Rationale and objectives: We developed a technique for estimating the volumes of distribution and intratumoral ethanol concentrations using computed tomography (CT) scanning in patients undergoing percutaneous ethanol injection (PEI) treatment of malignant hepatic tumors.
Methods: A phantom containing anhydrous ethanol diluted with deionized distilled water to concentrations of 0-100% ethanol was scanned by CT. Thirty-seven treatment sessions were performed on eight patients with malignant hepatic tumors undergoing PEI under CT guidance. The patients were scanned pre- and post-PEI, and a region of interest containing the treated hepatic tissue was selected for pixels between -250 and 15 Hounsfield units (H). The mean density of the pixels in this range was computed and the concentration of ethanol estimated. Volumes of distribution of ethanol and intratumoral concentration were then correlated with volume of ethanol injected during PEI. The ratios of volumes of distribution of ethanol to ethanol injected (adjusted in-range [IR]/volume injected) were compared for responders (n = 4) and nonresponders (n = 4).
Results: CT numbers in the phantom scaled linearly with ethanol concentration; 100% ethanol measured -234 H. On CT scans after PEI, the volume of distribution of ethanol correlated positively with the volume injected. Calculated intratumoral ethanol concentrations ranged from 4% to 31%. The adjusted IR/volume injected was significantly higher for responders than nonresponders (p < .5).
Conclusion: CT density data can be used to estimate volume of ethanol distribution in tissue; a larger relative intratumoral distribution of alcohol appears to correlate with a favorable response to PEI. However, CT measurement of intratumoral ethanol concentrations may require more complex computational techniques.