Artificial intelligence in digital pathology: a systematic review and meta-analysis of diagnostic test accuracy.

Ensuring diagnostic performance of artificial intelligence (AI) before introduction into clinical practice is essential. Growing numbers of studies using AI for digital pathology have been reported over recent years. The aim of this work is to examine the diagnostic accuracy of AI in digital pathology images for any disease. This systematic review and meta-analysis included diagnostic accuracy studies using any type of AI applied to whole slide images (WSIs) for any disease. The reference standard was diagnosis by histopathological assessment and/or immunohistochemistry. Searches were conducted in PubMed, EMBASE and CENTRAL in June 2022. Risk of bias and concerns of applicability were assessed using the QUADAS-2 tool. Data extraction was conducted by two investigators and meta-analysis was performed using a bivariate random effects model, with additional subgroup analyses also performed. Of 2976 identified studies, 100 were included in the review and 48 in the meta-analysis. Studies were from a range of countries, including over 152,000 whole slide images (WSIs), representing many diseases. These studies reported a mean sensitivity of 96.3% (CI 94.1-97.7) and mean specificity of 93.3% (CI 90.5-95.4). There was heterogeneity in study design and 99% of studies identified for inclusion had at least one area at high or unclear risk of bias or applicability concerns. Details on selection of cases, division of model development and validation data and raw performance data were frequently ambiguous or missing. AI is reported as having high diagnostic accuracy in the reported areas but requires more rigorous evaluation of its performance.

Immune subtyping of melanoma whole slide images using multiple instance learning.

Determining early-stage prognostic markers and stratifying patients for effective treatment are two key challenges for improving outcomes for melanoma patients. Previous studies have used tumour transcriptome data to stratify patients into immune subgroups, which were associated with differential melanoma specific survival and potential predictive biomarkers. However, acquiring transcriptome data is a time-consuming and costly process. Moreover, it is not routinely used in the current clinical workflow. Here, we attempt to overcome this by developing deep learning models to classify gigapixel haematoxylin and eosin (H&E) stained pathology slides, which are well established in clinical workflows, into these immune subgroups. We systematically assess six different multiple instance learning (MIL) frameworks, using five different image resolutions and three different feature extraction methods. We show that pathology-specific self-supervised models using 10x resolution patches generate superior representations for the classification of immune subtypes. In addition, in a primary melanoma dataset, we achieve a mean area under the receiver operating characteristic curve (AUC) of 0.80 for classifying histopathology images into 'high' or 'low immune' subgroups and a mean AUC of 0.82 in an independent TCGA melanoma dataset. Furthermore, we show that these models are able to stratify patients into 'high' and 'low immune' subgroups with significantly different melanoma specific survival outcomes (log rank test, P< 0.005). We anticipate that MIL methods will allow us to find new biomarkers of high importance, act as a tool for clinicians to infer the immune landscape of tumours and stratify patients, without needing to carry out additional expensive genetic tests.

Image analysis of cutaneous melanoma histology: a systematic review and meta-analysis.

The current subjective histopathological assessment of cutaneous melanoma is challenging. The application of image analysis algorithms to histological images may facilitate improvements in workflow and prognostication. To date, several individual algorithms applied to melanoma histological images have been reported with variations in approach and reported accuracies. Histological digital images can be created using a camera mounted on a light microscope, or through whole slide image (WSI) generation using a whole slide scanner. Before any such tool could be integrated into clinical workflow, the accuracy of the technology should be carefully evaluated and summarised. Therefore, the objective of this review was to evaluate the accuracy of existing image analysis algorithms applied to digital histological images of cutaneous melanoma. Database searching of PubMed and Embase from inception to 11th March 2022 was conducted alongside citation checking and examining reports from organisations. All studies reporting accuracy of any image analysis applied to histological images of cutaneous melanoma, were included. The reference standard was any histological assessment of haematoxylin and eosin-stained slides and/or immunohistochemical staining. Citations were independently deduplicated and screened by two review authors and disagreements were resolved through discussion. The data was extracted concerning study demographics; type of image analysis; type of reference standard; conditions included and test statistics to construct 2 × 2 tables. Data was extracted in accordance with our protocol and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Diagnostic Test Accuracy (PRISMA-DTA) Statement. A bivariate random-effects meta-analysis was used to estimate summary sensitivities and specificities with 95% confidence intervals (CI). Assessment of methodological quality was conducted using a tailored version of the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. The primary outcome was the pooled sensitivity and specificity of image analysis applied to cutaneous melanoma histological images. Sixteen studies were included in the systematic review, representing 4,888 specimens. Six studies were included in the meta-analysis. The mean sensitivity and specificity of automated image analysis algorithms applied to melanoma histological images was 90% (CI 82%, 95%) and 92% (CI 79%, 97%), respectively. Based on limited and heterogeneous data, image analysis appears to offer high accuracy when applied to histological images of cutaneous melanoma. However, given the early exploratory nature of these studies, further development work is necessary to improve their performance.

