Abstract:Chronic Obstructive Pulmonary Disease (COPD) is a chronic lung disease that causes airflow obstruction. Current methods can only detect COPD from prominent features in spirogram (Volume-Flow time series) but cannot predict future COPD risk from subtle data patterns. We propose a deep learning-based method, DeepSpiro, for early prediction of future COPD risk. DeepSpiro consists of four key components: SpiroSmoother for stabilizing the Volume-Flow curve, SpiroEncoder for capturing volume evolution through key patches of varying lengths, SpiroExplainer for integrating heterogeneous data and explaining predictions through volume attention, and SpiroPredictor for predicting the disease risk of undiagnosed high-risk patients based on key patch concavity, with prediction horizons of 1, 2, 3, 4, 5 years, or even longer. Evaluated on the UK Biobank dataset, DeepSpiro achieved an AUC of 0.8328 for COPD detection and demonstrated strong predictive performance for future COPD risk (p-value < 0.001). DeepSpiro effectively predicts the long-term progression of the disease.

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve This paper aims to address the early detection and future risk prediction of Chronic Obstructive Pulmonary Disease (COPD). Specifically, the paper proposes a deep learning-based method called DeepSpiro, which is used to detect and predict COPD from time-series data of pulmonary function tests (such as volume-flow curves). ### Background and Challenges 1. **Severity of COPD**: - COPD is a progressively worsening lung disease that leads to breathing difficulties, limited activity, and reduced quality of life. - As the disease progresses, COPD can also increase the risk of cardiovascular diseases and even lead to premature death. - Timely and accurate detection of COPD is crucial to reducing health risks for patients, especially in the early stages of the disease. 2. **Limitations of Existing Methods**: - Clinical diagnosis typically identifies COPD patients by determining whether the FEV1/FVC ratio is below 70%, but this method is not always accurate across different age groups. - Existing deep learning methods can identify features of COPD but cannot effectively predict an individual's future risk of COPD. - Deep learning models lack transparency, making it difficult to gain the trust of medical professionals and patients. 3. **Main Challenges**: - Generating stable volume-flow curves to determine the degree of airflow obstruction. - Handling volume-flow curves of different lengths, avoiding the introduction of noise or sacrificing important data dependencies. - Providing interpretable model results to enhance model transparency. - Early prediction of an individual's future probability of developing COPD. ### Solution To address the above challenges, the paper proposes DeepSpiro, a deep learning-based method for early prediction of COPD risk. DeepSpiro consists of four key components: 1. **SpiroSmoother**: - Stabilizes the volume-flow curves through a curve smoothing algorithm, preserving the physiological information in the original data. 2. **SpiroEncoder**: - Dynamically calculates the optimal number of "key segments" for each time-series data segment, unifying time-series representation and extracting key physiological information from high-dimensional dynamic sequences. 3. **SpiroExplainer**: - Combines demographic information such as age and gender, and explains model prediction results through a volume attention mechanism, improving model transparency and credibility. 4. **SpiroPredictor**: - Based on the concavity evolution of key segments, proposes a method for predicting the probability of undiagnosed high-risk patients developing the disease within 1-5 years or longer for the first time. ### Experimental Results - **Detection Performance**: - On the UK Biobank dataset, DeepSpiro achieved an AUC of 0.8328 in the COPD detection task, outperforming the baseline model ResNet18. - DeepSpiro outperformed the baseline model in all evaluation metrics (AUROC, AUPRC, F1-score). - **Future Risk Prediction**: - DeepSpiro can effectively predict an individual's future risk of developing COPD and accurately classify high-risk patients. - By analyzing volume-flow curves over different time periods, it was found that high-risk individuals are more likely to experience curve collapse in the early stages, while low-risk individuals experience curve collapse in the later stages. - **Subgroup Analysis**: - DeepSpiro's predictive performance was superior to other methods across subgroups of different ages, genders, and smoking statuses. - Smokers, males, and the elderly have a higher risk of COPD, which is consistent with clinical observations. ### Conclusion DeepSpiro effectively addresses the limitations of existing methods by stabilizing volume-flow curves, extracting key physiological information, providing interpretable prediction results, and predicting future COPD risk. It offers a new tool for the early detection and management of COPD.

Deep Learning for Detecting and Early Predicting Chronic Obstructive Pulmonary Disease from Spirogram Time Series

Deep Learning for Detecting and Early Predicting Chronic Obstructive Pulmonary Disease from Spirogram Time Series: A UK Biobank Study

Transformer-based Time-Series Biomarker Discovery for COPD Diagnosis

A Temporal Visualization of Chronic Obstructive Pulmonary Disease Progression Using Deep Learning and Unstructured Clinical Notes.

Fractional dynamics foster deep learning of COPD stage prediction

Early Diagnosis of Chronic Obstructive Pulmonary Disease from Chest X-Rays using Transfer Learning and Fusion Strategies

Screening and staging of chronic obstructive pulmonary disease with deep learning based on chest X-ray images and clinical parameters

Development and application of a deep learning-based comprehensive early diagnostic model for chronic obstructive pulmonary disease

Deep Learning Prediction of Survival in Patients with Chronic Obstructive Pulmonary Disease Using Chest Radiographs.

Detection and staging of chronic obstructive pulmonary disease using a computed tomography–based weakly supervised deep learning approach

Inference of chronic obstructive pulmonary disease with deep learning on raw spirograms identifies new genetic loci and improves risk models

Machine learning-enabled risk prediction of chronic obstructive pulmonary disease with unbalanced data

A Deep Learning Approach to Handling Temporal Variation in Chronic Obstructive Pulmonary Disease Progression

Chronic Obstructive Pulmonary Disease Prediction Using Deep Convolutional Network

Machine Learning-Assisted Diagnosis Model for Chronic Obstructive Pulmonary Disease

Machine Learning-based Characterization of COPD Progression from Spirometry Data

A Novel Approach for the Detection and Severity Grading of Chronic Obstructive Pulmonary Disease Based on Transformed Volumetric Capnography

Novel App-Based Portable Spirometer for the Early Detection of COPD

Deep Learning Assisted Diagnosis of Chronic Obstructive Pulmonary Disease Based on a Local-to-Global Framework

The application of impulse oscillometry system based on machine learning algorithm in the diagnosis of chronic obstructive pulmonary disease

Prediction of COPD Using Machine Learning, Clinical Summary Notes, and Vital Signs