Deep learning using multilayer perception improves the diagnostic acumen of spirometry: a single-centre Canadian study

Rationale Spirometry and plethysmography are the gold standard pulmonary function tests (PFT) for diagnosis and management of lung disease. Due to the inaccessibility of plethysmography, spirometry is often used alone but this leads to missed or misdiagnoses as spirometry cannot identify restrictive disease without plethysmography. We aimed to develop a deep learning model to improve interpretation of spirometry alone. Methods We built a multilayer perceptron model using full PFTs from 748 patients, interpreted according to international guidelines. Inputs included spirometry (forced vital capacity, forced expiratory volume in 1 s, forced mid-expiratory flow25–75), plethysmography (total lung capacity, residual volume) and biometrics (sex, age, height). The model was developed with 2582 PFTs from 477 patients, randomly divided into training (80%), validation (10%) and test (10%) sets, and refined using 1245 previously unseen PFTs from 271 patients, split 50/50 as validation (136 patients) and test (135 patients) sets. Only one test per patient was used for each of 10 experiments conducted for each input combination. The final model was compared with interpretation of 82 spirometry tests by 6 trained pulmonologists and a decision tree. Results Accuracies from the first 477 patients were similar when inputs included biometrics+spirometry+plethysmography (95%±3%) vs biometrics+spirometry (90%±2%). Model refinement with the next 271 patients improved accuracies with biometrics+pirometry (95%±2%) but no change for biometrics+spirometry+plethysmography (95%±2%). The final model significantly outperformed (94.67%±2.63%, p<0.01 for both) interpretation of 82 spirometry tests by the decision tree (75.61%±0.00%) and pulmonologists (66.67%±14.63%). Conclusions Deep learning improves the diagnostic acumen of spirometry and classifies lung physiology better than pulmonologists with accuracies comparable to full PFTs.

[1]  S. Stanojevic,et al.  ERS/ATS technical standard on interpretive strategies for routine lung function tests , 2021, European Respiratory Journal.

[2]  I. Sohn,et al.  Predicting Successes and Failures of Clinical Trials With Outer Product–Based Convolutional Neural Network , 2021, Frontiers in Pharmacology.

[3]  K. Gourgoulianis,et al.  Pulmonary function testing in COPD: looking beyond the curtain of FEV1 , 2021, npj Primary Care Respiratory Medicine.

[4]  Nadhir Al-Ansari,et al.  Influence of Data Splitting on Performance of Machine Learning Models in Prediction of Shear Strength of Soil , 2021 .

[5]  Byron C. Jaeger,et al.  Deep neural network analyses of spirometry for structural phenotyping of chronic obstructive pulmonary disease. , 2020, JCI insight.

[6]  J. Stoller,et al.  An Alternative Spirometric Measurement. Area under the Expiratory Flow–Volume Curve , 2020, Annals of the American Thoracic Society.

[7]  Kevin McCarthy,et al.  Standardization of Spirometry 2019 Update. An Official American Thoracic Society and European Respiratory Society Technical Statement , 2019, American journal of respiratory and critical care medicine.

[8]  Dimitris Spathis,et al.  Diagnosing asthma and chronic obstructive pulmonary disease with machine learning , 2019, Health Informatics J..

[9]  C. Vogelmeier,et al.  Artificial intelligence outperforms pulmonologists in the interpretation of pulmonary function tests , 2019, European Respiratory Journal.

[10]  Arie Nakhmani,et al.  New Spirometry Indices for Detecting Mild Airflow Obstruction , 2018, Scientific Reports.

[11]  L. Boulet,et al.  Underdiagnosis and Overdiagnosis of Asthma , 2018, American journal of respiratory and critical care medicine.

[12]  Yun Xu,et al.  On Splitting Training and Validation Set: A Comparative Study of Cross-Validation, Bootstrap and Systematic Sampling for Estimating the Generalization Performance of Supervised Learning , 2018, Journal of Analysis and Testing.

[13]  Dharm Singh Jat,et al.  Applications of statistical techniques and artificial neural networks: A review , 2018, Journal of Statistics and Management Systems.

[14]  Brian E. Ruttenberg,et al.  Causal Learning and Explanation of Deep Neural Networks via Autoencoded Activations , 2018, ArXiv.

[15]  R. Wood‐Baker,et al.  Improved spirometric detection of small airway narrowing: concavity in the expiratory flow–volume curve in people aged over 40 years , 2017, International journal of chronic obstructive pulmonary disease.

[16]  Anne E Carpenter,et al.  Opportunities and obstacles for deep learning in biology and medicine , 2017, bioRxiv.

[17]  M. Decramer,et al.  Automated Interpretation of Pulmonary Function Tests in Adults with Respiratory Complaints , 2017, Respiration.

[18]  Yongjiang Tang,et al.  The measurement of lung volumes using body plethysmography and helium dilution methods in COPD patients: a correlation and diagnosis analysis , 2016, Scientific Reports.

[19]  C. Chakraborty,et al.  Automated Screening Methodology for Asthma Diagnosis that Ensembles Clinical and Spirometric Information , 2016, Journal of Medical and Biological Engineering.

[20]  J. Walters,et al.  Diagnosis and early detection of COPD using spirometry. , 2014, Journal of thoracic disease.

[21]  S. Stanojevic,et al.  Multi-ethnic reference values for spirometry for the 3–95-yr age range: the global lung function 2012 equations , 2012, European Respiratory Journal.

[22]  Deniz Sahin,et al.  Diagnosis of Airway Obstruction or Restrictive Spirometric Patterns by Multiclass Support Vector Machines , 2010, Journal of Medical Systems.

[23]  J. Hankinson,et al.  Standardisation of spirometry , 2005, European Respiratory Journal.

[24]  I A Basheer,et al.  Artificial neural networks: fundamentals, computing, design, and application. , 2000, Journal of microbiological methods.

[25]  N. Zamel,et al.  Reference values of pulmonary function tests for Canadian Caucasians. , 2004, Canadian respiratory journal.