Development and validation of artificial intelligence-based prescreening of large-bowel biopsies taken in the UK and Portugal: a retrospective cohort study.

BACKGROUND: Histopathological examination is a crucial step in the diagnosis and treatment of many major diseases. Aiming to facilitate diagnostic decision making and improve the workload of pathologists, we developed an artificial intelligence (AI)-based prescreening tool that analyses whole-slide images (WSIs) of large-bowel biopsies to identify typical, non-neoplastic, and neoplastic biopsies. METHODS: This retrospective cohort study was conducted with an internal development cohort of slides acquired from a hospital in the UK and three external validation cohorts of WSIs acquired from two hospitals in the UK and one clinical laboratory in Portugal. To learn the differential histological patterns from digitised WSIs of large-bowel biopsy slides, our proposed weakly supervised deep-learning model (Colorectal AI Model for Abnormality Detection [CAIMAN]) used slide-level diagnostic labels and no detailed cell or region-level annotations. The method was developed with an internal development cohort of 5054 biopsy slides from 2080 patients that were labelled with corresponding diagnostic categories assigned by pathologists. The three external validation cohorts, with a total of 1536 slides, were used for independent validation of CAIMAN. Each WSI was classified into one of three classes (ie, typical, atypical non-neoplastic, and atypical neoplastic). Prediction scores of image tiles were aggregated into three prediction scores for the whole slide, one for its likelihood of being typical, one for its likelihood of being non-neoplastic, and one for its likelihood of being neoplastic. The assessment of the external validation cohorts was conducted by the trained and frozen CAIMAN model. To evaluate model performance, we calculated area under the convex hull of the receiver operating characteristic curve (AUROC), area under the precision-recall curve, and specificity compared with our previously published iterative draw and rank sampling (IDaRS) algorithm. We also generated heat maps and saliency maps to analyse and visualise the relationship between the WSI diagnostic labels and spatial features of the tissue microenvironment. The main outcome of this study was the ability of CAIMAN to accurately identify typical and atypical WSIs of colon biopsies, which could potentially facilitate automatic removing of typical biopsies from the diagnostic workload in clinics. FINDINGS: A randomly selected subset of all large bowel biopsies was obtained between Jan 1, 2012, and Dec 31, 2017. The AI training, validation, and assessments were done between Jan 1, 2021, and Sept 30, 2022. WSIs with diagnostic labels were collected between Jan 1 and Sept 30, 2022. Our analysis showed no statistically significant differences across prediction scores from CAIMAN for typical and atypical classes based on anatomical sites of the biopsy. At 0·99 sensitivity, CAIMAN (specificity 0·5592) was more accurate than an IDaRS-based weakly supervised WSI-classification pipeline (0·4629) in identifying typical and atypical biopsies on cross-validation in the internal development cohort (p<0·0001). At 0·99 sensitivity, CAIMAN was also more accurate than IDaRS for two external validation cohorts (p<0·0001), but not for a third external validation cohort (p=0·10). CAIMAN provided higher specificity than IDaRS at some high-sensitivity thresholds (0·7763 vs 0·6222 for 0·95 sensitivity, 0·7126 vs 0·5407 for 0·97 sensitivity, and 0·5615 vs 0·3970 for 0·99 sensitivity on one of the external validation cohorts) and showed high classification performance in distinguishing between neoplastic biopsies (AUROC 0·9928, 95% CI 0·9927-0·9929), inflammatory biopsies (0·9658, 0·9655-0·9661), and atypical biopsies (0·9789, 0·9786-0·9792). On the three external validation cohorts, CAIMAN had AUROC values of 0·9431 (95% CI 0·9165-0·9697), 0·9576 (0·9568-0·9584), and 0·9636 (0·9615-0·9657) for the detection of atypical biopsies. Saliency maps supported the representation of disease heterogeneity in model predictions and its association with relevant histological features. INTERPRETATION: CAIMAN, with its high sensitivity in detecting atypical large-bowel biopsies, might be a promising improvement in clinical workflow efficiency and diagnostic decision making in prescreening of typical colorectal biopsies. FUNDING: The Pathology Image Data Lake for Analytics, Knowledge and Education Centre of Excellence; the UK Government's Industrial Strategy Challenge Fund; and Innovate UK on behalf of UK Research and Innovation.

Screening of normal endoscopic large bowel biopsies with interpretable graph learning: a retrospective study.

OBJECTIVE: To develop an interpretable artificial intelligence algorithm to rule out normal large bowel endoscopic biopsies, saving pathologist resources and helping with early diagnosis. DESIGN: A graph neural network was developed incorporating pathologist domain knowledge to classify 6591 whole-slides images (WSIs) of endoscopic large bowel biopsies from 3291 patients (approximately 54% female, 46% male) as normal or abnormal (non-neoplastic and neoplastic) using clinically driven interpretable features. One UK National Health Service (NHS) site was used for model training and internal validation. External validation was conducted on data from two other NHS sites and one Portuguese site. RESULTS: Model training and internal validation were performed on 5054 WSIs of 2080 patients resulting in an area under the curve-receiver operating characteristic (AUC-ROC) of 0.98 (SD=0.004) and AUC-precision-recall (PR) of 0.98 (SD=0.003). The performance of the model, named Interpretable Gland-Graphs using a Neural Aggregator (IGUANA), was consistent in testing over 1537 WSIs of 1211 patients from three independent external datasets with mean AUC-ROC=0.97 (SD=0.007) and AUC-PR=0.97 (SD=0.005). At a high sensitivity threshold of 99%, the proposed model can reduce the number of normal slides to be reviewed by a pathologist by approximately 55%. IGUANA also provides an explainable output highlighting potential abnormalities in a WSI in the form of a heatmap as well as numerical values associating the model prediction with various histological features. CONCLUSION: The model achieved consistently high accuracy showing its potential in optimising increasingly scarce pathologist resources. Explainable predictions can guide pathologists in their diagnostic decision-making and help boost their confidence in the algorithm, paving the way for its future clinical adoption.

Role of AI and digital pathology for colorectal immuno-oncology.

Immunotherapy deals with therapeutic interventions to arrest the progression of tumours using the immune system. These include checkpoint inhibitors, T-cell manipulation, cytokines, oncolytic viruses and tumour vaccines. In this paper, we present a survey of the latest developments on immunotherapy in colorectal cancer (CRC) and the role of artificial intelligence (AI) in this context. Among these, microsatellite instability (MSI) is perhaps the most popular IO biomarker globally. We first discuss the MSI status of tumours, its implications for patient management, and its relationship to immune response. In recent years, several aspiring studies have used AI to predict the MSI status of patients from digital whole-slide images (WSIs) of routine diagnostic slides. We present a survey of AI literature on the prediction of MSI and tumour mutation burden from digitised WSIs of haematoxylin and eosin-stained diagnostic slides. We discuss AI approaches in detail and elaborate their contributions, limitations and key takeaways to drive future research. We further expand this survey to other IO-related biomarkers like immune cell infiltrates and alternate data modalities like immunohistochemistry and gene expression. Finally, we underline possible future directions in immunotherapy for CRC and promise of AI to accelerate this exploration for patient benefits.

Biomarkers of Castrate Resistance in Prostate Cancer: Androgen Receptor Amplification and T877A Mutation Detection by Multiplex Droplet Digital PCR.

Androgen Receptor (AR) alterations (amplification, point mutations, and splice variants) are master players in metastatic castration resistant prostate cancer (CRPC) progression and central therapeutic targets for patient management. Here, we have developed two multiplexed droplet digital PCR (ddPCR) assays to detect AR copy number (CN) and the key point mutation T877A. Overcoming challenges of determining gene amplification from liquid biopsies, these assays cross-validate each other to produce reliable AR amplification and mutation data from plasma cell free DNA (cfDNA) of advanced prostate cancer (PC) patients. Analyzing a mixed PC patient cohort consisting of CRPC and hormone sensitive prostate cancer (HSPC) patients showed that 19% (9/47) patients had AR CN amplification. As expected, only CRPC patients were positive for AR amplification, while interestingly the T877A mutation was identified in two patients still considered HSPC at the time. The ddPCR based analysis of AR alterations in cfDNA is highly economic, feasible, and informative to provide biomarker detection that may help to decide on the best follow-up therapy for CRPC patients.

