Tongue diagnosis system for quantitative assessment of tongue coating in patients with functional dyspepsia: A clinical trial

Ethnopharmacological relevance Tongue diagnosis is a significant procedure to examine the physiological and pathological changes of the human body in oriental medicine. However, the conventional method of tongue diagnosis including direct observation of tongue has limitations because of various external factors and subjective factors. Therefore, the current study investigated the usefulness of the tongue diagnosis system (TDS) as a diagnostic tool for evaluating tongue coating thickness (TCT) by assessing the agreement between the TDS and a gold standard established by assessors using the conventional method. Materials and methods The present study was designed as a prospective clinical trial including 60 patients with functional dyspepsia. The TDS was used to capture tongue images twice within a 30-min interval to assess its reproducibility. Tongue coating percentage was measured by the TDS, and TCT was classified as either no coating, thin coating, or thick coating according to the existing diagnostic criteria. After both TDS examinations, the weight of tongue coating was quantitatively measured, and the correlation between the actual quantity of tongue coating and the percentage of the coating measured by the TDS was analyzed using Pearson's correlation. After collecting tongue images in all 60 patients, tongue coating was evaluated using a conventional method by 5 well-trained assessors to establish the gold standard for evaluating TCT, which allowed us to assess the diagnostic agreement between the TDS and the gold standard. After 2 weeks, TCT evaluation was repeated by the same assessors using the same images but in a random order. Results The agreement between the TDS and the gold standard for evaluating TCT was almost perfect (weighted kappa, 0.840), as was the reproducibility of the TDS (weighted kappa, 0.851). The percentage of tongue coating measured by the TDS was significantly correlated with the weight of tongue coating (r=0.442, p <0.001). The levels of intra-rater reliability ranged from substantial to almost perfect (range of weighted kappa, 0.777-0.923). The inter-rater reliability of 5 assessors was moderate (weighted kappa, 0.563). Conclusions The present study demonstrated that the TDS can be used as a diagnostic tool for the objective and standardized evaluation of TCT in actual clinical practice.