Fischer AM, Varga-Szemes A, van Assen M, et al. Comparison of artificial Intelligence–Based fully automatic chest CT emphysema quantification to pulmonary function testing. Am J Roentgenol. 2020;214(5):1065-1071. https://doi.org/10.2214/AJR.19.21572.
OBJECTIVE. The purpose of this study was to evaluate an artificial intelligence (AI)-based prototype algorithm for fully automated quantification of emphysema on chest CT compared with pulmonary function testing (spirometry).
MATERIALS AND METHODS. A total of 141 patients (72 women, mean age ± SD of 66.46 ± 9.7 years [range, 23–86 years]; 69 men, mean age of 66.72 ± 11.4 years [range, 27–91 years]) who underwent both chest CT acquisition and spirometry within 6 months were retrospectively included. The spirometry-based Tiffeneau index (TI; calculated as the ratio of forced expiratory volume in the first second to forced vital capacity) was used to measure emphysema severity; a value less than 0.7 was considered to indicate airway obstruction. Segmentation of the lung based on two different reconstruction methods was carried out by using a deep convolution image-to-image network. This multilayer convolutional neural network was combined with multilevel feature chaining and depth monitoring. To discriminate the output of the network from ground truth, an adversarial network was used during training. Emphysema was quantified using spatial filtering and attenuation-based thresholds. Emphysema quantification and TI were compared using the Spearman correlation coefficient.
RESULTS. The mean TI for all patients was 0.57 ± 0.13. The mean percentages of emphysema using reconstruction methods 1 and 2 were 9.96% ± 11.87% and 8.04% ± 10.32%, respectively. AI-based emphysema quantification showed very strong correlation with TI (reconstruction method 1, ρ = −0.86; reconstruction method 2, ρ = −0.85; both p < 0.0001), indicating that AI-based emphysema quantification meaningfully reflects clinical pulmonary physiology.
CONCLUSION. AI-based, fully automated emphysema quantification shows good correlation with TI, potentially contributing to an image-based diagnosis and quantification of emphysema severity.
Optical coherence tomography (OCT) is a cross-sectional, micrometre-scale imaging modality with widespread clinical application. Typical OCT systems sacrifice lateral resolution to achieve long depths of focus for bulk tissue imaging, and therefore tend to have better axial than lateral resolution. Such anisotropic resolution can obscure fine ultrastructural features. Furthermore, conventional OCT suffers from refraction-induced image distortions. Here, we introduce optical coherence refraction tomography (OCRT), which extends the superior axial resolution to the lateral dimension, synthesizing undistorted cross-sectional image reconstructions from multiple conventional images acquired with angular diversity. In correcting refraction-induced distortions to register the OCT images, OCRT also achieves spatially resolved refractive index imaging. We demonstrate greater than threefold improvement in lateral resolution as well as speckle reduction in imaging the tissue ultrastructure, consistent with histology. With further optimization in optical designs to incorporate angular diversity into clinical instruments, OCRT could be widely applied as an enhancement over conventional OCT.
The tumor stroma, which accounts for a large part of the tumor mass, represents an attractive target for the delivery of diagnostic and therapeutic compounds. Here, the focus is notably on a subpopulation of stromal cells, known as cancer-associated fibroblasts, which are present in more than 90% of epithelial carcinomas, including pancreatic, colon, and breast cancer. Cancer-associated fibroblasts feature high expression of fibroblast activation protein (FAP), which is not detectable in adult normal tissue but is associated with a poor prognosis in cancer patients. Methods: We developed an iodinated and a DOTA-coupled radiotracer based on a FAP-specific enzyme inhibitor (FAPI) and evaluated them in vitro using uptake, competition, and efflux studies as well as confocal microscopy of a fluorescence-labeled variant. Furthermore, we performed imaging and biodistribution studies on tumor-bearing animals. Finally, proof of concept was realized by imaging patients with 68Ga-labeled FAPI. Results: Both FAPIs showed high specificity, affinity, and rapid internalization into FAP-expressing cells in vitro and in vivo. Biodistribution studies on tumor-bearing mice and on the first cancer patients demonstrated high intratumoral uptake of the tracer and fast body clearance, resulting in high-contrast images and negligible exposure of healthy tissue to radiation. A comparison with the commonly used radiotracer 18F-FDG in a patient with locally advanced lung adenocarcinoma revealed that the new FAP ligand was clearly superior. Conclusion: Radiolabeled FAPIs allow fast imaging with very high contrast in tumors having a high stromal content and may therefore serve as pantumor agents. Coupling of these molecules to DOTA or other chelators allows labeling not only with 68Ga but also with therapeutic isotopes such as 177Lu or 90Y.
