Discrimination of influenza virus-infected nasal fluids by Vis-NIR spectroscopy.

BACKGROUND

Influenza patients show a severe condition of the respiratory tract with high temperature. Efficient treatment of influenza requires early use of oseltamivir, and thus rapid diagnosis is needed. Recently, rapid diagnostic methods such as immunochromatography have been developed; however, immunochromatography is not an optimal technique because it is relatively expensive and has low sensitivity.

METHODS

Visible and near-infrared (Vis-NIR) spectroscopy in the region 600-1100 nm, combined with chemometrics analysis such as principal component analysis (PCA) or soft modeling of class analogy (SIMCA), was used to develop a potential diagnostic method for influenza based on nasal aspirates from infected patients.

RESULTS

The Vis-NIR spectra of nasal aspirates from 33 non-influenza patients and 34 influenza patients were subjected to PCA and SIMCA to develop multivariate models to discriminate between influenza and non-influenza patients. These models were further assessed by the prediction of 126 masked measurements [30 from non-influenza patients, 30 from influenza patients and 66 from patients infected with respiratory syncytial virus (RSV)]. The PCA model showed some discrimination of the masked samples. The SIMCA model correctly predicted 29 of 30 (96.7%) non-influenza patients, and 30 of 30 (100%) influenza patients from the Vis-NIR spectra of masked nasal aspirate samples. Nasal aspirates of RSV-infected patients were predicted as 50% non-influenza and 50% influenza by the SIMCA model, suggesting that discrimination between patients infected with influenza virus and those infected with RSV was difficult.

CONCLUSIONS

Although the study sample was small and there was difficulty in discriminating between influenza virus and RSV infection, these results suggest that Vis-NIR spectroscopy of nasal aspirates, combined with chemometrics analysis, might be a potential tool for diagnosis of influenza.