Machine learning analysis of confounding variables of a convolutional neural network specific for abdominal aortic aneurysms.

Objective: To identify confounding variables influencing the accuracy of a convolutional neural network (CNN) specific for infrarenal abdominal aortic aneurysms (AAAs) on computed tomography angiograms (CTAs). Methods: A Health Insurance Portability and Accountability Act-compliant, institutional review board-approved, retrospective study analyzed abdominopelvic CTA scans from 200 patients with infrarenal AAAs and 200 propensity-matched control patients. An AAA-specific trained CNN was developed by the application of transfer learning to the VGG-16 base model using model training, validation, and testing techniques. Model accuracy and area under the curve were analyzed based on data sets (selected, balanced, or unbalanced), aneurysm size, extra-abdominal extension, dissections, and mural thrombus. Misjudgments were analyzed by review of heatmaps, via gradient weighted class activation, overlaid on CTA images. Results: The trained custom CNN model reported high test group accuracies of 94.1%, 99.1%, and 99.6% and area under the curve of 0.9900, 0.9998, and 0.9993 in selected (n = 120), balanced (n = 3704), and unbalanced image sets (n = 31,899), respectively. Despite an eightfold difference between balanced and unbalanced image sets, the CNN model demonstrated high test group sensitivities (98.7% vs 98.9%) and specificities (99.7% vs 99.3%) in unbalanced and balanced image sets, respectively. For aneurysm size, the CNN model demonstrates decreasing misjudgments as aneurysm size increases: 47% (16/34) for aneurysms <3.3 cm, 32% (11/34) for aneurysms 3.3 to 5 cm, and 20% (7/34) for aneurysms >5 cm. Aneurysms containing measurable mural thrombus were over-represented within type II (false-negative) misjudgments compared with type I (false-positive) misjudgments (71% vs 15%, P < .05). Inclusion of extra-abdominal aneurysm extension (thoracic or iliac artery) or dissection flaps in these imaging sets did not decrease the model's overall accuracy, indicating that the model performance was excellent without the need to clean the data set of confounding or comorbid diagnoses. Conclusions: Analysis of an AAA-specific CNN model can accurately screen and identify infrarenal AAAs on CTA despite varying pathology and quantitative data sets. The highest anatomic misjudgments were with small aneurysms (<3.3 cm) or the presence of mural thrombus. Accuracy of the CNN model is maintained despite the inclusion of extra-abdominal pathology and imbalanced data sets.

Development of a convolutional neural network to detect abdominal aortic aneurysms.

Objective: We sought to train a foundational convolutional neural network (CNN) for screening computed tomography (CT) angiography (CTA) scans for the presence of infrarenal abdominal aortic aneurysms (AAAs) for future predictive modeling and other artificial intelligence applications. Methods: From January 2015 to January 2020, a HIPAA (Health Insurance and Accountability Act)-compliant, institutional review board-approved, retrospective clinical study analyzed contrast-enhanced abdominopelvic CTA scans from 200 patients with infrarenal AAAs and 200 propensity-matched control patients with non-aneurysmal infrarenal abdominal aortas. A CNN was trained to binary classification on the input. For model improvement and testing, transfer learning using the ImageNet database was applied to the VGG-16 base model. The image dataset was randomized to sets of 60%, 10%, and 30% for model training, validation, and testing, respectively. A stochastic gradient descent was used for optimization. The models were assessed by testing validation accuracy and the area under the receiver operating characteristic curve. Results: Preliminary data demonstrated a nonrandom pattern of accuracy and detectability. Iterations (≤10) of the model characteristics generated a final custom CNN model reporting an accuracy of 99.1% and area under the receiver operating characteristic curve of 0.99. Misjudgments were analyzed through review of the heat maps generated via gradient weighted class activation mapping overlaid on the original CT images. The greatest misjudgments were seen in small aneurysms (<3.3 cm) with mural thrombus. Conclusions: Preliminary data from a CNN model have shown that the model can accurately screen and identify CTA findings of infrarenal AAAs. This model serves as a proof-of-concept to proceed with potential future directions to include expansion to predictive modeling and other artificial intelligence-based applications.

Availability of a final abdominopelvic CT report before emergency department disposition: risk-adjusted outcomes in patients with abdominal pain.

