A Novel and Efficient Digital Pathology Classifier for Predicting Cancer Biomarkers Using Sequencer Architecture.

In digital pathology tasks, transformers have achieved state-of-the-art results, surpassing convolutional neural networks (CNNs). However, transformers are usually complex and resource intensive. This study developed a novel and efficient digital pathology classifier called DPSeq to predict cancer biomarkers through fine-tuning a sequencer architecture integrating horizontal and vertical bidirectional long short-term memory networks. Using hematoxylin and eosin-stained histopathologic images of colorectal cancer from two international data sets (The Cancer Genome Atlas and Molecular and Cellular Oncology), the predictive performance of DPSeq was evaluated in a series of experiments. DPSeq demonstrated exceptional performance for predicting key biomarkers in colorectal cancer (microsatellite instability status, hypermutation, CpG island methylator phenotype status, BRAF mutation, TP53 mutation, and chromosomal instability), outperforming most published state-of-the-art classifiers in a within-cohort internal validation and a cross-cohort external validation. In addition, under the same experimental conditions using the same set of training and testing data sets, DPSeq surpassed four CNNs (ResNet18, ResNet50, MobileNetV2, and EfficientNet) and two transformer (Vision Transformer and Swin Transformer) models, achieving the highest area under the receiver operating characteristic curve and area under the precision-recall curve values in predicting microsatellite instability status, BRAF mutation, and CpG island methylator phenotype status. Furthermore, DPSeq required less time for both training and prediction because of its simple architecture. Therefore, DPSeq appears to be the preferred choice over transformer and CNN models for predicting cancer biomarkers.

Characteristics and outcomes of COVID-19 patients with and without asthma from the United States, South Korea, and Europe.

Objective: Large international comparisons describing the clinical characteristics of patients with COVID-19 are limited. The aim of the study was to perform a large-scale descriptive characterization of COVID-19 patients with asthma.Methods: We included nine databases contributing data from January to June 2020 from the US, South Korea (KR), Spain, UK and the Netherlands. We defined two cohorts of COVID-19 patients ('diagnosed' and 'hospitalized') based on COVID-19 disease codes. We followed patients from COVID-19 index date to 30 days or death. We performed descriptive analysis and reported the frequency of characteristics and outcomes in people with asthma defined by codes and prescriptions.Results: The diagnosed and hospitalized cohorts contained 666,933 and 159,552 COVID-19 patients respectively. Exacerbation in people with asthma was recorded in 1.6-8.6% of patients at presentation. Asthma prevalence ranged from 6.2% (95% CI 5.7-6.8) to 18.5% (95% CI 18.2-18.8) in the diagnosed cohort and 5.2% (95% CI 4.0-6.8) to 20.5% (95% CI 18.6-22.6) in the hospitalized cohort. Asthma patients with COVID-19 had high prevalence of comorbidity including hypertension, heart disease, diabetes and obesity. Mortality ranged from 2.1% (95% CI 1.8-2.4) to 16.9% (95% CI 13.8-20.5) and similar or lower compared to COVID-19 patients without asthma. Acute respiratory distress syndrome occurred in 15-30% of hospitalized COVID-19 asthma patients.Conclusion: The prevalence of asthma among COVID-19 patients varies internationally. Asthma patients with COVID-19 have high comorbidity. The prevalence of asthma exacerbation at presentation was low. Whilst mortality was similar among COVID-19 patients with and without asthma, this could be confounded by differences in clinical characteristics. Further research could help identify high-risk asthma patients.[Box: see text]Supplemental data for this article is available online at https://doi.org/10.1080/02770903.2021.2025392 .

OpenDeID Pipeline for Unstructured Electronic Health Record Text Notes Based on Rules and Transformers: Deidentification Algorithm Development and Validation Study.

BACKGROUND: Electronic health records (EHRs) in unstructured formats are valuable sources of information for research in both the clinical and biomedical domains. However, before such records can be used for research purposes, sensitive health information (SHI) must be removed in several cases to protect patient privacy. Rule-based and machine learning-based methods have been shown to be effective in deidentification. However, very few studies investigated the combination of transformer-based language models and rules. OBJECTIVE: The objective of this study is to develop a hybrid deidentification pipeline for Australian EHR text notes using rules and transformers. The study also aims to investigate the impact of pretrained word embedding and transformer-based language models. METHODS: In this study, we present a hybrid deidentification pipeline called OpenDeID, which is developed using an Australian multicenter EHR-based corpus called OpenDeID Corpus. The OpenDeID corpus consists of 2100 pathology reports with 38,414 SHI entities from 1833 patients. The OpenDeID pipeline incorporates a hybrid approach of associative rules, supervised deep learning, and pretrained language models. RESULTS: The OpenDeID achieved a best F1-score of 0.9659 by fine-tuning the Discharge Summary BioBERT model and incorporating various preprocessing and postprocessing rules. The OpenDeID pipeline has been deployed at a large tertiary teaching hospital and has processed over 8000 unstructured EHR text notes in real time. CONCLUSIONS: The OpenDeID pipeline is a hybrid deidentification pipeline to deidentify SHI entities in unstructured EHR text notes. The pipeline has been evaluated on a large multicenter corpus. External validation will be undertaken as part of our future work to evaluate the effectiveness of the OpenDeID pipeline.

Seek COVER: using a disease proxy to rapidly develop and validate a personalized risk calculator for COVID-19 outcomes in an international network.

BACKGROUND: We investigated whether we could use influenza data to develop prediction models for COVID-19 to increase the speed at which prediction models can reliably be developed and validated early in a pandemic. We developed COVID-19 Estimated Risk (COVER) scores that quantify a patient's risk of hospital admission with pneumonia (COVER-H), hospitalization with pneumonia requiring intensive services or death (COVER-I), or fatality (COVER-F) in the 30-days following COVID-19 diagnosis using historical data from patients with influenza or flu-like symptoms and tested this in COVID-19 patients. METHODS: We analyzed a federated network of electronic medical records and administrative claims data from 14 data sources and 6 countries containing data collected on or before 4/27/2020. We used a 2-step process to develop 3 scores using historical data from patients with influenza or flu-like symptoms any time prior to 2020. The first step was to create a data-driven model using LASSO regularized logistic regression, the covariates of which were used to develop aggregate covariates for the second step where the COVER scores were developed using a smaller set of features. These 3 COVER scores were then externally validated on patients with 1) influenza or flu-like symptoms and 2) confirmed or suspected COVID-19 diagnosis across 5 databases from South Korea, Spain, and the United States. Outcomes included i) hospitalization with pneumonia, ii) hospitalization with pneumonia requiring intensive services or death, and iii) death in the 30 days after index date. RESULTS: Overall, 44,507 COVID-19 patients were included for model validation. We identified 7 predictors (history of cancer, chronic obstructive pulmonary disease, diabetes, heart disease, hypertension, hyperlipidemia, kidney disease) which combined with age and sex discriminated which patients would experience any of our three outcomes. The models achieved good performance in influenza and COVID-19 cohorts. For COVID-19 the AUC ranges were, COVER-H: 0.69-0.81, COVER-I: 0.73-0.91, and COVER-F: 0.72-0.90. Calibration varied across the validations with some of the COVID-19 validations being less well calibrated than the influenza validations. CONCLUSIONS: This research demonstrated the utility of using a proxy disease to develop a prediction model. The 3 COVER models with 9-predictors that were developed using influenza data perform well for COVID-19 patients for predicting hospitalization, intensive services, and fatality. The scores showed good discriminatory performance which transferred well to the COVID-19 population. There was some miscalibration in the COVID-19 validations, which is potentially due to the difference in symptom severity between the two diseases. A possible solution for this is to recalibrate the models in each location before use.

Colorectal Cancer Survival Prediction Using Deep Distribution Based Multiple-Instance Learning.

Most deep-learning algorithms that use Hematoxylin- and Eosin-stained whole slide images (WSIs) to predict cancer survival incorporate image patches either with the highest scores or a combination of both the highest and lowest scores. In this study, we hypothesize that incorporating wholistic patch information can predict colorectal cancer (CRC) cancer survival more accurately. As such, we developed a distribution-based multiple-instance survival learning algorithm (DeepDisMISL) to validate this hypothesis on two large international CRC WSIs datasets called MCO CRC and TCGA COAD-READ. Our results suggest that combining patches that are scored based on percentile distributions together with the patches that are scored as highest and lowest drastically improves the performance of CRC survival prediction. Including multiple neighborhood instances around each selected distribution location (e.g., percentiles) could further improve the prediction. DeepDisMISL demonstrated superior predictive ability compared to other recently published, state-of-the-art algorithms. Furthermore, DeepDisMISL is interpretable and can assist clinicians in understanding the relationship between cancer morphological phenotypes and a patient's cancer survival risk.

Statin Prescription Patterns and Associations with Subclinical Inflammation.

Background and Objectives: Statins have been extensively utilised in atherosclerotic cardiovascular disease (ASCVD) prevention and can inhibit inflammation. However, the association between statin therapy, subclinical inflammation and associated health outcomes is poorly understood in the primary care setting. Materials and Methods: Primary care electronic health record (EHR) data from the electronic Practice-Based Research Network (ePBRN) from 2012−2019 was used to assess statin usage and adherence in South-Western Sydney (SWS), Australia. Independent determinants of elevated C-reactive protein (CRP) were determined. The relationship between baseline CRP levels and hospitalisation rates at 12 months was investigated. Results: The prevalence of lipid-lowering medications was 14.0% in all adults and 44.6% in the elderly (≥65 years). The prevalence increased from 2012 to 2019 despite a drop in statin use between 2013−2015. A total of 55% of individuals had good adherence (>80%). Hydrophilic statin use and higher intensity statin therapy were associated with elevated CRP levels. However, elevated CRP levels were not associated with all-cause or ASCVD hospitalisations after adjusting for confounders. Conclusions: The prevalence and adherence patterns associated with lipid-lowering medications highlighted the elevated ASCVD-related burden in the SWS population, especially when compared with the Australian general population. Patients in SWS may benefit from enhanced screening protocols, targeted health literacy and promotion campaigns, and timely incorporation of evidence into ASCVD clinical guidelines. This study, which used EHR data, did not support the use of CRP as an independent marker of future short-term hospitalisations.

A Machine Learning Model to Predict Length of Stay and Mortality among Diabetes and Hypertension Inpatients.

Background and Objectives: Taiwan is among the nations with the highest rates of Type 2 Diabetes Mellitus (T2DM) and Hypertension (HTN). As more cases are reported each year, there is a rise in hospital admissions for people seeking medical attention. This creates a burden on hospitals and affects the overall management and administration of the hospitals. Hence, this study aimed to develop a machine learning (ML) model to predict the Length of Stay (LoS) and mortality among T2DM and HTN inpatients. Materials and Methods: Using Taiwan's National Health Insurance Research Database (NHIRD), this cohort study consisted of 58,618 patients, where 25,868 had T2DM, 32,750 had HTN, and 6419 had both T2DM and HTN. We analyzed the data with different machine learning models for the prediction of LoS and mortality. The evaluation was done by plotting descriptive statistical graphs, feature importance, precision-recall curve, accuracy plots, and AUC. The training and testing data were set at a ratio of 8:2 before applying ML algorithms. Results: XGBoost showed the best performance in predicting LoS (R2 0.633; RMSE 0.386; MAE 0.123), and RF resulted in a slightly lower performance (R2 0.591; RMSE 0.401; MAE 0.027). Logistic Regression (LoR) performed the best in predicting mortality (CV Score 0.9779; Test Score 0.9728; Precision 0.9432; Recall 0.9786; AUC 0.97 and AUPR 0.93), closely followed by Ridge Classifier (CV Score 0.9736; Test Score 0.9692; Precision 0.9312; Recall 0.9463; AUC 0.94 and AUPR 0.89). Conclusions: We developed a robust prediction model for LoS and mortality of T2DM and HTN inpatients. Linear Regression showed the best performance for LoS, and Logistic Regression performed the best in predicting mortality. The results showed that ML algorithms can not only help healthcare professionals in data-driven decision-making but can also facilitate early intervention and resource planning.

COVID-19 in patients with autoimmune diseases: characteristics and outcomes in a multinational network of cohorts across three countries.

