Estimation of severity of lung disease on CT scans has been shown to have a high degree of inter-observer variability. We propose using quantification of CT densities as a modality to more accurately and reproducibly gage the clinical status of heart failure patients via the measurement of mean lung densities and lung cross sectional area.
CT results were retrospectively obtained from patients at Newark Beth Israel Medical Center followed for heart failure. Patients were enrolled in this study if two CT scans from the same patient within a two year period were available. Patients were excluded if the CT modalities differed or if there were qualitative defects. Cross sectional area and mean lung density were calculated using software we wrote to analyze a single slice of the CT scan taken directly below the carina. The difference in cross sectional area between CT scans was calculated. The same was done for mean lung densities. Paired T-tests were performed to determine the significance of these differences.
We obtained a sample size of 14 total patients, from which 2 patients were excluded secondary to motion artifacts. T scores obtained for the mean difference in lung density was 3.52 and the mean difference in cross sectional area was -4.62. P-values for both tests were less than 0.05 (df=11).
Quantitative analysis of serial CT imaging is an accurate and reliable method for following disease progression and remission in heart failure patients.
Utilization of this tool to compare serial CT scans allows clinicians to follow pulmonary effects of heart failure in an objective fashion. Considering the subjective and variable interpretation of CTs as done by radiologists, this modality may provide a more consistent means of tracking disease. This software is not intended to replace a radiologist reading of the scan but to instead quantify particular characteristics that clinicians may use to assist in their patient assessment.
Irtza Sharif, No Financial Disclosure Information; No Product/Research Disclosure Information