Fully automated 3D machine learning model for HPV status characterization in oropharyngeal squamous cell carcinomas based on CT images.

BACKGROUND: Human papillomavirus (HPV) status plays a major role in predicting oropharyngeal squamous cell carcinoma (OPSCC) survival. This study assesses the accuracy of a fully automated 3D convolutional neural network (CNN) in predicting HPV status using CT images. METHODS: Pretreatment CT images from OPSCC patients were used to train a 3D DenseNet-121 model to predict HPV-p16 status. Performance was evaluated by the ROC Curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and F1 score. RESULTS: The network achieved a mean AUC of 0.80 ± 0.06. The best-preforming fold had a sensitivity of 0.86 and specificity of 0.92 at the Youden's index. The PPV, NPV, and F1 scores are 0.97, 0.71, and 0.82, respectively. CONCLUSIONS: A fully automated CNN can characterize the HPV status of OPSCC patients with high sensitivity and specificity. Further refinement of this algorithm has the potential to provide a non-invasive tool to guide clinical management.

AI-based differential diagnosis of dementia etiologies on multimodal data.

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an AI model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations, and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a micro-averaged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the micro-averaged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in various clinical settings and drug trials, with promising implications for person-level management.

Endovascular Therapy for Acute Stroke: New Evidence and Indications.

Endovascular therapy (EVT) has revolutionized the treatment of acute ischemic stroke. In the past few years, endovascular treatment indications have expanded to include patients being treated in the extended window, with large ischemic core infarction, basilar artery occlusion (BAO) thrombectomy, as demonstrated by several randomized clinical trials. Intravenous thrombolysis (IVT) bridging to mechanical thrombectomy has also been studied via several randomized clinical trials, with the overall results indicating that IVT should not be skipped in patients who are candidates for both IVT and EVT. Simplification of neuroimaging protocols in the extended window to permit non-contrast CT, CTA collaterals have also expanded access to mechanical thrombectomy, particularly in regions across the world where access to advanced imaging may not be available. Ongoing study of areas to develop include rescue stenting in patients with failed thrombectomy, medium vessel occlusion thrombectomy, and carotid tandem occlusions. In this narrative review, we summarize recent trials and key data in the treatment of patients with large ischemic core infarct, simplification of neuroimaging protocols for the treatment of patients presenting in the late window, bridging thrombolysis, and BAO EVT evidence. We also summarize areas of ongoing study including medium and distal vessel occlusion.

Utilization of a two-material decomposition from a single-source, dual-energy CT in acute traumatic vertebral fractures.

Purpose: The purpose of this study is to utilize a two-material decomposition to quantify bone marrow edema on a dual-energy computed tomography (DECT) scanner at the cervical, thoracic, and lumbar spine acute fractures in correlation with short tau inversion recovery (STIR) hyperintensity on magnetic resonance imaging (MRI) in comparison with the normal bone marrow. Materials and methods: This retrospective institutional review board-approved study gathered patients over 18 years old who had acute cervical, thoracic, or lumbar spinal fractures scanned on a DECT scanner. Those who had a spinal MRI done with bone marrow STIR hyperintensity within 3 weeks of the DECT were included. The water (calcium) and fat (calcium) density (mg/cm3) measurements of the region of interest of the bone marrow were obtained at a normal anatomic equivalent site and at the fracture site where STIR hyperintensity was noted on MRI. A statistical analysis was performed using the paired t-test and Wilcoxon signed rank test (p > 0.05). Results: A total of 20 patients met the inclusion criteria (males n = 17 males, females n = 3). A total of 32 fractures were analyzed: 19 cervical and 13 thoracolumbar. There were statistically significant differences in the water (43 ± 24 mg/cm3) and fat (36 ± 31 mg/cm3) density (mg/cm3) at the acute thoracic and lumbar spine fractures in correlation with edema on STIR images (both paired t-test <0.001, both Wilcoxon signed ranked test p < 0.01). There were no significant differences in the water (-10 ± 46 mg/cm3) or fat (+7 ± 50 mg/cm3) density (mg/cm3) at the cervical spine fractures. Conclusion: The DECT two-material decomposition using water (calcium) and fat (calcium) analyses has the ability to quantify a bone marrow edema at the acute fracture site in the thoracic and lumbar spine.

