Role of Artificial Intelligence for Diagnosing Tuberculosis

Tuberculosis (TB) is a constant illness brought about by the microbes Mycobacterium TB (MTB) and is a transferable sickness that spreads starting with one individual then onto the next through air. It largely influences the lungs; hence itis also known as pulmonary tuberculosis (PTB); however, it can likewise influence other body parts, for example, brain, bones, intestines, skin, kidneys, or the spine—hence, it is also referred to as extra-pulmonary tuberculosis (EPTB). Hereditary investigations have indicated that MTB has existed for around 15,000 years. TB is a deadly sickness. World Health Organization (WHO), in 2018, introduced insights that expressed around 10 million individuals endured with TB, and about 1.5 million lost their lives because of this infection in 2018. In 2018, around 10 million individuals were affected by TB around the world. Starting at 2018, India represents the world’s most noteworthy number of individuals experiencing the infection. In 2018, 21.5 lakh TB cases were found in India, out of which 25% were from the private segment. Eighty-nine percent of the TB cases are found in the age bunch between 15 and69 years. For a low-asset setting like India, there is a critical need to have reasonable, convenient, and quick tests for TB diagnosis. Thus, various computer-aided technologies can be used for the accurate and timely detection of this disease. Lately, with the fast advancement of data innovation and the developing regard for interdisciplinary practices, artificial intelligence (AI) has become another zone of enthusiasm for clinical experts. A systematic literature review has been conducted in all the relevant areas to analyze the role of technology for the diagnosis of TB. The vast majority of the examination works have been done to analyze PTB. There is no specific framework structured so far to distinguish and analyze EPTB, which accounts for 16% of 2.1 million cases in 2013, comparing to 336,000 individuals with EPTB. Moreover, in the literature, researchers had taken less symptoms and attributes to diagnose PTB. However, there are a number of other symptoms that play a vital role in the detection of complete TB. The restrictions of the current frameworks rely upon the less accessibility of the dataset. With the increase in the dataset, the sensitivity and the accuracy of the systems designed so far decrease. A lot of work has been carried out for the automatic detection of TB but still there is less literature available for the diagnosis of EPTB. The work that has been done so far is not being utilized by clinicians or clinical professionals in real life. This can be done with coordinating AI-based strategies with Internet of Things (IoT) so that patients can be checked by clinicians, in actuality.