The eNose (Netherlands)

<b>Rationale:</b> The gold standard to rule out acute pulmonary embolism (PE), computed tomography pulmonary angiography (CTPA), is expensive, requires harmful irradiation and intravenous administration of contrast agents. Of the performed CTPA for PE, approximately 75% have a negative result. Electronic nose technology might aid in the decision to perform a CTPA in PE suspected patients. <b>Methods:</b> Patients seen at the emergency department that received a CTPA scan due to suspected PE (based on increased D-dimer or Wells-score) breathed normally into an electronic nose (the Aeonose^TM (The eNose Company, Zutphen, the Netherlands)) for five minutes while wearing a nose clip. The Aeonose^TM measurement data were combined with the CTPA result as the gold standard for training an artificial neural network. Resulting Aeonose^TM classifications into PE or no PE were obtained using Leave-10%-Out cross validation. <b>Results:</b> A total of 61 subjects (mean age 61.8±15.5 years; 56% male) were included in the training set and PE was confirmed on CTPA in 20. Median D-dimer was 1427 (IQR 704 – 3244) µg/L. Median Wells-score was 3.0 (IQR 3.0 – 4.5). The area under the ROC curve was 0.73. The optimal threshold to rule out PE resulted in a sensitivity of 90% with a specificity of 54%, positive predictive value of 49% and negative predictive value of 92%. Thus, 24 subjects (39%) were classified as not having PE of which 2 cases did have PE on CTPA (false-omission rate: 8.3%). <b>Conclusion:</b> The data suggest that in subjects with suspected PE the Aeonose^TM could reduce the need for a CTPA scan in 39% at the costs of a false-omission rate of 8.3%. Caution is warranted as all data were used for training the neural network. Prospective validation of the model is imperative.

Introduction. Only 15% of lung cancer cases present with potentially curable disease. Therefore, there is much interest in a fast, non-invasive tool to detect lung cancer earlier. Exhaled breath analysis using electronic nose technology measures volatile organic compounds (VOCs) in exhaled breath that are associated with lung cancer. Methods. The diagnostic accuracy of the Aeonose™ is currently being studied in a multi-centre, prospective study in 210 subjects suspected for lung cancer, where approximately half will have a confirmed diagnosis and the other half will have a rejected diagnosis of lung cancer. We will also include 100–150 healthy control subjects. The eNose Company (provider of the Aeonose™) uses a software program, called Aethena, comprising pre-processing, data compression and neural networks to handle big data analyses. Each individual exhaled breath measurement comprises a data matrix with thousands of conductivity values. This is followed by data compression using a Tucker3-like algorithm, resulting in a vector. Subsequently, model selection takes place after entering vectors with different presets in an artificial neural network to train and evaluate the results. Next, a ‘judge model’ is formed, which is a combination of models for optimizing performance. Finally, two types of cross-validation, being ‘leave-10%-out’ cross-validation and ‘bagging’, are used when recalculating the judge models. These judge models are subsequently used to classify new, blind measurements. Discussion. Data analysis in eNose technology is principally based on generating prediction models that need to be validated internally and externally for eventual use in clinical practice. This paper describes the analysis of big data, captured by eNose technology in lung cancer. This is done by means of generating prediction models with Aethena, a data analysis program specifically developed for analysing VOC data.

A set of 15 Italian (Lolium multiflorum), three hybrid (L. boucheanum) and 33 perennial ryegrass (L. perenne) cultivars were evaluated for their rust susceptibility in the field. The cultivars were grown in 2001, 2004 and 2007 at 29, 32 and 27 European sites, respectively. Rust incidence was scored during different growth cycles using a scale from one (no rust) to nine (susceptible). Crown rust (Puccinia coronata f. sp. lolii) was the most frequently observed rust species on Italian, hybrid and perennial ryegrass during all three experimental years. There was a highly significant difference in mean crown rust scores among Italian and perennial ryegrass cultivars, respectively, in each of the 3 years. The ranking of the cultivars was very consistent across the different sites within each year (with some exceptions), despite a significant interaction of cultivar response with sites. Moreover, the rank order correlations of mean cultivar rust scores were highly significant (r > 0.9) between the years. Stem rust (P. graminis f. sp. graminicola) occurred almost exclusively on perennial ryegrass. The cultivars showed significant variation in stem rust susceptibility. Rank order correlations of mean cultivar stem rust scores were mostly significant between sites within each year and highly significant between the years. The ranking of perennial ryegrass cultivars was different depending on whether crown or stem rust was scored. However, the rank order correlation between the mean disease scores of the cultivars for the two pathogens was low but still significant. At a particular site the assessment of crown or stem rust was in most cases very consistent over the period of experimentation. Therefore there was no evidence that rust resistance of an individual cultivar was overcome by the rust pathogen at a particular site over the 7 years of experimentation.