Multiparametric machine learning algorithm for human papillomavirus status and survival prediction in oropharyngeal cancer patients.

BACKGROUND: Human papillomavirus (HPV) status influences prognosis in oropharyngeal cancer (OPC). Identifying high-risk patients are critical to improving treatment. We aim to provide a noninvasive opportunity for managing OPC patients by training multiple machine learning pipelines to determine the best model for characterizing HPV status and survival. METHODS: Multi-parametric algorithms were designed using a 492 OPC patient database. HPV status incorporated age, sex, smoking/drinking habits, cancer subsite, TNM, and AJCC 7th edition staging. Survival considered HPV model inputs plus HPV status. Patients were split 4:1 training: testing. Algorithm efficacy was assessed through accuracy and area under the receiver operator characteristic curve (AUC). RESULTS: From 31 HPV status models, ensemble yielded 0.83 AUC and 78.7% accuracy. From 38 survival models, ensemble yielded 0.91 AUC and 87.7% accuracy. CONCLUSION: Results reinforce artificial intelligence's potential to use tumor imaging and patient characterizations for HPV status and outcome prediction. Utilizing these algorithms can optimize clinical guidance and patient care noninvasively.

Robotic Interventional Neuroradiology: Progress, Challenges, and Future Prospects.

Advances in robotic technology have improved standard techniques in numerous surgical and endovascular specialties, offering more precision, control, and better patient outcomes. Robotic-assisted interventional neuroradiology is an emerging field at the intersection of interventional neuroradiology and biomedical robotics. Endovascular robotics can automate maneuvers to reduce procedure times and increase its safety, reduce occupational hazards associated with ionizing radiations, and expand networks of care to reduce gaps in geographic access to neurointerventions. To date, many robotic neurointerventional procedures have been successfully performed, including cerebral angiography, intracranial aneurysm embolization, carotid stenting, and epistaxis embolization. This review aims to provide a survey of the state of the art in robotic-assisted interventional neuroradiology, consider their technical and adoption limitations, and explore future developments critical for the widespread adoption of robotic-assisted neurointerventions.

Quantitative Ultrasound Assessment of Hepatic Steatosis.

Background/Aims: Non-alcoholic fatty liver disease (NAFLD) is widespread chronic disease of the live in humans with the prevalence of 30% of the United States population.1,2 The goal of the study is to validate the performance of quantitative ultrasound algorithms in the assessment of hepatic steatosis in patients with suspected NAFLD. Methods: This prospective study enrolled a total of 31 patients with clinical suspicion of NAFLD to receive liver fat measurements by quantitative ultrasound and reference MRI measurements (proton density fat-fraction, PDFF). The following ultrasound (US) parameters based on both raw ultrasound RF (Radio Frequency) data and 2D B-mode images of the liver were analyzed with subsequent correlation with MRI-PDFF: hepatorenal index, acoustic attenuation coefficient, Nakagami coefficient parameter, shear wave viscosity, shear wave dispersion and shear wave elasticity. Ultrasound parameters were also correlated with the presence of hypertension and diabetes. Results: The mean (± SD) age and body mass index of the patients were 49.03 (± 12.49) and 30.12 (± 6.15), respectively. Of the aforementioned ultrasound parameters, the hepatorenal index and acoustic attenuation coefficient showed a strong correlation with MRI-PDFF derivations of hepatic steatosis, with r-values of 0.829 and 0.765, respectively. None of the remaining US parameters showed strong correlations with PDFF. Significant differences in Nakagami parameters and acoustic attenuation coefficients were found in those patients with and without hypertension. Conclusions: Hepatorenal index and acoustic attenuation coefficient correlate well with MRI-PDFF-derived measurements of hepatic steatosis. Quantitative ultrasound is a promising tool for the diagnosis and assessment of patients with NAFLD.

Noncontrast Computed Tomography vs Computed Tomography Perfusion or Magnetic Resonance Imaging Selection in Late Presentation of Stroke With Large-Vessel Occlusion.