OBJECTIVE: To determine whether availability of a final radiologist report versus an experienced senior resident preliminary report prior to disposition affects major care outcomes in emergency department (ED) patient presenting with abdominal pain undergoing abdominopelvic CT. MATERIALS AND METHODS: This single-institution, IRB-approved, HIPAA-compliant retrospective cohort study included 5019 ED patients with abdominal pain undergoing abdominopelvic CT from October 2015 to April 2019. Patients were categorized as being dispositioned after either an experienced senior resident preliminary report (i.e., overnight model) or the final attending radiologist interpretation (i.e., daytime model) of the CT was available. Multivariable regression models were built accounting for demographic data, clinical factors (vital signs, ED triage score, laboratory data), and disposition timing to analyze the impact on four important patient outcomes: inpatient admission (primary outcome), readmission (within 30 days), second operation within 30 days, and death. RESULTS: In the setting of an available experienced senior resident preliminary report, timing of the final radiologist report (before vs. after disposition) was not a significant multivariable predictor of inpatient admission (p = 0.63), readmission within 30 days (p = 0.66), second operation within 30 days (p = 0.09), or death (p = 0.63). Unadjusted event rates for overnight vs daytime reports, respectively, were 37.2% vs. 38.0% (inpatient admission), 15.9% vs. 16.5% (30-day readmission), 0.65% vs. 0.3% (second operation within 30 days), and 0.85% vs. 1.3% (death). CONCLUSION: Given the presence of an experienced senior resident preliminary report, availability of a final radiology report prior to ED disposition did not affect four major clinical care outcomes of patients with abdominal pain undergoing abdominopelvic CT.

The Next Step in the Treatment of Stroke.

Although many patients do not receive reperfusion therapy because of delayed presentation and/or severity and location of infarct, new reperfusion approaches are expanding the window of intervention. Novel application of neuroprotective agents in combination with the latest methods of reperfusion provide a path to improved stroke intervention outcomes. We examine why neuroprotective agents have failed to translate to the clinic and provide suggestions for new approaches. New developments in recanalization therapy in combination with therapeutics evaluated in parallel animal models of disease will allow for novel, intra-arterial deployment of therapeutic agents over a vastly expanded therapeutic time window and with greater likelihood success. Although the field of neuronal, endothelial, and glial protective therapies has seen numerous large trials, the application of therapies in the context of newly developed reperfusion strategies is still in its infancy. Given modern imaging developments, evaluation of the penumbra will likely play a larger role in the evolving management of stroke. Increasingly more patients will be screened with neuroimaging to identify patients with adequate collateral blood supply allowing for delayed rescue of the penumbra. These patients will be ideal candidates for therapies such as reperfusion dependent therapeutic agents that pair optimally with cutting-edge reperfusion techniques.

Hydrogen gas therapy improves survival rate and neurological deficits in subarachnoid hemorrhage rats: a pilot study.

The high morbidity, high mortality, and significant shortage of effective therapies for subarachnoid hemorrhage (SAH) have created an urgency to discover novel therapies. Human studies in Asia have established the safety of hydrogen gas in the treatment of hepatic, renal, pulmonary, and cardiac diseases. Mechanistically, hydrogen gas has been shown to affect oxidative stress, inflammation, and apoptosis. We hypothesized that hydrogen therapy would improve neurological function and increase survival rate in SAH. High dose hydrogen gas (66% at 3 L/min) was administered for 2 hours at 0.5, 8, and 18 hours after SAH. This treatment increased 72-hour survival rate and provided 24-hour neuroprotection after SAH in rats. To our knowledge, this is the first report demonstrating that high dose hydrogen gas therapy reduces mortality and improves outcome after SAH. Our results correlate well with the proposed mechanisms of hydrogen gas therapy within the literature. We outline four pathways and downstream targets of hydrogen gas potentially responsible for our results. A potentially complex network of pathways responsible for the efficacy of hydrogen gas therapy, along with a limited mechanistic understanding of these pathways, justifies further investigation to provide a basis for clinical trials and the advancement of hydrogen gas therapy in humans. This study was approved by the Institutional Animal Care and Use Committee of Loma Linda University, USA (Approval No. 8160016) in May 2016.

Effects of omega-3 polyunsaturated fatty-acid supplementation on neuropathic pain symptoms and sphingosine levels in Mexican-Americans with type 2 diabetes.

