A deep learning approach to case prioritisation of colorectal biopsies.

AIMS: To create and validate a weakly supervised artificial intelligence (AI) model for detection of abnormal colorectal histology, including dysplasia and cancer, and prioritise biopsies according to clinical significance (severity of diagnosis). MATERIALS AND METHODS: Triagnexia Colorectal, a weakly supervised deep learning model, was developed for the classification of colorectal samples from haematoxylin and eosin (H&E)-stained whole slide images. The model was trained on 24 983 digitised images and assessed by multiple pathologists in a simulated digital pathology environment. The AI application was implemented as part of a point and click graphical user interface to streamline decision-making. Pathologists assessed the accuracy of the AI tool, its value, ease of use and integration into the digital pathology workflow. RESULTS: Validation of the model was conducted on two cohorts: the first, on 100 single-slide cases, achieved micro-average model specificity of 0.984, micro-average model sensitivity of 0.949 and micro-average model F1 score of 0.949 across all classes. A secondary multi-institutional validation cohort, of 101 single-slide cases, achieved micro-average model specificity of 0.978, micro-average model sensitivity of 0.931 and micro-average model F1 score of 0.931 across all classes. Pathologists reflected their positive impressions on the overall accuracy of the AI in detecting colorectal pathology abnormalities. CONCLUSIONS: We have developed a high-performing colorectal biopsy AI triage model that can be integrated into a routine digital pathology workflow to assist pathologists in prioritising cases and identifying cases with dysplasia/cancer versus non-neoplastic biopsies.

Future of biomarker evaluation in the realm of artificial intelligence algorithms: application in improved therapeutic stratification of patients with breast and prostate cancer.

Clinical workflows in oncology depend on predictive and prognostic biomarkers. However, the growing number of complex biomarkers contributes to costly and delayed decision-making in routine oncology care and treatment. As cancer is expected to rank as the leading cause of death and the single most important barrier to increasing life expectancy in the 21st century, there is a major emphasis on precision medicine, particularly individualisation of treatment through better prediction of patient outcome. Over the past few years, both surgical and pathology specialties have suffered cutbacks and a low uptake of pathology specialists means a solution is required to enable high-throughput screening and personalised treatment in this area to alleviate bottlenecks. Digital imaging in pathology has undergone an exponential period of growth. Deep-learning (DL) platforms for hematoxylin and eosin (H&E) image analysis, with preliminary artificial intelligence (AI)-based grading capabilities of specimens, can evaluate image characteristics which may not be visually apparent to a pathologist and offer new possibilities for better modelling of disease appearance and possibly improve the prediction of disease stage and patient outcome. Although digital pathology and AI are still emerging areas, they are the critical components for advancing personalised medicine. Integration of transcriptomic analysis, clinical information and AI-based image analysis is yet an uncultivated field by which healthcare professionals can make improved treatment decisions in cancer. This short review describes the potential application of integrative AI in offering better detection, quantification, classification, prognosis and prediction of breast and prostate cancer and also highlights the utilisation of machine learning systems in biomarker evaluation.

Evaluation of the Use of Single- and Multi-Magnification Convolutional Neural Networks for the Determination and Quantitation of Lesions in Nonclinical Pathology Studies.

Digital pathology platforms with integrated artificial intelligence have the potential to increase the efficiency of the nonclinical pathologist's workflow through screening and prioritizing slides with lesions and highlighting areas with specific lesions for review. Herein, we describe the comparison of various single- and multi-magnification convolutional neural network (CNN) architectures to accelerate the detection of lesions in tissues. Different models were evaluated for defining performance characteristics and efficiency in accurately identifying lesions in 5 key rat organs (liver, kidney, heart, lung, and brain). Cohorts for liver and kidney were collected from TG-GATEs open-source repository, and heart, lung, and brain from internally selected R&D studies. Annotations were performed, and models were trained on each of the available lesion classes in the available organs. Various class-consolidation approaches were evaluated from generalized lesion detection to individual lesion detections. The relationship between the amount of annotated lesions and the precision/accuracy of model performance is elucidated. The utility of multi-magnification CNN implementations in specific tissue subtypes is also demonstrated. The use of these CNN-based models offers users the ability to apply generalized lesion detection to whole-slide images, with the potential to generate novel quantitative data that would not be possible with conventional image analysis techniques.

Using Deep Learning Artificial Intelligence Algorithms to Verify N-Nitroso-N-Methylurea and Urethane Positive Control Proliferative Changes in Tg-RasH2 Mouse Carcinogenicity Studies.

In Tg-rasH2 carcinogenicity mouse models, a positive control group is treated with a carcinogen such as urethane or N-nitroso-N-methylurea to test study validity based on the presence of the expected proliferative lesions in the transgenic mice. We hypothesized that artificial intelligence-based deep learning (DL) could provide decision support for the toxicologic pathologist by screening for the proliferative changes, verifying the expected pattern for the positive control groups. Whole slide images (WSIs) of the lungs, thymus, and stomach from positive control groups were used for supervised training of a convolutional neural network (CNN). A single pathologist annotated WSIs of normal and abnormal tissue regions for training the CNN-based supervised classifier using INHAND criteria. The algorithm was evaluated using a subset of tissue regions that were not used for training and then additional tissues were evaluated blindly by 2 independent pathologists. A binary output (proliferative classes present or not) from the pathologists was compared to that of the CNN classifier. The CNN model grouped proliferative lesion positive and negative animals at high concordance with the pathologists. This process simulated a workflow for review of these studies, whereby a DL algorithm could provide decision support for the pathologists in a nonclinical study.

Measurement of the IgM and IgG Autoantibody Immune Responses in Human Serum has High Predictive Value for the Presence of Colorectal Cancer.

INTRODUCTION: Colorectal cancer is a major public health issue, with incidences continuing to rise owing to the growing and aging world population. Current screening strategies for colorectal cancer diagnosis suffer from various limitations, including invasiveness and poor uptake. Consequently, there is an unmet clinical need for a minimally invasive, sensitive, and specific method for detecting the presence of colorectal cancer and pre-malignant lesions. PATIENTS AND METHODS: An indirect enzyme-linked immunosorbent assay was used to measure the primary (IgM) and secondary (IgG) adaptive humoral immune responses to a panel of previously identified cancer antigens in the sera of normal and adenoma samples, and sera from patients with colorectal cancer. RESULTS: An optimal panel of 7 biomarkers capable of identifying patients with colorectal cancer as distinct from both normal and adenoma samples is identified. The cumulative sensitivity and specificity of the assay are 70.8% and 86.5%, respectively. The positive and negative predictive values of the cohort are 77.3% and 82.1%. This assay was not able to accurately discriminate between normal and adenoma samples. Patients whose serum was positive for the presence of anti-ICLN IgM autoantibodies had a significantly poorer 5-year survival than patients whose serum was negative (P = .004). CONCLUSION: This study describes a novel minimally invasive enzyme-linked immunosorbent assay-based method, capable of identifying patients with colorectal cancer as distinct from both normal and adenoma samples. Patients are likely to be far more amenable to a blood-based test such as the one described herein, rather than a fecal-based test, likely leading to increased patient uptake.

Understanding microcystin-LR antibody binding interactions using in silico docking and in vitro mutagenesis.

Microcystins (MCs) are a group of highly potent cyanotoxins that are becoming more widely distributed due to increased global temperatures and climate change. Microcystin-leucine-arginine (MC-LR) is the most potent and most common variant, with a guideline limit of 1 µg/l in drinking water. We previously developed a novel avian single-chain fragment variable (scFv), designated 2G1, for use in an optical-planar waveguide detection system for microcystin determination. This current work investigates interactions between 2G1 and MC-LR at the molecular level through modelling with an avian antibody template and molecular docking by AutoDock Vina to identify key amino acid (AA) residues involved. These potential AA interactions were investigated in vitro by targeted mutagenesis, specifically, by alanine scanning mutations. Glutamic acid (E) was found to play a critical role in the 2G1-MC-LR binding interaction, with the heavy chain glutamic acid (E) 102 (H-E102) forming direct bonds with the arginine (R) residue of MC-LR. In addition, alanine mutation of light chain residue aspartic acid 57 (L-D57) led to an improvement in antigen-binding observed using enzyme-linked immunosorbent assay (ELISA), and was confirmed by surface plasmon resonance (SPR). This work will contribute to improving the binding of recombinant anti-MC-LR to its antigen and aid in the development of a higher sensitivity harmful algal toxin diagnostic.

Novel Microfluidic Analytical Sensing Platform for the Simultaneous Detection of Three Algal Toxins in Water.

Globally, the need for "on-site" algal-toxin monitoring has become increasingly urgent due to the amplified demand for fresh-water and for safe, "toxin-free" shellfish and fish stocks. Herein, we describe the first reported, Lab-On-A-Disc (LOAD) based-platform developed to detect microcystin levels in situ, with initial detectability of saxitoxin and domoic acid also reported. Using recombinant antibody technology, the LOAD platform combines immunofluorescence with centrifugally driven microfluidic liquid handling to achieve a next-generation disposable device capable of multianalyte sampling. A low-complexity "LED-photodiode" based optical sensing system was tailor-made for the platform, which allows the fluorescence signal of the toxin-specific reaction to be quantified. This system can rapidly and accurately detect the presence of microcystin-LR, domoic acid, and saxitoxin in 30 min, with a minimum of less than 5 min end-user interaction for maximum reproducibility. This method provides a robust "point of need" diagnostic alternative to the current laborious and costly methods used for qualitative toxin monitoring.

Immunoaffinity Chromatography: Concepts and Applications.

Antibody-based separation methods, such as immunoaffinity chromatography (IAC), are powerful purification and isolation techniques. Antibodies isolated using these techniques have proven highly efficient in applications ranging from clinical diagnostics to environmental monitoring. Immunoaffinity chromatography is an efficient antibody separation method which exploits the binding efficiency of a ligand to an antibody. Essential to the successful design of any IAC platform is the optimization of critical experimental parameters such as (a) the biological affinity pair, (b) the matrix support, (c) the immobilization coupling chemistry, and (d) the effective elution conditions. These elements and the practicalities of their use are discussed in detail in this review. At the core of all IAC platforms is the high affinity interactions between antibodies and their related ligands; hence, this review entails a brief introduction to the generation of antibodies for use in immunoaffinity chromatography and also provides specific examples of their potential applications.

Purification of Antibodies Using Affinity Chromatography.

Affinity chromatography permits the isolation of a target analyte from a complex mixture and can be utilized to purify proteins, carbohydrates, drugs, haptens, or any analyte of interest once an affinity pair is available. It involves the exploitation of specific interactions between a binding affinity pair, such as those between an antibody and its associated antigen, or between any ligand and its associated binding receptor/protein. With the discovery of protein A in 1970, and, subsequently protein G and L, immuno-affinity chromatography has grown in popularity and is now the standard methodology for the purification of antibodies which may be implemented for a selection of different applications such as immunodiagnostics. This chapter is designed to inform the researcher about the basic techniques involved in the affinity chromatography-based purification of monoclonal, polyclonal, and recombinant antibodies. Examples are provided for the use of protein A and G. In addition, tables are provided that allow the reader to select the most appropriate protein for use in the isolation of their antibody.

Protein engineering with artificial chemical nucleases.

Herein we report the application of oxidative artificial chemical nucleases as novel agents for protein engineering. The complex ion [Cu(Phen)2(H2O)](2+) (CuPhen; Phen = 1,10-phenanthroline) was applied under Fenton-type conditions against a recombinant antibody fragment specific for prostate-specific antigen (PSA) and compared against traditional DNA shuffling using DNase I for the generation of recombinant mutagenesis libraries. We show that digestion and re-annealment of single chain variable fragment (scFv) coding DNA is possible using CuPhen. Results indicate recombinant library generation in this manner may generate novel clones­not accessible through the use of DNase I­with CuPhen producing highly PSA-specific binding antibodies identified by surface plasmon resonance.

Diagnostic potential of zinc finger protein-specific autoantibodies and associated linear B-cell epitopes in colorectal cancer.

Colorectal cancer is one of the most common cancers worldwide with almost 700,000 deaths every year. Detection of colorectal cancer at an early stage significantly improves patient survival. Cancer-specific autoantibodies found in sera of cancer patients can be used for pre-symptomatic detection of the disease. In this study we assess the zinc finger proteins ZNF346, ZNF638, ZNF700 and ZNF768 as capture antigens for the detection of autoantibodies in colorectal cancer. Sera from 96 patients with colorectal cancer and 35 control patients with no evidence of cancer on colonoscopy were analysed for the presence of ZNF-specific autoantibodies using an indirect ELISA. Autoantibodies to individual ZNF proteins were detected in 10-20% of colorectal cancer patients and in 0-5.7% of controls. A panel of all four ZNF proteins resulted in an assay specificity of 91.4% and sensitivity of 41.7% for the detection of cancer patients in a cohort of non-cancer controls and colorectal cancer patients. Clinicopathological and survival analysis revealed that ZNF autoantibodies were independent of disease stage and did not correlate with disease outcome. Since ZNF autoantibodies were shared between patients and corresponding ZNF proteins showed similarities in their zinc finger motifs, we performed an in silico epitope sequence analysis. Zinc finger proteins ZNF700 and ZNF768 showed the highest sequence similarity with a bl2seq score of 262 (E-value 1E-81) and their classical C2H2 ZNF motifs were identified as potential epitopes contributing to their elevated immunogenic potential. Our findings show an enhanced and specific immunogenicity to zinc finger proteins, thereby providing a multiplexed autoantibody assay for minimally invasive detection of colorectal cancer.

Rapid Simultaneous Detection of Anti-protozoan Drugs Using a Lateral-Flow Immunoassay Format.

This research describes the development of a multi-analyte lateral-flow immunoassay intended for the simultaneous detection of three anti-protozoan drugs (coccidiostats). These drugs, namely, halofuginone, toltrazuril and diclazuril, are used in the treatment of Eimeria spp. infections in cattle, pigs, chickens and turkeys. Coloured carboxylated microspheres were coated with each of the detection antibodies and employed in a lateral-flow assay format for detection of these residues in eggs. Using this approach, halofuginone was detectable at a limit of 10 ng/mL or greater, toltrazuril at 100 ng/mL and, similarly, diclazuril had a detection limit of 100 ng/mL, which is below the maximum allowed residue limit for all three as outlined by EU regulation. This simple cost-efficient assay and analysis method could pave the way for more efficient simultaneous monitoring of small-molecule residues in the future.

Immunoaffinity chromatography.

Antibody-based separation methods, such as immunoaffinity chromatography (IAC), are powerful purification and isolation techniques. Antibodies isolated using these techniques have proven highly efficient in applications ranging from clinical diagnostics to environmental monitoring. IAC is an efficient antibody separation method which exploits the binding efficiency of a ligand to an antibody. Essential to the successful design of any IAC platform is the optimisation of critical experimental parameters such as: (a) the biological affinity pair, (b) the matrix support, (c) the immobilisation coupling chemistry, and (d) the effective elution conditions. These elements and the practicalities of their use are discussed in detail in this review. At the core of all IAC platforms is the high-affinity interactions between antibodies and their related ligands; hence, this review entails a brief introduction to the generation of antibodies for use in IAC and also provides specific examples of their potential applications.

Purification of antibodies using affinity chromatography.

Affinity chromatography permits the isolation of a target analyte from a complex mixture and can be utilised to purify proteins, carbohydrates, drugs, haptens, or any analyte of interest once an affinity pair is available. It involves the exploitation of specific interactions between a binding affinity pair, such as those between an antibody and its associated antigen, or between any ligand and its associated binding receptor/protein. With the discovery of protein A in 1970, and, subsequently proteins G and L, immuno-affinity chromatography has grown in popularity and is now the standard methodology for the purification of antibodies which may be implemented for a selection of different applications such as immunodiagnostics. This chapter is designed to inform the researcher about the basic techniques involved in the affinity chromatography-based purification of monoclonal, polyclonal, and recombinant antibodies. Examples are provided for the use of proteins A and G. In addition, tables are provided that allow the reader to select the most appropriate protein for use in the isolation of their antibody.

Light-chain shuffling from an antigen-biased phage pool allows 185-fold improvement of an anti-halofuginone single-chain variable fragment.

Halofuginone is an antiprotozoal drug used in the treatment of coccidiosis in poultry, a contagious enteric disease caused by parasites of the Eimeria spp. To ensure that food is free from any halofuginone residues and safe for human consumption, a rapid method to detect these residues below the maximum residue limits (MRLs) in a variety of matrices is necessary. To address this need, we constructed an immune single-chain variable fragment (scFv) library from the RNA of a halofuginone-immunized chicken and selected halofuginone-specific scFv by phage display. The best clone isolated from the library had a limit of detection of 30 ng/ml as determined by enzyme-linked immunosorbent assay (ELISA). However, the minimum MRL for halofuginone in certain foodstuffs can be as low as 1 ng/ml, well below the sensitivity of the selected antibody. The selected antibody was then affinity maturated by light-chain shuffling to further improve the antibody's assay performance. The halofuginone-specific heavy-chain pool of the biopanned library was assembled with the light-chain repertoire amplified from the original prepanned library. This resulted in a heavy-chain-biased library from which an scFv with the potential to detect halofuginone residues as low as 80 pg/ml was isolated, a 185-fold improvement over the original scFv. This new chain-shuffled scFv was incorporated into a validated ELISA (according to Commission Regulation 2002/657/EC) for the sensitive detection of halofuginone in spiked processed egg samples.