The utility of teledermatology in the evaluation of skin lesions.

INTRODUCTION: Past studies have shown mixed results about the accuracy of store-and-forward (SAF) teledermatology in the evaluation of skin lesions. The objective of this study is to determine the accuracy of SAF teledermatology in the diagnosis of skin lesions and biopsy decision compared to in-person clinical evaluation. METHODS: Histories and photographs of skin lesions gathered at clinic visits were sent as SAF consults to teledermatologists, whose diagnoses and biopsy decisions were recorded and compared statistically to the clinic data.Results and Discussion: We enrolled 206 patients with 308 lesions in the study. The study population was composed of 50% males (n = 104), and most patients were white (n = 179, 87%) and not Hispanic/Latino (n = 167, 81%). There was good concordance for biopsy decision between the clinic dermatologist (CD) and teledermatologist (TD) (Cohen's kappa (κ) = 0.51), which did not significantly differ when melanocytic lesions were excluded (κ = 0.54). The sensitivity and specificity of teledermatology based on biopsy decision was 0.71 and 0.85, respectively. Overall concordance in first diagnosis between the CD and TD was good (κ = 0.60). While there was no difference between CD and TD in proportion of correct diagnoses compared to histopathology, two skin cancers presentations were missed by TD. Study limitations included sample size, enrolment bias and differing amounts of teledermatologist case experience. Teledermatology has good concordance in diagnosis and biopsy decision when compared to clinic dermatology. Teledermatology may be utilized in the evaluation of skin lesions to expand access to dermatologic care.

Association of Tumor Characteristics With Insurance Type Among Patients Undergoing Mohs Micrographic Surgery for Nonmelanoma Skin Cancer.

Importance: Little is known about the association between insurance type and tumor or treatment characteristics among patients undergoing Mohs micrographic surgery (MMS) for nonmelanoma skin cancer (NMSC). Objective: To investigate whether there are differences in tumor and treatment characteristics among patients undergoing MMS for NMSC by insurance type. Design, Setting, and Participants: This retrospective cohort study included patients with NMSC who presented for surgery at an academic MMS practice between May 2017 and May 2019. Main Outcomes and Measures: Preoperative and postoperative tumor diameters, number of MMS stages, type of closure, and number of high-risk tumors were compared based on insurance type among uninsured and underinsured patients and those with private insurance, Medicare, and Veterans Affairs (VA) insurance. Results: A total of 1397 patients with NMSC (978 [70%] male; mean [SD] age, 68.5 [12.4] years) underwent 1916 MMS procedures. Of these patients, 868 (45%) had Medicare, 570 (30%) had private insurance, 299 (16%) had VA insurance, and 179 (9%) were treated at a safety net clinic or were uninsured. Compared with patients with private insurance, uninsured and underinsured patients had significantly larger preoperative tumor bed diameters (difference, 28%; 95% CI, 14%-43%; P < .001) and postoperative defect sizes (difference, 28%, 95% CI, 16%-41%; P < .001). Patients with Medicare and VA insurance did not have significantly different preoperative tumor bed diameters compared with patients with private insurance. Patients with VA insurance had larger postoperative defect sizes than patients with private insurance (difference, 12%; 95% CI, 2%-23%; P = .02). The number of MMS stages and type of closure did not significantly differ based on insurance type. Conclusions and Relevance: In this cohort study of patients undergoing MMS for NMSC, larger preoperative tumor and postoperative defect sizes were associated with being uninsured or underinsured compared with privately insured. Future studies are required to determine why these differences exist to deliver optimal care to all patients.