PURPOSE: To determine whether dietary supplementation with omega-3 polyunsaturated fatty acids (PUFAs) reduces neuropathic pain symptoms in Mexican-Americans with type 2 diabetes. METHODS: Forty volunteers with type 2 diabetes enrolled in the "En Balance-PLUS" program, which provided weekly nutrition-diabetes education and daily supplementation with 1,000 mg docosahexaenoic acid (DHA)-200 mg eicosapentaenoic acid over 3 months. The study assessed self-reported neuropathic pain symptoms pre/postintervention using the short-form McGill Pain Questionnaire (SF-MPQ), monitored clinical laboratory values at baseline and 3 months, and performed baseline and 3-month metabolomic analysis of plasma samples. RESULTS: A total of 26 participants self-reported neuropathic pain symptoms at baseline. After 3 months of omega-3 PUFA supplementation, participants reported significant improvement in SF-MPQ scores (sensory, affective, and visual analogue scale; P<0.001, P=0.012, and P<0.001, respectively). Untargeted metabolomic analysis revealed that participants in the moderate-high SF-MPQ group had the highest relative plasma sphingosine levels at baseline compared to the low SF-MPQ group (P=0.0127) and the nonpain group (P=0.0444). Omega-3 PUFA supplementation increased plasma DHA and reduced plasma sphingosine levels in participants reporting neuropathic pain symptoms (P<0.001 and P<0.001, respectively). Increased plasma DHA levels significantly correlated with improved SF-MPQ sensory scores (r=0.425, P=0.030). Improved SF-MPQ scores, however, did not correlate with clinical/laboratory parameters. CONCLUSION: The data suggest that omega-3 PUFAs dietary supplementation may reduce neuropathic pain symptoms in individuals with type 2 diabetes and correlates with sphingosine levels in the plasma.

Intranasal wnt3a Attenuates Neuronal Apoptosis through Frz1/PIWIL1a/FOXM1 Pathway in MCAO Rats.

After ischemic stroke, apoptosis of neurons is a primary factor in determining outcome. Wnt3a is a naturally occurring protein that has been shown to have protective effects in the brain for traumatic brain injury. Although wnt3a has been investigated in the phenomena of neurogenesis, anti-apoptosis, and anti-inflammation, it has never been investigated as a therapy for stroke. We hypothesized that the potential neuroprotective agent wnt3a would reduce infarction and improve behavior following ischemic stroke by attenuating neuronal apoptosis and promoting cell survival through the Frizzled-1/PIWI1a/FOXM1 pathway in middle cerebral artery occlusion (MCAO) rats. A total of 229 Sprague Dawley rats were assigned to male, female, and 9-month-old male MCAO or sham groups followed by reperfusion 2 h after MCAO. Animals assigned to MCAO were either given wnt3a or its control. To explore the downstream signaling of wnt3a, the following interventions were given: Frizzled-1 siRNA, PIWI1a siRNA, and PIWI1a-clustered regularly interspaced short palindromic repeats, along with the appropriate controls. Post-MCAO assessments included neurobehavioral tests, infarct volume, Western blot, and immunohistochemistry. Endogenous levels of wnt3a and Frizzled-1/PIWI1a/FOXM1 were lowered after MCAO. The administration of intranasal wnt3a, 1 h after MCAO, increased PIWIL1a and FOXM1 expression through Frizzled-1, reducing brain infarction and neurological deficits at 24 and 72 h. Frizzled-1 and PIWI1a siRNAs reversed the protective effects of wnt3a after MCAO. Restoration of PIWI1a after knockdown of Frizzled-1 increased FOXM1 survival protein and reduced cleaved caspase-3 levels. In summary, wnt3a decreases neuronal apoptosis and improves neurological deficits through Frizzled-1/PIWI1a/FOXM1 pathway after MCAO in rats. Therefore, wnt3a is a novel intranasal approach to decrease apoptosis after stroke.SIGNIFICANCE STATEMENT Only 5% of patients receive recombinant tissue plasminogen activator after stroke, and few qualify for mechanical thrombectomy. No neuroprotective agents have been successfully translated to promote neuronal survival in stroke. Thus, using a clinically relevant rat model of stroke, middle cerebral artery occlusion, we explored a novel intranasal administration of wnt3a. wnt3a naturally occurs in the body and crosses the blood-brain barrier, supporting the clinically translatable approach of intranasal administration. Significant neuronal apoptosis occurs during stroke, and wnt3a shows promise due to its antiapoptotic effects. We investigated whether wnt3a mediates its poststroke effects via Frizzled-1 and the impact on its downstream signaling molecules, PIWI1a and FOXM1, in apoptosis. Elucidating the mechanism of wnt3a will identify additional pharmacological targets and further understanding of stroke.