COVID-19 post-mortem findings: how the departed can teach us.

We present an asthmatic 39 year old female frontline healthcare worker in contact with COVID-19 patients, who suffered from respiratory problems for a few days, after which she was found dead by her spouse. Post-mortem (PM) examination showed lung consolidation and histological evidence of diffuse alveolar damage (DAD), in form of hyaline membrane disease, as well as atypical cells, thrombi and fibrin plugs inside pulmonary capillaries. Swab sample testing for SARS-CoV-2 confirmed the infection status. SARS-CoV-2 is a novel coronavirus which first emerged in Wuhan, China, and PM is contributing immensely to uncovering the pathophysiological mechanism of this disease. SARS-CoV-2 enters host cells via the angiotensin-converting enzyme 2 (ACE2), and can lead to a fatal pneumonia characterized histologically by DAD (the most consistent PM finding). This is due to invasion by the virus of type II pneumocytes, which leads to the death of both type I and II pneumocytes and increased capillary permeability. This compromises gas exchange which can lead eventually to cardiorespiratory failure and death.

The Prospect of Identifying Resistance Mechanisms for Castrate-Resistant Prostate Cancer Using Circulating Tumor Cells: Is Epithelial-to-Mesenchymal Transition a Key Player?

Prostate cancer (PCa) is initially driven by excessive androgen receptor (AR) signaling with androgen deprivation therapy (ADT) being a major therapeutic approach to its treatment. However, the development of drug resistance is a significant limitation on the effectiveness of both first-line and more recently developed second-line ADTs. There is a need then to study AR signaling within the context of other oncogenic signaling pathways that likely mediate this resistance. This review focuses on interactions between AR signaling, the well-known phosphatidylinositol-3-kinase/AKT pathway, and an emerging mediator of these pathways, the Hippo/YAP1 axis in metastatic castrate-resistant PCa, and their involvement in the regulation of epithelial-mesenchymal transition (EMT), a feature of disease progression and ADT resistance. Analysis of these pathways in circulating tumor cells (CTCs) may provide an opportunity to evaluate their utility as biomarkers and address their importance in the development of resistance to current ADT with potential to guide future therapies.

Knowledge of Pharmacists about Anti-epileptic Drugs in a Developing Country.

BACKGROUND: Pharmacists play an essential role in educating the epileptic patients about their disease and their medications. Improving the patient's awareness may lead to improve their compliance and decrease drug-drug interaction and ultimately improve their quality of life. OBJECTIVE: This study aimed to assess the pharmacist's knowledge about anti-epileptic drugs in Khartoum State, Sudan. METHODS: We conducted a descriptive cross-sectional study in Khartoum State, Sudan. Proportionate stratified sampling was used to determine the targeted Pharmacies, and all pharmacists who were present in the selected pharmacy at the time of data collection and fulfilled our selection criteria were included in the study. A structure closed-ended questionnaire was used to collect quantitative data from candidates. RESULTS: Majority of participants were female (66.9%), less than 30 years old (66.7%) and have less than 5 years of experience (62.1%). Unfortunately, the majority of the participants (85.3%) had poor knowledge, and only (14.7%) of them had good knowledge. Furthermore, the study revealed that age (p =.030), years of experience (p =.026) and the degree in pharmacy (p = .003) were significantly associated with knowledge level. CONCLUSION: Majority of the pharmacists in Khartoum State have poor knowledge about anti-epileptic drugs. Further research is needed to investigate the actual factors behind this knowledge gap and to propose interventions to improve the pharmacist's knowledge and practice aiming to improve the quality of health care provided to the patients.

Effects of school-based health education on attitudes of female students towards female genital mutilation in Sudan.

BACKGROUND: The practice of female genital mutilation (FGM) is widespread in Sudan. Over the years, the government, civil society and the international community implemented multiple interventions to address the issue. However, due to a number of cultural and educational factors, this harmful practice continues. AIMS: This study aimed to assess the effects of a secondary school-based health education intervention on the knowledge and attitude of female students towards FGM in Sudan. METHODS: We conducted a quasi-experimental study in Karary Locality, Khartoum State, Sudan. A multistage sampling technique was used to determine targeted schools. Within the schools, students of two randomly selected classes received the intervention. The study included three phases; in the pre-intervention phase, data were collected from the totality of students (154 students) using a pre-tested questionnaire, after which students received health education sessions. The same questionnaire was used to re-collect the data in a post-intervention phase 6 weeks later. RESULTS: The participants were between 14 to 17 years old, 30.3% of which were subjected to FGM. The main source of information about FGM was family and friends (41.1%). The majority of participants had a negative attitude towards FGM. The means of knowledge and attitude scores increased from 8.63 (SD=2.562) and 5.76 pre-intervention (SD=1.937) to 11.99 (SD=2.264) and 6.53 post-intervention (SD=1.164), respectively. CONCLUSIONS: School-based health education has a positive impact on both knowledge and attitude of female students towards FGM in Sudan. As such, introducing health education about the complications of FGM in curricula of secondary schools in Sudan has the potential to improve students' knowledge and attitude. Ultimately, such interventions can help reduce the prevalence of the practice when students become responsible for future families.

Detection of AR-V7 in Liquid Biopsies of Castrate Resistant Prostate Cancer Patients: A Comparison of AR-V7 Analysis in Circulating Tumor Cells, Circulating Tumor RNA and Exosomes.

Detection of androgen receptor (AR) variant 7 (AR-V7) is emerging as a clinically important biomarker in castrate resistant prostate cancer (CRPC). Detection is possible from tumor tissue, which is often inaccessible in the advanced disease setting. With recent progress in detecting AR-V7 in circulating tumor cells (CTCs), circulating tumor RNA (ctRNA) and exosomes from prostate cancer patients, liquid biopsies have emerged as an alternative to tumor biopsy. Therefore, it is important to clarify whether these approaches differ in sensitivity in order to achieve the best possible biomarker characterization for the patient. In this study, blood samples from 44 prostate cancer patients were processed for CTCs and ctRNA with subsequent AR-V7 testing, while exosomal RNA was isolated from 16 samples and tested. Detection of AR and AR-V7 was performed using a highly sensitive droplet digital PCR-based assay. AR and AR-V7 RNA were detectable in CTCs, ctRNA and exosome samples. AR-V7 detection from CTCs showed higher sensitivity and has proven specificity compared to detection from ctRNA and exosomes. Considering that CTCs are almost always present in the advanced prostate cancer setting, CTC samples should be considered the liquid biopsy of choice for the detection of this clinically important biomarker.

Monoallelic characteristic-bearing heterozygous L1053X in BRCA2 gene among Sudanese women with breast cancer.

BACKGROUND: Breast cancer (BC) is the most common type of cancer in women. Among many risk factors of BC, mutations in BRCA2 gene were found to be the primary cause in 5-10% of cases. The majority of deleterious mutations are frameshift or nonsense mutations. Most of the reported BRCA2 mutations are protein truncating mutations. METHODS: The study aimed to describe the pattern of mutations including single nucleotide polymorphisms (SNPs) and variants of the BRCA2 (exon11) gene among Sudanese women patients diagnosed with BC. In this study a specific region of BRCA2 exon 11 was targeted using PCR and DNA sequencing. RESULTS: Early onset cases 25/45 (55.6%) were premenopausal women with a mean age of 36.6 years. Multiparity was more frequent within the study amounting to 30 cases (66.6%), with a mean parity of 4.1. Ductal type tumor was the predominant type detected in 22 cases (48.8%) among the reported histotypes. A heterozygous monoallelic nonsense mutation at nucleotide 3385 was found in four patients out of 9, where TTA codon was converted into the stop codon TGA. CONCLUSION: This study detected a monoallelic nonsense mutation in four Sudanese female patients diagnosed with early onset BC from different families. Further work is needed to demonstrate its usefulness in screening of BC.

