Objective
There has been a large amount of research in the field of artificial intelligence (AI) as applied to clinical radiology. However, these studies vary in design and quality and systematic ...reviews of the entire field are lacking.This systematic review aimed to identify all papers that used deep learning in radiology to survey the literature and to evaluate their methods. We aimed to identify the key questions being addressed in the literature and to identify the most effective methods employed.
Methods
We followed the PRISMA guidelines and performed a systematic review of studies of AI in radiology published from 2015 to 2019. Our published protocol was prospectively registered.
Results
Our search yielded 11,083 results. Seven hundred sixty-seven full texts were reviewed, and 535 articles were included. Ninety-eight percent were retrospective cohort studies. The median number of patients included was 460. Most studies involved MRI (37%). Neuroradiology was the most common subspecialty. Eighty-eight percent used supervised learning. The majority of studies undertook a segmentation task (39%). Performance comparison was with a state-of-the-art model in 37%. The most used established architecture was UNet (14%). The median performance for the most utilised evaluation metrics was Dice of 0.89 (range .49–.99), AUC of 0.903 (range 1.00–0.61) and Accuracy of 89.4 (range 70.2–100). Of the 77 studies that externally validated their results and allowed for direct comparison, performance on average decreased by 6% at external validation (range increase of 4% to decrease 44%).
Conclusion
This systematic review has surveyed the major advances in AI as applied to clinical radiology.
Key Points
• While there are many papers reporting expert-level results by using deep learning in radiology, most apply only a narrow range of techniques to a narrow selection of use cases.
• The literature is dominated by retrospective cohort studies with limited external validation with high potential for bias.
• The recent advent of AI extensions to systematic reporting guidelines and prospective trial registration along with a focus on external validation and explanations show potential for translation of the hype surrounding AI from code to clinic.
Background
During the current COVID-19 health crisis virtual geriatric clinics have become increasingly utilised to complete outpatient consultations, although concerns exist about feasibility of ...such virtual consultations for older people. The aim of this rapid review is to describe the satisfaction, clinic productivity, clinical benefit, and costs associated with the virtual geriatric clinic model of care.
Methods
A rapid review of PubMed, MEDLINE and CINAHL databases was conducted up to April 2020. Two independent reviewers extracted the information. Four subdomains were focused on: satisfaction with the virtual geriatric clinic, clinic productivity, clinical benefit to patients, costs and any challenges associated with the virtual clinic process.
Results
Nine studies with 975 patients met our inclusion criteria. All were observational studies. Seven studies reported patients were satisfied with the virtual geriatric clinic model of care. Productivity outcomes included reports of cost-effectiveness, savings on transport, and improved waiting list metrics. Clinical benefits included successful polypharmacy reviews, and reductions in acute hospitalisation rates. Varying challenges were reported for both clinicians and patients in eight of the nine studies. Hearing impairments and difficulty with technology added to anxieties experienced by patients. Physicians missed the added value of a thorough physical examination and had concerns about confidentiality.
Conclusion
Virtual geriatric clinics demonstrate evidence of productivity, benefit to patients, cost effectiveness and patient satisfaction with the treatment provided. In the current suboptimal pandemic climate, virtual geriatric clinics may allow Geriatricians to continue to provide an outpatient service, despite the encountered inherent challenges.
Abstract
Identification of gene-by-environment interactions (GxE) is crucial to understand the interplay of environmental effects on complex traits. However, current methods evaluating GxE on ...biobank-scale datasets have limitations. We introduce MonsterLM, a multiple linear regression method that does not rely on model specification and provides unbiased estimates of variance explained by GxE. We demonstrate robustness of MonsterLM through comprehensive genome-wide simulations using real genetic data from 325,989 individuals. We estimate GxE using waist-to-hip-ratio, smoking, and exercise as the environmental variables on 13 outcomes (N = 297,529-325,989) in the UK Biobank. GxE variance is significant for 8 environment-outcome pairs, ranging from 0.009 – 0.071. The majority of GxE variance involves SNPs without strong marginal or interaction associations. We observe modest improvements in polygenic score prediction when incorporating GxE. Our results imply a significant contribution of GxE to complex trait variance and we show MonsterLM to be well-purposed to handle this with biobank-scale data.
