Diagnosis of small hepatocellular carcinoma--imaging diagnosis and significance of tumor biopsy.

2000 
Background/Aims: To compare the effectiveness of different imaging modalities and the significance of tumor biopsy for diagnosing small hepatocellular carcinoma. Methodology: Nodules (n=352) with diameters of 30mm or less newly detected by periodic ultrasonography and computed tomography in 234 patients with chronic liver disease were investigated with magnetic resonance imaging and digital subtraction angiography. These findings were compared with histologic findings. Histologic diagnoses were dysplastic nodule (n=23), well-differentiated hepatocellular carcinoma (n=163), moderately differentiated hepatocellular carcinoma (n= 159), and poorly differentiated hepatocellular carcinoma (n=7). We compared three groups based on-diameters of 10, 11-20, and 21-30mm. Nodules were diagnosed as hepatocellular carcinoma if they had hypervascular staining on digital subtraction angiography, hyperintensity on magnetic resonance T2-weighted images, arterial phase enhancement on enhanced magnetic resonance imaging, or low-high-low density on enhanced computed tomography. Results: Imaging alone was sufficient to diagnose hepatocellular carcinoma in 66.3% of the well-differentiated nodules and 91.6% of the moderately and poorly differentiated nodules (P<0.001) The size of the nodule influenced the diagnosis of hepatocellular carcinoma by imaging alone in 65.5% (≤10mm), 77.2% (11-20mm), and 92.3% (21-30mm) (≤10 vs. 21-30: P<0.0001, 11-20 vs. 21-30: P<0.0005). It was impossible to determine the degree of differentiation of the hepatocellular carcinoma by imaging alone. Conclusions: The effectiveness of imaging for the diagnosis of hepatocellular carcinoma improved with decreasing differentiation and increasing diameter of the nodules. Tumor biopsy was required to make a histological accurate diagnosis.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []