Explainable Machine Learning Model for Predicting First-Time Acute Exacerbation in Patients with Chronic Obstructive Pulmonary Disease

Machine learning approaches for practical predicting outpatient near-future AECOPD based on nationwide electronic medical records

In this research, we aimed to harness machine learning to predict the imminent risk of acute exacerbation in chronic obstructive pulmonary disease (AECOPD) patients. Utilizing retrospective data from electronic medical records of two Taiwanese hospitals, we identified 26 critical features. To predict 3- and 6-month AECOPD occurrences, we deployed five distinct machine learning algorithms alongside ensemble learning. The 3-month risk prediction was best realized by the XGBoost model, achieving an AUC of 0.795, whereas the XGBoost was superior for the 6-month prediction with an AUC of 0.813. We conducted an explainability analysis and found that the episode of AECOPD, mMRC score, CAT score, respiratory rate, and the use of inhaled corticosteroids were the most impactful features. Notably, our approach surpassed predictions that relied solely on CAT or mMRC scores. Accordingly, we designed an interactive prediction system that provides physicians with a practical tool to predict near-term AECOPD risk in outpatients.

DeepXplainer: An interpretable deep learning based approach for lung cancer detection using explainable artificial intelligence

BACKGROUND AND OBJECTIVE

Artificial intelligence (AI) has several uses in the healthcare industry, some of which include healthcare management, medical forecasting, practical making of decisions, and diagnosis. AI technologies have reached human-like performance, but their use is limited since they are still largely viewed as opaque black boxes. This distrust remains the primary factor for their limited real application, particularly in healthcare. As a result, there is a need for interpretable predictors that provide better predictions and also explain their predictions.

METHODS

This study introduces "DeepXplainer", a new interpretable hybrid deep learning-based technique for detecting lung cancer and providing explanations of the predictions. This technique is based on a convolutional neural network and XGBoost. XGBoost is used for class label prediction after "DeepXplainer" has automatically learned the features of the input using its many convolutional layers. For providing explanations or explainability of the predictions, an explainable artificial intelligence method known as "SHAP" is implemented.

RESULTS

The open-source "Survey Lung Cancer" dataset was processed using this method. On multiple parameters, including accuracy, sensitivity, F1-score, etc., the proposed method outperformed the existing methods. The proposed method obtained an accuracy of 97.43%, a sensitivity of 98.71%, and an F1-score of 98.08. After the model has made predictions with this high degree of accuracy, each prediction is explained by implementing an explainable artificial intelligence method at both the local and global levels.

CONCLUSIONS

A deep learning-based classification model for lung cancer is proposed with three primary components: one for feature learning, another for classification, and a third for providing explanations for the predictions made by the proposed hybrid (ConvXGB) model. The proposed "DeepXplainer" has been evaluated using a variety of metrics, and the results demonstrate that it outperforms the current benchmarks. Providing explanations for the predictions, the proposed approach may help doctors in detecting and treating lung cancer patients more effectively.

Machine learning and deep learning predictive models for long-term prognosis in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis

BACKGROUND

Machine learning and deep learning models have been increasingly used to predict long-term disease progression in patients with chronic obstructive pulmonary disease (COPD). We aimed to summarise the performance of such prognostic models for COPD, compare their relative performances, and identify key research gaps.

METHODS

We conducted a systematic review and meta-analysis to compare the performance of machine learning and deep learning prognostic models and identify pathways for future research. We searched PubMed, Embase, the Cochrane Library, ProQuest, Scopus, and Web of Science from database inception to April 6, 2023, for studies in English using machine learning or deep learning to predict patient outcomes at least 6 months after initial clinical presentation in those with COPD. We included studies comprising human adults aged 18-90 years and allowed for any input modalities. We reported area under the receiver operator characteristic curve (AUC) with 95% CI for predictions of mortality, exacerbation, and decline in forced expiratory volume in 1 s (FEV1). We reported the degree of interstudy heterogeneity using Cochran's Q test (significant heterogeneity was defined as p≤0·10 or I2>50%). Reporting quality was assessed using the TRIPOD checklist and a risk-of-bias assessment was done using the PROBAST checklist. This study was registered with PROSPERO (CRD42022323052).

FINDINGS

We identified 3620 studies in the initial search. 18 studies were eligible, and, of these, 12 used conventional machine learning and six used deep learning models. Seven models analysed exacerbation risk, with only six reporting AUC and 95% CI on internal validation datasets (pooled AUC 0·77 [95% CI 0·69-0·85]) and there was significant heterogeneity (I2 97%, p<0·0001). 11 models analysed mortality risk, with only six reporting AUC and 95% CI on internal validation datasets (pooled AUC 0·77 [95% CI 0·74-0·80]) with significant degrees of heterogeneity (I2 60%, p=0·027). Two studies assessed decline in lung function and were unable to be pooled. Machine learning and deep learning models did not show significant improvement over pre-existing disease severity scores in predicting exacerbations (p=0·24). Three studies directly compared machine learning models against pre-existing severity scores for predicting mortality and pooled performance did not differ (p=0·57). Of the five studies that performed external validation, performance was worse than or equal to regression models. Incorrect handling of missing data, not reporting model uncertainty, and use of datasets that were too small relative to the number of predictive features included provided the largest risks of bias.

INTERPRETATION

There is limited evidence that conventional machine learning and deep learning prognostic models demonstrate superior performance to pre-existing disease severity scores. More rigorous adherence to reporting guidelines would reduce the risk of bias in future studies and aid study reproducibility.

FUNDING

None.

Automated detection of airflow obstructive diseases: A systematic review of the last decade (2013-2022)

BACKGROUND AND OBJECTIVE

Obstructive airway diseases, including asthma and Chronic Obstructive Pulmonary Disease (COPD), are two of the most common chronic respiratory health problems. Both of these conditions require health professional expertise in making a diagnosis. Hence, this process is time intensive for healthcare providers and the diagnostic quality is subject to intra- and inter- operator variability. In this study we investigate the role of automated detection of obstructive airway diseases to reduce cost and improve diagnostic quality.

METHODS

We investigated the existing body of evidence and applied Preferred Reporting Items for Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to search records in IEEE, Google scholar, and PubMed databases. We identified 65 papers that were published from 2013 to 2022 and these papers cover 67 different studies. The review process was structured according to the medical data that was used for disease detection. We identified six main categories, namely air flow, genetic, imaging, signals, and miscellaneous. For each of these categories, we report both disease detection methods and their performance.

RESULTS

We found that medical imaging was used in 14 of the reviewed studies as data for automated obstructive airway disease detection. Genetics and physiological signals were used in 13 studies. Medical records and air flow were used in 9 and 7 studies, respectively. Most papers were published in 2020 and we found three times more work on Machine Learning (ML) when compared to Deep Learning (DL). Statistical analysis shows that DL techniques achieve higher Accuracy (ACC) when compared to ML. Convolutional Neural Network (CNN) is the most common DL classifier and Support Vector Machine (SVM) is the most widely used ML classifier. During our review, we discovered only two publicly available asthma and COPD datasets. Most studies used private clinical datasets, so data size and data composition are inconsistent.

CONCLUSIONS

Our review results indicate that Artificial Intelligence (AI) can improve both decision quality and efficiency of health professionals during COPD and asthma diagnosis. However, we found several limitations in this review, such as a lack of dataset consistency, a limited dataset and remote monitoring was not sufficiently explored. We appeal to society to accept and trust computer aided airflow obstructive diseases diagnosis and we encourage health professionals to work closely with AI scientists to promote automated detection in clinical practice and hospital settings.

Blood monocyte levels predict the risk of acute exacerbations of chronic obstructive pulmonary disease: a retrospective case-control study

Monocytes were critical cells in the innate immune system. Monocyte recruitment to the lungs is a crucial process of pathophysiology in chronic obstructive pulmonary disease (COPD). Current evidence on the association between the occurrence of acute exacerbations of COPD (AECOPD) and monocytes was unclear. This study aimed to examine whether blood monocytes are associated with the occurrence of AECOPD and to determine the specific blood monocyte level to predict AECOPD. A retrospective case-control study was conducted at Changhua Christian Hospital. A total of 444 eligible patients with COPD were included between January 2017 and December 2019. Restricted cubic splines were used to analyze the nonlinear relationships between continuous white blood cell values and the occurrence of AECOPD. The association between monocytes and the occurrence of AECOPD was assessed using the logistic, lasso, and ridge regression models. Restricted cubic splines revealed nonlinear associations among the monocyte level, the continuous value of the eosinophil-to-lymphocyte ratio, and the occurrence of AECOPD. The lowest risk of occurrence of AECOPD ranged from 7.4 to 10%; < 7.4% with an absolute count < 0.62 or > 10% indicated significant risk. No significant association was noted between the eosinophil-to-lymphocyte ratio categories in the tertiles (< 0.049, 0.049 to < 0.122, and ≥ 0.122) and the risk of AECOPD. A significantly higher risk was noted in the association of the occurrence of AECOPD with the CAT score; mMRC score; wheezing cough; preexisting chronic pulmonary disease; hypertension and malignancy; use of dual- and triple, and oral long-acting bronchodilators for COPD treatment; and WBC count. We reported a nonlinear relationship between monocytes and the occurrence of AECOPD. Patients with monocyte percentage of > 10% or < 7.4% with an absolute count < 0.62 had higher risk of occurrence of AECOPD. Overall, our study demonstrated the specific value of monocytes in identifying high risks of the occurrence of AECOPD; this value is an easy-to-obtain, inexpensive biomarker in patients with AECOPD and should be further investigated in future prospective clinical studies.

Application of explainable artificial intelligence for healthcare: A systematic review of the last decade (2011-2022)

BACKGROUND AND OBJECTIVES

Artificial intelligence (AI) has branched out to various applications in healthcare, such as health services management, predictive medicine, clinical decision-making, and patient data and diagnostics. Although AI models have achieved human-like performance, their use is still limited because they are seen as a black box. This lack of trust remains the main reason for their low use in practice, especially in healthcare. Hence, explainable artificial intelligence (XAI) has been introduced as a technique that can provide confidence in the model's prediction by explaining how the prediction is derived, thereby encouraging the use of AI systems in healthcare. The primary goal of this review is to provide areas of healthcare that require more attention from the XAI research community.

METHODS

Multiple journal databases were thoroughly searched using PRISMA guidelines 2020. Studies that do not appear in Q1 journals, which are highly credible, were excluded.

RESULTS

In this review, we surveyed 99 Q1 articles covering the following XAI techniques: SHAP, LIME, GradCAM, LRP, Fuzzy classifier, EBM, CBR, rule-based systems, and others.

CONCLUSION

We discovered that detecting abnormalities in 1D biosignals and identifying key text in clinical notes are areas that require more attention from the XAI research community. We hope this is review will encourage the development of a holistic cloud system for a smart city.

Explainable Machine Learning-Based Risk Prediction Model for In-Hospital Mortality after Continuous Renal Replacement Therapy Initiation

In this study, we established an explainable and personalized risk prediction model for in-hospital mortality after continuous renal replacement therapy (CRRT) initiation. This retrospective cohort study was conducted at Changhua Christian Hospital (CCH). A total of 2932 consecutive intensive care unit patients receiving CRRT between 1 January 2010, and 30 April 2021, were identified from the CCH Clinical Research Database and were included in this study. The recursive feature elimination method with 10-fold cross-validation was used and repeated five times to select the optimal subset of features for the development of machine learning (ML) models to predict in-hospital mortality after CRRT initiation. An explainable approach based on ML and the SHapley Additive exPlanation (SHAP) and a local explanation method were used to evaluate the risk of in-hospital mortality and help clinicians understand the results of ML models. The extreme gradient boosting and gradient boosting machine models exhibited a higher discrimination ability (area under curve [AUC] = 0.806, 95% CI = 0.770–0.843 and AUC = 0.823, 95% CI = 0.788–0.858, respectively). The SHAP model revealed that the Acute Physiology and Chronic Health Evaluation II score, albumin level, and the timing of CRRT initiation were the most crucial features, followed by age, potassium and creatinine levels, SPO2, mean arterial pressure, international normalized ratio, and vasopressor support use. ML models combined with SHAP and local interpretation can provide the visual interpretation of individual risk predictions, which can help clinicians understand the effect of critical features and make informed decisions for preventing in-hospital deaths.