AI Methods for Lung Disease Diagnosis: A Comprehensive Survey

Continuous exposure from breathing in coal dust, and other hazardous airborne particulates generally has a way of causing many serious lung diseases such as pneumoconiosis and chronic obstructive pulmonary disorder (COPD) in coal miners. Thus, identifying risks as early as we can is essential, so that remedial or preventative steps protecting at risk employees can be taken. To find out the risk of these diseases in coal miners, this paper reviews 75 studies that explore various predictive methods. The studies reviewed use a range of research designs, from the conventional statistical models to more advanced machine learning algorithms, including neural networks, logistic regression, support vector machines, and random forests. The target diseases are primarily COPD and pneumoconiosis, and the data sources are clinical test data, environmental exposure records, and biochemical indicators in blood. The amount of time a miner faces riskier environments, specific characteristics of their environment (like angle of the coal seams), the measurements they collected on their lung activity (like flow-volume loops and vital capacity), and biological indicators (like cortisol levels, and plasma antioxidant/reductant activity) would need to be addressed to make disease predictions. The models that were reviewed were 70% to 91% accurate. However, many of the models specified were developed using the original datasets, and consequently are not applicable to a larger audience. Additionally, the review highlights major deficiencies in moving forward by addressing the limitations of comparing multiple data pertaining to diverse components of their health and working contexts. There is a need for valid predictive models that are potentially generalizable to larger groups [11],[12]. By synthesizing clinical data with both environmental exposure and biochemical data, lung health may be preserved for coal miners; and the potential to improve the detection of early disease is promising.