Evaluation of Medicare Claims for the Development of Heart Failure Diagnostics.

BACKGROUND: Although claims data provide a large and efficient source of clinical events, validation is needed prior to use in heart failure (HF) diagnostic development. METHODS AND RESULTS: Data from the Multisensor Chronic Evaluations in Ambulatory Heart Failure Patients (MultiSENSE) study, used to create the HeartLogic HF diagnostic, were linked with fee-for-service (FFS) Medicare claims. Events were matched by patient ID and date, and agreement was calculated between claims primary HF diagnosis codes and study event adjudication. HF events (HFEs) were defined as inpatient visits, or outpatient visits with intravenous decongestive therapy. Diagnostic performance was measured as HFE-detection sensitivity and false-positive rate (FPR). Linkage of 791 MultiSENSE subjects returned 320 FFS patients with an average follow-up duration of 0.94 years. Although study and claims deaths matched exactly (n = 14), matching was imperfect between study hospitalizations and acute inpatient claims events. Of 239 total events, 165 study hospitalizations (69%) matched inpatient claims events, 28 hospitalizations matched outpatient claims events (12%), 14 hospitalizations were study-unique (6%), and 32 inpatient events were claims-unique (13%). Inpatient HF classification had substantial agreement with study adjudication (κ = 0.823). Diagnostic performance was not different between claims and study events (sensitivity = 75.6% vs 77.6% and FPR = 1.539 vs 1.528 alerts/patient-year). HeartLogic-detected events contributed to > 90% of the HFE costs used for evaluation. CONCLUSIONS: Acceptable event matching, good agreement of claims diagnostic codes with adjudication, and equivalent diagnostic performance support the validity of using claims for HF diagnostic development.

Objective image characterization of a spectral CT scanner with dual-layer detector.

This work evaluated the performance of a detector-based spectral CT system by obtaining objective reference data, evaluating attenuation response of iodine and accuracy of iodine quantification, and comparing conventional CT and virtual monoenergetic images in three common phantoms. Scanning was performed using the hospital's clinical adult body protocol. Modulation transfer function (MTF) was calculated for a tungsten wire and visual line pair targets were evaluated. Image noise power spectrum (NPS) and pixel standard deviation were calculated. MTF for monoenergetic images agreed with conventional images within 0.05 lp cm-1. NPS curves indicated that noise texture of 70 keV monoenergetic images is similar to conventional images. Standard deviation measurements showed monoenergetic images have lower noise except at 40 keV. Mean CT number and CNR agreed with conventional images at 75 keV. Measured iodine concentration agreed with true concentration within 6% for inserts at the center of the phantom. Performance of monoenergetic images at detector based spectral CT is the same as, or better than, that of conventional images. Spectral acquisition and reconstruction with a detector based platform represents the physical behaviour of iodine as expected and accurately quantifies the material concentration.

Virtual assessment of stereoscopic viewing of digital breast tomosynthesis projection images.

Digital breast tomosynthesis (DBT) acquires a series of projection images from different angles as an x-ray source rotates around the breast. Such imaging geometry lends DBT naturally to stereoscopic viewing as two projection images with a reasonable separation angle can easily form a stereo pair. This simulation study assessed the efficacy of stereo viewing of DBT projection images. Three-dimensional computational breast phantoms with realistically shaped synthetic lesions were scanned by three simulated DBT systems. The projection images were combined into a sequence of stereo pairs and presented to a stereomatching-based model observer for deciding lesion presence. Signal-to-noise ratio was estimated, and the detection performance with stack viewing of reconstructed slices was the benchmark. We have shown that: (1) stereo viewing of projection images may underperform stack viewing of reconstructed slices for current DBT geometries; (2) DBT geometries may impact the efficacy of the two viewing modes differently: narrow-arc and wide-arc geometries may be better for stereo viewing and stack viewing, respectively; (3) the efficacy of stereo viewing may be more robust than stack viewing to reductions in dose. While in principle stereo viewing is potentially effective for visualizing DBT data, effective stereo viewing may require specifically optimized DBT image acquisition.

Model observer for assessing digital breast tomosynthesis for multi-lesion detection in the presence of anatomical noise.

