Effect of Ultrasound Images Based on Filtering Algorithm on the Diagnosis Efficiency of Cervical Lymph Node Tuberculosis

Objective: This paper uses filtering algorithms to explore and analyze the characteristics of cervical tuberculous lymph nodes and metastatic lymph nodes in ultrasound sonograms. Methods: 160 lymph nodes of 105 patients who obtained pathological results were retrospectively analyzed in this paper, including 76 tuberculous lymph nodes and 84 metastatic lymph nodes. The following ultrasound findings of lymph nodes in the two groups were compared: L/S, border, lymphatic door, internal echo, calcification, necrosis and liquefaction, and their characteristics were analyzed. Results: Both calcifications and liquid echoless areas were present in both groups. The internal calcifications of metastatic lymph nodes were mostly micro-calcifications, which were clustered, and the tuberculous lymph nodes were mostly scattered coarse calcifications. The difference was statistically significant (P < 0.01). The liquefaction area in metastatic lymph nodes was mostly eccentric and the area ratio was less than 50%. The liquefaction area in tuberculous lymph nodes was mostly in the center and the area ratio was more than 50%. The difference was statistically significant (P < 0.01). Conclusion: Ultrasound imaging features such as the size and distribution of calcification, and the location and area ratio of liquefaction zone are helpful to distinguish tuberculous lymph nodes from metastatic lymph nodes.