Prospective machine learning CT quantitative evaluation of idiopathic pulmonary fibrosis in patients undergoing anti-fibrotic treatment using low- and ultra-low-dose CT.

AIM: To compare the machine learning computed tomography (CT) quantification tool, Computer-Aided Lung Informatics for Pathology Evaluation and Ratings (CALIPER) to pulmonary function testing (PFT) in assessing idiopathic pulmonary fibrosis (IPF) for patients undergoing treatment and determine the effects of limited (LD) and ultra-low dose (ULD) CT on CALIPER performance. MATERIALS AND METHODS: Thirty-eight IPF patients underwent PFT and standard, LD, and ULD CT. CALIPER classified each CT voxel into either vessel-related structures (VRS), normal, reticular (R), honeycomb (HC) or ground-glass (GG) features. CALIPER-derived interstitial lung disease (ILD) extent represented the sum of GG, R and HC values. Repeated-measures correlation coefficient (ρrm) and 95% confidence interval (CI) evaluated CALIPER features correlation with PFT. Lin's concordance correlation coefficient (CCC) assessed concordance of CALIPER parameters across different CT dosages. RESULTS: Twenty patients completed 12 months of follow-up. CALIPER ILD correlated significantly with percent predicted (%) forced vital capacity (FVC) and forced expiratory volume in 1 second (%FEV1; p=0.004, ρrm -0.343, 95% CI [-0.547, -0.108] and 0.008, -0.321, [-0.518, -0.07], respectively). VRS significantly correlated with %FVC and %FEV1 (p=0.000, ρrm -0.491, 95% CI [-0.685, -0.251] and -0.478, 0.000, [-0.653, -0.231], respectively). There was near perfect LD and moderate ULD concordance with standard dose CT for both ILD (CCC 0.995, 95% CI 0.988-0.999 and 0.9, 0.795-0.983, respectively) and VRS (CCC 0.989, 95% CI 0.963-0.997 and 0.915, 0.806-0.956, respectively). CONCLUSIONS: CALIPER parameters correlate well with PFTs for evaluation of IPF in patients undergoing anti-fibrotic treatment without being influenced by dose variation. CALIPER may serve as a robust, objective adjunct to PFTs in assessing anti-fibrotic treatment related changes.

A quick, reliable, and automated method for fat cell sizing.

Mean diameters of fat cells from abdominal tissues from 31 volunteers were determined by three methods based on fat cell isolation after collagenase digestion and methylene blue staining. The three methods were direct microscopy (Micro), manual measurement of diameters from digital images by using the public domain NIH Image program (Scion), and automated measurement of diameters from digital images using a customized program developed by Biomedical Imaging Resource at Mayo Clinic (AdCount). There was excellent agreement between the methods' measurement of mean abdominal fat cell diameter (concordance correlation coefficient >0.84). The Scion method gave slightly but systematically lower mean abdominal fat cell diameters than did either AdCount or Micro. The AdCount approach produced results that are comparable to those from Micro. Comparison of AdCount and Micro in measuring diameters of fat cells from thigh confirmed the good comparability between the two methods independent of fat depot. AdCount is very reliable, and the quickest and most objective of the three methods in measuring fat cell diameters from various depots.

Analysis of cardiac function by MRI and stereology.

Design-based stereology and phase contrast magnetic resonance imaging (MRI) were combined to monitor changes in the volume of the four chambers of the human heart during the cardiac cycle. The data set consisted of 18 adjacent slices (or 'scanning levels') of 0.5 cm thickness, perpendicular to the long axis of the body, and encompassing the whole heart of a healthy volunteer. At each scanning level, a cardiac gated MR image was obtained at each of 16 equally spaced time frames within the cardiac cycle. Given stationarity with respect to time, absence of image artefacts and appropriate definition of chamber boundaries, for each time frame unbiased estimates of total blood volume in the relevant heart chambers were efficiently obtained using the Cavalieri method and point counting. Combined with a proper MRI acquisition, modern stereological methods constitute an efficient and reliable tool to quantify cardiac function noninvasively.

