Constance Boissin

Background: Patients with ER-positive/HER2-negative early breast cancer of low or intermediate grade are generally considered to have a favourable prognosis, yet a number will recur. This risk is poorly captured by standard clinicopathological factors. AI-based medical devices for whole slide image (WSI) analyses used for risk-stratification often predict on slide-level thereby overlooking spatial localisation of the biomarkers. We hypothesise that the spatial distribution of a previously reported AI-based marker provides independent prognostic value beyond its global score. Patients and Methods: This Swedish multicentre cohort study includes 6106 patients with primary invasive breast cancer: 3078 from the Stockholm region (training) and 3028 from Skåne (independent test). Tile-level high-grade morphology scores were generated with ViT_DG, a vision-transformer adaptation of DeepGrade, and the presence of clusters of high-risk tiles localised at the invasive tumour front was defined as a binary biomarker. Primary analyses were pre-specified in the ER-positive/HER2-negative/lymph-node-negative and Grade 1-2 subgroup. Association with breast-cancer recurrence-free interval (BCRFI) and distant metastasis were assessed using multivariable Cox proportional hazards models (HR). Results: In the primary subgroup (n=1536 training; n=1403 test), tumour front clusters were present in 37% and 39% of patients. Cluster-positive status was associated with a 2.30-fold increased risk in the external test. Cluster-negative patients achieved a 10-year BCRFI of 95.3%, equivalent to Grade 1 patients and better than Grade 2 (92.4%), whereas cluster-positive patients (90.2%) did worse than the Grade 3 reference. The association was retained in patients classified as low-risk by ViT_DG (test set multivariable HR 2.86), identifying high-risk patients not captured by the global score. Conclusions: The absence of high-risk morphology clusters at the tumour front is an independent prognostic indicator, identifying a low-risk group independently of the global AI biomarker classification. More broadly, preserving the spatial distribution of AI-derived predictions enables prognostic stratification beyond slide-level aggregation. ClinicalTrials.gov identifier: NCT02306096

Background: Burn automated diagnosis may be instrumental for accurate and timely decision-making at point-of-care, helping to ensure that the right patients are triaged to burns centers. This is particularly important in resource-poor settings. Objective: We studied the intention of nonspecialized clinicians to engage in automated diagnosis in burn care as well as their perceptions toward clinical risks. Methods: A self-administered survey was used among a purposive sample of first contact clinicians (n=56) and burns specialists (n=35). The survey had 2 main parts: 1 measuring the intention to use automated diagnosis as per 7 constructs of the Automation Acceptance Model (yielding 8 hypotheses) and 1 on clinical risk perceptions (likelihood and severity of 7 risks). Structural Equation Modelling was used to test the hypotheses among first contact clinicians, and the Mann-Whitney U test was used to measure differences in risk perceptions between the two clinical groups. Results: Many first contact clinicians would intend to use automated diagnosis for burns should the technology be made available in their departments (41/56, 73%). The Automation Acceptance Model concepts contributed moderately to explain what the intention to use automated diagnosis rests on (R²=0.432), with 5 out of 8 hypotheses being supported. The intention to use automated diagnosis was associated with perceived usefulness but not with attitudes toward using it. Of the 7 risks studied, the 1 that was most often considered as high risk of occurring was that of complex burns not being recognized (n=23, 29%). The 2 groups differed significantly in their concern regarding both the likelihood of happening and the severity of 2 risks: the undermanagement of severe burns and the overmanagement of minor burns. Specifically, a larger proportion of first contact clinicians were more concerned than burns specialists (n=13, 27% versus 6% and n=11, 23% versus 6% for undermanagement and overmanagement, respectively). Conclusions: Almost three-quarters of first contact clinicians were inclined to seek automated advice for burn diagnosis. The proposed model contributes to explaining the intention to use with 5 hypotheses supported. When seeking additional determinants, clinical risk perception is a dimension that should be considered in any artificial intelligence implementation process, to help ensure sustainability.

Assessment of burn extent and depth are critical and require very specialized diagnosis. Automated image-based algorithms could assist in performing wound detection and classification. We aimed to develop two deep-learning algorithms that respectively identify burns, and classify whether they require surgery. An additional aim assessed the performances in different Fitzpatrick skin types. Annotated burn (n = 1105) and background (n = 536) images were collected. Using a commercially available platform for deep learning algorithms, two models were trained and validated on 70% of the images and tested on the remaining 30%. Accuracy was measured for each image using the percentage of wound area correctly identified and F1 scores for the wound identifier; and area under the receiver operating characteristic (AUC) curve, sensitivity, and specificity for the wound classifier. The wound identifier algorithm detected an average of 87.2% of the wound areas accurately in the test set. For the wound classifier algorithm, the AUC was 0.885. The wound identifier algorithm was more accurate in patients with darker skin types; the wound classifier was more accurate in patients with lighter skin types. To conclude, image-based algorithms can support the assessment of acute burns with relatively good accuracy although larger and different datasets are needed.

