In this study, we compared 103 OralCDx results with the histological findings of 96 clinical sites in 80 patients (33 females; 64.3+/-13.7 years and 47 males; 53.2+/-11.5 years). The histological findings were classified as follows: compatible with oral leukoplakia (OL; n = 60) or oral lichen planus (OLP; n = 17), both without dysplasia; dysplasia in OL or OLP (n = 9); and oral squamous cell carcinoma (OSCC; n = 17). There were seven (6.8%) specimens with an inadequate cell count. Overall, the sensitivity of the OralCDx technique to detect dysplasia and OSCC was 92.3% (95% CI: 74.9-99.1%), and the specificity was 94.3% (95% CI: 86.0-98.4%). The positive likelihood ratio (LR+) was 16.2 (95% CI: 6.2-42.1) and the negative likelihood ratio (LR-) was 0.08 (95% CI: 0.02-0.31). In conclusion, these figures are in agreement with previously published data and support the use of OralCDx as a screening tool of oral lesions, but further trials are still necessary.
