Deep learning-based risk stratification of preoperative breast biopsies using digital whole slide images.

BACKGROUND: Nottingham histological grade (NHG) is a well established prognostic factor in breast cancer histopathology but has a high inter-assessor variability with many tumours being classified as intermediate grade, NHG2. Here, we evaluate if DeepGrade, a previously developed model for risk stratification of resected tumour specimens, could be applied to risk-stratify tumour biopsy specimens. METHODS: A total of 11,955,755 tiles from 1169 whole slide images of preoperative biopsies from 896 patients diagnosed with breast cancer in Stockholm, Sweden, were included. DeepGrade, a deep convolutional neural network model, was applied for the prediction of low- and high-risk tumours. It was evaluated against clinically assigned grades NHG1 and NHG3 on the biopsy specimen but also against the grades assigned to the corresponding resection specimen using area under the operating curve (AUC). The prognostic value of the DeepGrade model in the biopsy setting was evaluated using time-to-event analysis. RESULTS: Based on preoperative biopsy images, the DeepGrade model predicted resected tumour cases of clinical grades NHG1 and NHG3 with an AUC of 0.908 (95% CI: 0.88; 0.93). Furthermore, out of the 432 resected clinically-assigned NHG2 tumours, 281 (65%) were classified as DeepGrade-low and 151 (35%) as DeepGrade-high. Using a multivariable Cox proportional hazards model the hazard ratio between DeepGrade low- and high-risk groups was estimated as 2.01 (95% CI: 1.06; 3.79). CONCLUSIONS: DeepGrade provided prediction of tumour grades NHG1 and NHG3 on the resection specimen using only the biopsy specimen. The results demonstrate that the DeepGrade model can provide decision support to identify high-risk tumours based on preoperative biopsies, thus improving early treatment decisions.

Real-world clinical characterization, healthcare resource utilization and productivity loss in chronic graft versus host patients exposed to extracorporeal photopheresis in Sweden.

BACKGROUND: Extracorporeal photopheresis (ECP) is frequently used to treat moderate-severe chronic graft versus host disease (cGVHD), however limited data exists describing ECP treatment effects on healthcare and societal costs. We aimed to characterize clinical and health economic outcomes and productivity loss in cGVHD patients exposed to ECP. METHODS: We identified 2708 patients aged ≥ 18 years with a record of allogeneic hematopoietic stem cell transplantation (HSCT) in the Swedish Patient Register between 2006 and 2020. Patients exposed to ECP from 3-months post HSCT (index) were included (n= 183). Data was linked to the Prescribed Drug Register, the Cause of Death Register, and the Longitudinal Integrated Database for Health Insurance and Labor Market Studies (LISA). RESULTS: The median patient age at index was 51 years (IQR1-3; 38-61). In the 3-month period before ECP initiation compared to 9-12 months post-ECP, the cumulative three-month dose per patient decreased prednisolone/prednisone (1,381 mg vs. 658 mg, p < 0.001) and cyclosporin (12,242 mg vs. 3,501 mg, p < 0.001). Infection incidence also decreased over the same period (79.2% vs 59.1%, p < 0.001). Time spent in healthcare decreased from 68.9% to 22.1% from the first and fifth follow-up year respectively, and corresponding annual healthcare cost reduced from €27,719 to €1,981. Among patients < 66 years of age, sickness-related workplace absence decreased from 73.2% to 31.9% between the first and fifth follow-up year, with median annual productivity loss decreasing from €20,358 to €7,211 per patient. CONCLUSIONS: ECP was associated with reduced use of corticosteroids, immunosuppressive agents, and fewer infections. Furthermore, cost and healthcare utilization decreased over time.

Image-based teleconsultation using smartphones or tablets: qualitative assessment of medical experts.