OBJECTIVE: Patients with autoimmune diseases were advised to shield to avoid coronavirus disease 2019 (COVID-19), but information on their prognosis is lacking. We characterized 30-day outcomes and mortality after hospitalization with COVID-19 among patients with prevalent autoimmune diseases, and compared outcomes after hospital admissions among similar patients with seasonal influenza. METHODS: A multinational network cohort study was conducted using electronic health records data from Columbia University Irving Medical Center [USA, Optum (USA), Department of Veterans Affairs (USA), Information System for Research in Primary Care-Hospitalization Linked Data (Spain) and claims data from IQVIA Open Claims (USA) and Health Insurance and Review Assessment (South Korea). All patients with prevalent autoimmune diseases, diagnosed and/or hospitalized between January and June 2020 with COVID-19, and similar patients hospitalized with influenza in 2017-18 were included. Outcomes were death and complications within 30 days of hospitalization. RESULTS: We studied 133 589 patients diagnosed and 48 418 hospitalized with COVID-19 with prevalent autoimmune diseases. Most patients were female, aged ≥50 years with previous comorbidities. The prevalence of hypertension (45.5-93.2%), chronic kidney disease (14.0-52.7%) and heart disease (29.0-83.8%) was higher in hospitalized vs diagnosed patients with COVID-19. Compared with 70 660 hospitalized with influenza, those admitted with COVID-19 had more respiratory complications including pneumonia and acute respiratory distress syndrome, and higher 30-day mortality (2.2-4.3% vs 6.32-24.6%). CONCLUSION: Compared with influenza, COVID-19 is a more severe disease, leading to more complications and higher mortality.

Cohort selection for clinical trials using multiple instance learning.

Identifying patients eligible for clinical trials using electronic health records (EHRs) is a challenging task usually requiring a comprehensive analysis of information stored in multiple EHRs of a patient. The goal of this study is to investigate different methods and their effectiveness in identifying patients that meet specific eligibility selection criteria based on patients' longitudinal records. An unstructured dataset released by the n2c2 cohort selection for clinical trials track was used, each of which included 2-5 records manually annotated to thirteen pre-defined selection criteria. Unlike the other studies, we formulated the problem as a multiple instance learning (MIL) task and compared the performance with that of the rule-based and the single instance-based classifiers. Our official best run achieved an average micro-F score of 0.8765 which was ranked as one of the top ten results in the track. Further experiments demonstrated that the performance of the MIL-based classifiers consistently yield better performance than their single-instance counterparts in the criteria that require the overall comprehension of the information distributed among all of the patient's EHRs. Rule-based and single instance learning approaches exhibited better performance in criteria that don't require a consideration of several factors across records. This study demonstrated that cohort selection using longitudinal patient records can be formulated as a MIL problem. Our results exhibit that the MIL-based classifiers supplement the rule-based methods and provide better results in comparison to the single instance learning approaches.

Enhanced functionalities for annotating and indexing clinical text with the NCBO Annotator.

Summary: Second use of clinical data commonly involves annotating biomedical text with terminologies and ontologies. The National Center for Biomedical Ontology Annotator is a frequently used annotation service, originally designed for biomedical data, but not very suitable for clinical text annotation. In order to add new functionalities to the NCBO Annotator without hosting or modifying the original Web service, we have designed a proxy architecture that enables seamless extensions by pre-processing of the input text and parameters, and post processing of the annotations. We have then implemented enhanced functionalities for annotating and indexing free text such as: scoring, detection of context (negation, experiencer, temporality), new output formats and coarse-grained concept recognition (with UMLS Semantic Groups). In this paper, we present the NCBO Annotator+, a Web service which incorporates these new functionalities as well as a small set of evaluation results for concept recognition and clinical context detection on two standard evaluation tasks (Clef eHealth 2017, SemEval 2014). Availability and implementation: The Annotator+ has been successfully integrated into the SIFR BioPortal platform-an implementation of NCBO BioPortal for French biomedical terminologies and ontologies-to annotate English text. A Web user interface is available for testing and ontology selection (http://bioportal.lirmm.fr/ncbo_annotatorplus); however the Annotator+ is meant to be used through the Web service application programming interface (http://services.bioportal.lirmm.fr/ncbo_annotatorplus). The code is openly available, and we also provide a Docker packaging to enable easy local deployment to process sensitive (e.g. clinical) data in-house (https://github.com/sifrproject). Contact: andon.tchechmedjiev@lirmm.fr. Supplementary information: Supplementary data are available at Bioinformatics online.

Statistical supervised meta-ensemble algorithm for medical record linkage.