A Point-of-Use Quality Assurance Tool for Digital Pathology Remote Working.

Pathology services are facing pressures due to the COVID-19 pandemic. Digital pathology has the capability to meet some of these unprecedented challenges by allowing remote diagnoses to be made at home, during periods of social distancing or self-isolation. However, while digital pathology allows diagnoses to be made on standard computer screens, unregulated home environments may not be conducive for optimal viewing conditions. There is also a paucity of experimental evidence available to support the minimum display requirements for digital pathology. This study presents a Point-of-Use Quality Assurance (POUQA) tool for remote assessment of viewing conditions for reporting digital pathology slides. The tool is a psychophysical test combining previous work from successfully implemented quality assurance tools in both pathology and radiology to provide a minimally intrusive display screen validation task, before viewing digital slides. The test is specific to pathology assessment in that it requires visual discrimination between colors derived from hematoxylin and eosin staining, with a perceptual difference of ±1 delta E (dE). This tool evaluates the transfer of a 1 dE signal through the digital image display chain, including the observers' contrast and color responses within the test color range. The web-based system has been rapidly developed and deployed as a response to the COVID-19 pandemic and may be used by anyone in the world to help optimize flexible working conditions at: http://www. virtualpathology.leeds.ac.uk/res earch/systems/pouqa/.

The prevalence of symptoms in 24,410 adults infected by the novel coronavirus (SARS-CoV-2; COVID-19): A systematic review and meta-analysis of 148 studies from 9 countries.

BACKGROUND: To limit the spread of SARS-CoV-2, an evidence-based understanding of the symptoms is critical to inform guidelines for quarantining and testing. The most common features are purported to be fever and a new persistent cough, although the global prevalence of these symptoms remains unclear. The aim of this systematic review is to determine the prevalence of symptoms associated with COVID-19 worldwide. METHODS: We searched PubMed, Embase, CINAHL, AMED, medRxiv and bioRxiv on 5th April 2020 for studies of adults (>16 years) with laboratory test confirmed COVID-19. No language or publication status restrictions were applied. Data were independently extracted by two review authors into standardised forms. All datapoints were independently checked by three other review authors. A random-effects model for pooling of binomial data was applied to estimate the prevalence of symptoms, subgrouping estimates by country. I2 was used to assess inter-study heterogeneity. RESULTS: Of 851 unique citations, 148 articles were included which comprised 24,410 adults with confirmed COVID-19 from 9 countries. The most prevalent symptoms were fever (78% [95% CI 75%-81%]; 138 studies, 21,701 patients; I2 94%), a cough (57% [95% CI 54%-60%]; 138 studies, 21,682 patients; I2 94%) and fatigue (31% [95% CI 27%-35%]; 78 studies, 13,385 patients; I2 95%). Overall, 19% of hospitalised patients required non-invasive ventilation (44 studies, 6,513 patients), 17% required intensive care (33 studies, 7504 patients), 9% required invasive ventilation (45 studies, 6933 patients) and 2% required extra-corporeal membrane oxygenation (12 studies, 1,486 patients). The mortality rate was 7% (73 studies, 10,402 patients). CONCLUSIONS: We confirm that fever and cough are the most prevalent symptoms of adults infected by SARS-CoV-2. However, there is a large proportion of infected adults which symptoms-alone do not identify.

Display evaluation for primary diagnosis using digital pathology.