Treatment of IgG4-related pachymeningitis in a patient with steroid intolerance: The role of early use of rituximab.

IgG4-related pachymeningitis is a serious inflammatory condition that can present with symptoms of mass effect and focal deficits. The first-line therapy is steroids and second-line is chemotherapy (methotrexate, azathioprine, etc.). We describe a patient with IgG4-related pachymeningitis in whom steroid use was contraindicated and methotrexate was ineffective. During the course of treatment, the patient presented to the emergency department with receptive and expressive aphasia, slurred speech, right-sided neglect, and loss of sensation. After a single infusion of rituximab and anticonvulsants, her symptoms resolved. Our unique case suggests that patients with IgG4-related pachymeningitis might benefit from early initiation of rituximab.

Acute Hyperglycemia Does Not Affect Brain Swelling or Infarction Volume After Middle Cerebral Artery Occlusion in Rats.

Stroke disproportionally affects diabetic and hyperglycemic patients with increased incidence and is associated with higher morbidity and mortality due to brain swelling. In this study, the intraluminal suture middle cerebral artery occlusion (MCAO) model was used to examine the effects of blood glucose on brain swelling and infarct volume in acutely hyperglycemic rats and normo-glycemic controls. Fifty-four rats were distributed into normo-glycemic sham surgery, hyperglycemic sham surgery, normo-glycemic MCAO, and hyperglycemic MCAO. To induce hyperglycemia, 15 min before MCAO surgery, animals were injected with 50 % dextrose. Animals were subjected to 90 min of MCAO and sacrificed 24 h after reperfusion for hemispheric brain swelling and infarct volume calculations using standard equations. While normo-glycemic and hyperglycemic animals after MCAO presented with significantly higher brain swelling and larger infarcts than their respective controls, no statistical difference was observed for either brain swelling or infarct volume between normo-glycemic shams and hyperglycemic shams or normo-glycemic MCAO animals and hyperglycemic MCAO animals. The findings of this study suggest that blood glucose does not have any significant effect on hemispheric brain swelling or infarct volume after MCAO in rats.

Contaminated open fracture and crush injury: a murine model.

Modern warfare has caused a large number of severe extremity injuries, many of which become infected. In more recent conflicts, a pattern of co-infection with Acinetobacter baumannii and methicillin-resistant Staphylococcus aureus has emerged. We attempted to recreate this pattern in an animal model to evaluate the role of vascularity in contaminated open fractures. Historically, it has been observed that infected bones frequently appear hypovascular, but vascularity in association with bone infection has not been examined in animal models. Adult rats underwent femur fracture and muscle crush injury followed by stabilization and bacterial contamination with A. baumannii complex and methicillin-resistant Staphylococcus aureus. Vascularity and perfusion were assessed by microCT angiography and SPECT scanning, respectively, at 1, 2 and 4 weeks after injury. Quantitative bacterial cultures were also obtained. Multi-bacterial infections were successfully created, with methicillin-resistant S. aureus predominating. There was overall increase in blood flow to injured limbs that was markedly greater in bacteria-inoculated limbs. Vessel volume was greater in the infected group. Quadriceps atrophy was seen in both groups, but was greater in the infected group. In this animal model, infected open fractures had greater perfusion and vascularity than non-infected limbs.

Functional radiography in examination of spondylolisthesis.

OBJECTIVE: Despite the predominant use of standing flexion-extension radiography for quantifying instability in isthmic and degenerative spondylolisthesis, other functional radio-graphic techniques have been presented in the literature. CONCLUSION: The current evidence reported in the literature is insufficient to influence how the results of these other functional radiographic techniques should affect clinical management; however, it does raise doubts regarding the accuracy and reliability of standing flexion-extension radiography in this setting. Based on the currently available evidence and until randomized studies are performed to assess the efficacy of functional radiographic techniques in directing clinical decision making, positioning schemes other than traditional standing flexion-extension may be considered as options in the evaluation of patients with symptomatic isthmic and degenerative spondylolisthesis in which standard flexion-extension radiographs fail to show pathologic instability.