Volumetric evaluation of dual-energy perfusion CT for the assessment of intrapulmonary clot burden.

AIM: To evaluate the volumetric values of intrapulmonary clots (IPCs) using 64-section dual-energy perfusion computed tomography (DEpCT). MATERIALS AND METHODS: A total of 174 patients suspected of having acute pulmonary embolism (PE) underwent DEpCT, and acute PE was diagnosed in 48 of these patients. DEpCT images were three-dimensionally reconstructed with four threshold ranges: 1-120 HU (V120), 1-15 HU (V15), 1-10 HU (V10), and 1-5 HU (V5). Each relative value per V120 was expressed as %V15, %V10 and %V5. These values were compared with the d-dimer, pulmonary arterial (PA) pressure, right ventricular (RV) diameter, RV/left ventricular diameter ratio, PA diameter, and CT angiographic obstruction index (CTOI). RESULTS: In patients with IPCs, PA pressure, d-dimer and volumetric values of DEpCT were significantly higher (p < 0.001). Relative volumetric values at DEpCT had better correlations with the PA pressure, PA diameter, and CTOI than absolute ones, and %V5 especially had good correlations with PA pressure (r = 0.44, p = 0.02), PA diameter (r = 0.40, p = 0.005), and CTOI (r = 0.38, p = 0.009). CONCLUSION: The relative volumetric evaluation of DEpCT images with a lower attenuation threshold range may be helpful for assessing right heart strain, because these values had good correlation with CTOI, pulmonary pressure, and diameter in suggesting right heart load.

Low-cost soft-copy display accuracy in the detection of pulmonary nodules by single-exposure dual-energy subtraction: comparison with hard-copy viewing.

This study endeavored to clarify the usefulness of single-exposure dual-energy subtraction computed radiography (CR) of the chest and the ability of soft-copy images to detect low-contrast simulated pulmonary nodules. Conventional and bone-subtracted CR images of 25 chest phantom image sets with a low-contrast nylon nodule and 25 without a nodule were interpreted by 12 observers (6 radiologists, 6 chest physicians) who rated each on a continuous confidence scale and marked the position of the nodule if one was present. Hard-copy images were 7 x 7-inch laser-printed CR films, and soft-copy images were displayed on a 21-inch noninterlaced color CRT monitor with an optimized dynamic range. Soft-copy images were adjusted to the same size as hard-copy images and were viewed under darkened illumination in the reading room. No significant differences were found between hard- and soft-copy images. In conclusion, the soft-copy images were found to be useful in detecting low-contrast simulated pulmonary nodules.

Comparison between computer-aided diagnosis and radiologists: assessment of pulmonary blood flow on chest radiographs.

To evaluate the performance of a computer-aided diagnosis (CAD) scheme for estimating increased pulmonary blood flow on chest radiographs, we compared computerized assessment with findings by radiologists. Our CAD scheme extracts selectively linear opacities corresponding to vessels in regions of interest (ROIs) in the right upper and lower lung zones on digitized chest radiographs, and then calculates a radiographic index as a physical measure that reflects the area of the extracted opacities in selected ROIs. As a measure of increased pulmonary blood flow, the upper/lower radiographic index ratio was calculated for each patient. Seven radiologists estimated the degree of increased pulmonary blood flow for the same images of ROI sets presented on a cathode-ray tube monitor in a randomized order. Between the normal-pulmonary capillary wedge pressure (PCWP) group and increased-PCWP groups, there was no significant difference in performance between CAD and radiologists (p = 0.105). However, when the normal and mild PCWP groups were compared, the performance of CAD was superior to that of radiologists (p = 0.001). This study indicates that our CAD scheme is promising for quantitative estimation of increased pulmonary blood flow, especially in mild cases.

Computerized radiographic assessment of pulmonary blood flow in patients with and those without mitral stenosis.

The authors investigated whether a computerized analysis system can be used with chest radiography for estimating the redistribution of pulmonary blood flow in patients with and those without mitral stenosis (MS). Their system uses four-directional Laplacian-Gaussian filtering and binarization. As a physical measure, the radiographic index, which reflects the area of opacity in selected regions of interest (ROIs), was used. Fifteen men with MS and 15 men without MS were included in the study. ROIs were selected in the right upper and lower lung zones on radiographs, and the images were processed with the system. The radiographic indexes were determined in each ROI. As a measure of pulmonary blood flow redistribution, the upper-to-lower lung zone radiographic index ratio (U/L) was calculated. The U/L values correlated with pulmonary capillary wedge pressure values (r = .405, P = .025). The mean U/L value of the MS group was significantly higher than that of the non-MS group (1.16 +/- 0.14 vs 1.05 +/- 0.09, respectively; P = .016). These preliminary results indicate that this system can be used for quantitative estimation of the redistribution of pulmonary blood flow.

[Assessment of the severity of interstitial lung diseases in chest radiographs using a computer-aided analysis system. A comparison with X-ray CT].

A new computer-aided system to assess the severity of interstitial lung diseases in chest radiographs using proposed parameters, radiographic indices, was developed. The radiographic index is the normalized percent area of shadows in a ROI selected in the lung fields of chest radiographs. In the right lungs of 82 patients, 164 ROIs were processed by computerized methods such as filtering, binarization, determination and subtraction of linear shadows. In subtraction method, the image processed with determination of linear shadows is subtracted from the image filtered by 4-directional Laplacian-Gaussian filters. Then, the radiographic indices, Dbin (binarization), Dlin (linear shadow determination) and Dsub (subtraction), were obtained for the corresponding images. The combined radiographic index "Dcom" was calculated from "Dlin" and "Dsub". Values of these indices of the ROIs were compared with the grade (0-4) of the severity of interstitial involvement evaluated by CT. The results were obtained as follows: 1) Dbin, Dlin and Dcom provided good correlation with the severity throughout 5 grades, while Dsub did not. 2) As a parameter for detection of interstitial lung diseases, Dcom was significantly superior to the other indices (p less than .05). 3) As a parameter for assessment of severity of the diseases, Dlin showed the best performance and Dcom was the second best. Dlin and Dcom were superior to Dbin, and Dsub was inferior to Dbin. These results indicate that a new system is useful for assessment of the severity of interstitial lung diseases in chest radiographs.