Identifying unique patients across multiple care facilities or services is a major challenge in providing continuous care and undertaking health research. Identifying and linking patients without compromising privacy and security is an emerging issue in the big data era. The large quantity and complexity of the patient data emphasize the need for effective linkage methods that are both scalable and accurate. In this study, we aim to develop and evaluate an ensemble classification method using the three most typically used supervised learning methods, namely support vector machines, logistic regression and standard feed-forward neural networks, to link records that belong to the same patient across multiple service locations. Our ensemble method is the combination of bagging and stacking. Each base learner's critical hyperparameters were selected through grid search technique. Two synthetic datasets were used in this study namely FEBRL and ePBRN. ePBRN linkage dataset was based on linkage errors noticed in the Australian primary care setting. The overall linkage performance was determined by assessing the blocking performance and classification performance. Our ensemble method outperformed the base learners in all evaluation metrics on one dataset. More specifically, the precision, which is average of individual precision scores in case of base learners increased from 90.70% to 94.85% in FEBRL, and from 62.17% to 99.28% in ePBRN. Similarly, the F-score increased from 94.92% to 98.18% in FEBRL, and from 72.99% to 91.72% in ePBRN. Our experiments suggest that we can significantly improve the linkage performance of individual algorithms by employing ensemble strategies.

Exploring associations of clinical and social parameters with violent behaviors among psychiatric patients.

Evidence has revealed interesting associations of clinical and social parameters with violent behaviors of patients with psychiatric disorders. Men are more violent preceding and during hospitalization, whereas women are more violent than men throughout the 3days following a hospital admission. It has also been proven that mental disorders may be a consistent risk factor for the occurrence of violence. In order to better understand violent behaviors of patients with psychiatric disorders, it is important to investigate both the clinical symptoms and psychosocial factors that accompany violence in these patients. In this study, we utilized a dataset released by the Partners Healthcare and Neuropsychiatric Genome-scale and RDoC Individualized Domains project of Harvard Medical School to develop a unique text mining pipeline that processes unstructured clinical data in order to recognize clinical and social parameters such asage, gender, history of alcohol use, and violent behaviors, and explored the associations between these parameters and violent behaviors of patients with psychiatric disorders. The aim of our work was to demonstrate the feasibility of mining factors that are strongly associated with violent behaviors among psychiatric patients from unstructured psychiatric evaluation records using clinical text mining. Experiment results showed that stimulants, followed by a family history of violent behavior, suicidal behaviors, and financial stress were strongly associated with violent behaviors. Key aspects explicated in this paper include employing our text mining pipeline to extract clinical and social factors linked with violent behaviors, generating association rules to uncover possible associations between these factors and violent behaviors, and lastly the ranking of top rules associated with violent behaviors using statistical analysis and interpretation.

A context-aware approach for progression tracking of medical concepts in electronic medical records.

Electronic medical records (EMRs) for diabetic patients contain information about heart disease risk factors such as high blood pressure, cholesterol levels, and smoking status. Discovering the described risk factors and tracking their progression over time may support medical personnel in making clinical decisions, as well as facilitate data modeling and biomedical research. Such highly patient-specific knowledge is essential to driving the advancement of evidence-based practice, and can also help improve personalized medicine and care. One general approach for tracking the progression of diseases and their risk factors described in EMRs is to first recognize all temporal expressions, and then assign each of them to the nearest target medical concept. However, this method may not always provide the correct associations. In light of this, this work introduces a context-aware approach to assign the time attributes of the recognized risk factors by reconstructing contexts that contain more reliable temporal expressions. The evaluation results on the i2b2 test set demonstrate the efficacy of the proposed approach, which achieved an F-score of 0.897. To boost the approach's ability to process unstructured clinical text and to allow for the reproduction of the demonstrated results, a set of developed .NET libraries used to develop the system is available at https://sites.google.com/site/hongjiedai/projects/nttmuclinicalnet.

Coronary artery disease risk assessment from unstructured electronic health records using text mining.

