Background and purpose: Internal mammary lymph node (IMLN) has a similar predictive value of distant metastasis and long-term recurrence as that of axillary lymph node. IMLN can not be removed under modified radical mastectomy

so the blin

d radiotherapy can sometimes be applied to the patients with suspected metastasis in the internal mammary region clinically. How to select the breast cancer patients who are suitable for IMLN biopsy or resection is an urgent clinical problem. This study aimed to investigate the feasibility of internal mammary sentinel lymph node (SLN) mapping using magnetic resonance lymphography (MR-LG) in VX2 rabbit breast cancer model. Methods: Fifty-five purebred female New Zealand white rabbits were used to establish the model of VX2 rabbit breast cancer

then MR-LG examinations were performed. The raw images were transferred to Philips IntelliSpace Portal (ISP) workstation for post-processing. The first one or several lymph nodes along the lymph duct draining from the injection sites to axilla or internal mammary were defined as SLN. The SLN and draining lymphatic vessels were recorded

then compared with the results of SLN biopsy (SLNB). Results: The success rate of model establishment was 98.2% (54/55)

and 51 rabbits were enrolled in this study finally. On MR-LG

compared with the mapping of SLN and lymph ducts immediately after administration of the contrast agent

more internal mammary lymph ducts (36 vs 25

P=0.021) were shown

and more rabbits with both axillary SLN and lymph ducts mapping were detected (45 vs 37

P= 0.040) 5 minutes after administration of the contrast agent. On the internal mammary MR-LG

the long diameter

the short diameter of SLN and the diameter of mapping lymph ducts were (2.64±0.59) mm

(2.24±0.54) mm and 2.18 mm

respectively. The size of tumor and the number of axillary lymph nodes had significantly different effects on the mapping of internal mammary lymph ducts (P=0.032 and 0.040

respectively). Among 15 internal mammary SLNs detected on 11 cancer-bearing rabbits by SLNB

6 SLNs were located at the 2

nd

and the 3

rd

intercostal space

5 SLNs at the 1

st

and the 2

nd

intercostal space

2 SLNs at the 3

rd

and the 4

th

intercostal s

pace

and one SLN at the 10

th

intercostal space and xiphoid process. On MR-LG

11 SLNs were detected in 9 cancer-bearing rabbits

and 7 rabbits (77.8%

7/9) showed the same results on SLNB. Both the long and short diameters of SLNs measured by SLNB were bigger than those of SLNs measured by MR-LG

which showed significant differences (both P0.05). Conclusion: MR-LG can be used for mapping internal mammary SLN of VX2 rabbit breast cancer. However

the mapping rate should be improved by further study.