Search

Purpose: Recently developed machine-learning algorithms have demonstrated strong performance in the detection of intracranial hemorrhage (ICH) and large vessel occlusion (LVO). However, their generalizability is often limited by geographic bias of studies. The aim of this study was to validate a commercially available deep learning-based tool in the detection of both ICH and LVO across multiple hospital sites and vendors throughout the U.S.Materials and Methods: This was a retrospective and multicenter study using anonymized data from two institutions. Eight hundred fourteen non-contrast CT cases and 378 CT angiography cases were analyzed to evaluate ICH and LVO, respectively. The tool's ability to detect and quantify ICH, LVO, and their various subtypes was assessed among multiple CT vendors and hospitals across the United States. Ground truth was based off imaging interpretations from two board-certified neuroradiologists.Results: There were 255 positive and 559 negative ICH cases. Accuracy was 95.6%, sensitivity was 91.4%, and specificity was 97.5% for the ICH tool. ICH was further stratified into the following subtypes: intraparenchymal, intraventricular, epidural/subdural, and subarachnoid with true positive rates of 92.9, 100, 94.3, and 89.9%, respectively. ICH true positive rates by volume [small (25 mL)] were 71.8, 100, and 100%, respectively. There were 156 positive and 222 negative LVO cases. The LVO tool demonstrated an accuracy of 98.1%, sensitivity of 98.1%, and specificity of 98.2%. A subset of 55 randomly selected cases were also assessed for LVO detection at various sites, including the distal internal carotid artery, middle cerebral artery M1 segment, proximal middle cerebral artery M2 segment, and distal middle cerebral artery M2 segment with an accuracy of 97.0%, sensitivity of 94.3%, and specificity of 97.4%.Conclusion: Deep learning tools can be effective in the detection of both ICH and LVO across a wide variety of hospital systems. While some limitations were identified, specifically in the detection of small ICH and distal M2 occlusion, this study highlights a deep learning tool that can assist radiologists in the detection of emergent findings in a variety of practice settings.

The aim of this study is to explore the feasibility of 3D subject-specific skeletal reconstructions of lower limb in children using stereoradiography, and to assess uncertainty of clinical and anatomical parameters for children with cerebral palsy and for healthy children. The stereoradiography technique, using the EOS(®) system (Eos-imaging(®)), is based on the acquisition of two simultaneous digital anteroposterior and lateral X-rays, from head to feet in standing position and at low radiation dose. This technique allows subject-specific skeletal 3D reconstructions. Five children with cerebral palsy (CP) and 5 typically developing children (TD) were included in the study. Two operators performed the lower limb reconstructions twice. Tridimensional reconstructions were feasible for children over the age of 5 years. The study of reproducibility of anatomical parameters defining skeletal alignment showed uncertainties under 3° for the neck shaft angle, the femoral mechanical angle, and for the femoral and tibial torsions. The maximum degree of uncertainty was obtained for the femoral tibial rotation (4° for healthy children and 3.5° for children with CP).

PURPOSE: Quantitative assessment of 3D clinical indices may be crucial for elbow surgery planning. 3D parametric modeling from bi-planar radiographs was successfully proposed for spine and lower limb clinical investigation as an alternative for CT-scan. The aim of this study was to adapt this method to the upper limb with a preliminary validation. METHODS: CT-scan 3D models of humerus, radius and ulna were obtained from 20 cadaveric upper limbs and yielded parametric models made of geometric primitives. Primitives were defined by descriptor parameters (diameters, angles...) and correlations between these descriptors were found. Using these correlations, a semi-automated reconstruction method of humerus using bi-planar radiographs was achieved: a 3D personalized parametric model was built, from which clinical parameters were computed [orientation and projections on bone surface of trochlea sulcus to capitulum (CTS) axis, trochlea sulcus anterior offset and width of distal humeral epiphysis]. This method was evaluated by accuracy compared to CT-scan and reproducibility. RESULTS: Points-to-surface mean distance was 0.9 mm (2 RMS = 2.5 mm). For clinical parameters, mean differences were 0.4-1.9 mm and from 1.7° to 2.3°. All parameters except from angle formed by CTS axis and bi-epicondylar axis in transverse plane were reproducible. Reconstruction time was about 5 min. CONCLUSIONS: The presented method provides access to morphological upper limb parameters with very low level of radiation. Preliminary in vitro validation for humerus showed that it is fast and accurate enough to be used in clinical daily practice as an alternative to CT-scan for total elbow arthroplasty pre operative evaluation.

In clinical routine, lower limb analysis relies on conventional X-ray (2D view) or computerised tomography (CT) Scan (lying position). However, these methods do not allow 3D analysis in standing position. The aim of this study is to propose a fast and accurate 3D-reconstruction-method based on parametric models and statistical inferences from biplanar X-rays with clinical measurements' (CM) assessment in standing position for a clinical routine use. For the reproducibility study, the 95% CI was under 2.7° for all lower limbs' angular measurements except for tibial torsion, femoral torsion and tibiofemoral rotation ( 5°). The 95% CI were under 2.5 mm for lower limbs' lengths and 1.5 to 3° for the pelvis' CM. Comparisons between X-rays and CT-scan based 3D shapes in vitro showed mean differences of 1.0 mm (95% CI = 2.4 mm). Comparisons of 2D lower limbs' and 3D pelvis' CM between standing 'Shifted-Feet' and 'Non-Shifted-Feet' position showed means differences of 0.0 to 1.4°. Significant differences were found only for pelvic obliquity and rotation. The reconstruction time was about 5 min.