Rebecca J. Mieloszyk

Automated Quantitative Analysis of Capnogram Shape for COPD–Normal and COPD–CHF Classification

We develop an approach to quantitative analysis of carbon dioxide concentration in exhaled breath, recorded as a function of time by capnography. The generated waveform-or capnogram-is currently used in clinical practice to establish the presence of respiration as well as determine respiratory rate and end-tidal CO2 concentration. The capnogram shape also has diagnostic value, but is presently assessed qualitatively, by visual inspection. Prior approaches to quantitatively characterizing the capnogram shape have explored the correlation of various geometric parameters with pulmonary function tests. These studies attempted to characterize the capnogram in normal subjects and patients with cardiopulmonary disease, but no consistent progress was made, and no translation into clinical practice was achieved. We apply automated quantitative analysis to discriminate between chronic obstructive pulmonary disease (COPD) and congestive heart failure (CHF), and between COPD and normal. Capnograms were collected from 30 normal subjects, 56 COPD patients, and 53 CHF patients. We computationally extract four physiologically based capnogram features. Classification on a hold-out test set was performed by an ensemble of classifiers employing quadratic discriminant analysis, designed through cross validation on a labeled training set. Using 80 exhalations of each capnogram record in the test set, performance analysis with bootstrapping yields areas under the receiver operating characteristic (ROC) curve of 0.89 (95% CI: 0.72-0.96) for COPD/CHF classification, and 0.98 (95% CI: 0.82-1.0) for COPD/normal classification. This classification performance is obtained with a run time sufficiently fast for realtime monitoring.

Model-based estimation of pulmonary compliance and resistance parameters from time-based capnography.

We propose a highly-simplified single-alveolus mechanistic model of lung mechanics and gas mixing that leads to an analytical solution for carbon dioxide partial pressure in exhaled breath, as measured by time-based capnography. Using this solution, we estimate physiological parameters of the lungs on a continuous, breath-by-breath basis. We validate our model with capnograms from 15 subjects responding positively (>20% FEV1 drop from baseline) to methacholine challenge, and subsequently recovering with bronchodilator treatment. Our results suggest that parameter estimates from capnography may provide discriminatory value for lung function comparable to spirometry, thus warranting more detailed study.

Clustering of capnogram features to track state transitions during procedural sedation

Procedural sedation has allowed many painful interventions to be conducted outside the operating room. During such procedures, it is important to maintain an appropriate level of sedation to minimize the risk of respiratory depression if patients are over-sedated and added pain or anxiety if under-sedated. However, there is currently no objective way to measure the patients evolving level of sedation during a procedure. We investigated the use of capnography-derived features as an objective measure of sedation level. Time-based capnograms were recorded from 30 patients during sedation for cardioversion. Through causal k-means clustering of selected features, we sequentially assigned each exhalation to one of three distinct clusters, or states. Transitions between these states correlated to events during sedation (drug administration, procedure start and end, and clinical interventions). Similar clustering of capnogram recordings from 26 healthy, non-sedated subjects did not reveal distinctly separated states.

Statistical analysis of the age dependence of the normal capnogram

The age dependence of the time-based capnogram from normal, healthy subjects has not been quantitatively characterized. The existence of age dependence would impact the development and operation of automated quantitative capnographic tools. Here, we quantitatively assess the relationship between normal capnogram shape and age. Capnograms were collected from healthy subjects, and physiologically-based features (exhalation duration, end-tidal CO2 and time spent at this value, normalized time spent at end-tidal CO2, end-exhalation slope, and instantaneous respiratory rate) were computationally extracted. The mean values of the individual features over 30 exhalations were linearly regressed against subject age, accounting for inter-feature correlation. After data collection, 154 of 178 subjects were eligible for analysis, with an age range of 3–78 years (mean age 39, std. dev. 20 years). The Bonferroni-corrected joint 95% confidence intervals (CIs) of the regression line slopes contained the origin for five of six features (the remaining CI was only slightly offset from the origin). The associated individual r2 values for the regressions were all below 0.07. We conclude that age is not a significant explanatory factor in describing variations in the shape of the normal capnogram. This finding could be exploited in the design of automated methods for quantitative capnogram analysis across a range of ages.

Model-Based Estimation of Respiratory Parameters from Capnography, With Application to Diagnosing Obstructive Lung Disease

Objective: We use a single-alveolar-compartment model to describe the partial pressure of carbon dioxide in exhaled breath, as recorded in time-based capnography. Respiratory parameters are estimated using this model, and then related to the clinical status of patients with obstructive lung disease. Methods: Given appropriate assumptions, we derive an analytical solution of the model, describing the exhalation segment of the capnogram. This solution is parametrized by alveolar CO2 concentration, dead-space fraction, and the time constant associated with exhalation. These quantities are estimated from individual capnogram data on a breath-by-breath basis. The model is applied to analyzing datasets from normal (n = 24) and chronic obstructive pulmonary disease (COPD) (n = 22) subjects, as well as from patients undergoing methacholine challenge testing for asthma (n = 22). Results: A classifier based on linear discriminant analysis in logarithmic coordinates, using estimated dead-space fraction and exhalation time constant as features, and trained on data from five normal and five COPD subjects, yielded an area under the receiver operating characteristic curve (AUC) of 0.99 in classifying the remaining 36 subjects as normal or COPD. Bootstrapping with 50 replicas yielded a 95% confidence interval of AUCs from 0.96 to 1.00. For patients undergoing methacholine challenge testing, qualitatively meaningful trends were observed in the parameter variations over the course of the test. Significance: A simple mechanistic model allows estimation of underlying respiratory parameters from the capnogram, and may be applied to diagnosis and monitoring of chronic and reversible obstructive lung disease.