Importance: Advanced imaging for patient selection in mechanical thrombectomy is not widely available. Objective: To compare the clinical outcomes of patients selected for mechanical thrombectomy by noncontrast computed tomography (CT) vs those selected by computed tomography perfusion (CTP) or magnetic resonance imaging (MRI) in the extended time window. Design, Setting, and Participants: This multinational cohort study included consecutive patients with proximal anterior circulation occlusion stroke presenting within 6 to 24 hours of time last seen well from January 2014 to December 2020. This study was conducted at 15 sites across 5 countries in Europe and North America. The duration of follow-up was 90 days from stroke onset. Exposures: Computed tomography with Alberta Stroke Program Early CT Score, CTP, or MRI. Main Outcomes and Measures: The primary end point was the distribution of modified Rankin Scale (mRS) scores at 90 days (ordinal shift). Secondary outcomes included the rates of 90-day functional independence (mRS scores of 0-2), symptomatic intracranial hemorrhage, and 90-day mortality. Results: Of 2304 patients screened for eligibility, 1604 patients were included, with a median (IQR) age of 70 (59-80) years; 848 (52.9%) were women. A total of 534 patients were selected to undergo mechanical thrombectomy by CT, 752 by CTP, and 318 by MRI. After adjustment of confounders, there was no difference in 90-day ordinal mRS shift between patients selected by CT vs CTP (adjusted odds ratio [aOR], 0.95 [95% CI, 0.77-1.17]; P = .64) or CT vs MRI (aOR, 0.95 [95% CI, 0.8-1.13]; P = .55). The rates of 90-day functional independence (mRS scores 0-2 vs 3-6) were similar between patients selected by CT vs CTP (aOR, 0.90 [95% CI, 0.7-1.16]; P = .42) but lower in patients selected by MRI than CT (aOR, 0.79 [95% CI, 0.64-0.98]; P = .03). Successful reperfusion was more common in the CT and CTP groups compared with the MRI group (474 [88.9%] and 670 [89.5%] vs 250 [78.9%]; P < .001). No significant differences in symptomatic intracranial hemorrhage (CT, 42 [8.1%]; CTP, 43 [5.8%]; MRI, 15 [4.7%]; P = .11) or 90-day mortality (CT, 125 [23.4%]; CTP, 159 [21.1%]; MRI, 62 [19.5%]; P = .38) were observed. Conclusions and Relevance: In patients undergoing proximal anterior circulation mechanical thrombectomy in the extended time window, there were no significant differences in the clinical outcomes of patients selected with noncontrast CT compared with those selected with CTP or MRI. These findings have the potential to widen the indication for treating patients in the extended window using a simpler and more widespread noncontrast CT-only paradigm.

Mechanical thrombectomy beyond the circle of Willis: efficacy and safety of different techniques for M2 occlusions.

BACKGROUND: M2 segment occlusions represent approximately one-third of non-lacunar ischemic stroke and can lead to permanent neurological deficits. Various techniques are available for mechanical thrombectomy beyond the circle of Willis, but data evaluating their effectiveness and safety are lacking. METHODS: A retrospective review of patients with ischemic stroke undergoing mechanical thrombectomy for M2 occlusions from 13 centers in North American and Europe was performed. Tandem or multiple-territory occlusions were excluded. The primary outcome was 90-day modified Rankin Scale and reperfusion rates across stent-retriever, direct aspiration and combined techniques. RESULTS: There were 465 patients (mean age 71.48±14.03 years, 53.1% female) with M2 occlusions who underwent mechanical thrombectomy. Stent-retriever alone was used in 133 (28.6%), direct aspiration alone in 93 (20.0%) and the combined technique in 239 (51.4%) patients. Successful reperfusion was achieved with the combined technique in 198 (82.2%; OR 2.6 (1.1-6.9)), with stent-retriever alone in 112 (84.2%; OR 9.2 (1.9-44.6)) and with direct aspiration alone in 62 (66.7%; referencecategory). Intraprocedural subarachnoid hemorrhages (iSAH) were 36 (7.7%) and were more likely to occur in patients treated with the stent-retrievers (OR 5.0 (1.1-24.3)) and combined technique (OR 4.6 (1.1-20.9)). Good clinical outcome was achieved in 260 (61.8%) patients, while 59 (14.0%) patients died. Older age, higher baseline NIHSS (National Institutes of Health Stroke Scale), parenchymal hemorrhage and iSAH were associated with poor outcome while successful recanalization and higher baseline ASPECTS (Alberta Stroke Program Early CT Score) were associated with good outcome. No differences were found among the three techniques in terms of clinical outcome. CONCLUSION: Stent-retrievers and a combined approach for M2 occlusions seem more effective than direct aspiration, but with higher rates of iSAH. This leads to no detectable difference in clinical outcome at 3 months.

Acute occlusion of the fetal posterior cerebral artery: diagnosis and management paradigms.

BACKGROUND AND PURPOSE: The diagnosis and management of acute fetal posterior cerebral artery occlusion are challenging. While endovascular treatment is established for anterior circulation large vessel occlusion stroke, little is known about the course of acute fetal posterior cerebral artery occlusions. We report the clinical course, radiological findings and management considerations of acute fetal posterior cerebral artery occlusion stroke. METHODS: We performed a retrospective review of consecutive patients presenting with acute large vessel occlusion who underwent cerebral angiogram and/or mechanical thrombectomy between January 2015 and January 2021. Patients diagnosed with fetal posterior cerebral artery occlusion were included. Demographic data, clinical presentation, imaging findings and management strategies were reviewed. RESULTS: Between January 2015 and January 2021, three patients with fetal posterior cerebral artery occlusion were identified from 400 patients who underwent angiogram and/or mechanical thrombectomy for acute stroke (0.75%). The first patient presented with concomitant fetal posterior cerebral artery and middle cerebral artery occlusions. Thrombectomy was performed with recanalisation of the fetal posterior cerebral artery but the patient died from malignant oedema. The second patient presented with isolated fetal posterior cerebral artery occlusion. No endovascular intervention was performed and the patient was disabled from malignant posterior cerebral artery infarct. The third patient presented with carotid occlusion and was found to have fetal posterior cerebral artery occlusion after internal carotid artery recanalisation. No further intervention was performed. The patient was left with residual contralateral homonymous hemianopia and mild left sided weakness. CONCLUSION: Fetal posterior cerebral artery occlusion is a rare, but potentially disabling, cause of ischaemic stroke. Endovascular treatment is feasible. Further investigation is needed to compare the efficacy of medical versus endovascular management strategies.

Repeated Mechanical Endovascular Thrombectomy for Recurrent Large Vessel Occlusion: A Multicenter Experience.

BACKGROUND AND PURPOSE: Mechanical thrombectomy (MT) is now the standard of care for large vessel occlusion (LVO) stroke. However, little is known about the frequency and outcomes of repeat MT (rMT) for patients with recurrent LVO. METHODS: This is a retrospective multicenter cohort of patients who underwent rMT at 6 tertiary institutions in the United States between March 2016 and March 2020. Procedural, imaging, and outcome data were evaluated. Outcome at discharge was evaluated using the modified Rankin Scale. RESULTS: Of 3059 patients treated with MT during the study period, 56 (1.8%) underwent at least 1 rMT. Fifty-four (96%) patients were analyzed; median age was 64 years. The median time interval between index MT and rMT was 2 days; 35 of 54 patients (65%) experienced recurrent LVO during the index hospitalization. The mechanism of stroke was cardioembolism in 30 patients (56%), intracranial atherosclerosis in 4 patients (7%), extracranial atherosclerosis in 2 patients (4%), and other causes in 18 patients (33%). A final TICI recanalization score of 2b or 3 was achieved in all 54 patients during index MT (100%) and in 51 of 54 patients (94%) during rMT. Thirty-two of 54 patients (59%) experienced recurrent LVO of a previously treated artery, mostly the pretreated left MCA (23 patients, 73%). Fifty of the 54 patients (93%) had a documented discharge modified Rankin Scale after rMT: 15 (30%) had minimal or no disability (modified Rankin Scale score ≤2), 25 (50%) had moderate to severe disability (modified Rankin Scale score 3-5), and 10 (20%) died. CONCLUSIONS: Almost 2% of patients treated with MT experience recurrent LVO, usually of a previously treated artery during the same hospitalization. Repeat MT seems to be safe and effective for attaining vessel recanalization, and good outcome can be expected in 30% of patients.

Cerebrovascular events and outcomes in hospitalized patients with COVID-19: The SVIN COVID-19 Multinational Registry.

BACKGROUND: Severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) has been associated with a significant risk of thrombotic events in critically ill patients. AIM: To summarize the findings of a multinational observational cohort of patients with SARS-CoV-2 and cerebrovascular disease. METHODS: Retrospective observational cohort of consecutive adults evaluated in the emergency department and/or admitted with coronavirus disease 2019 (COVID-19) across 31 hospitals in four countries (1 February 2020-16 June 2020). The primary outcome was the incidence rate of cerebrovascular events, inclusive of acute ischemic stroke, intracranial hemorrhages (ICH), and cortical vein and/or sinus thrombosis (CVST). RESULTS: Of the 14,483 patients with laboratory-confirmed SARS-CoV-2, 172 were diagnosed with an acute cerebrovascular event (1.13% of cohort; 1130/100,000 patients, 95%CI 970-1320/100,000), 68/171 (40.5%) were female and 96/172 (55.8%) were between the ages 60 and 79 years. Of these, 156 had acute ischemic stroke (1.08%; 1080/100,000 95%CI 920-1260/100,000), 28 ICH (0.19%; 190/100,000 95%CI 130-280/100,000), and 3 with CVST (0.02%; 20/100,000, 95%CI 4-60/100,000). The in-hospital mortality rate for SARS-CoV-2-associated stroke was 38.1% and for ICH 58.3%. After adjusting for clustering by site and age, baseline stroke severity, and all predictors of in-hospital mortality found in univariate regression (p < 0.1: male sex, tobacco use, arrival by emergency medical services, lower platelet and lymphocyte counts, and intracranial occlusion), cryptogenic stroke mechanism (aOR 5.01, 95%CI 1.63-15.44, p < 0.01), older age (aOR 1.78, 95%CI 1.07-2.94, p = 0.03), and lower lymphocyte count on admission (aOR 0.58, 95%CI 0.34-0.98, p = 0.04) were the only independent predictors of mortality among patients with stroke and COVID-19. CONCLUSIONS: COVID-19 is associated with a small but significant risk of clinically relevant cerebrovascular events, particularly ischemic stroke. The mortality rate is high for COVID-19-associated cerebrovascular complications; therefore, aggressive monitoring and early intervention should be pursued to mitigate poor outcomes.

Influence of the COVID-19 Pandemic on Treatment Times for Acute Ischemic Stroke: The Society of Vascular and Interventional Neurology Multicenter Collaboration.

BACKGROUND AND PURPOSE: The pandemic caused by the novel coronavirus disease 2019 (COVID-19) has led to an unprecedented paradigm shift in medical care. We sought to evaluate whether the COVID-19 pandemic may have contributed to delays in acute stroke management at comprehensive stroke centers. METHODS: Pooled clinical data of consecutive adult stroke patients from 14 US comprehensive stroke centers (January 1, 2019, to July 31, 2020) were queried. The rate of thrombolysis for nontransferred patients within the Target: Stroke goal of 60 minutes was compared between patients admitted from March 1, 2019, and July 31, 2019 (pre-COVID-19), and March 1, 2020, to July 31, 2020 (COVID-19). The time from arrival to imaging and treatment with thrombolysis or thrombectomy, as continuous variables, were also assessed. RESULTS: Of the 2955 patients who met inclusion criteria, 1491 were admitted during the pre-COVID-19 period and 1464 were admitted during COVID-19, 15% of whom underwent intravenous thrombolysis. Patients treated during COVID-19 were at lower odds of receiving thrombolysis within 60 minutes of arrival (odds ratio, 0.61 [95% CI, 0.38-0.98]; P=0.04), with a median delay in door-to-needle time of 4 minutes (P=0.03). The lower odds of achieving treatment in the Target: Stroke goal persisted after adjustment for all variables associated with earlier treatment (adjusted odds ratio, 0.55 [95% CI, 0.35-0.85]; P<0.01). The delay in thrombolysis appeared driven by the longer delay from imaging to bolus (median, 29 [interquartile range, 18-41] versus 22 [interquartile range, 13-37] minutes; P=0.02). There was no significant delay in door-to-groin puncture for patients who underwent thrombectomy (median, 83 [interquartile range, 63-133] versus 90 [interquartile range, 73-129] minutes; P=0.30). Delays in thrombolysis were observed in the months of June and July. CONCLUSIONS: Evaluation for acute ischemic stroke during the COVID-19 period was associated with a small but significant delay in intravenous thrombolysis but no significant delay in thrombectomy time metrics. Taking steps to reduce delays from imaging to bolus time has the potential to attenuate this collateral effect of the pandemic.