underwent WB-DWI examinations between April 2007 and August 2007 in our hospital. Before WB-DWI examination, the primary cancers of all the patients were confirmed by pathology, and the TNM-stage was assessed with conventional magnetic resonance imaging (MRI) or computed tomography (CT). WB-DWI was performed using short TI inversion recovery echo-planar imaging (STIR-EPI) sequence. Abnormal high signal intensities on WB-DWI were considered as metastases. The results of WB-DWI were compared with other imaging modalities. For the assessment of the diagnostic capability of WB-DWI, WB-DWI were compared with CT for demonstrating mediastinal lymph node metastases and lung metastases, and with conventional MRI for demonstrating metastases in other locations.

RESULTSWB-DWI demonstrated 143 focuses, 14 of which were diagnosed to be benign lesions in routine imaging. The number of bone metastases depicted on WB-DWI and routine imaging was 85 and 86; lymph node metastases was 17 and 18; liver metastases was 14 and 14; lung metastases was 4 and 8; and brain metastases was 6 and 8, respectively. WB-DWI failed to detect 12 metastatic lesions including 3 osteoplastic bone metastases, 4 lung metastases, 3 mediastinal lymph node metastases, and 2 brain metastases. Four metastatic lesions including 2 deltopectoral lymph nodes and 2 rib metastases were detected with WB-DWI alone, all of which evolved greatly during clinical follow-up for more than 6 months. WB-DWI had higher detection rates for metastatic lesions in liver, bone, and lymph nodes than those in lung and brain (chi2=30, P<0.001).

CONCLUSIONSWB-DWI could detect most of metastatic lesions that were diagnosed with conventional MRI and CT. The limitations of WB-DWI might be had high false-positive rate and low efficiency in detecting mediastinal lymph node, brain, and lung metastases.