[Comparative study of CT versus gross pathology in rabbit VX2 colorectal cancer model].

Objective: To establish rabbit VX2 colorectal cancer(CRC) model and to compare CT images with gross pathology in order to offer help for TNM staging in patients with CRC.

Methods: VX2 tumor pieces were implanted into colonic wall in 9 New Zealand white rabbits and rectal wall in 2 New Zealand white rabbits. Four weeks after inoculation, Ultravist(370 mg/ml) was injected through ear marginal vein with high pressure injector for stage 3 scanning of chest, abdomen and pelvis, and enhanced CT (collimation 0.5 mm mm × 320, pitch factor 0.828, bulb rotation speed 0.5 s/cycle, 120 kV, automatic ma, range 80 to 100 mAs) was performed to determine the presence of CRC or metastasis once a week for 4-6 weeks. Once inoculated CRC or metastases occurred or 6 weeks after implantation, the rabbits were sacrificed regardless of the presence or absence of CRC or metastasis on the CT images. One rabbit was used for gross anatomy observation. Others were placed in wood boxes with -80centi-degree for 24 hours, then samples of 3 mm thickness were cut using a motorized saw to make macropathology. Each cutting surface of the specimens was photographed in serial number. If certain or suspected lesions were found on the slices, such part was labeled and then placed in 10% phosphate-buffered formaldehyde numbered box for subsequent pathological examination. CT image postprocessing was performed referring to the gross slice specimens and all findings were compared with the pathological reports.

Results: Among 11 rabbits, tumor was successfully established in 8 rabbits. Pathology showed that single lung metastasis (7 to 10 mm) was found in 2 rabbits and liver metastasis (9 mm) in 1 rabbit. Number of lymph node located around the inoculated tumor was 22 and that around mesenteric vessels was 13 with diameter of 2 to 16 mm. Among these 35 lymph nodes from 8 successful rabbits, 9 nodes were positive, including 7 around inoculated tumor and 2 around mesenteric vessels. CT identified above 8 primary inoculated tumors, 2 lung metastatic lesions and 1 liver metastatic lesion, with detection rate of 100%. For the detection of lymph node in CT, 27 nodes were identified in the pericolorectal region (17 nodes) and perimesenteric vessels (10 nodes), in which 6 were positive metastasis (ring-shaped enrichment and central low density necrosis), resulting in a detection rate of 77.1%(27/35 nodes), and positive detection rate of 66.7% (6/9 nodes), respectively.

Conclusion: Living rabbit CT-gross pathological slice(3 mm-cut) of VX2 CRC model can be applied in image evaluation of small metastatic lesion.