OBJECTIVETo compare the diagnostic value of ¹⁸F-fluorodeoxyglucose-positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) and large-scale diffusion weighted imaging (DWI) for evaluation of non-Hodgkin lymphoma (NHL) bone marrow (BM) infiltration.

METHODSA total of 79 patients with pathologically diagnosed NHL underwent ¹⁸F-FDG PET/CT, large scale DWI and BM pathological examination. BM examination as the "gold standard", the performance (the sensitivity, specificity, accuracy, positive and negative predictive value) of ¹⁸F-FDG PET/CT and large scale DWI for evaluation of BM infiltration was compared and the risk of BM infiltration of different subtypes and sources of NHL was analyzed.

RESULTS25 of 79 cases were diagnosed as BM infiltration by pathological examination with 57 BM sites. Abnormal high BM metabolisms were identified in 22 cases with 56 BM sites by ¹⁸F-FDG PET/CT and 25 cases with 58 BM sites by large-scale DWI. The sensitivity, specificity, accuracy, positive and negative predictive value of ¹⁸F-FDG PET/CT were 80.0%, 96.3%, 91.1%, 90.9%, 91.2%, respectively. And they were 84.0%, 92.6%, 89.9%, 84.0%, and 92.6% by large-scale DWI, respectively. A receiver operating characteristic (ROC) analysis demonstrated that there was no statistical difference in ¹⁸F-FDG PET/CT and large-scale DWI (P>0.05). The area under ROC curve for ¹⁸F-FDG PET/CT and large-scale DWI were 0.911 and 0.883 respectively. The incidences of BM infiltration in aggressive NHL patients by ¹⁸F-FDG PET/CT (21/69, 30.4%) and large-scale DWI (23/69, 33.3%) were higher than those (PET/CT: 10.0%; large-scale DWI: 20.0%; P>0.05) in indolent NHL patients.

CONCLUSION¹⁸F-FDG PET/CT and large-scale DWI had important clinical value in diagnosing BM infiltration of NHL. A combination of ¹⁸F-FDG PET/CT, large-scale DWI and pathological examination could improve the positive rate of BM infiltration in NHL.