Search

To facilitate both the detection and the interpretation of findings in chest X-rays, comparison with a previous image of the same patient is very valuable to radiologists. Today, the most common approach for deep learning methods to automatically inspect chest X-rays disregards the patient history and classifies only single images as normal or abnormal. Nevertheless, several methods for assisting in the task of comparison through image registration have been proposed in the past. However, as we illustrate, they tend to miss specific types of pathological changes like cardiomegaly and effusion. Due to assumptions on fixed anatomical structures or their measurements of registration quality, they produce unnaturally deformed warp fields impacting visualization of differences between moving and fixed images. We aim to overcome these limitations, through a new paradigm based on individual rib pair segmentation for anatomy penalized registration. Our method proves to be a natural way to limit the folding percentage of the warp field to 1/6 of the state of the art while increasing the overlap of ribs by more than 25%, implying difference images showing pathological changes overlooked by other methods. We develop an anatomically penalized convolutional multi-stage solution on the National Institutes of Health (NIH) data set, starting from less than 25 fully and 50 partly labeled training images, employing sequential instance memory segmentation with hole dropout, weak labeling, coarse-to-fine refinement and Gaussian mixture model histogram matching. We statistically evaluate the benefits of our method and highlight the limits of currently used metrics for registration of chest X-rays., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 Elsevier Ltd. All rights reserved.)

Purpose: To investigate the accuracy and total assessment time (TAT) of the "All-in-one" (AIO)-window/level setting for whole-body computed tomography (CT) image compared to multiple tissue-specific window/level settings conventionally used for detection of traumatic injuries., Method: Contrast-enhanced chest, abdomen, and pelvic CT scans of 50 patients who presented to our emergency department (ED) for major trauma were retrospectively selected. In a simulation of a "wet read" performed at the CT scanner console, 6 readers with different levels of experience had up to 3 min to describe any traumatic finding identified on the CTs. The readers reviewed each patient in two different sessions separated by a washout period to suppress any recall bias from one session to the next. Each scan was reviewed once using the AIO-window/level setting and another time using the conventional bone, lung, and soft tissue window/level display settings, in a randomized order. The CT reports were used as reference standard., Results: Overall, there was no statistically significant difference in the assessment accuracy of the review based on the AIO or the conventional window/level settings (0.89 ± 0.09 vs 0.90 ± 0.08). Using the AIO-window/level settings, TAT was 14.3 s faster when compared with the conventional window/level settings (2.33 ± 0.63 vs 2.57 ± 0.51 min; p < 0.001)., Conclusions: In a time-delimited image review, similar diagnostic accuracy was reached faster using the AIO vs the conventional window/level settings. When providing a "wet read" at the CT console, the ability to identify traumatic injury using a single AIO-window/level may help expedite patient management., Competing Interests: Declaration of Competing Interest None., (Copyright © 2022. Published by Elsevier Inc.)

Purpose: Projections acquired with continuous gantry rotation may suffer from blurring effects, depending on the rotation speed and the exposure time of each projection. This leads to blurred reconstructions if conventional reconstruction algorithms are applied. In this paper, the authors propose a reconstruction method for fast acquisitions based on a continuously moving and continuously emitting x-ray source. They study the trade-off between total acquisition time and reconstruction quality and compare with conventional reconstructions using projections acquired with a stepwise moving x-ray source., Methods: The authors introduce the algebraic reconstruction technique with angular integration concept, which models the angular integration due to the relative motion of the x-ray source during the projection., Results: Compared to conventional reconstruction from projections acquired with pulsed x-ray emission, the proposed method results in substantially improved reconstruction quality around the center of rotation. Outside this region, the proposed method results in improved radial resolution and a decreased tangential resolution. For a fixed reconstruction quality of this region of interest, the proposed method enables a lower number of projections and thus a faster acquisition., Conclusions: The modeling of the continuous gantry rotation in the proposed method substantially improves the reconstruction quality in a region of interest around the rotation center. The proposed method shows potential for fast region of interest tomography.