System for analyzing high-resolution three-dimensional coronary angiograms.

Three-dimensional (3-D) high-resolution coronary angiograms offer a means for visualizing the entire coronary arterial tree from any orientation and for detecting and quantitating coronary arterial stenoses. Previously, a skilled operator had to perform several hours of tedious manual analysis using an interactive graphical user-interface (GUI) system (Tree Trace) to analyze a 3-D angiogram. The authors have devised an improved GUI system, consisting of three tools for analyzing 3-D angiograms. The Artery Extractor first performs automatic image-analysis operations to extract the central axes of the arterial tree. Next, using the Artery Display tool and results from the Artery Extractor, the operator can visualize structures in the angiogram and compute various measurements. Finally, the aforementioned Tree Trace tool can be used to manually correct irregularities in the automatically generated results of the Artery Extractor. The system greatly reduces operator analysis time, gives exactly reproducible results, uses true 3-D image-processing operations, and provides a comprehensive interface for visualizing and quantifying features of the 3-D coronary arteries.

EEG scalp electrode projection onto three-dimensional surface rendered images of the brain.

A technique is described for generating magnetic resonance-based, surface rendered images of the brain with electroencephalographic (EEG) scalp electrode positions projected onto the cortical surface. This technique (EEG electrode projection) was used in 10 patients who subsequently underwent surgery for medically intractable frontal lobe epilepsy. In most cases of intractable epilepsy, successful surgery entails the resection of electrophysiologically abnormal cortical tissue rather than an identifiable mass lesion. EEG electrode projection is a unique and useful surgical tool because it provides images that spatially correlate the surface anatomy of the brain and the electrophysiologic abnormality recorded at the scalp. Excellent correlation was found between cortical topography delineated by the surface rendered images and cortical anatomy at surgery. Agreement between EEG electrode projection and electrocorticography as to the location of the electrophysiologic abnormality increases confidence that appropriate cortical areas have been identified for resection. The technique provides new and unique insight into important anatomic-electrophysiologic relationships and aids in formulation of surgical strategy.

Analyze: a comprehensive, operator-interactive software package for multidimensional medical image display and analysis.

A comprehensive software package, called ANALYZE, has been developed (1) which permits detailed investigation and evaluation of multidimensional biomedical images. ANALYZE can be used with 3-D imaging modalities based on x-ray computed tomography, radionuclide emission tomography, ultrasound tomography, and magnetic resonance imaging. The software is written entirely in "C" and runs on standard UNIX workstations. The ANALYZE package features integrated, complimentary tools for fully interactive display, manipulation and measurement of multidimensional image data. The software architecture permits systematic enhancements and upgrades which has fostered development of a readily expandable package. It provides an effective shell for custom software prototyping and turnkey applications. This paper provides a general description of this software as well as specific details on the methodology employed to develop it, both conceptual and technical. Applications of the software are illustrated.

Estimation of tissue volume from serial tomographic sections. A statistical random marking method.

Computer operator-interactive planimetric methods that have been employed to measure volume of a selected tissue in serial computed tomographic sections are time consuming, and the accuracy of the measurement is limited by the operator's skill in outlining the tissue. We introduce an alternate method that is 5 to 6 times faster than planimetry and is independent of the operator's outlining dexterity. This method involves the random marking of voxels in a three-dimensional array of known size and subsequent identification of the proportion of those voxels that are within the tissue of interest. Tissue of the kidney and left ventricular myocardium were imaged in situ with a high speed, volume scanning, computerized x-ray tomographic imaging system. Volumetric measurements made by the random marking method correlated highly (r = 0.99) with measurements made by planimetry and with postmortem weight. These data indicate that the random marking method provides a rapid and accurate means of estimating tissue volume in serial tomographic sections.