BACKGROUND: Mobile health has promising potential in improving healthcare delivery by facilitating access to expert advice. Enabling experts to review images on their smartphone or tablet may save valuable time. This study aims at assessing whether images viewed by medical specialists on handheld devices such as smartphones and tablets are perceived to be of comparable quality as when viewed on a computer screen. METHODS: This was a prospective study comparing the perceived quality of 18 images on three different display devices (smartphone, tablet and computer) by 27 participants (4 burn surgeons and 23 emergency medicine specialists). The images, presented in random order, covered clinical (dermatological conditions, burns, ECGs and X-rays) and non-clinical subjects and their perceived quality was assessed using a 7-point Likert scale. Differences in devices' quality ratings were analysed using linear regression models for clustered data adjusting for image type and participants' characteristics (age, gender and medical specialty). RESULTS: Overall, the images were rated good or very good in most instances and more so for the smartphone (83.1%, mean score 5.7) and tablet (78.2%, mean 5.5) than for a standard computer (70.6%, mean 5.2). Both handheld devices had significantly higher ratings than the computer screen, even after controlling for image type and participants' characteristics. Nearly all experts expressed that they would be comfortable using smartphones (n=25) or tablets (n=26) for image-based teleconsultation. CONCLUSION: This study suggests that handheld devices could be a substitute for computer screens for teleconsultation by physicians working in emergency settings.

Accuracy of acute burns diagnosis made using smartphones and tablets: a questionnaire-based study among medical experts.

BACKGROUND: Remote assistance for burns by medical experts can support nurses and general physicians in emergency care with diagnostic and management advice. Previous studies indicate a high diagnostic accuracy based on images viewed on a computer screen, but whether image-based analysis by experts using handheld devices is accurate remains to be determined. METHOD: A review of patient data from eight emergency centres in the Western Cape, South Africa, revealed 10 typical cases of burns commonly seen in children and adults. A web-based questionnaire was created with 51 images of burns representing those cases. Burns specialists from two countries (South Africa and Sweden (n = 8 and 7 respectively)) and emergency medicine specialists from South Africa (n = 11) were contacted by email and asked to assess each burn's total body surface area (TBSA) and depth using a smartphone or tablet. The accuracy and inter-rater reliability of the assessments were measured using intraclass correlation coefficients (ICC), both for all cases aggregated and for paediatric and adult burn cases separately. Eight participants repeated the questionnaire on a computer and intra-rater reliability was calculated. RESULTS: The assessments of TBSA are of high accuracy all specialists aggregated (ICC = 0.82 overall and 0.81 for both child and adult cases separately) and remain high for all three participant groups separately. The burn depth assessments have low accuracy all specialists aggregated, with ICCs of 0.53 overall, 0.61 for child and 0.46 for adult cases. The most accurate assessments of depth are among South African burns specialists (reaching acceptable for child cases); the other two groups' ICCs are low in all instances. Computer-based assessments were similar to those made on handheld devices. CONCLUSION: As was the case for computer-based studies, burns images viewed on handheld devices may be a suitable means of seeking expert advice even with limited additional information when it comes to burn size but less so in the case of burn depth. Familiarity with the type of cases presented could facilitate image-based diagnosis of depth.

Can We Trust the Use of Smartphone Cameras in Clinical Practice? Laypeople Assessment of Their Image Quality.

BACKGROUND: Smartphone cameras are rapidly being introduced in medical practice, among other devices for image-based teleconsultation. Little is known, however, about the actual quality of the images taken, which is the object of this study. MATERIALS AND METHODS: A series of nonclinical objects (from three broad categories) were photographed by a professional photographer using three smartphones (iPhone(®) 4 [Apple, Cupertino, CA], Samsung [Suwon, Korea] Galaxy S2, and BlackBerry(®) 9800 [BlackBerry Ltd., Waterloo, ON, Canada]) and a digital camera (Canon [Tokyo, Japan] Mark II). In a Web survey a convenience sample of 60 laypeople "blind" to the types of camera assessed the quality of the photographs, individually and best overall. We then measured how each camera scored by object category and as a whole and whether a camera ranked best using a Mann-Whitney U test for 2×2 comparisons. RESULTS: There were wide variations between and within categories in the quality assessments for all four cameras. The iPhone had the highest proportion of images individually evaluated as good, and it also ranked best for more objects compared with other cameras, including the digital one. The ratings of the Samsung or the BlackBerry smartphone did not significantly differ from those of the digital camera. CONCLUSIONS: Whereas one smartphone camera ranked best more often, all three smartphones obtained results at least as good as those of the digital camera. Smartphone cameras can be a substitute for digital cameras for the purposes of medical teleconsulation.