Specialized care is crucial for severe burn injuries whereas minor burns should be handled at point-of-care. Misdiagnosis is common which leads to overburdening the system and to a lack of treatmen...

Although they are a common type of injury worldwide, burns are challenging to diagnose, not least by untrained point-of-care clinicians. Given their visual nature, developments in artificial intelligence (AI) have sparked growing interest in the automated diagnosis of burns. This review aims to appraise the state of evidence thus far, with a focus on the identification and severity classification of acute burns. Three publicly available electronic databases were searched to identify peer-reviewed studies on the automated diagnosis of acute burns, published in English since 2005. From the 20 identified, three were excluded on the grounds that they concerned animals, older burns or lacked peer review. The remaining 17 studies, from nine different countries, were classified into three AI generations, considering the type of algorithms developed and the images used. Whereas the algorithms for burn identification have not gained much in accuracy across generations, those for severity classification improved substantially (from 66.2% to 96.4%), not least in the latest generation (n = 8). Those eight studies were further assessed for methodological bias and results applicability, using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. This highlighted the feasibility nature of the studies and their detrimental dependence on online databases of poorly documented images, at the expense of a substantial risk for patient selection and limited applicability in the clinical setting. In moving past the pilot stage, future development work would benefit from greater input from clinicians, who could contribute essential point-of-care knowledge and perspectives.

In high-income countries of the Arabian Peninsula, including the Sultanate of Oman, motorization has been extremely rapid. As a result, road traffic crashes are by far the highest cause of premature mortality, and speeding is an acknowledged key risk factor. Theory-based interventions are needed to target prevention of this unsafe practice. This study sheds light on determinants of speeding among new generations of Omani drivers applying the Theory of Planned Behaviour (TPB). A questionnaire covering all five main constructs of the TPB was first contextualized and administered to two target groups: male drivers of all ages (n = 1107) approached in person when renewing their driving license and university students drivers (men and women) reached through internet contact (n = 655). Multiple, stepwise linear regression analyses were used to explore factors associated with speeding. Results indicate that driving fast and not respecting the posted speed limits was common in both groups of drivers, although rates were higher among students; 41.8% reported driving a bit faster than other drivers and 24.1% faster than the posted speed limit compared with 31.4% and 14.2% in male drivers of all ages. In both groups the TPB model predicted to a limited extent the determinants of speeding behaviour. However, the intention to speed was associated with a negative attitude towards the respect of rules for men of all ages (β = -0.30 (p<0.001)) and for students (β = -0.26 (p<0.001)); a positive view regarding subjective norms (β = 0.25 (p<0.001) and β = 0.28 (p<0.001) respectively), and behavioural control (β = 0.15 (p<0.001) and β = 0.20 (p<0.001) respectively). Intention was the only significant predictor of speeding behaviour (β = 0.48 (p<0.001); and β = 0.64 (p<0.001)). To conclude, speeding is widespread among Omani drivers of all ages and the intention to respect posted speed limits meets a range of barriers that need greater consideration in order to achieve a better safety culture in the country.

Background: Acute burns are complex to diagnose and erroneous assessments impact on the victim’s mortality and morbidity. Specialists operate from a small number of burns centres and the capacity to provide timely assistance to front line clinicians is therefore limited. Diagnostic assistance through artificial intelligence could be an option to provide a timely and equitable diagnosis. This project sheds light on the feasibility of the development of an artificial intelligence algorithm for assisted diagnosis of acute burns and clarifies challenges faced along the way. Methods: A bank of images has been built from a number of burn centres in South Africa and is continuously being updated (currently about 1200 images). Attempts have been made to train deep learning algorithms to diagnose the burn depth, an element that is challenging both at bedside and using image-based teleconsultation. We came across methodological challenges that need further consideration. Results: Some challenges are clinical, i.e. inherent to the complexity of burn diagnosis, and imply the need of having an accurate diagnosis for the burn images on which the algorithm is trained. Other challenges pertain to the actual development of an algorithm. Prior to identifying burn depth a complex task relates to the feasibility of finding the burn itself in images of varying body parts and backgrounds. Further, training an algorithm for diagnosing burn depth also requires, large numbers of varying cases, the accurate labelling of the wound area for training, and decisions to be made as regards the best outcome to train upon. Current preliminary results indicate satisfactory identification of the burn area and promising results with regards burn depth diagnosis. Conclusions: Development of artificial intelligence algorithms require strong collaborations and discussions between technical and clinical experts but are showing promising results. Key messages: Development of clinical image-based automated diagnosis involve a number of critical challenges on both the technical and clinical sides that need to be addressed prior to optimization. Once optimally developed, deep learning algorithms are a potential solution to assist with the reduction of the burden of burns on the health services by providing timely, and cost-effective advice.