Comparison of classification methods in breath analysis by electronic nose

Abstract

Currently, many different methods are being used for pre-processing, statistical analysis and validation of data obtained by electronic nose technology from exhaled air. These various methods, however, have never been thoroughly compared. We aimed to empirically evaluate and compare the influence of different dimension reduction, classification and validation methods found in published studies on the diagnostic performance in several datasets. Our objective was to facilitate the selection of appropriate statistical methods and to support reviewers in this research area. We reviewed the literature by searching Pubmed up to the end of 2014 for all human studies using an electronic nose and methodological quality was assessed using the QUADAS-2 tool tailored to our review. Forty-six studies were evaluated regarding the range of different approaches to dimension reduction, classification and validation. From forty-six reviewed articles only seven applied external validation in an independent dataset, mostly with a case-control design. We asked their authors to share the original datasets with us. Four of the seven datasets were available for re-analysis. Published statistical methods for eNose signal analysis found in the literature review were applied to the training set of each dataset. The performance (area under the receiver operating characteristics curve (ROC-AUC)) was calculated for the training cohort (in-set) and after internal validation (leave-one-out cross validation). The methods were also applied to the external validation set to assess the external validity of the performance. Risk of bias was high in most studies due to non-random selection of patients. Internal validation resulted in a decrease in ROC-AUCs compared to in-set performance: -0.15,-0.14,-0.1,-0.11 in dataset 1 through 4, respectively. External validation resulted in lower ROC-AUC compared to internal validation in dataset 1 (-0.23) and 3 (-0.09). ROC-AUCs did not decrease in dataset 2 (+0.07) and 4 (+0.04). No single combination of dimension reduction and classification methods gave consistent results between internal and external validation sets in this sample of four datasets. This empirical evaluation showed that it is not meaningful to estimate the diagnostic performance on a training set alone, even after internal validation. Therefore, we recommend the inclusion of an external validation set in all future eNose projects in medicine.

Laser spectroscopy for breath analysis: towards clinical implementation

“…Recent advances in nanotechnologies, optoelectronics, and photonics integration reinforce the demand for compactness and miniaturization of the devices. Non-specific sensors assembled into arrays that respond to different odors by generating a complex signal, e.g., the electronic nose or semiconductor-based sensor array, have become popular and have been used in clinical applications [144][145][146]. They are relatively inexpensive, small, easy-to-use, but lack both sensitivity and selectivity, require frequent calibrations, drift over time, have memory effects, are sensitive to changes in humidity and temperature and cannot identify individual compounds [25,147].…”

“…They are relatively inexpensive, small, easy-to-use, but lack both sensitivity and selectivity, require frequent calibrations, drift over time, have memory effects, are sensitive to changes in humidity and temperature and cannot identify individual compounds [25,147]. Moreover, since a limited number of specific sensorial elements are used, it is mandatory to know which compounds will be targeted and/or the composition matrix of the background [144][145][146]. Optical techniques are by far more sensitive and selective, are virtually maintenance free and can operate continuously for long periods of time.…”

Breathomics from exhaled volatile organic compounds in pediatric asthma

“…For this, specialist machine learning methods such as logistic regression, support vector machine learning, random forest, neural networks, or linear discriminant analysis, can be used. These methods have been recently compared in published eNose data sets by Leopold et al After the classification algorithm has been build, internal cross‐validation (eg, by bootstrapping, split‐half, or leave‐one‐out) is required as an intermediate step. In addition, external validation (evaluation of the performance of the model, in data that were not used to develop the model) in a new validation group that includes a newly recruited population is strongly advised, and preferably show high sensitivity and specificity .…”

Diagnosis of pulmonary tuberculosis and assessment of treatment response through analyses of volatile compound patterns in exhaled breath samples

“…Significance was defined as a value of p < 0.05. Principal component analyses (PCA) and pattern recognition analysis was employed to study the relationship between the sensors data and the pulmonary TB diagnosis (18). First, an exploratory evaluation using PCA was performed to investigate data structure and similarities between subjects and between.…”

“…Those changes are then quantified into in a unique breath signal (or “breathprint”). Pattern recognition algorithms may then sort the data into classes discriminating those pertaining to specific microorganisms or diseases (18). In principle, identification of breath signals does not require identification of individual molecular constituents.…”