[O-134] MRI FEATURES OF HCC DO NOT PREDICT MICROVASCULAR INVASION IN LIVER TRANSPLANT FOR HEPATOCELLULAR CARCINOMA.

Emma Robinson, Hersh Chandarana, Leonid Droxhinin, Cristina Hajdu, Ruliang Xu, Lewis Teperman, Bachir Taouli Departments of Radiology, Pathology, and Surgery, NYU Langone Medical Center, New York, NY, USA

Microvascular invasion (MVI) predicts poor outcome after liver transplantation for hepatocellular carcinoma (HCC). Pre-operative prediction of MVI is difficult, as biopsies are not always feasible and correlate poorly with final tumor grade and MVI. This study sought to identify MRI features that correlate pre-operatively with MVI.
Methods: This retrospective study identified 24 patients with MRI and liver transplant within 90 days. Pathologically identified HCCs were seen on MRI in 21/24 patients (88%). Four patients were excluded because all HCCs were over 90% necrotic from chemoembolization. The study group comprised 17 patients (M/F 14/3, mean age 58 y). Alfa-fetoprotein (AFP), number of HCC, and tumor differentiation were noted. Routine liver MRI included in- and out-of-phase T1 GRE, axial T2 fat suppressed, and 3D VIBE post gadolinium images in the arterial, portal venous, and equilibrium phases. For the largest HCC in each patient, the size, T1 and T2 signal intensity relative to liver, margin (well defined or ill-defined), presence of a capsule, peritumoral enhancement, and the distance to the closest vessel were noted by 2 readers in consensus. ROIs were drawn on the largest lesion pre-contrast and copied to all three phases of enhancement to obtain signal intensity (SI) ratios of HCCs: SIpost-SIpre/SIpre.
Results:Patient age and number of HCCs at explant correlated with MVI (p less than 0.05). Patients with MVI were more likely to be young (57 vs. 63 y). As the number of HCC increased, the likelihood of MVI increased (mean 3.7 tumors in MVI group vs. 2.3 tumors in non-MVI group). No MRI imaging parameters of the HCCs correlated with the presence of MVI.

Results
ParameterP value
Age0.049
Sex0.219
AFP0.116
Number of HCC0.043
Size of HCC0.482
Differentiation of HCC1.0
T1 signal intensity0.244
T2 signal intensity0.250
HCC margin0.603
Presence of a capsule1.0
Peritumoral enhancement0.519
Distance to nearest vessel0.506
SI ration Arterial0.965
SI ratio Venous0.386
SI ratio Equilibrium0.514


Conclusion: Other than number of lesions, no other predictive signs of MVI were seen on MRI. Given the prognostic significance of MVI, further work to predict MVI non-invasively remains warranted.

Date: Friday, July 10, 2009
Session Info: Radiology (4:45 PM-6:15 PM)
Presentation Time: 5:45 PM
Room: Regeant Parlor, Level 2

 

Close Window