Objective To investigate the feasibility of magnetic resonance （MR） diffusion weighted imaging （DWI） in discriminating inflammatory from VX2 carcinoma metastatic lymph nodes in rabbit model. Methods Twenty New Zealand white rabbits were randomly divided into 2 groups. Complete Freund’s adjuvant was injected into the bilateral dorsal footpads to set up ipsilateral lymphadenitis model （n = 10）

and the other 10 rabbits received a subcutaneous implantation of VX2 tumor cell suspension （1.5×107 cells/mL） in both thighs to set up metastatic lymph node model. MR imaging scan covering the popliteal fossa and iliac fossa including short time inversion recovery echo-planar imaging DWI （STIR-EPI-DWI）

T1-weighted imaging （T1WI） and T2-weighted imaging （T2WI） was performed 2 weeks after injection. T2WI signal intensity （SI）

DWI SI

long/short axial ratio （LSR） and apparent diffusion coefficient （ADC） values of the lymph nodes were evaluated in all cases. Right after MR imaging scan

popliteal and iliac fossa lymph nodes were collected for hematoxylin-eosin staining. Results Totally 33 lymph nodes larger than 5 mm

including 22 inflammatory and 11 metastatic ones

were successfully isolated and taken into pathological analysis. LSR showed no significant difference between the inflammatory and malignant lymph nodes （P > 0.05）. Both benign and malignant lymph nodes appeared iso-intense on T1WI and hyperintense on both T2WI and DWI images with an even lower T1WI and higher T2WI SI core at the hilum. Both T2WI and DWI SI showed no significant difference between two pathological groups （P > 0.01） in popli- teal fossa. The mean ADC value of inflammatory nodes [（1.199 ± 0.281） ×10-3 mm2/s] was significantly higher than that of metastatic nodes [（0.858 ± 0.090） ×10-3 mm2/s

P < 0.01]. On ADC map

a high ADC value central area could be seen in most of the lymph nodes no matter benign or malignant. ADC value gave out the largest area under curve （AUCADC = 0.955） compared with other three indexes （AUCLSR = 0.488; AUCT2WI SI = 0.727; AUCDWI SI = 0.822） and gave the best sensitivity and specificity in lymph node differential diagnosis compared with the other three indexes. Conclusions High quality DWI image can be obtained using STIR-EPI-DWI sequence in rabbit model. DWI is a new promising technique for differentiating inflammatory from metastatic lymph nodes. Compared with routine MR sequence

DWI could provide more useful physiological and functional information for diagnosis.