Coronary artery disease (CAD) often leads to myocardial infarction, which may be fatal. Risk factors can be used to predict CAD, which may subsequently lead to prevention or early intervention. Patient data such as co-morbidities, medication history, social history and family history are required to determine the risk factors for a disease. However, risk factor data are usually embedded in unstructured clinical narratives if the data is not collected specifically for risk assessment purposes. Clinical text mining can be used to extract data related to risk factors from unstructured clinical notes. This study presents methods to extract Framingham risk factors from unstructured electronic health records using clinical text mining and to calculate 10-year coronary artery disease risk scores in a cohort of diabetic patients. We developed a rule-based system to extract risk factors: age, gender, total cholesterol, HDL-C, blood pressure, diabetes history and smoking history. The results showed that the output from the text mining system was reliable, but there was a significant amount of missing data to calculate the Framingham risk score. A systematic approach for understanding missing data was followed by implementation of imputation strategies. An analysis of the 10-year Framingham risk scores for coronary artery disease in this cohort has shown that the majority of the diabetic patients are at moderate risk of CAD.

Visit-to-Visit Blood Pressure Variability in Cardiovascular Disease.

Visit-to-visit blood pressure variability (BPV) is associated with cardiovascular disease (CVD), independently of mean blood pressure (BP). However, in real world clinical practice, this phenomenon is frequently considered as random fluctuation. This review aimed to investigate the differences among studies investigating visit-to-visit BPV and CVD using electronic health record (EHR) and clinical trial data. Our review suggests that BP values in clinical trial data are derived using a stricter protocol compared to EHR data. Furthermore, there was no consensus on metrics used in estimation of BPV.

Strategies to Improve Statin Medication Adherence Among Patients at Risk of Cardiovascular Disease Identified Through Electronic Health Records: A Literature Review.

Statin is a group of lipid/cholesterol-lowering medications that is commonly used for primary and secondary prevention of cardiovascular diseases (CVD). In Australia, this is the first line of pharmacological therapy for CVD risk management. High-risk patients who do not adhere to lipid-modifying medicines have an increased risk of CVD mortality, hospitalization, and revascularization. However, studies show that 67% of patients are non-adherent to statins. As such, improving statin adherence through various strategies is very important. This literature review delves into the studies from the past 10 years to identify the various strategies used and their effectiveness to improve statin adherence. The initial search results on PubMed showed 157 articles and based on the inclusion and exclusion criteria, 7 articles were finally used for this review. The patients in the studies were identified through electronic health records. The findings suggest that education, counselling and motivation through face-to-face interaction, phone calls or text messages, reminder messages and frequent follow-up visits are good strategies to improve statin adherence. Alongside these, simplifying regimens, switching combinations of medicines, or using alternate dosing have also been shown to improve statin adherence. In summary, counselling and face-to-face interaction are effective methods for improving statin adherence. The use of electronic health record (EHR) systems combined with targeted interventions delivered to patients identified to be non-adherent to statin may further improve statin adherence.

Uptake of pneumococcal vaccines in older Australian adults before and after universal public funding of PCV13.

BACKGROUND: In 2020 Australia changed the funded universal older adult pneumococcal vaccination program from use of the 23-valent pneumococcal polysaccharide vaccine (PPV23) at age 65 to the 13-valent pneumococcal conjugate vaccine (PCV13) at age 70 years. We investigated uptake of both PCV13 and PPV23 in older adults before and after the program change. METHODS: We analysed a national dataset of records of patients attending general practices (GPs). We included regular attendees aged 65 or above in 2020. Cumulative uptake of PCV13 and monthly uptake of PPV23 was compared for the two periods before (January 2019 to June 2020) and after (July 2020 to May 2021) the program change on 1 July 2020, by age groups and presence of comorbid conditions. RESULTS: Our study included data from 192,508 patients (mean age in 2020: 75.1 years, 54.2 % female, 46.1 % with at least one comorbidity). Before July 2020, for all adults regardless of underlying comorbidities, the cumulative uptake of PCV13 was < 1 % but by May 2021, eleven months after the program changes, cumulative uptake of PCV13 had increased among those aged 70-79 years (without comorbidity: 16.3 %; with comorbidity: 21.1 %) and 80 + years (without comorbidity: 13.5 %; with comorbidity: 17.7 %), but not among those aged 65-69 years (without comorbidity: 1.3 %; with comorbidity: 3 %). Monthly uptake of PPV23 dropped following the program change across all age groups. CONCLUSIONS: Changes in uptake of PCV13 and PPV23 among those aged 70 + years were consistent with program changes. However, PCV13 uptake was still substantially lower in individuals aged 65-69 years overall and in those with comorbidities.

End-to-end prognostication in colorectal cancer by deep learning: a retrospective, multicentre study.

BACKGROUND: Precise prognosis prediction in patients with colorectal cancer (ie, forecasting survival) is pivotal for individualised treatment and care. Histopathological tissue slides of colorectal cancer specimens contain rich prognostically relevant information. However, existing studies do not have multicentre external validation with real-world sample processing protocols, and algorithms are not yet widely used in clinical routine. METHODS: In this retrospective, multicentre study, we collected tissue samples from four groups of patients with resected colorectal cancer from Australia, Germany, and the USA. We developed and externally validated a deep learning-based prognostic-stratification system for automatic prediction of overall and cancer-specific survival in patients with resected colorectal cancer. We used the model-predicted risk scores to stratify patients into different risk groups and compared survival outcomes between these groups. Additionally, we evaluated the prognostic value of these risk groups after adjusting for established prognostic variables. FINDINGS: We trained and validated our model on a total of 4428 patients. We found that patients could be divided into high-risk and low-risk groups on the basis of the deep learning-based risk score. On the internal test set, the group with a high-risk score had a worse prognosis than the group with a low-risk score, as reflected by a hazard ratio (HR) of 4·50 (95% CI 3·33-6·09) for overall survival and 8·35 (5·06-13·78) for disease-specific survival (DSS). We found consistent performance across three large external test sets. In a test set of 1395 patients, the high-risk group had a lower DSS than the low-risk group, with an HR of 3·08 (2·44-3·89). In two additional test sets, the HRs for DSS were 2·23 (1·23-4·04) and 3·07 (1·78-5·3). We showed that the prognostic value of the deep learning-based risk score is independent of established clinical risk factors. INTERPRETATION: Our findings indicate that attention-based self-supervised deep learning can robustly offer a prognosis on clinical outcomes in patients with colorectal cancer, generalising across different populations and serving as a potentially new prognostic tool in clinical decision making for colorectal cancer management. We release all source codes and trained models under an open-source licence, allowing other researchers to reuse and build upon our work. FUNDING: The German Federal Ministry of Health, the Max-Eder-Programme of German Cancer Aid, the German Federal Ministry of Education and Research, the German Academic Exchange Service, and the EU.

Artificial intelligence based personalized predictive survival among colorectal cancer patients.

BACKGROUND AND OBJECTIVE: Colorectal cancer is a major health concern. It is now the third most common cancer and the fourth leading cause of cancer mortality worldwide. The aim of this study was to evaluate the performance of machine learning algorithms for predicting survival of colorectal cancer patients 1 to 5 years after diagnosis, and identify the most important variables. METHODS: A sample of 1236 patients diagnosed with colorectal cancer and 118 predictor variables has been used. The outcome of interest was a binary variable indicating whether the patient survived the number of years in question or not. 20 predictor variables were selected using mutual information score with the outcome. We implemented 11 machine learning algorithms and evaluated their performance with a 5 by 2-fold cross-validation with stratified folds and with paired Student's t-tests. We compared the results with the Kaplan-Meier estimator and Cox's proportional hazard regression. RESULTS: Using the 20 most important predictor variables for each of the survival years, the logistic regression algorithm achieved an area under the receiver operating characteristic curve of 0.850 (0.014 SD, 0.840-0.860 95 % CI) for the 1-year, and 0.872 (0.014 SD, 0.861-0.882 95% CI) for the 5-year survival prediction. Using only the 5 most important predictor variables, the corresponding values are 0.793 (0.020 SD, 0.778-0.807 95% CI) and 0.794 (0.011 SD, 0.785-0.802 95% CI). The most important variables for 1-year prediction were number of R residual, M distant metastasis, overall stage, probable recurrence within 5 years, and tumour length, whereas for 5-year prediction the most important were probable recurrence within 5 years, R residual, M distant metastasis, number of positive lymph nodes, and palliative chemotherapy. Biomarkers do not appear among the top 20 most important ones. For all survival intervals, the probability of the top model agrees with the Kaplan-Meier estimate, both in the interval of one standard deviation and in the 95% confidence interval. CONCLUSIONS: The findings suggest that machine learning algorithms can predict the survival probability of colorectal cancer patients and can be used to inform the patients and assist decision-making in clinical care management. In addition, this study unveils the most essential variables for estimating survival short- and long-term among patients with Colorectal cancer.