Purpose: As pathology departments around the world contemplate digital microscopy for primary diagnosis, making an informed choice regarding display procurement is very challenging in the absence of defined minimum standards. In order to help inform the decision, we aimed to conduct an evaluation of displays with a range of technical specifications and sizes. Approach: We invited histopathologists within our institution to take part in a survey evaluation of eight short-listed displays. Pathologists reviewed a single haematoxylin and eosin whole slide image of a benign nevus on each display and gave a single score to indicate their preference in terms of image quality and size of the display. Results: Thirty-four pathologists took part in the display evaluation experiment. The preferred display was the largest and had the highest technical specifications (11.8-MP resolution, 2100 cd / m 2 maximum luminance). The least preferred display had the lowest technical specifications (2.3-MP resolution, 300 cd / m 2 maximum luminance). A trend was observed toward an increased preference for displays with increased luminance and resolution. Conclusions: This experiment demonstrates a preference for large medical-grade displays with the high luminance and high resolution. As cost becomes implicated in procurement, significantly less expensive medical-grade displays with slightly lower technical specifications may be the most cost-effective option.

Development and Evaluation of a Novel Point-of-Use Quality Assurance Tool for Digital Pathology.

CONTEXT.­: Flexible working at diverse or remote sites is a major advantage when reporting using digital pathology, but currently there is no method to validate the clinical diagnostic setting within digital microscopy. OBJECTIVE.­: To develop a preliminary Point-of-Use Quality Assurance (POUQA) tool designed specifically to validate the diagnostic setting for digital microscopy. DESIGN.­: We based the POUQA tool on the red, green, and blue (RGB) values of hematoxylin-eosin. The tool used 144 hematoxylin-eosin-colored, 5×5-cm patches with a superimposed random letter with subtly lighter RGB values from the background color, with differing levels of difficulty. We performed an initial evaluation across 3 phases within 2 pathology departments: 1 in the United Kingdom and 1 in Sweden. RESULTS.­: In total, 53 experiments were conducted across all phases resulting in 7632 test images viewed in all. Results indicated that the display, the user's visual system, and the environment each independently impacted performance. Performance was improved with reduction in natural light and through use of medical-grade displays. CONCLUSIONS.­: The use of a POUQA tool for digital microscopy is essential to afford flexible working while ensuring patient safety. The color-contrast test provides a standardized method of comparing diagnostic settings for digital microscopy. With further planned development, the color-contrast test may be used to create a "Verified Login" for diagnostic setting validation.

Community screening for visual impairment in older people.

