Validation of a Computerized Module for the Interpretation of Cardiopulmonary Exercise Stress Test

Background:
Cardiopulmonary exercise testing (CPET) is an important modality for the evaluation and management of patients with a diverse array of medical problems or symptoms. However, interpreting these tests requires significant expertise, and time consuming. The aim of the present study was to develop a Computer-Aided Interpretation System (CAIx) to help physicians and physiologists in objectively evaluating CPET test results.

Material and methods:
239 consecutive CPET tests from the Institutes of Lung and Heart Diseases at the Sheba Medical Center in Ramat-Gan, Israel were used for this study. For the Support Vector Machine (SVM) learning stage, 150 CPET tests of healthy participants and patients with moderate-to-severe heart failure (CHF) or chronic obstructive pulmonary disease (COPD) were used. The remaining 89 tests from a similar population were used for the SVM’s disease, type and level of exercise intolerance and module validation. The performance of the SVM was compared with the clinical diagnosis using distribution analysis.

Results:
The CAIx successfully classified normal participants and CHF and COPD patients. The positive predictive values (PPV) were 94%, 92% and 79% respectively for the normal, CHF and COPD patients. The negative predictive values (NPV) were correspondingly 96%, 97% and 91%. Consequently, the sensitivity, specificity and overall preciseness of the proposed module were 89%, 94% and 89% respectively. An Exercise Limitation Classification Module (ELCx) contained within the SVM was able to accurately classify the type and level of the exercise limitations with 95% sensitivity, 97% specificity and 95% precision.

Conclusions:
The new CAIx and the ELCx modules developed herein for CPET analyses are highly sensitive and specific. Their use may reduce the complexity, increase the objectivity and be time saving of CPET interpretation in clinical settings.