1 January 2022 in Digital Health
Objective: Medical students, as clinicians and healthcare leaders of the future, are key stakeholders in the clinical roll-out of artificial intelligence-driven technologies. The authors aim to provide the first report on the state of artificial intelligence in medical education globally by exploring the perspectives of medical students. Methods: The authors carried out a mixed-methods study of focus groups and surveys with 128 medical students from 48 countries. The study explored knowledge around artificial intelligence as well as what students wished to learn about artificial intelligence and how they wished to learn this. A combined qualitative and quantitative analysis was used. Results: Support for incorporating teaching on artificial intelligence into core curricula was ubiquitous across the globe, but few students had received teaching on artificial intelligence. Students showed knowledge on the applications of artificial intelligence in clinical medicine as well as on artificial intelligence ethics. They were interested in learning about clinical applications, algorithm development, coding and algorithm appraisal. Hackathon-style projects and multidisciplinary education involving computer science students were suggested for incorporation into the curriculum. Conclusions: Medical students from all countries should be provided teaching on artificial intelligence as part of their curriculum to develop skills and knowledge around artificial intelligence to ensure a patient-centred digital future in medicine. This teaching should focus on the applications of artificial intelligence in clinical medicine. Students should also be given the opportunity to be involved in algorithm development. Students in low- and middle-income countries require the foundational technology as well as robust teaching on artificial intelligence to ensure that they can drive innovation in their healthcare settings.
16 July 2020 in International Journal of Computer Assisted Radiology and Surgery
Purpose: Management of vestibular schwannoma (VS) is based on tumour size as observed on T1 MRI scans with contrast agent injection. The current clinical practice is to measure the diameter of the tumour in its largest dimension. It has been shown that volumetric measurement is more accurate and more reliable as a measure of VS size. The reference approach to achieve such volumetry is to manually segment the tumour, which is a time intensive task. We suggest that semi-automated segmentation may be a clinically applicable solution to this problem and that it could replace linear measurements as the clinical standard.Methods: Using high-quality software available for academic purposes, we ran a comparative study of manual versus semi-automated segmentation of VS on MRI with 5 clinicians and scientists. We gathered both quantitative and qualitative data to compare the two approaches; including segmentation time, segmentation effort and segmentation accuracy.Results: We found that the selected semi-automated segmentation approach is significantly faster (167 s vs 479 s, $$p<0.001$$ ), less temporally and physically demanding and has approximately equal performance when compared with manual segmentation, with some improvements in accuracy. There were some limitations, including algorithmic unpredictability and error, which produced more frustration and increased mental effort in comparison with manual segmentation.Conclusion: We suggest that semi-automated segmentation could be applied clinically for volumetric measurement of VS on MRI. In future, the generic software could be refined for use specifically for VS segmentation, thereby improving accuracy.
16 August 2024 in Nature Communications
Functional coactivation between human brain regions is partly explained by white matter connections; however, how the structure-function relationship varies by function remains unclear. Here, we reference large data repositories to compute maps of structure-function correspondence across hundreds of specific functions and brain regions. We use natural language processing to accurately predict structure-function correspondence for specific functions and to identify macroscale gradients across the brain that correlate with structure-function correspondence as well as cortical thickness. Our findings suggest structure-function correspondence unfolds along a sensory-fugal organizational axis, with higher correspondence in primary sensory and motor cortex for perceptual and motor functions, and lower correspondence in association cortex for cognitive functions. Our study bridges neuroscience and natural language to describe how structure-function coupling varies by region and function in the brain, offering insight into the diversity and evolution of neural network properties.
30 September 2020 in Medical Science Educator
We read with great interest the article by Goldenberg et al. [1] and would like to offer our perspective on virtual learning experiences as medical students from four UK universities. During the COVID-19 pandemic, our learning was shifted online with clinical teaching largely suspended. We believe that innovative solutions to virtual teaching can facilitate valuable and safe educational experiences. Furthermore, these can support the message that engaging with real patients is integral, not extraneous, to becoming a physician. During the COVID-19 pandemic, medical students have expressed dissatisfaction due to a lack of clinical teaching and patient interaction [2]. The need for novel ways to engage with patients remotely and safely is evident. As the authors highlighted, online resources can be limited in their ability to help students develop clinical skills and professionalism. We agree that traditional online teaching resources cannot fully substitute in-person learning. However, there are a number of innovative virtual solutions which could provide the opportunity for safe and meaningful interaction with clinical teams and real patients. We would like to outline some here for consideration by the authors.
17 January 2022 in Epilepsia Open
Objective: To evaluate the role of intracranial electroencephalography monitoring in diagnosing and directing the appropriate therapy for MRI‐negative epilepsy and to present the surgical outcomes of patients following treatment.Methods: Retrospective chart review between 2015‐2021 at a single institution identified 48 patients with no lesion on MRI, who received surgical intervention for their epilepsy. The outcomes assessed were the surgical treatment performed and the International League Against Epilepsy seizure outcomes at 1 year of follow‐up.Results: Eleven patients underwent surgery without invasive monitoring, including vagus nerve stimulation (10%), deep brain stimulation (8%), laser interstitial thermal therapy (2%), and callosotomy (2%). The remaining 37 patients received invasive monitoring followed by resection (35%), responsive neurostimulation (21%), and deep brain stimulation (15%) or no treatment (6%). At 1 year postoperatively, 39% were Class 1‐2, 36% were Class 3‐4 and 24% were Class 5. More patients with Class 1‐2 or 3‐4 outcomes underwent invasive monitoring (100% and 83% respectively) compared with those with poor outcomes (25%,P < .001). Patients with Class 1‐2 outcomes more commonly underwent resection or responsive neurostimulation: 69% and 31%, respectively (P < .001).Significance: The optimal management of MRI‐negative focal epilepsy may involve invasive monitoring followed by resection or responsive neurostimulation in most cases, as these treatments were associated with the best seizure outcomes in our cohort. Unless multifocal onset is clear from the noninvasive evaluation, invasive monitoring is preferred before pursuing deep brain stimulation or vagal nerve stimulation directly.
29 June 2021 in Neurology