OBJECTIVE: To evaluate the relationship between purple-bluish tongue and platelet counts, and further to examine their associations with the recurrence of epithelial ovarian cancer. METHODS: A total of 82 epithelial ovarian cancer patients were enrolled in this study. Cluster analysis was used for grouping patients' P(RGB) (Red-R; Green-G; Blue-B; Average percentage of RGB, P(RGB)) values. Receiver operating characteristic (ROC) curve was performed for detecting the diagnostic standard of purple-bluish tongue. Χ2 test was used to assess the relationship between purple-bluish tongue and platelet counts, and the recurrence of epithelial ovarian cancer. The perioperative (preoperative) platelet level was examined with tongue image and disease recurrence. RESULTS: Tongue images were classified into two groups basing on P(RGB) values of images by cluster analysis. The numbers of cases in cluster "1" (normal color tongue) was 16 and cluster "2" (purple-bluish tongue) was 66. Two groups of P(RGB) values, classified by cluster analysis, were significantly correlated with vision-based tongue color recognition (Kappa = 0.852, P < 0.001). ROC curve showed that the ratio of P(B) to P(R) had the highest diagnostic value. The sensitivity and the specificity of the ratio of P(B) to P(R) were 95.3% and 88.9% respectively and the optimal cut-off point was 0.71. Purple-bluish tongue was significantly correlated with increased platelet counts (P < 0.001). Both the increased platelet counts (P = 0.01) and purple-bluish tongue were associated with recurrence of epithelial ovarian cancer (P < 0.001). CONCLUSION: The ratio of P(B) to P(R) greater than 0.71 could serve as an indicator for purple-bluish tongue diagnosing used in symptom pattern identification in Traditional Chinese Medicine. Purple-bluish tongue, associated with increased platelet counts, was also closely correlated with the recurrence of epithelial ovarian cancer.
OBJECTIVE: To evaluate the relationship between purple-bluish tongue and platelet counts, and further to examine their associations with the recurrence of epithelial ovarian cancer. METHODS: A total of 82 epithelial ovarian cancerpatients were enrolled in this study. Cluster analysis was used for grouping patients' P(RGB) (Red-R; Green-G; Blue-B; Average percentage of RGB, P(RGB)) values. Receiver operating characteristic (ROC) curve was performed for detecting the diagnostic standard of purple-bluish tongue. Χ2 test was used to assess the relationship between purple-bluish tongue and platelet counts, and the recurrence of epithelial ovarian cancer. The perioperative (preoperative) platelet level was examined with tongue image and disease recurrence. RESULTS: Tongue images were classified into two groups basing on P(RGB) values of images by cluster analysis. The numbers of cases in cluster "1" (normal color tongue) was 16 and cluster "2" (purple-bluish tongue) was 66. Two groups of P(RGB) values, classified by cluster analysis, were significantly correlated with vision-based tongue color recognition (Kappa = 0.852, P < 0.001). ROC curve showed that the ratio of P(B) to P(R) had the highest diagnostic value. The sensitivity and the specificity of the ratio of P(B) to P(R) were 95.3% and 88.9% respectively and the optimal cut-off point was 0.71. Purple-bluish tongue was significantly correlated with increased platelet counts (P < 0.001). Both the increased platelet counts (P = 0.01) and purple-bluish tongue were associated with recurrence of epithelial ovarian cancer (P < 0.001). CONCLUSION: The ratio of P(B) to P(R) greater than 0.71 could serve as an indicator for purple-bluish tongue diagnosing used in symptom pattern identification in Traditional Chinese Medicine. Purple-bluish tongue, associated with increased platelet counts, was also closely correlated with the recurrence of epithelial ovarian cancer.