Multiscale Three‐Dimensional Features and Spatial Feature Evaluation of Human Pulmonary Tuberculosis

ABSTRACT The low detection rate of Mycobacterium tuberculosis in clinical practice leads to a high rate of missed diagnoses for pulmonary tuberculosis (PTB). This study aimed to assess the imaging and pathological characteristics of PTB lesions from different multiple dimensions, with a focus on evaluating their three‐dimensional(3D) and spatial features. This study employed multiple methods to evaluate the three‐dimensional characteristics of PTB. CT was used to visually assess the density and spatial positioning of PTB lesions, and acid‐fast staining was used to evaluate the two‐dimensional histological features of PTB. Using fMOST technology, a total of 2399 consecutive single‐cell resolution images of human PTB tissue were obtained. These images were subsequently reconstructed in 3D to evaluate the pathological characteristics of PTB in three dimensions. The 3D imaging precisely extracted the distribution of different CT values (HU values) and accurately obtained the spatial location information of the lesions, achieving precise localization. Using fMOST technology, we clearly identified the microscopic structures within both normal lung tissue and PTB lesions, revealing the loose structure, continuous alveolar septa, and clearly visible blood vessels of normal lung tissue. In contrast, typical characteristics of PTB lesions included the destruction of normal lung structure, tissue proliferation, necrosis, and inflammatory infiltration, with a significant increase in overall density. 3D observations of the necrotic areas showed high tissue density but low cellular density, primarily composed of necrotic tissue, consistent with the histological characteristics commonly seen in PTB lesions. This enhanced our understanding of the spatial distribution of PTB lesions. The 3D visualization of imaging and pathology enables a more comprehensive identification of the pathological features of PTB lesions. The multiscale model based on the fMOST system provides more detailed structural information and displays the spatial distribution of lesions more accurately. This is particularly beneficial in the evaluation of complex lesions, demonstrating its potential for optimizing diagnostic methods and supporting clinical decision‐making.