In silico analysis of single nucleotide polymorphisms (SNPs) in human FOXC2 gene.

Introduction: Lymphedema is an abnormal accumulation of interstitial fluid, due to inefficient uptake and reduced flow, leading to swelling and disability, mostly in the extremities. Hereditary lymphedema usually occurs as an autosomal dominant trait with allelic heterogeneity. Methods: We identified single nucleotide polymorphisms (SNPs) in the FOXC2 gene using dbSNP, analyzed their effect on the resulting protein using VEP and Biomart, modelled the resulting protein using Project HOPE, identified gene - gene interactions using GeneMANIA and predicted miRNAs affected and the resulting effects of SNPs in the 5' and 3' regions using PolymiRTS. Results: We identified 473 SNPs - 429 were nsSNPs and 44 SNPs were in the 5' and 3' UTRs. In total, 2 SNPs - rs121909106 and rs121909107 - have deleterious effects on the resulting protein, and a 3D model confirmed those effects. The gene - gene interaction network showed the involvement of FOXC2 protein in the development of the lymphatic system. hsa-miR-6886-5p, hsa-miRS-6886-5p, hsa-miR-6720-3p, which were affected by the SNPs rs201118690, rs6413505, rs201914560, respectively, were the most important miRNAs affected, due to their high conservation score. Conclusions: rs121909106 and rs121909107 were predicted to have the most harmful effects, while hsa-miR-6886-5p, hsa-miR-6886-5p and hsa-miR-6720-3p were predicted to be the most important miRNAs affected. Computational biology tools have advantages and disadvantages, and the results they provide are predictions that require confirmation using methods such as functional studies.

Risk factors of diabetic cardiac autonomic neuropathy in patients with type 1 diabetes mellitus: a meta-analysis.

OBJECTIVES: We aimed to stratify the possible risk factors for diabetic cardiac autonomic neuropathy (CAN). METHODS: We did a meta-analysis of risk factors of CAN. We did a web-based search for literature in MEDLINE/PubMed, Scopus database and CENTRAL database up to August 2015. We included clinical trials or cohort studies that provide data about relationship between CAN and variables of interest. Our risk factors of interest were age, sex, duration of diabetes, body mass index (BMI), systolic blood pressure (sBP) and diastolic blood pressure (dBP), glycated haemoglobin (HbA1c), high-density lipoprotein and low-density lipoprotein (HDL and LDL), triglycerides, retinopathy and nephropathy. We generated Forest plots, χ2 test and I2 as tests for heterogeneity, risk ratio (RR), mean difference (MD), CIs and p values by ReVMan V.5.3 software. RESULTS: We found a total of 882 related items. We excluded 873 studies from the title and abstract and 4 studies after review of full reports. Four studies were included. Our meta-analysis showed significant association between CAN and age (MD=4.94 (3.46 to 6.42)), duration of diabetes (MD=4.51 (2.51 to 6.52)), HbA1c (MD=0.48 (0.28 to 0.67)), BMI (MD=0.55 (0.08 to 1.01)), serum triglycerides (MD=0.09 (0.01 to 0.17)), proliferative retinopathy (RR=3.69 (1.20 to 11.34)), microalbuminuria (RR=2.47 (1.43 to 4.29)), hypertension (RR=4.18 (2.52 to 6.91)) and sBP (MD=4.10 (2.20 to 6.00)). We neither discovered the absence of significant association between the development of CAN and male sex (RR=1.57 (0.45 to 5.39)), dBP (MD=0.89 (-0.36 to 2.14)), cholesterol level (MD=1.19 (-0.99 to 3.36)), LDL (MD=0.12 (-0.15 to 0.39)), nor HDL level (MD=-0.28 (-0.58 to 0.03)). CONCLUSIONS: Age, duration of diabetes, HbA1c, BMI, serum triglycerides, proliferative retinopathy, microalbuminuria, hypertension and sBP are directly related to the risk of development of diabetic CAN.