Thrombus Distribution in Vaccine-induced Immune Thrombotic Thrombocytopenia after ChAdOx1 nCov-19 Vaccination.

Vaccination strategies have been at the forefront of controlling the COVID-19 pandemic. An association between vaccine-induced immune thrombotic thrombocytopenia (VITT) and one of these vaccines, the ChAdOx1 nCov-19 vaccine, is now recognized. The purpose of this study was to investigate the frequency and location of thrombosis in each vascular system using CT, MRI, and US to identify additional sites of thrombus in a United Kingdom-wide sample of patients with confirmed VITT. Thirty-two radiology centers identified through the national collaborative Radiology Academic Network for Trainees were invited from the United Kingdom; seven of these contributed to this study. All patients with confirmed VITT ¬between February 3 and May 12, 2021, who met the inclusion criteria were included. The location and extent of thrombi were evaluated using CT, MRI, and US. A total of 40 patients (median age, 41 years [IQR, 32-52]; 22 [55%] men) with confirmed vaccine-induced immune thrombotic thrombocytopenia after administration of their first ChAdOx1 nCov-19 vaccine were included. Thirty-two patients (80%) developed symptoms within the first 14 days, and eight (20%) developed symptoms within 14-28 days. Twenty-nine patients (72%) experienced neurologic symptoms and were confirmed to have cerebral venous sinus thrombosis, 12 (30%) had clinical deterioration and repeat imaging demonstrated extension of their primary thrombus, and eight (20%) died. Twenty-five of 30 patients (83%) who underwent additional imaging had occult thrombosis. In conclusion, patients with VITT are likely to have multiple sites of thrombosis, with the most frequent being cerebral venous sinus thrombosis in combination with pulmonary embolism and portomesenteric venous thrombosis. Whole-body imaging with contrast-enhanced CT can be used to identify occult thrombosis.

Deep learning to automate the labelling of head MRI datasets for computer vision applications.

OBJECTIVES: The purpose of this study was to build a deep learning model to derive labels from neuroradiology reports and assign these to the corresponding examinations, overcoming a bottleneck to computer vision model development. METHODS: Reference-standard labels were generated by a team of neuroradiologists for model training and evaluation. Three thousand examinations were labelled for the presence or absence of any abnormality by manually scrutinising the corresponding radiology reports ('reference-standard report labels'); a subset of these examinations (n = 250) were assigned 'reference-standard image labels' by interrogating the actual images. Separately, 2000 reports were labelled for the presence or absence of 7 specialised categories of abnormality (acute stroke, mass, atrophy, vascular abnormality, small vessel disease, white matter inflammation, encephalomalacia), with a subset of these examinations (n = 700) also assigned reference-standard image labels. A deep learning model was trained using labelled reports and validated in two ways: comparing predicted labels to (i) reference-standard report labels and (ii) reference-standard image labels. The area under the receiver operating characteristic curve (AUC-ROC) was used to quantify model performance. Accuracy, sensitivity, specificity, and F1 score were also calculated. RESULTS: Accurate classification (AUC-ROC > 0.95) was achieved for all categories when tested against reference-standard report labels. A drop in performance (ΔAUC-ROC > 0.02) was seen for three categories (atrophy, encephalomalacia, vascular) when tested against reference-standard image labels, highlighting discrepancies in the original reports. Once trained, the model assigned labels to 121,556 examinations in under 30 min. CONCLUSIONS: Our model accurately classifies head MRI examinations, enabling automated dataset labelling for downstream computer vision applications. KEY POINTS: • Deep learning is poised to revolutionise image recognition tasks in radiology; however, a barrier to clinical adoption is the difficulty of obtaining large labelled datasets for model training. • We demonstrate a deep learning model which can derive labels from neuroradiology reports and assign these to the corresponding examinations at scale, facilitating the development of downstream computer vision models. • We rigorously tested our model by comparing labels predicted on the basis of neuroradiology reports with two sets of reference-standard labels: (1) labels derived by manually scrutinising each radiology report and (2) labels derived by interrogating the actual images.

Cerebral venous sinus thrombosis and thrombocytopenia after COVID-19 vaccination - A report of two UK cases.

Recent reports have highlighted rare, and sometimes fatal, cases of cerebral venous sinus thrombosis (CVST) and thrombocytopenia following the Vaxzevria vaccine. An underlying immunological mechanism similar to that of spontaneous heparin-induced thrombocytopenia (HIT) is suspected, with the identification of antibodies to platelet factor-4 (PF4), but without previous heparin exposure. This unusual mechanism has significant implications for the management approach used, which differs from usual treatment of CVST. We describe the cases of two young males, who developed severe thrombocytopenia and fatal CVST following the first dose of Vaxzevria. Both presented with a headache, with subsequent rapid neurological deterioration. One patient underwent PF4 antibody testing, which was positive. A rapid vaccination programme is essential in helping to control the COVID-19 pandemic. Hence, it is vital that such COVID-19 vaccine-associated events, which at this stage appear to be very rare, are viewed through this lens. However, some cases have proved fatal. It is critical that clinicians are alerted to the emergence of such events to facilitate appropriate management. Patients presenting with CVST features and thrombocytopenia post-vaccination should undergo PF4 antibody testing and be managed in a similar fashion to HIT, in particular avoiding heparin and platelet transfusions.

