Título: OPTICAL BIOPSY AS POINT OF CARE FOR DIABETES DIAGNOSIS
Nome do Apresentador: Sara Maria Santos Dias da SILVA
Categoria do Trabalho: Painel de pesquisa científica (PPC)
Área Temática: Estomatologia
Resumo: INTRODUCTION: Diabetes is a multifactorial endocrine disease that can have several consequences if not diagnosed at an early stage. Currently, the gold standard diagnostic method is the fasting blood glucose and glycated hemoglobin test, both with a delay of approximately 7 days for diagnosis. FT-IR spectroscopy can be used to aid the diagnosis of patients with real-time results. OBJECTIVE: this study aimed to obtain spectral signatures of saliva for comparation between diabetic and healthy patients. STUDY DESIGN: The study consisted of a sample n of 40, with 20 healthy patients (control group) and 20 patients with diabetes. For the analysis of this saliva, a Bruker ATR-FTIR spectrometer equipped with a diamond crystal was used. Data analysis was performed in a MATLAB routine, the spectral regions analyzed were the "fingerprint" between 600-1800 cm-1. After data pre-processing statistical methods RESULTS: statistical methods involving artificial intelligence and machine learning were used. The results showed that after the machine learning technique - Suppor Vector Machine was the one that best contributed to discrimination between samples with 86% specificity and 92% sensitivity. CONCLUSION: We conclude that the technique was effective for group discrimination, and can contribute to its solidification as a point of care diagnosis.
Autor 1: Sara Maria Santos Dias da SILVA
E-mail 1: [email protected]
Autor 2: Jaqueline Maria Brandão Rizzato
E-mail 2: [email protected]
Autor 3 : Maria Clara de Moura Santos Coelho Ferreira
E-mail 3: [email protected]
Autor 4: Monique Vieira Furukawa
E-mail 4: [email protected]
Autor 5: Emanuel da Silva Rovai
E-mail 5: [email protected]
Autor 6: Marcelo Saito Nogueira
E-mail 6: [email protected]
Autor 7: Luis Felipe das Chagas e Silva de Carvalho
E-mail 7: [email protected]
Para baixar o aplicativo, escolha abaixo: