Diagnóstico de anemia usando o smartphone

Abstract

INTRODUCTION: Anemia is a public health problem that affects approximately one third of the world's population. OBJECTIVES: Create a way to analyze hemoglobin levels using a smartphone by taking photos of the palpebral conjunctiva using an app, camera and smartphone flashlight, demonstrating the possibility of a non-invasive diagnosis of anemia. MATERIAL AND METHODS: 19 patients treated at the Santa Casa de Misericórdia de São Carlos, at the UFSCar University Hospital, and at the Dr. Maricondi Medical Laboratory who underwent a blood count test were invited to participate in the research. We captured a photograph of each patient's palpebral conjunctiva using the RGB Color Detector application with the smartphone's rear camera with the “flash” activated. After obtaining the photo, we selected the central region of the palpebral conjunctiva and quantitatively measured the color intensity of the RGB components (red, green and blue). Then, with the intention of filtering the signal obtained, reducing light interference from the environment, we subtracted the green channel from the red channel, since the green channel has most of the reflections and lighter colored noise, and the red channel is the one we want to analyze, since hemoglobin is a red pigmented molecule. RESULTS: For each blood count there was a respective image, which we compared the measured color intensity with hemoglobin values, then we compared whether there is a proportional relationship between color intensity x hemoglobin value. We analyzed the three different channels (red, green and blue) and realized that comparing only the color intensity of the red channel and the respective hemoglobin value there was no directly proportional relationship, however, when we subtracted the green channel from the red channel, the relationship between the colorimetric signal and the hemoglobin value becomes directly proportional, demonstrating that the green channel possibly contains most of the noise, such as light reflections, which are clearer signals. After defining the Cut-off at 93 color intensity, we obtained sensitivity = 0.875, specificity = 1, accuracy = 0.947. CONCLUSION: Our method of non-invasive analysis of anemia using just a photograph of the palpebral conjunctiva and a smartphone application showed promise, and future steps will involve an analysis of a larger number of patients, other types of colorimetric analyzes to obtain a more accurate signal. pure, as well as the creation of a smartphone application capable of instantly providing the presence of anemia with a simple photo.