Human organisms have a diverse benefic microbial population which protects the organism from diseases. Once the homeostasis is unbalanced, it can cause infections that comprise the human well being. It is the cause of tonsillitis, an inflammation of the tonsils, commonly caused by bacteria inducing respiratory infection. Standardized laboratory technique to identify bacterial infection is phenotyping, a technique that may take many days to release concrete results.

The objective of this project is to differentiate bacterial genera/species, isolated from child with tonsilitis, as well as to identify the antimicrobial resistance of these isolated, using hyperspectral images. To achieve that, the bilateral cooperation between the research groups from Brazil and Italy is essential in order to combine the expertise from biological experiments to the SWIR-HSI and computational analysis.

The brazilian group, led by prof. Dr. Lilian Carneiro at Instituto de Patologia Tropical e Saúde Pública (IPTSP) of Federal University of Goiás (UFG) and prof. Dr. Clarimar J. Coelho at Pontifical Catholic University (PUC Goiás) will provide the biological experiments and SWIR-HSI imaging acquisition from different bacteria. The italian group, led by prof. Dr.  Claudio Cusano (UNIPV) will collaborate with the SWIR-HSI imaging analysis and computational models to differentiate bacterial genera/species. This collaborative force intends to provide an epidemiological study of the microorganisms from the upper respiratory tract and its antimicrobial resistance in order to indicate fast and efficient medical treatment. 

Antibiotic therapy is the most frequent solution to control the infection, Cochrane Database analysis from 2013 showed the effectiveness of different antibiotics in acute (Streptococcal) tonsillitis (Van Driel, 2013). Penicillin and its drug generations still an effective solution for the tonsillitis treatment. Although, a not precise antibiotics treatment reduces the efficacy and result in more critical side effects (Chiappini, 2012 e Chiappini, 2011). Therefore, a precise diagnostic of the pathogen is very important to guide doctors to indicate the ideal antibiotic treatment.    

The laboratory routine to identify precisely the pathogen is timed and resourceful costly. Faced with this problem, the diagnostic methods do not meet the speed and precision needed to minimize patient discomfort, especially if they are children. There are currently different methods for bacterial identification, such as classical manual methodology, PCR (DNA polymerase chain reaction), automation (Vitek and Maldi-tof) (Manaka et al., 2017; Febbraro et al., 2016). In general, manual diagnosis is slow and subject to inaccuracies, (Martinez and Tadei, 2005). However, in the last years the diagnosis by image analysis has reached a high level of quality (Ganguly, et al., 2010). Modern imaging applies computer-aided diagnosis, determining accuracy and speed in the release of results, cautioning strategies for health promotion and disease prevention (Doi, 2014).