Abstract Background Contact Endoscopy is a non invasive tool to visualise alterations in cell architecture in vivo. In this study we investigated the diagnostic accuracy of Contact Endoscopy in detecting malignancy in oral mucosal lesions. Methods 76 patients with oral mucosal lesions requiring biopsy were included. Contact Endoscopy was performed by Otolaryngologist before biopsy and findings recorded. The lesion was then biopsied and sent for histopathological examination by Pathologist who was blinded to Contact Endoscopy findings. Findings of Contact Endoscopy were compared with histopathological findings taking the latter as the gold standard. Two biopsies were reported as ‘inconclusive’ on histopathological examination and hence excluded from the final analysis. Data of the remaining 74 patients is presented here. Results Clinically lesions were diagnosed as ulcero-proliferative lesions in 34 patients, Leukoplakia in 19, Erythroplakia in 9, Lichen planus in 5 and Submucous fibrosis in 7 patients. Histopathological examination revealed presence of malignancy in 97.06% of ulcero-proliferative mucosal lesions, 10.53% of leukoplakia and 33.33% of erythroplakia while corresponding figures on Contact Endoscopy were 94.12%, 5.26% and 11.11% respectively. No malignancy was detected in lichen planus and submucous fibrosis by either technique. When compared with histopathological examination, CE showed sensitivity of 84.21%, specificity of 94.44% and accuracy of 89.19%. No adverse effects on the patients were seen due to the procedure or stain. Conclusion Contact Endoscopy may be useful in determining cellular structure in vivo without biopsy to detect oral malignancy early. Further studies are suggested.