Guidelines recommend increased salt intake as a first-line recommendation in the management of symptomatic orthostatic hypotension and recurrent syncope. There have been no systematic reviews of this ...intervention. We sought to summarize the evidence for increased salt intake in patients with orthostatic intolerance syndromes.
We conducted a systematic review and meta-analysis of studies in PubMed, EMBASE, and CINAHL. Interventional studies that increased salt intake in individuals with orthostatic intolerance syndromes were included. Primary outcome measures included incidence of falls and injuries, and rates of syncope and presyncope. Secondary outcome measures included other orthostatic intolerance symptoms, blood pressure, and heart rate.
A total of 14 studies were eligible, including participants with orthostatic hypotension, syncope, postural orthostatic tachycardia syndrome, and idiopathic orthostatic tachycardia (n = 391). Mean age was 35.6 (± 15) years. All studies were small and short-term (<60 mins-90 days). No study reported on the effect of increased salt intake on falls or injuries. Meta-analysis demonstrated that during head-up tilt, mean time to presyncope with salt intake increased by 1.57 minutes (95% confidence interval CI, 1.26-1.88), mean systolic blood pressure increased by 12.27 mm Hg (95% CI, 10.86-13.68), and mean heart rate decreased by −3.97 beats per minute (95% CI, −4.08 to −3.86), compared with control. Increased salt increased supine blood pressure by 1.03 mm Hg (95% CI, 0.81 to 1.25). Increased salt intake resulted in an improvement or resolution of symptoms in 62.3% (95% CI, 51.6 to 72.6) of participants in short-term follow-up studies (mean follow-up of 44.3 days, 6 studies; n=91). Methodological quality of studies were low with high statistical heterogeneity in all meta-analyses.
Our meta-analysis provides low-quality evidence of a short-term improvement in orthostatic intolerance with increased salt intake. There were no clinical trials demonstrating the efficacy and safety of increased salt intake on long-term clinical outcomes. Overall, there is a paucity of clinical trial evidence to support a cornerstone recommendation in the management of orthostatic intolerance syndromes.
Abstract
Background
To compare functional and health related quality of life outcomes post-transcatheter aortic valve implantation (TAVI) and surgical aortic valve replacement (SAVR) in patients with ...critical aortic stenosis (AS) across low to high-risk surgical candidates. These patient-centred factors will be compared between both groups in the short to medium term time frames and will aid in shared decision making between patients and healthcare workers.
Materials and methods
We conducted a systematic review and meta-analysis of randomised controlled trials which compared TAVI with SAVR and reported on quality of life (QoL) and functional scores.
The scores used were the Kansas City Cardiomyopathy Questionnaire (KCCQ), Euroqol-5DL (EQ5DL), the short form-36/12 (SF-36/12) and the NYHA.
Results
We identified eight trials with a total of 8898 participants. Both groups showed improvements from baseline at one month. At one month there was a statistically significant difference in standardised mean difference (SMD) in favour of TAVI for EQ5DL (SMD 0.37, 95% CI 0.26,0.49), KCCQ (SMD 0.53,95% CI 0.48, 0.58), SF physical summary (SMD 0.55, 95% CI 0.31 – 0.78) and SF mental summary (SMD 0.34, 95% CI 0.27 – 0.40). At one year there was no statistically significant difference between any of these QoL metrics. For NYHA, no significant difference in odds ratio of class III/IV was observed at one month between TAVI and SAVR (OR 0.94, 95% CI 0.83, 1.07), however, TAVI was associated with reduced odds ratio of NYHA class I/II at one year (OR 0.87, 95% CI 0.78, 0.98).
Conclusion
Both groups were associated with improvements in QoL and functional outcomes with TAVI reporting more significant improvements in QoL at one-month post-procedures. No significant improvements between groups were seen at one year. This is the largest meta-analysis comparing post-operative health-related quality of life outcomes post SAVR and TAVI and has major implications in shared decision making for the treatment of aortic stenosis.
To date no research has examined the potential influence of acute stress symptoms (ASD) on subsequent development of post-traumatic stress disorder (PTSD) symptoms in stroke survivors. Our objective ...was to examine whether acute stress symptoms measured 1-2 weeks post-stroke predicted the presence of post-traumatic stress symptoms measured 6-12 weeks later. Fifty four participants who completed a measure of acute stress disorder at 1-2 weeks following stroke (time 1) and 31 of these participants completed a measure of posttraumatic stress disorder 6-12 weeks later (time 2). Participants also completed measures of stroke severity, functional impairment, cognitive impairment, depression, anxiety, pre-morbid intelligence and pain across both time points. Some 22% met the criteria for ASD at baseline and of those, 62.5% went on to meet the criteria for PTSD at follow-up. Meanwhile two of the seven participants (28.6%) who met the criteria for PTSD at Time 2, did not meet the ASD criteria at Time 1 (so that PTSD developed subsequently). A hierarchical multiple regression analysis indicated that the presence of acute stress symptoms at baseline was predictive of post-traumatic stress symptoms at follow-up (R.sup.2 = .26, p < .01). Less severe stroke was correlated with higher levels of post-traumatic stress symptoms at Time 2 (rho = .42, p < .01). The results highlight the importance of early assessment and identification of acute stress symptoms in stroke survivors as a risk factor for subsequent PTSD. Both ASD and PTSD were prevalent and the presence of both disorders should be assessed.
Depression has been reported to be a risk factor of acute stroke, based largely on studies in high-income countries. In the INTERSTROKE study, we explored the contribution of depressive symptoms to ...acute stroke risk and 1-month outcome across regions of the world, within subpopulations and by stroke type.
The INTERSTROKE is an international case-control study of risk factors of first acute stroke, conducted in 32 countries. Cases were patients with CT- or MRI-confirmed incident acute hospitalized stroke, and controls were matched for age, sex, and within sites. Standardized questions asked about self-reported depressive symptoms during the previous 12 months and the use of prescribed antidepressant medications were recorded. Multivariable conditional logistic regression was used to determine the association of prestroke depressive symptoms with acute stroke risk. Adjusted ordinal logistic regression was used to explore the association of prestroke depressive symptoms with poststroke functional outcome, measured with the modified Rankin scale at 1 month after stroke.
Of 26,877 participants, 40.4% were women, and the mean age was 61.7 ± 13.4 years. The prevalence of depressive symptoms within the last 12 months was higher in cases compared with that in controls (18.3% vs 14.1%,
< 0.001) and differed by region (
interaction <0.001), with lowest prevalence in China (6.9% in controls) and highest in South America (32.2% of controls). In multivariable analyses, prestroke depressive symptoms were associated with greater odds of acute stroke (odds ratio OR 1.46, 95% CI 1.34-1.58), which was significant for both intracerebral hemorrhage (OR 1.56, 95% CI 1.28-1.91) and ischemic stroke (OR 1.44, 95% CI 1.31-1.58). A larger magnitude of association with stroke was seen in patients with a greater burden of depressive symptoms. While preadmission depressive symptoms were not associated with a greater odds of worse baseline stroke severity (OR 1.02, 95% CI 0.94-1.10), they were associated with a greater odds of poor functional outcome at 1 month after acute stroke (OR 1.09, 95% CI 1.01-1.19).
In this global study, we recorded that depressive symptoms are an important risk factor of acute stroke, including both ischemic and hemorrhagic stroke. Preadmission depressive symptoms were associated with poorer functional outcome, but not baseline stroke severity, suggesting an adverse role of depressive symptoms in poststroke recovery.
Therapeutic antibiotic dose monitoring can be particularly challenging in septic patients requiring renal replacement therapy. Our aim was to conduct an exploratory population pharmacokinetic (PK) ...analysis on PK of vancomycin following intermittent infusion in critically ill patients receiving continuous venovenous haemodiafiltration (CVVHDF); focussing on the influence of dialysis-related covariates.
