ABSTRACT
Identification of ventilatory constraint is a key objective of clinical exercise testing. Expiratory flow-limitation (EFL) is a well-known type of ventilatory constraint. However, EFL is difficult to measure, and commercial metabolic carts do not readily identify or quantify EFL. Deep machine learning might provide a new approach for identifying EFL. The objective of this study was to determine if a convolutional neural network (CNN) could accurately identify EFL during exercise in adults in whom baseline airway function varied from normal to mildly obstructed. 2931 spontaneous exercise flow-volume loops (eFVL) were placed within the baseline maximal expiratory flow-volume curves (MEFV) from 22 adults (15 M, 7 F; age, 32 yrs) in whom lung function varied from normal to mildly obstructed. Each eFVL was coded as EFL or non-EFL, where EFL was defined by eFVLs with expired airflow meeting or exceeding the MEFV curve. A CNN with seven hidden layers and a 2-neuron softmax output layer was used to analyze the eFVLs. Three separate analyses were conducted: (1) all subjects (n = 2931 eFVLs, [GRALL]), (2) subjects with normal spirometry (n = 1921 eFVLs [GRNORM]), (3) subjects with mild airway obstruction (n = 1010 eFVLs, [GRLOW]). The final output of the CNN was the probability of EFL or non-EFL in each eFVL, which is considered EFL if the probability exceeds 0.5 or 50%. Baseline forced expiratory volume in 1 s/forced vital capacity was 0.77 (94% predicted) in GRALL, 0.83 (100% predicted) in GRNORM, and 0.69 (83% predicted) in GRLOW. CNN model accuracy was 90.6, 90.5, and 88.0% in GRALL, GRNORM and GRLOW, respectively. Negative predictive value (NPV) was higher than positive predictive value (PPV) in GRNORM (93.5 vs. 78.2% for NPV vs. PPV). In GRLOW, PPV was slightly higher than NPV (89.5 vs. 84.5% for PPV vs. NPV). A CNN performed very well at identifying eFVLs with EFL during exercise. These findings suggest that deep machine learning could become a viable tool for identifying ventilatory constraint during clinical exercise testing.
Subject(s)
Exhalation , Lung , Humans , Adult , Lung/physiology , Exhalation/physiology , Forced Expiratory Volume/physiology , Exercise/physiology , Vital Capacity/physiology , PyrinABSTRACT
We determined the effect of exercise-induced bronchoconstriction (EIB) on the shape of the maximal expiratory flow-volume (MEFV) curve in asthmatic adults. The slope-ratio index (SR) was used to quantitate the shape of the MEFV curve. We hypothesized that EIB would be accompanied by increases in SR and thus increased curvilinearity of the MEFV curve. Adult asthmatic ( n = 10) and non-asthmatic control subjects ( n = 9) cycled for 6-8 min at 85% of peak power. Following exercise, subjects remained on the ergometer and performed a maximal forced exhalation every 2 min for a total 20 min. In each MEFV curve, the slope-ratio index (SR) was calculated in 1% volume increments beginning at peak expiratory flow (PEF) and ending at 20% of forced vital capacity (FVC). Baseline spirometry was lower in asthmatics compared to control subjects (FEV1 % predicted, 89.1 ± 14.3 vs. 96.5 ± 12.2% [SD] in asthma vs. control; p < 0.05). In asthmatic subjects, post-exercise FEV1 decreased by 29.9 ± 13.2% from baseline (3.48 ± 0.74 and 2.24 ± 0.59 [SD] L for baseline and post-exercise nadir; p < 0.001). At baseline and at all timepoints after exercise, average SR between 80 and 20% of FVC was larger in asthmatic than control subjects (1.48 ± 0.02 vs. 1.23 ± 0.02 [SD] for asthma vs. control; p < 0.005). This averaged SR did not change after exercise in either subject group. In contrast, post-exercise SR between PEF and 75% of FVC was increased from baseline in subjects with asthma, suggesting that airway caliber heterogeneity increases with EIB. These findings suggest that the SR-index might provide useful information on the physiology of acute airway narrowing that complements traditional spirometric measures.
