Current gold standard imaging in the assessment of lymphoma is 2-deoxy-2-(18F) fluoro-D-glucose positron emission tomography fused with computer tomography (PET-CT) scan. Whole body MRI (WB-MRI), integrating structural and functional MRI sequences, offers an alternative radiation-free imaging method. However, currently, there is no consensus on appropriate MRI sequences for WB-MRI protocols with conflicting reports on diagnostic performances of different protocols [1, 2, 3]. We evaluated multiple MRI sequences as part of a WB-MRI protocol for staging of newly diagnosed lymphoma patients.Twenty newly diagnosed patients (Male/female 10/10, median age 31.5, range 22-88) with histologically proven lymphoma were prospectively recruited and underwent WB-MR imaging using a 3.0 T scanner. Axial T2-weighted turbo-spin echo (TSE) and diffusion-weighted imaging (DWI) were supplemented by coronal pre- and post-contrast 2-point mDixon imaging, covering head to mid-thigh (figure 1). For each patient, two radiologists, blinded to other imaging investigations, independently reviewed 3 set of images (1) T2-weighted only, (2) b1000 DWI+ T1 pre-contrast in-phase mDixon (IP) and (3) T1 post-contrast water-only mDixon (Post-C). Each set of images was reviewed in a random paradigm with a minimum of 7 days wash out period between readings for the same patient. Involvement of nodal sites (cervical/supraclavicular, subpectoral/axillary, iliac, inguinal, mesenteric/retroperitoneal, mediastinal, splenic hilar and liver hilar) and extra-nodal sites (lung, chest wall, liver, spleen, stomach and bone marrow) were highlighted and transferred to study specific proforma and final Ann-Arbor staging for each sequence derived. All patients underwent PET-CT imaging at baseline, which in conjunction with follow-up imaging and multi-disciplinary team meeting outcome established the reference standard against which MRI sequences were tested. Cohen’s kappa statistic was used to assess agreement of final staging for each sequence between each reader and PET-CT, and between the two reporting radiologists. Agreement for nodal and extra-nodal staging for each sequence between each reader and PET-CT was derived using Cohen’s kappa test and sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) was calculated for each sequence. Kappa agreement between each reader and PET-CT and between two readers is tabulated in figure 2. There were 7, 6, 2 and 5 patients with Ann-Arbor stage 1, 2, 3 and 4, respectively. For staging, DWI+T1-IP and T2-TSE showed the highest (k=0.71) and lowest kappa agreement (k=0.44) between the two readers. For reader 1, DWI+T1-IP and T1 post-C showed the highest and lowest agreement compared to PET-CT staging (k=0.65 and 0.58, respectively). For reader 2 substantial agreement observed for both post-C T1 and T2-TSE (k=0.79 for both), with moderate agreement for DWI+T1-IP (k=0.57). Cohen’s kappa statistic for nodal and extra-nodal staging between each reader and PET-CT is summarised in Figure 3. The highest agreement for nodal staging for both readers was achieved using T2-TSE (k= 0.88 and 0.88, for reader 1 and 2, respectively). Controversially, the same sequence showed fair agreement for extra-nodal staging for reader 1 (k= 0.29). Kappa agreement for extra-nodal assessment for reader 2 was comparable on all sequences (k=0.56-0.65). On T2-TSE, over-staging and under-staging due to discrepancies in the assessment of skeletal metastasis occurred in 15% and 10% of cases for reader 1 and 10% and 5% for reader 2 (figure 4). The sensitivity, specificity, PPV and NPV for nodal and extra-nodal staging are summarised in figure 5. T2-TSE showed the highest sensitivity for nodal assessment for both readers (88.9% and 94.6%, for reader 1 and 2, respectively), but low sensitivity for extra-nodal evaluation (44.4% and 66.7% for readers 1 and 2, respectively). DWI+T1-IP showed the lowest sensitivity for extra-nodal assessment for both readers (44.4%), but higher sensitivity for nodal evaluation (78.1% and 80.6% for readers 1 and 2, respectively). Post-C T1 showed the highest sensitivity for extra-nodal assessment (70.0% and 66.7% for readers 1 and 2, respectively), with sensitivity of 73.7% and 81.1% for nodal assessment for readers 1 and 2, respectively. In our cohort, 3.0T WB-MR imaging showed a comparable diagnostic performance to PET-CT for nodal staging on T2-TSE only for both readers. However, our results demonstrate lower sensitivity for depicting extra-nodal site involvements on all sequences, specifically T2-TSE. As previously reported [4] anatomical imaging alone could provide a nearly excellent overall accuracy for identifying involved nodal sites. Furthermore, in line with previous publications [5], we found a moderate agreement for depicting extra-nodal site involvement in our cohort using any of the sequences. Our initial result supports adoption of a combined MRI protocol consisting of a morphological sequence (i.e. T2-TSE) with post-contrast imaging that could provide a higher overall accuracy for both nodal and extra-nodal staging using WB-MRI.