BACKGROUND: Visual problems in older people are common and frequently under-reported. The effects of poor vision in older people are wide reaching and include falls, confusion and reduced quality of life. Much of the visual impairment in older ages can be treated (e.g. cataract surgery, correction of refractive error). Vision screening may therefore reduce the number of older people living with sight loss. OBJECTIVES: The objective of this review was to assess the effects on vision of community vision screening of older people for visual impairment. SEARCH METHODS: We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (which contains the Cochrane Eyes and Vision Trials Register) (2017, Issue 10); Ovid MEDLINE; Ovid Embase; the ISRCTN registry; ClinicalTrials.gov and the ICTRP. The date of the search was 23 November 2017. SELECTION CRITERIA: We included randomised controlled trials (RCTs) that compared vision screening alone or as part of a multi-component screening package as compared to no vision screening or standard care, on the vision of people aged 65 years or over in a community setting. We included trials that used self-reported visual problems or visual acuity testing as the screening tool. DATA COLLECTION AND ANALYSIS: We used standard methods expected by Cochrane. We graded the certainty of the evidence using GRADE. MAIN RESULTS: Visual outcome data were available for 10,608 people in 10 trials. Four trials took place in the UK, two in Australia, two in the United States and two in the Netherlands. Length of follow-up ranged from one to five years. Three of these studies were cluster-randomised trials whereby general practitioners or family physicians were randomly allocated to undertake vision screening or no vision screening. All studies were funded by government agencies. Overall we judged the studies to be at low risk of bias and only downgraded the certainty of the evidence (GRADE) for imprecision.Seven trials compared vision screening as part of a multi-component screening versus no screening. Six of these studies used self-reported vision as both screening tool and outcome measure, but did not directly measure vision. One study used a combination of self-reported vision and visual acuity measurement: participants reporting vision problems at screening were treated by the attending doctor, referred to an eye care specialist or given information about resources that were available to assist with poor vision. There was a similar risk of "not seeing well" at follow-up in people screened compared with people not screened in meta-analysis of six studies (risk ratio (RR) 1.05, 95% confidence interval (CI) 0.97 to 1.14, 4522 participants high-certainty evidence). One trial reported "improvement in vision" and this occurred slightly less frequently in the screened group (RR 0.85, 95% CI 0.52 to 1.40, 230 participants, moderate-certainty evidence).Two trials compared vision screening (visual acuity testing) alone with no vision screening. In one study, distance visual acuity was similar in the two groups at follow-up (mean difference (MD) 0.02 logMAR, 95% CI -0.02 to 0.05, 532 participants, high-certainty evidence). There was also little difference in near acuity (MD 0.02 logMAR, 95% CI -0.03 to 0.07, 532 participants, high-certainty evidence). There was no evidence of any important difference in quality of life (MD -0.06 National Eye Institute 25-item visual function questionnaire (VFQ-25) score adjusted for baseline VFQ-25 score, 95% CI -2.3 to 1.1, 532 participants, high-certainty evidence). The other study could not be included in the data analysis as the number of participants in each of the arms at follow-up could not be determined. However the authors stated that there was no significant difference in mean visual acuity in participants who had visual acuity assessed at baseline (39 letters) as compared to those who did not have their visual acuity assessed (35 letters, P = 0.25, 121 participants).One trial compared a detailed health assessment including measurement of visual acuity (intervention) with a brief health assessment including one question about vision (standard care). People given the detailed health assessment had a similar risk of visual impairment (visual acuity worse than 6/18 in either eye) at follow-up compared with people given the brief assessment (RR 1.07, 95% CI 0.84 to 1.36, 1807 participants, moderate-certainty evidence). The mean composite score of the VFQ-25 was 86.0 in the group that underwent visual acuity screening compared with 85.6 in the standard care group, a difference of 0.40 (95% CI -1.70 to 2.50, 1807 participants, high-certainty evidence). AUTHORS' CONCLUSIONS: The evidence from RCTs undertaken to date does not support vision screening for older people living independently in a community setting, whether in isolation or as part of a multi-component screening package. This is true for screening programmes involving questions about visual problems, or direct measurements of visual acuity.The most likely reason for this negative review is that the populations within the trials often did not take up the offered intervention as a result of the vision screening and large proportions of those who did not have vision screening appeared to seek their own intervention. Also, trials that use questions about vision have a lower sensitivity and specificity than formal visual acuity testing. Given the importance of visual impairment among older people, further research into strategies to improve vision of older people is needed. The effectiveness of an optimised primary care-based screening intervention that overcomes possible factors contributing to the observed lack of benefit in trials to date warrants assessment; trials should consider including more dependent participants, rather than those living independently in the community.

Colour in digital pathology: a review.

Colour is central to the practice of pathology because of the use of coloured histochemical and immunohistochemical stains to visualize tissue features. Our reliance upon histochemical stains and light microscopy has evolved alongside a wide variation in slide colour, with little investigation into the implications of colour variation. However, the introduction of the digital microscope and whole-slide imaging has highlighted the need for further understanding and control of colour. This is because the digitization process itself introduces further colour variation which may affect diagnosis, and image analysis algorithms often use colour or intensity measures to detect or measure tissue features. The US Food and Drug Administration have released recent guidance stating the need to develop a method of controlling colour reproduction throughout the digitization process in whole-slide imaging for primary diagnostic use. This comprehensive review introduces applied basic colour physics and colour interpretation by the human visual system, before discussing the importance of colour in pathology. The process of colour calibration and its application to pathology are also included, as well as a summary of the current guidelines and recommendations regarding colour in digital pathology.