Model observers are widely used in task-based assessments of medical image quality. The presence of multiple abnormalities in a single set of images, such as in multifocal multicentric breast cancer (MFMC), has an immense clinical impact on treatment planning and survival outcomes. Detecting multiple breast tumors is challenging as MFMC is relatively uncommon, and human observers do not know the number or locations of tumors a priori. Digital breast tomosynthesis (DBT), in which an x-ray beam sweeps over a limited angular range across the breast, has the potential to improve the detection of multiple tumors. However, prior studies of DBT image quality all focus on unifocal breast cancers. In this study, we extended our 2D multi-lesion (ML) channelized Hotelling observer (CHO) into a 3D ML-CHO that detects multiple lesions from volumetric imaging data. Then we employed the 3D ML-CHO to identify optimal DBT acquisition geometries for detection of MFMC. Digital breast phantoms with multiple embedded synthetic lesions were scanned by simulated DBT scanners of different geometries (wide/narrow angular span, different number of projections per scan) to simulate MFMC cases. With new implementations of 3D partial least squares (PLS) and modified Laguerre-Gauss (LG) channels, the 3D ML-CHO made detection decisions based upon the overall information from individual DBT slices and their correlations. Our evaluation results show that: (1) the 3D ML-CHO could achieve good detection performance with a small number of channels, and 3D PLS channels on average outperform the counterpart LG channels; (2) incorporating locally varying anatomical backgrounds and their correlations as in the 3D ML-CHO is essential for multi-lesion detection; (3) the most effective DBT geometry for detection of MFMC may vary when the task of clinical interest changes, and a given DBT geometry may not yield images that are equally informative for detecting MF, MC, and unifocal cancers.

Model observer design for multi-signal detection in the presence of anatomical noise.

As psychophysical studies are resource-intensive to conduct, model observers are commonly used to assess and optimize medical imaging quality. Model observers are typically designed to detect at most one signal. However, in clinical practice, there may be multiple abnormalities in a single image set (e.g. multifocal multicentric (MFMC) breast cancer), which can impact treatment planning. Prevalence of signals can be different across anatomical regions, and human observers do not know the number or location of signals a priori. As new imaging techniques have the potential to improve multiple-signal detection (e.g. digital breast tomosynthesis may be more effective for diagnosis of MFMC than mammography), image quality assessment approaches addressing such tasks are needed. In this study, we present a model observer to detect multiple signals in an image dataset. A novel implementation of partial least squares (PLS) was developed to estimate different sets of efficient channels directly from the images. The PLS channels are adaptive to the characteristics of signals and the background, and they capture the interactions among signal locations. Corresponding linear decision templates are employed to generate both image-level and location-specific scores on the presence of signals. Our results show that: (1) the model observer can achieve high performance with a reasonably small number of channels; (2) the model observer with PLS channels outperforms that with benchmark modified Laguerre-Gauss channels, especially when realistic signal shapes and complex background statistics are involved; (3) the tasks of clinical interest, and other constraints such as sample size would alter the optimal design of the model observer.

Comparative study of computational visual attention models on two-dimensional medical images.

Computational modeling of visual attention is an active area of research. These models have been successfully employed in applications such as robotics. However, most computational models of visual attention are developed in the context of natural scenes, and their role with medical images is not well investigated. As radiologists interpret a large number of clinical images in a limited time, an efficient strategy to deploy their visual attention is necessary. Visual saliency maps, highlighting image regions that differ dramatically from their surroundings, are expected to be predictive of where radiologists fixate their gaze. We compared 16 state-of-art saliency models over three medical imaging modalities. The estimated saliency maps were evaluated against radiologists' eye movements. The results show that the models achieved competitive accuracy using three metrics, but the rank order of the models varied significantly across the three modalities. Moreover, the model ranks on the medical images were all considerably different from the model ranks on the benchmark MIT300 dataset of natural images. Thus, modality-specific tuning of saliency models is necessary to make them valuable for applications in fields such as medical image compression and radiology education.

Computational assessment of visual search strategies in volumetric medical images.

When searching through volumetric images [e.g., computed tomography (CT)], radiologists appear to use two different search strategies: "drilling" (restrict eye movements to a small region of the image while quickly scrolling through slices), or "scanning" (search over large areas at a given depth before moving on to the next slice). To computationally identify the type of image information that is used in these two strategies, 23 naïve observers were instructed with either "drilling" or "scanning" when searching for target T's in 20 volumes of faux lung CTs. We computed saliency maps using both classical two-dimensional (2-D) saliency, and a three-dimensional (3-D) dynamic saliency that captures the characteristics of scrolling through slices. Comparing observers' gaze distributions with the saliency maps showed that search strategy alters the type of saliency that attracts fixations. Drillers' fixations aligned better with dynamic saliency and scanners with 2-D saliency. The computed saliency was greater for detected targets than for missed targets. Similar results were observed in data from 19 radiologists who searched five stacks of clinical chest CTs for lung nodules. Dynamic saliency may be superior to the 2-D saliency for detecting targets embedded in volumetric images, and thus "drilling" may be more efficient than "scanning."