BACKGROUNDRetroperitoneal fibrosis (RPF) and lymphoma presenting as retroperitoneal mass may closely resemble each other and misdiagnosis may occur. This study investigated the differential imaging features of RPF and lymphoma which presented as a retroperitoneal soft tissue using multidetector-row computed tomography (MDCT).

METHODSThe 42 consecutive patients were included in this retrospective review, including 19 RPF patients (45.2%; including 13 males and 6 females; mean age: 56.7 ± 6.2 years) and 23 patients with lymphoma (54.8%; including 14 males and 9 females; mean age: 57.4 ± 12.3 years). An array of qualitative computed tomography (CT) features of lesions in 42 consecutive patients with newly diagnosed untreated RPF and lymphoma were retrospectively analyzed. The quantitative size of the lesion at the para-aortic region and attenuation in the precontrast, arterial, and portal phases were calculated in regions of interest and compared between the patients with newly diagnosed untreated RPF and with lymphoma. Receiver operating characteristic curve analysis was used to assess the potential diagnostic value of each quantitative parameter. Inter-reader concordance was also calculated.

RESULTSMean ages between patients with RPF and lymphoma were not significantly different (56.7 ± 6.2 years vs. 57.4 ± 12.3 years P = 0.595). Compared to those in patients with lymphoma, homogeneous enhancement (65.2% vs. 94.7%, P = 0.027) and pelvic extension (52.2% vs. 89.5%, P = 0.017) were significantly more common while the involvement of additional nodes (78.3% vs. 5.3%, P < 0.001), suprarenal extension (60.9% vs. 15.8%, P = 0.004), and aortic displacement (43.5% vs. 5.3%, P = 0.006) were significantly less common in patients with RPF. Lesion size at the para-aorta was significantly greater in patients with lymphoma, compared with RPF patients (3.9 ± 1.2 cm vs. 1.8 ± 0.6 cm; P < 0.001). The attenuation values in three phases were not significantly different between patients with RPF and lymphoma. Inter-reader concordance for subjective features ranged from very good to excellent (range: 85.7-100.0%).

CONCLUSIONSThis study showed that MDCT can help differentiate between untreated RPF and lymphoma on the basis of qualitative CT features and lesion sizes. Differentiating RPF from lymphoma on the basis of attenuation values in the precontrast, arterial, and portal phases was difficult to accomplish.