The role of imaging techniques in left double lumen tube applications: A diagnostic agreement study

This study aimed to evaluate the accuracy of left double-lumen tube (LDLT) size selection based on tracheal diameter measured by ultrasound (USG), tracheal diameter measured by computed tomography (CT), and left main bronchus (LMB) diameter measured by CT, using dimensions suggested in the literature, and to determine whether there were significant differences in accuracy among these methods. One hundred patients requiring LDLT intubation were analyzed. LDLT sizes were determined using three different measurement methods: CT measurement of tracheal diameter, CT measurement of LMB diameter, and USG measurement of tracheal diameter. The accuracy rates and agreement of the LDLT sizes determined by these methods were analyzed. Tracheal diameter measurements obtained by USG (18.10 ± 2.97 mm) and CT (18.65 ± 3.34 mm) were comparable, with no statistically significant difference ( p = 0.237). Exact size-matching rates were relatively low, with clinically appropriate predictions achieved in 13% of cases using tracheal USG, 28% using LMB-CT, and 32% using tracheal CT measurements. For exact size agreement, unweighted kappa values indicated agreement no better than chance (unweighted kappa: − 0.031 for tracheal USG, 0.089 for tracheal CT, and 0.094 for LMB-CT). However, applying a clinically acceptable margin (± 1 size) markedly improved prediction accuracy to 87% for tracheal CT, 77% for LMB-CT, and 52% for tracheal USG. In the weighted kappa analysis, agreement levels increased; tracheal USG (κ = 0.32; p < 0.001), LMB-CT (κ = 0.42; p < 0.001), and tracheal CT (κ = 0.56; p < 0.001) showed moderate agreement. USG provides tracheal diameter measurements similar to those obtained by CT. Therefore, USG may be a practical method for measuring tracheal diameter when CT is unavailable but should be regarded primarily as a complementary or screening tool rather than a stand-alone method. Imaging-based methods showed limited accuracy and poor agreement for exact LDLT size selection, although they may be useful for identifying a clinically acceptable size range. However, methodological standardization and refinement of threshold values are needed to improve their ability to predict the exact tube size.