The ACROBAT 2022 challenge: Automatic registration of breast cancer tissue.

The alignment of tissue between histopathological whole-slide-images (WSI) is crucial for research and clinical applications. Advances in computing, deep learning, and availability of large WSI datasets have revolutionised WSI analysis. Therefore, the current state-of-the-art in WSI registration is unclear. To address this, we conducted the ACROBAT challenge, based on the largest WSI registration dataset to date, including 4,212 WSIs from 1,152 breast cancer patients. The challenge objective was to align WSIs of tissue that was stained with routine diagnostic immunohistochemistry to its H&E-stained counterpart. We compare the performance of eight WSI registration algorithms, including an investigation of the impact of different WSI properties and clinical covariates. We find that conceptually distinct WSI registration methods can lead to highly accurate registration performances and identify covariates that impact performances across methods. These results provide a comparison of the performance of current WSI registration methods and guide researchers in selecting and developing methods.

Development and evaluation of deep learning algorithms for assessment of acute burns and the need for surgery.

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.

A Multi-Stain Breast Cancer Histological Whole-Slide-Image Data Set from Routine Diagnostics.

The analysis of FFPE tissue sections stained with haematoxylin and eosin (H&E) or immunohistochemistry (IHC) is essential for the pathologic assessment of surgically resected breast cancer specimens. IHC staining has been broadly adopted into diagnostic guidelines and routine workflows to assess the status of several established biomarkers, including ER, PGR, HER2 and KI67. Biomarker assessment can also be facilitated by computational pathology image analysis methods, which have made numerous substantial advances recently, often based on publicly available whole slide image (WSI) data sets. However, the field is still considerably limited by the sparsity of public data sets. In particular, there are no large, high quality publicly available data sets with WSIs of matching IHC and H&E-stained tissue sections from the same tumour. Here, we publish the currently largest publicly available data set of WSIs of tissue sections from surgical resection specimens from female primary breast cancer patients with matched WSIs of corresponding H&E and IHC-stained tissue, consisting of 4,212 WSIs from 1,153 patients.

Clinical decision-support for acute burn referral and triage at specialized centres - Contribution from routine and digital health tools.

BACKGROUND: 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 treatment for others due to resources shortage. OBJECTIVES: The overarching aim was to evaluate four decision-support tools for diagnosis, referral, and triage of acute burns injuries in South Africa and Sweden: referral criteria, mortality prediction scores, image-based remote consultation and automated diagnosis. METHODS: Study I retrospectively assessed adherence to referral criteria of 1165 patients admitted to the paediatric burns centre of the Western Cape of South Africa. Study II assessed mortality prediction of 372 patients admitted to the adults burns centre by evaluating an existing score (ABSI), and by using logistic regression. In study III, an online survey was used to assess the diagnostic accuracy of burn experts' image-based estimations using their smartphone or tablet. In study IV, two deep-learning algorithms were developed using 1105 acute burn images in order to identify the burn, and to classify burn depth. RESULTS: Adherence to referral criteria was of 93.4%, and the age and severity criteria were associated with patient care. In adults, the ABSI score was a good predictor of mortality which affected a fifth of the patients and which was associated with gender, burn size and referral status. Experts were able to diagnose burn size, and burn depth using handheld devices. Finally, both a wound identifier and a depth classifier algorithm could be developed with relatively high accuracy. CONCLUSIONS: Altogether the findings inform on the use of four tools along the care trajectory of patients with acute burns by assisting with the diagnosis, referral and triage from point-of-care to burns centres. This will assist with reducing inequities by improving access to the most appropriate care for patients.