Factors affecting the labelling accuracy of brain MRI studies relevant for deep learning abnormality detection.

Unlocking the vast potential of deep learning-based computer vision classification systems necessitates large data sets for model training. Natural Language Processing (NLP)-involving automation of dataset labelling-represents a potential avenue to achieve this. However, many aspects of NLP for dataset labelling remain unvalidated. Expert radiologists manually labelled over 5,000 MRI head reports in order to develop a deep learning-based neuroradiology NLP report classifier. Our results demonstrate that binary labels (normal vs. abnormal) showed high rates of accuracy, even when only two MRI sequences (T2-weighted and those based on diffusion weighted imaging) were employed as opposed to all sequences in an examination. Meanwhile, the accuracy of more specific labelling for multiple disease categories was variable and dependent on the category. Finally, resultant model performance was shown to be dependent on the expertise of the original labeller, with worse performance seen with non-expert vs. expert labellers.

Improving stroke pathways using an adhesive ambulatory ECG patch: reducing time for patients to ECGs and subsequent results.

Three south-London hospital trusts undertook a feasibility study, comparing data from 93 patients who received the 14-day adhesive ambulatory electrocardiography (ECG) patch Zio XT with retrospective data from 125 patients referred for 24-hour Holter for cryptogenic stroke and transient ischaemic attack following negative 12-lead ECG. As the ECG patch was fitted the same day as the clinical decision for ambulatory ECG monitoring was made, median time to the patient having the monitor fitted was significantly reduced in all three hospital trusts compared with 24-hour Holter being ordered and fitted. Hospital visits reduced by a median of two for patients receiving Zio XT. This project supports that it is feasible to use a patch as part of routine clinical care with a positive impact on care pathways.

Comparing APLS and oranges: caution with the use of direct-acting oral anticoagulants (DOACs) instead of warfarin in the treatment of antiphospholipid syndrome (APLS).

A 63-year-old woman with known antiphospholipid syndrome (APLS) presented with catastrophic APLS and multiorgan dysfunction after a change in her anticoagulation from warfarin to rivaroxaban. Evidence suggests direct-acting oral anticoagulants (DOACs) like rivaroxaban may be less effective than warfarin in secondary prevention of thrombotic events in high-risk APLS patients.

The Interaction of Genetic Mutations in PARK2 and FA2H Causes a Novel Phenotype in a Case of Childhood-Onset Movement Disorder.

Mutations in the PARK2 gene have been implicated in the pathogenesis of early-onset Parkinson's disease. We present a case of movement disorder in a 4-year-old child from consanguineous parents and with a family history of Dopamine responsive dystonia, who was diagnosed with early-onset Parkinson's disease based on initial identification of a pathogenic PARK2 mutation. However, the evolution of the child's clinical picture was unusually rapid, with a preponderance of pyramidal rather than extrapyramidal symptoms, leading to re-investigation of the case with further imaging and genetic sequencing. Interestingly, a second homozygous mutation in the FA2H gene, implicated in Hereditary spastic paraplegia, was revealed, appearing to have contributed to the novel phenotype observed, and highlighting a potential interaction between the two mutated genes.

Autism spectrum disorder: prospects for treatment using gene therapy.

Autism spectrum disorder (ASD) is characterised by the concomitant occurrence of impaired social interaction; restricted, perseverative and stereotypical behaviour; and abnormal communication skills. Recent epidemiological studies have reported a dramatic increase in the prevalence of ASD with as many as 1 in every 59 children being diagnosed with ASD. The fact that ASD appears to be principally genetically driven, and may be reversible postnatally, has raised the exciting possibility of using gene therapy as a disease-modifying treatment. Such therapies have already started to seriously impact on human disease and particularly monogenic disorders (e.g. metachromatic leukodystrophy, SMA type 1). In regard to ASD, technical advances in both our capacity to model the disorder in animals and also our ability to deliver genes to the central nervous system (CNS) have led to the first preclinical studies in monogenic ASD, involving both gene replacement and silencing. Furthermore, our increasing awareness and understanding of common dysregulated pathways in ASD have broadened gene therapy's potential scope to include various polygenic ASDs. As this review highlights, despite a number of outstanding challenges, gene therapy has excellent potential to address cognitive dysfunction in ASD.