Screening is a key tool for early cancer detection/prevention and potentially saves lives. Oral mucosal vascular aberrations and color changes have been reported in hereditary nonpolyposis colorectal cancer patients, possibly reflecting a subclinical extracellular matrix abnormality implicated in the general process of cancer development. Reasoning that physicochemical changes of a tissue should affect its optical properties, we investigated the diagnostic ability of oral mucosal color to identify patients with several types of cancer. A total of 67 patients with several histologically proven malignancies at different stages were enrolled along with a group of 60 healthy controls of comparable age and sex ratio. Oral mucosal color was measured in selected areas, and then univariate, cluster, and principal component analyses were carried out. Lower red and green and higher blue values were significantly associated with evidence of cancer (all P<0.0001), and efficiently discriminated patients from controls. The blue color coordinate showed significantly higher sensitivity and specificity (96.66±2.77 and 97.16±3.46%, respectively) compared with the red and green coordinates. Likewise, the second principal component coordinate of the red–green clusters discriminated patients from controls with 98.2% sensitivity and 95% specificity (cut-off criterion≤0.4547; P=0.0001). The scatterplots of the chrominances revealed the formation of two well separated clusters, separating cancer patients from controls with a 99.4% probability of correct classification. These findings highlight the ability of oral color to encode clinically relevant biophysical information. In the near future, this low-cost and noninvasive method may become a useful tool for early cancer detection.