[A comparative study of MRI manifestations and pathological changes in hepatocellular carcinoma treated by transcatheter arterial chemoembolization with lipiodol].

OBJECTIVES

To analyze the MRI manifestations and pathological changes of hepatocellular carcinoma (HCC) after transcatheter arterial chemoembolization (TACE) with lipiodol.

METHODS

23 patients with 31 HCC lesions treated by TACE underwent MRI examination within 1 week before their surgical resections. MRI was performed with SE sequence (T1WI and FSE T2WI) and FMPSPGR sequence dynamic multi-phase contrast scans. All resected specimens were cut into 5-10mm thick slices, corresponding to the same plane as that of MRI scans. The specimens were wholly embedded in paraffin, serial sections made and stained with hematoxylin and eosin. The MRI findings were thus compared with the pathology of the specimen sections.

RESULTS

(1) MRI findings: In all 31 lesions, the signal intensity of lesions varied and was mostly heterogeneous on SE T1WI and T2WI images. Three lesions were inhomogeneous hyper-intensity and the other 28 lesions were iso- or hypo-intensity on FMPSPGR plain scannings. Twenty-two lesions were enhanced on early-phase dynamic scanning, and no enhancement was found in the other 9 lesions. Partial enhancement was also seen in 6 lesions on delay-phase dynamic scanning. (2) Pathologically, no coagulation necrosis was found in 2 specimens, but 6 lesions showed complete coagulation necrosis and 23 showed various degrees of it. The other pathological changes found included intra-tumoral hemorrhage (n=10), intra-lesional fibrotic septa formation (n=5), capsule-like fibrotic tissue proliferation around the lesions (n=12), inflammatory infiltration (n=28), focal mucoid degeneration (n=2), focal hyaline degeneration (n=2), and lipiodol retention (n=6). (3) Radiological-pathological correlation study: hyper-intense areas on T1WI corresponded to areas of coagulation necrosis with or without hemorrhage and of residual viable tumor; iso- and hypo-intense corresponded to areas of coagulation necrosis or residual viable tumor. Hyper-intense areas on T2WI corresponded to those of residual viable tumor or coagulation necrosis with hemorrhage, and iso-intense areas corresponded to those of coagulation necrosis, small residual viable tumor or intra-lesional fibrotic septa formation, and hypo-intense areas corresponded to those of coagulation necrosis or intra-lesional fibrotic septa formation. Areas of enhancement within the lesions on the early-phase dynamic-contrast images corresponded to areas of residual viable tumors, while areas of no enhancement were those of coagulation necrosis, hemorrhage, intra-lesional fibrotic septa formation or small residual viable tumors. Areas of enhancement on the delay-phase dynamic scanning were those of residual viable tumors or intra-lesional fibrotic septa formation, while no enhancement corresponded to the areas of residual viable tumors, coagulation necrosis, and hemorrhage. Areas of enhancement on the delay-phase dynamic scanning corresponded to those areas of fibrosis tissue or residual viable tumors. Inflammatory infiltration was found in areas of different signal intensity on MRI images.

CONCLUSIONS

(1) Different pathological changes in HCCs after TACE are represented by various signal intensities on SE sequence images. The only area of hypo-intensity on T2WI has a specificity in representing coagulation necrosis. (2) FMPSPGR sequence dynamic MRI is superior to SE sequence in demonstrating and determining the necrosis and residual viable tumor. Enhanced areas within the lesions on the early-phase dynamic-contrast images represent residual viable tumors and the enhancement of capsule on early-phase dynamic-contrast images also represent subcapsular residual viable tumors. (3) MRI can demonstrate accurately the areas of necrosis and residual viable HCC tissues after TACE and evaluate the effect of TACE.