To describe the initial neuroradiology findings in a cohort of military service members with primarily chronic mild traumatic brain injury (TBI) from blast by using an integrated magnetic resonance (MR) imaging protocol.
Materials and Methods
This study was approved by the Walter Reed National Military Medical Center institutional review board and is compliant with HIPAA guidelines. All participants were military service members or dependents recruited between August 2009 and August 2014. There were 834 participants with a history of TBI and 42 participants in a control group without TBI (not explicitly age- and sex-matched). MR examinations were performed at 3 T primarily with three-dimensional volume imaging at smaller than 1 mm3 voxels for the structural portion of the examination. The structural portion of this examination, including T1-weighted, T2-weighted, before and after contrast agent administrtion T2 fluid attenuation inversion recovery, and susceptibility-weighted images, was evaluated by neuroradiologists by using a modified version of the neuroradiology TBI common data elements (CDEs). Incident odds ratios (ORs) between the TBI participants and a comparison group without TBI were calculated.
The 834 participants were diagnosed with predominantly chronic (mean, 1381 days; median, 888 days after injury) and mild (92% [768 of 834]) TBI. Of these participants, 84.2% (688 of 817) reported one or more blast-related incident and 63.0% (515 of 817) reported loss of consciousness at the time of injury. The presence of white matter T2-weighted hyperintense areas was the most common pathologic finding, observed in 51.8% (432 of 834; OR, 1.75) of TBI participants. Cerebral microhemorrhages were observed in a small percentage of participants (7.2% [60 of 834]; OR, 6.64) and showed increased incidence with TBI severity (P < .001, moderate and severe vs mild). T2-weighted hyperintense areas and microhemorrhages did not collocate by visual inspection. Pituitary abnormalities were identified in a large proportion (29.0% [242 of 834]; OR, 16.8) of TBI participants.
Blast-related injury and loss of consciousness is common in military TBI. Structural MR imaging demonstrates a high incidence of white matter T2-weighted hyperintense areas and pituitary abnormalities, with a low incidence of microhemorrhage in the chronic phase.
Radiology; 278;1 (Ahead of Print), Gerard Riedy, MD, PhD , Justin S. Senseney, MS , Wei Liu, DSc , John Ollinger, PhD , Elyssa Sham, BA , Pavel Krapiva, MD , Jigar B. Patel, MD , Alice Smith, MD , Ping-Hong Yeh, PhD , John Graner, PhD , Dominic Nathan, PhD , Jesus Caban, PhD , Louis M. French, PsyD , Jamie Harper, MPH , Victoria Eskay, BA , John Morissette , Terrence R. Oakes, PhD. DOI: http://dx.doi.org/10.1148/radiol.2015150438
To facilitate MRI-guided catheterization procedures, we present an MRI-compatible remote catheter navigation system that allows remote navigation of steerable catheters with 3 degrees of freedom.
The system consists of a user interface (master), a robot (slave), and an ultrasonic motor control servomechanism. The interventionalist applies conventional motions (axial, radial and plunger manipulations) on an input catheter in the master unit; this user input is measured and used by the servomechanism to control a compact catheter manipulating robot, such that it replicates the interventionalist’s input motion on the patient catheter. The performance of the system was evaluated in terms of MRI compatibility (SNR and artifact), feasibility of remote navigation under real-time MRI guidance, and motion replication accuracy.
Real-time MRI experiments demonstrated that catheter was successfully navigated remotely to desired target references in all 3 degrees of freedom. The system had an absolute value error of <1 mm in axial catheter motion replication over 30 mm of travel and 3∘±2∘ for radial catheter motion replication over 180∘. The worst case SNR drop was observed to be <3 %; the robot did not introduce any artifacts in the MR images.
An MRI-compatible compact remote catheter navigation system has been developed that allows remote navigation of steerable catheters with 3 degrees of freedom. The proposed system allows for safe and accurate remote catheter navigation, within conventional closed-bore scanners, without degrading MR image quality.
M. A. Tavallaei, International Journal of Computer Assisted Radiology and Surgery, pp 1-9, First online: