Validation of a deep learning model for traumatic brain injury detection and NIRIS grading on non-contrast CT: a multi-reader study with promising results and opportunities for improvement.

PURPOSE: This study aimed to assess and externally validate the performance of a deep learning (DL) model for the interpretation of non-contrast computed tomography (NCCT) scans of patients with suspicion of traumatic brain injury (TBI). METHODS: This retrospective and multi-reader study included patients with TBI suspicion who were transported to the emergency department and underwent NCCT scans. Eight reviewers, with varying levels of training and experience (two neuroradiology attendings, two neuroradiology fellows, two neuroradiology residents, one neurosurgery attending, and one neurosurgery resident), independently evaluated NCCT head scans. The same scans were evaluated using the version 5.0 of the DL model icobrain tbi. The establishment of the ground truth involved a thorough assessment of all accessible clinical and laboratory data, as well as follow-up imaging studies, including NCCT and magnetic resonance imaging, as a consensus amongst the study reviewers. The outcomes of interest included neuroimaging radiological interpretation system (NIRIS) scores, the presence of midline shift, mass effect, hemorrhagic lesions, hydrocephalus, and severe hydrocephalus, as well as measurements of midline shift and volumes of hemorrhagic lesions. Comparisons using weighted Cohen's kappa coefficient were made. The McNemar test was used to compare the diagnostic performance. Bland-Altman plots were used to compare measurements. RESULTS: One hundred patients were included, with the DL model successfully categorizing 77 scans. The median age for the total group was 48, with the omitted group having a median age of 44.5 and the included group having a median age of 48. The DL model demonstrated moderate agreement with the ground truth, trainees, and attendings. With the DL model's assistance, trainees' agreement with the ground truth improved. The DL model showed high specificity (0.88) and positive predictive value (0.96) in classifying NIRIS scores as 0-2 or 3-4. Trainees and attendings had the highest accuracy (0.95). The DL model's performance in classifying various TBI CT imaging common data elements was comparable to that of trainees and attendings. The average difference for the DL model in quantifying the volume of hemorrhagic lesions was 6.0 mL with a wide 95% confidence interval (CI) of - 68.32 to 80.22, and for midline shift, the average difference was 1.4 mm with a 95% CI of - 3.4 to 6.2. CONCLUSION: While the DL model outperformed trainees in some aspects, attendings' assessments remained superior in most instances. Using the DL model as an assistive tool benefited trainees, improving their NIRIS score agreement with the ground truth. Although the DL model showed high potential in classifying some TBI CT imaging common data elements, further refinement and optimization are necessary to enhance its clinical utility.

Association of Cerebrovascular and Alzheimer Disease Biomarkers With Cholinergic White Matter Degeneration in Cognitively Unimpaired Individuals.

BACKGROUND AND OBJECTIVES: Several pathologic processes might contribute to the degeneration of the cholinergic system in aging. We aimed to determine the contribution of amyloid, tau, and cerebrovascular biomarkers toward the degeneration of cholinergic white matter (WM) projections in cognitively unimpaired individuals. METHODS: The contribution of amyloid and tau pathology was assessed through CSF levels of the Aß42/40 ratio and phosphorylated tau (p-tau). CSF Aß38 levels were also measured. Cerebrovascular pathology was assessed using automatic segmentations of WM lesions (WMLs) on MRI. Cholinergic WM projections (i.e., cingulum and external capsule pathways) were modeled using tractography based on diffusion tensor imaging data. Sex and APOE ε4 carriership were also included in the analysis as variables of interest. RESULTS: We included 203 cognitively unimpaired individuals from the H70 Gothenburg Birth Cohort Studies (all individuals aged 70 years, 51% female). WM lesion burden was the most important contributor to the degeneration of both cholinergic pathways (increase in mean square error [IncMSE] = 98.8% in the external capsule pathway and IncMSE = 93.3% in the cingulum pathway). Levels of Aß38 and p-tau also contributed to cholinergic WM degeneration, especially in the external capsule pathway (IncMSE = 28.4% and IncMSE = 23.4%, respectively). The Aß42/40 ratio did not contribute notably to the models (IncMSE<3.0%). APOE ε4 carriers showed poorer integrity in the cingulum pathway (IncMSE = 21.33%). Women showed poorer integrity of the external capsule pathway (IncMSE = 21.55%), which was independent of amyloid status as reflected by the nonsignificant differences in integrity when comparing amyloid-positive vs amyloid-negative women participants (T201 = -1.55; p = 0.123). DISCUSSION: In cognitively unimpaired older individuals, WMLs play a central role in the degeneration of cholinergic pathways. Our findings highlight the importance of WM lesion burden in the elderly population, which should be considered in the development of prevention programs for neurodegeneration and cognitive impairment.

Brain Atrophy Subtypes and the ATN Classification Scheme in Alzheimer's Disease.

INTRODUCTION: We investigated the association between atrophy subtypes of Alzheimer's disease (AD), the ATN classification scheme, and key demographic and clinical factors in 2 cohorts with different source characteristics (a highly selective research-oriented cohort, the Alzheimer's Disease Neuroimaging Initiative [ADNI]; and a naturalistic heterogeneous clinically oriented cohort, Karolinska Imaging Dementia Study [KIDS]). METHODS: A total of 382 AD patients were included. Factorial analysis of mixed data was used to investigate associations between AD subtypes based on brain atrophy patterns, ATN profiles based on cerebrospinal fluid biomarkers, and age, sex, Mini Mental State Examination (MMSE), cerebrovascular disease (burden of white matter signal abnormalities, WMSAs), and APOE genotype. RESULTS: Older patients with high WMSA burden, belonging to the typical AD subtype and showing A+T+N+ or A+T+N- profiles clustered together and were mainly from ADNI. Younger patients with low WMSA burden, limbic-predominant or minimal atrophy AD subtypes, and A+T-N- or A+T-N+ profiles clustered together and were mainly from KIDS. APOE ε4 carriers more frequently showed the A+T-N- and A+T+N- profiles. CONCLUSIONS: Our findings align with the recent framework for biological subtypes of AD: the combination of risk factors, protective factors, and brain pathologies determines belonging of AD patients to distinct subtypes.

Contemporary outcomes of infrapopliteal atherectomy with angioplasty versus balloon angioplasty alone for critical limb ischemia.

OBJECTIVE: Limited data exist comparing atherectomy (At) with balloon angioplasty for infrapopliteal peripheral arterial disease. The objective of this study was to compare the outcomes of infrapopliteal At with angioplasty vs angioplasty alone in patients with critical limb ischemia. METHODS: This is a retrospective, single-center, longitudinal study comparing patients undergoing either infrapopliteal At with angioplasty or angioplasty alone for critical limb ischemia, between January 2014 and October 2017. The primary outcome was primary patency rates. Secondary outcomes were reintervention rates, assisted primary patency, secondary patency, major adverse cardiac events, major adverse limb events, amputation-free survival, overall survival, and wound healing rates. Data were analyzed in multivariate generalized linear models with log rank tests to determine survival in Kaplan-Meier curves. RESULTS: There were 342 infrapopliteal interventions, 183 percutaneous balloon angioplasty (PTA; 54%), and 159 atherectomies (At) with PTA (46%) performed on 290 patients, with a mean age of 67 ± 12 years; 61% of the patients were male. The PTA and At/PTA groups had similar demographics, tissue loss (79% vs 84%; P = .26), ischemic rest pain (21% vs 16%; P = .51), mean follow-up (19 ± 9 vs 20 ± 9 months; P = .32), mean number of vessels treated (1.7 ± 0.8 vs 1.9 ± 0.8; P = .08) and the mean lesion length treated (6.55 ± 5.00 cm vs 6.02 ± 4.00 cm; P = .08), respectively. Similar 3-month (96 ± 1% vs 94 ± 1%), 6-month (85 ± 2% vs 86 ± 3%), 12-month (68 ± 3% vs 69 ± 4%), and 18-month (57 ± 4% vs 62 ± 4%) primary patency rates were seen in the two groups (P = .87). At/PTA patients had significantly higher reintervention rates as compared with the PTA patients (28% vs 16%; P = .02). Similar assisted primary patency rates (67 ± 4% vs 69 ± 4%; P = .78) and secondary patency rates (61 ± 4% vs 66 ± 4%; P = .98) were seen in the PTA and At/PTA groups at 18 months. The 30-days major adverse cardiac event rates (3% vs 2%; P = .13) and 30-day major adverse limb event rates (5% vs 4%; P = .2) were similar in both groups. Wound healing rates (72 ± 3% vs 75 ± 2%; P = .12), 1-year amputation-free survival (68 ± 4.1% vs 70 ± 2%; P = .5), and 1-year overall survival (76 ± 4% vs 78 ± 4%; P = .39) rates did not differ in the PTA and At/PTA groups. THE At/PTA group had higher local complication rates (7 [4%] vs 1 [0.5%]; P = .03) CONCLUSIONS: At with angioplasty provides similar patency rates compared with angioplasty alone for infrapopliteal peripheral arterial disease, but associated with higher reintervention and local complication rates. Further appropriately designed studies are required to determine the exact role of At in this subset of patients.

Management of Thoracic and Lumbar Spine Fractures: Is MRI Necessary in Patients without Neurological Deficits?

MRI after a CT scan for thoracolumbar spine (TLS) trauma has become commonplace because of the concerns for detection of posterior ligamentous complex injuries in the absence of substantial scientific evidence to support its use. We hypothesized that MRI scans were not necessary in the clinical management of TLS fractures. A prospective study was conducted at our Level I trauma center. A total of 39 neurologically intact patients with TLS fracture on CT were enrolled. The patients' CT scan and neurological examination were reviewed by a senior neurosurgeon, who determined clinical management based on these data. Assessment was repeated after an MRI of the spine was performed, and a second clinical plan was devised. The two treatment schemes were then compared. MRI resulted in a change in clinical management in 15 per cent of patients. Ten per cent of patients changed from requiring a brace to no brace and merely observation alone. In no patient planned for nonoperative care was surgery deemed necessary after completion of MRI. Among five patients with initial plans for operative intervention, two avoided surgery after the MRI. MRI has little impact on the management of patients with CT-proven thoracic and lumbar spine fractures. Only when surgery is planned based on CT studies does an MRI seem to assist with determining optimal care.

Mutually exclusive redox forms of HMGB1 promote cell recruitment or proinflammatory cytokine release.

Tissue damage causes inflammation, by recruiting leukocytes and activating them to release proinflammatory mediators. We show that high-mobility group box 1 protein (HMGB1) orchestrates both processes by switching among mutually exclusive redox states. Reduced cysteines make HMGB1 a chemoattractant, whereas a disulfide bond makes it a proinflammatory cytokine and further cysteine oxidation to sulfonates by reactive oxygen species abrogates both activities. We show that leukocyte recruitment and activation can be separated. A nonoxidizable HMGB1 mutant in which serines replace all cysteines (3S-HMGB1) does not promote cytokine production, but is more effective than wild-type HMGB1 in recruiting leukocytes in vivo. BoxA, a HMGB1 inhibitor, interferes with leukocyte recruitment but not with activation. We detected the different redox forms of HMGB1 ex vivo within injured muscle. HMGB1 is completely reduced at first and disulfide-bonded later. Thus, HMGB1 orchestrates both key events in sterile inflammation, leukocyte recruitment and their induction to secrete inflammatory cytokines, by adopting mutually exclusive redox states.