This was a retrospective single-centre tertiary level intensive care unit (ICU) study, which included patients treated concurrently with vancomycin and CVVHDF between January 2015 and July 2016. We extracted clinical, laboratory and dialysis data from the electronic healthcare record (EHR), using strict inclusion criteria. A population PK analysis was conducted with a one-compartment model using the PMetrics population PK modelling package. A base structural model was developed, with further analyses including clinical and dialysis-related data to improve model prediction through covariate inclusion. The final selected model simulated patient concentrations using probability of target attainment (PTA) plots to investigate the probability of different dosing regimens achieving target therapeutic concentrations.
A total of 106 vancomycin dosing intervals (155 levels) in 24 patients were examined. An acceptable 1-compartment base model was produced (Plots of observed vs. population predicted concentrations (Obs-Pred) R
= 0.78). No continuous covariates explored resulted in a clear improvement over the base model. Inclusion of anticoagulation modality and vasopressor use as categorical covariates resulted in similar PK parameter estimates, with a trend towards lower parameter estimate variability when using regional citrate anti-coagulation or without vasopressor use. Simulations using PTA plots suggested that a 2 g loading dose followed by 750 mg 12 hourly as maintenance dose, commencing 12 h after loading, is required to achieve adequate early target trough concentrations of at least 15 mg/L.
PTA simulations suggest that acceptable trough vancomycin concentrations can be achieved early in treatment with a 2 g loading dose and maintenance dose of 750 mg 12 hourly for critically ill patients on CVVHDF.
Mitochondrial DNA copy number (mtDNA-CN) is an accessible blood-based measurement believed to capture underlying mitochondrial (MT) function. The specific biological processes underpinning its ...regulation, and whether those processes are causative for disease, is an area of active investigation.
We developed a novel method for array-based mtDNA-CN estimation suitable for biobank-scale studies, called 'automatic mitochondrial copy (AutoMitoC).' We applied AutoMitoC to 395,781 UKBiobank study participants and performed genome- and exome-wide association studies, identifying novel common and rare genetic determinants. Finally, we performed two-sample Mendelian randomization to assess whether genetically low mtDNA-CN influenced select MT phenotypes.
Overall, genetic analyses identified 71 loci for mtDNA-CN, which implicated several genes involved in rare mtDNA depletion disorders, deoxynucleoside triphosphate (dNTP) metabolism, and the MT central dogma. Rare variant analysis identified
mutation carriers as having higher mtDNA-CN (beta = 0.23 SDs; 95% CI, 0.18-0.29; p=2.6 × 10
), a potential therapeutic target for patients with mtDNA depletion disorders, but at increased risk of breast cancer (OR = 1.91; 95% CI, 1.52-2.40; p=2.7 × 10
). Finally, Mendelian randomization analyses suggest a causal effect of low mtDNA-CN on dementia risk (OR = 1.94 per 1 SD decrease in mtDNA-CN; 95% CI, 1.55-2.32; p=7.5 × 10
).
Altogether, our genetic findings indicate that mtDNA-CN is a complex biomarker reflecting specific MT processes related to mtDNA regulation, and that these processes are causally related to human diseases.
No funds supported this specific investigation. Awards and positions supporting authors include: Canadian Institutes of Health Research (CIHR) Frederick Banting and Charles Best Canada Graduate Scholarships Doctoral Award (MC, PM); CIHR Post-Doctoral Fellowship Award (RM); Wellcome Trust Grant number: 099313/B/12/A; Crasnow Travel Scholarship; Bongani Mayosi UCT-PHRI Scholarship 2019/2020 (TM); Wellcome Trust Health Research Board Irish Clinical Academic Training (ICAT) Programme Grant Number: 203930/B/16/Z (CJ); European Research Council COSIP Grant Number: 640580 (MO); E.J. Moran Campbell Internal Career Research Award (MP); CISCO Professorship in Integrated Health Systems and Canada Research Chair in Genetic and Molecular Epidemiology (GP).