Proposal for a New Liver Tumor Classification Method in MRI

Motivation: Liver tumors could be classified with the help of machine learning or other methods based solely on changes in Gd-EOB-DTPA uptake over time. Goal(s): To evaluate the possibility of classifying liver tumor types using changes in liver and tumor contrast (Q-LTC) over time. Approach: Liver tumors (HCC, metastasis, and hemangiomas) were classified. The rate of change in Q-LTC were calculated using images obtained at 3, 10, and 15 min after Gd-EOB-DTPA administration. Results: The rate of change in Q-LTC over time tended to be higher in HCC, metastasis, and hemangioma, in that order; therefore, its potential use in liver tumor classification. Impact: To reduce the burden on patients caused by extended examination time, we performed liver tumors classification using simple liver and tumor contrast based on the liver function, during routine clinical studies without requiring additional specialized imaging.