Admission factors associated with the in-hospital mortality of burns patients in resource-constrained settings: A two-year retrospective investigation in a South African adult burns centre.

OBJECTIVE: Little is known concerning the factors associated with in-hospital mortality of trauma patients in resource-constrained settings, not least in burns centres. We investigated this question in the adult burns centre at Tygerberg Hospital in Cape Town. We further assessed whether the Abbreviated Burn Severity Index (ABSI) is an accurate predictive score of mortality in this setting. METHODS: Medical records of all patients admitted with fresh burns over a two-year period (2015 and 2016) were scrutinized to obtain data on patient, injury and admission-related characteristics. Association with in-hospital mortality was investigated for flame burns using logistic regressions and expressed as odds ratios (ORs). The mortality prediction of the ABSI score was assessed using sensitivity and specificity analyses. RESULTS: Overall the in-hospital mortality was 20.4%. For the 263 flame burns, while crude ORs suggested gender, burn depth, burn size, inhalation injury, and referral status were all individually significantly associated with mortality, only the association with female gender, not being referred and burn size remained significant after adjustments (adjusted ORs = 3.79, 2.86 and 1.11 (per percentage increase in size) respectively). For the ABSI score, sensitivity and specificity were 84% and 86% respectively. CONCLUSION: In this specialised centre, mortality occurs in one in five patients. It is associated with a few clinical parameters, and can be predicted using the ABSI score.

Determinants of speeding among new generations of car drivers from the Arabian Peninsula. An investigation based among Omani drivers using the theory of planned behaviour.

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 (ß = -0.30 (p<0.001)) and for students (ß = -0.26 (p<0.001)); a positive view regarding subjective norms (ß = 0.25 (p<0.001) and ß = 0.28 (p<0.001) respectively), and behavioural control (ß = 0.15 (p<0.001) and ß = 0.20 (p<0.001) respectively). Intention was the only significant predictor of speeding behaviour (ß = 0.48 (p<0.001); and ß = 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.

Adherence to Referral Criteria at Admission and Patient Management at a Specialized Burns Centre: The Case of the Red Cross War Memorial Children's Hospital in Cape Town, South Africa.

Referral guidelines for burn care are meant to assist in decision-making as regards patient transfer and admissions to specialized units. Little is known, however, concerning how closely they are followed and whether they are linked to patient care. This is the object of the current study, focused on the paediatric burns centre of the Red Cross War Memorial Children's Hospital in Cape Town, South Africa. All patients admitted to the centre during the winters of 2011-2015 (n = 1165) were included. The patient files were scrutinized to clarify whether the referral criteria in place were identified (seven in total) and to compile data on patient and injury characteristics. A case was defined as adherent to the criteria when at least one criterion was fulfilled and adherence was expressed as a percentage with 95% confidence intervals, for all years aggregated as well as by year and by patient or injury characteristics. The association between adherence to any individual criterion and hospital care (surgery or longer length of stay) was measured using logistic regressions. The overall adherence was 93.4% (100% among children under 2 years of age and 86% among the others) and it did not vary remarkably over time. The two criteria of "injury sustained at a specific anatomical site" (85.2%) and "young age" (51.9%) were those most often identified. Children aged 2 years or older were more likely to undergo surgery or to stay longer than those of young age (although a referral criterion) and so were those with higher injury severity (a referral criterion). In this specialized paediatric burns centre, children are admitted mainly according to the guidelines. However, given the high prevalence of paediatric burns in the region and the limited resources at the burns centre, adherence to the guidelines need to be further studied at all healthcare levels in the province.