Reliability of non-contact tongue diagnosis for Sjögren's syndrome using machine learning method

Abstract

Sjögren's syndrome (SS) is an autoimmune disease characterized by dry mouth. The cause of SS is unknown, and its diverse symptoms make diagnosis difficult. The Saxon test, an intraoral examination, is used as the primary diagnostic method for SS, however, the risk of salivary infection is problematic. Therefore, we investigate the possibility of diagnosing SS by non-contact and imaging observation of the tongue surface. In this study, we obtained tongue photographs of 60 patients at the Tsurumi University School of Dentistry outpatient clinic to clarify the relationship between the features of the tongue and SS. We divided the tongue into four regions, and the color of each region was transformed into CIE1976L*a*b* space and statistically analyzed. To clarify experimentally the possibility of SS diagnosis using tongue color, we employed three machine-learning models: logistic regression, support vector machine, and random forest. In addition, we constructed diagnostic prediction models based on the Bagging and Stacking methods combined with three machine-learning models for comparative evaluation. This analysis used dimensionality compression by principal component analysis to eliminate redundancy in tongue color information. We found a significant difference between the a* value of the rear part of the tongue and the b* value of the middle part of the tongue in SS and non-SS patients. In addition to the principal component scores of tongue color, the support vector machine was trained using age, and achieved high accuracy (71.3%) and specificity (78.1%). The results indicate that the prediction of SS diagnosis by tongue color reaches a level comparable to machine learning models trained using the Saxon test. This is the first study using machine learning to predict SS diagnosis by non-contact tongue observation. Our proposed method can potentially support early SS detection simply and conveniently, eliminating the risk of infection at diagnosis, and it should be validated and optimized in clinical practice.

Figure 1

Image of the Tongue Image Analyzing System (TIAS). The TIAS used in this study equipped with a chin rest and a forehead rest for fixing the face of the subject.

Figure 2

Definition of the areas. (1) Red corresponds to the tongue edge, (2) Green corresponds to the tongue posterior, (3) Blue corresponds to the tongue middle, and (4) Orange corresponds to the tongue apex.

Figure 3

The basis of SVM classification.

Figure 4

The basis of RF classification.

Figure 5

Ensemble of machine learning algorithms. (a) Bagging: Train several classifiers for each feature and combine the results by the soft voting. (b) Stacking: Use the predictions from multiple weak classifiers to train a model.

Figure 6

Comparison of the tongues of SS patients to those of non-SS patients. (a) a1-4 and (b) b1-4 indicate the a* and the b* of each tongue area defined in Fig. 2.

Figure 7

Result of the principal component analysis (PCA) of the tongue color. (a) Cumulative contribution rate of the principal components (PCs). (b) The distribution of factor loadings for the first and second principal components of the 12 color values is shown, where L, a, and b represent L*, a*, and b* values, and the numbers represent tongue regions (1:edge, 2:posterior, 3:middle, 4:apex).