Deconvolution technique for measuring tissue perfusion by dynamic CT: Application to normal and metastatic liver

2002 
Because many diseases induce early changes in tissue hemodynamic status, quantitative tissue perfusion imaging could have the capacity to characterize pathologic states, establish a diagnosis, and map the response to treatment. Since the first attempt from Kety (1) to measure perfusion, calculations of perfusion have been made using different measurement techniques and different models, all based on the conservation of mass (2–4). With its subsecond scan time and high resolution rate, slip-ring computed tomography (CT) appears well suited for tissue perfusion measurements, and fast CT has already been used to measure perfusion (5) in the cerebral (6), myocardial (7,8), renal (9), and hepatic (10–14) circulations. By scanning sequentially at a fixed level, the passage of bolus of contrast medium may be tracked by changes in CT number within the tissue and afferent vessels. Time-attenuation curves extracted from vessel and parenchymal regions of interest (ROIs) are then used for perfusion measurement. The algorithms used are generally based on transit time measurement (the central volume theorem) or on compartmental analysis. Both types of modeling, however, are limited by the fact that the venous output is usually not measurable (15). Many attempts have been made to measure the hepatic blood flow (HBF). The liver’s dual circulation is a considerable obstacle to its successful investigation; indeed, the liver enhancement after an injection of contrast medium results from an initial arterial input and a subsequent portal venous input. We have developed a deconvolution method to analyze the functional CT data of the liver, and tested the method in normal rats and rats bearing overt colorectal metastases, both with a classical iodinated contrast medium and a macromolecular contrast medium. The deconvolution method yielded reasonable results in normal liver and showed large hemodynamic changes in tumors. As expected, the comparison between the conventional and the macromolecular contrast media showed that, although the mean transit time (MTT) and the distribution volume in the tissue (DVT) were different, the calculated values of HBF remained unchanged.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    25
    References
    23
    Citations
    NaN
    KQI
    []