RESUMO
Tuberculosis (TB) remains an impactful infectious disease, leading to millions of deaths every year. Mycobacterium tuberculosis causes the formation of granulomas, which will determine, through the host-pathogen relationship, if the infection will remain latent or evolve into active disease. Early TB diagnosis is life-saving, especially among immunocompromised individuals, and leads to proper treatment, preventing transmission. This review addresses different approaches to diagnosing TB, from traditional methods such as sputum smear microscopy to more advanced molecular techniques. Integrating these techniques, such as polymerase chain reaction (PCR) and loop-mediated isothermal amplification (LAMP), has significantly improved the sensitivity and specificity of M. tuberculosis identification. Additionally, exploring novel biomarkers and applying artificial intelligence in radiological imaging contribute to more accurate and rapid diagnosis. Furthermore, we discuss the challenges of existing diagnostic methods, including limitations in resource-limited settings and the emergence of drug-resistant strains. While the primary focus of this review is on TB diagnosis, we also briefly explore the challenges and strategies for diagnosing non-tuberculous mycobacteria (NTM). In conclusion, this review provides an overview of the current landscape of TB diagnostics, emphasizing the need for ongoing research and innovation. As the field evolves, it is crucial to ensure that these advancements are accessible and applicable in diverse healthcare settings to effectively combat tuberculosis worldwide.
RESUMO
In recent decades, many studies have been published on the use of automatic smear microscopy for diagnosing pulmonary tuberculosis (TB). Most of them deal with a preliminary step of the diagnosis, the bacilli detection, whereas sputum smear microscopy for diagnosis of pulmonary TB comprises detecting and reporting the number of bacilli found in at least 100 microscopic fields, according to the 5 grading scales (negative, scanty, 1+, 2+ and 3+) endorsed by the World Health Organization (WHO). Pulmonary TB diagnosis in bright-field smear microscopy, however, depends upon the attention of a trained and motivated technician, while the automated TB diagnosis requires little or no interpretation by a technician. As far as we know, this work proposes the first automatic method for pulmonary TB diagnosis in bright-field smear microscopy, according to the WHO recommendations. The proposed method comprises a semantic segmentation step, using a deep neural network, followed by a filtering step aiming to reduce the number of false positives (false bacilli): color and shape filtering. In semantic segmentation, different configurations of encoders are evaluated, using depth-wise separable convolution layers and channel attention mechanism. The proposed method was evaluated with a large, robust, and annotated image dataset designed for this purpose, consisting of 250 testing sets, 50 sets for each of the 5 TB diagnostic classes. The following performance metrics were obtained for automatic pulmonary TB diagnosis by smear microscopy: mean precision of 0.894, mean recall of 0.896, and mean F1-score of 0.895.