This study rapidly reviews deep learning applications for tuberculosis analysis in computer tomography (CT) images. A total number of 16 articles were carefully selected from the topmost databases. The databases used included IEEE Xplore, Web of Science, PubMed and Science Direct. The search strategy is developed using a combination of keywords. The data extracted from the article are publication type, data sources, number of samples used, tasks performed, deep learning method used, and the performance evaluation metrics measured. The performance evaluation metrics from the selected papers were compared. The evaluation results from the metrics showed the highest accuracy (ACC) and Area under curve (AUC) of 99.20% and 99.00% respectively. The highest sensitivity (SEN) is 96.40%, while the highest specificity (SPEC) is 81.24%. All the reviewed papers employed deep learning methods for detection, segmentation, and classification, while other studies combined both segmentation and classification tasks.