Autofluorescence-guided surveillance for suspicious lesions of the oral cavity: Identification of biomarker profiles and chemopreventive targets
Frustino, Jennifer Limina
MetadataShow full item record
Problem under Investigation: Prevention and early detection of oral and oropharyngeal cancers is the key to timely and effective treatment, decreased morbidity, and increased survival. Although the oral cavity is easy to examine, the 5-year survival rate is only around 50% when these cancers are detected late. New modalities like autofluorescence visualization (AFV) have emerged for early detection. It is important to validate such devices and identify relevant biomarkers and targets for chemoprevention. Objectives/Hypotheses: It is hypothesized that the addition of AFV to conventional white light examination (WLE) of the oral cavity will increase the sensitivity and specificity of detecting oral potentially malignant disorders (OPMDs) and cancers in a high-risk cohort and sensitivity and specificity may vary among anatomical subsites. It is also hypothesized that known biomarkers in head and neck cancer (p53, p16, ki-67) and the chemopreventive targets vitamin D receptor (VDR) and epidermal growth factor receptor (EGFR) will change expression with increasing grade of tissue pathology. Previous Findings/ Preliminary Results: In the first 60 patients, the addition of AFV provided greater sensitivity than WLE alone in detecting low-grade lesions (LGLs) (75% vs. 44%), high-grade lesions (HGLs) (100% vs. 71%) and cancers (100% vs. 80%). From this cohort, 29 patients had biopsies from the oropharynx subsite. The addition of AFV+WLE improved sensitivity in detecting LGLs (95% vs. 35%) and HGLs (100% vs. 75%) compared to WLE alone. Research Design/Materials Methods: Over 170 high-risk patients have undergone WLE and AFV screening with biopsies and histopathologic diagnosis. Tissue was stratified based on histologic grade and underwent immunohistochemistry for known biomarkers: p53, p16, ki-67, VDR, and EGFR. Methods of Data Analysis: Sensitivity, specificity, negative and positive predictive values for the addition of AFV to WLE was analyzed and compared over all sites and subsites. Biomarker expressions were quantified and analyzed across histologic groups using Fisher's exact test. ROC analysis was used to determine the AUC for biomarkers predicting histologic grade. Results: In 169 patients who provided 546 biopsies, the addition of AFV to WLE showed increased sensitivity in identifying OPMDs and cancers. WLE alone vs. AFV alone vs. AFV+ WLE detected 30% vs. 63% vs. 71% of LGLs respectively. WLE alone vs. AFV alone vs. the addition of AFV+ WLE detected 52% vs. 85% vs. 88% of HGLs respectively. WLE alone vs. AFV alone vs. the addition of AFV+ WLE detected 66% vs. 94% vs. 97% of cancers respectively. The specificity for the addition of AFV to WLE remained at 47% for all lesion grades. This was a reduction from 54% for AFV alone and 77% for WLE alone across all three histology grades. There was a statistically significant difference in the sensitivity of WL detection of LGLs in the oral tongue (21%) when compared to the gingiva (50%, p = 0.01), floor of mouth (47%, p = 0.03) and RMT/alveolar ridge (48%, p = 0.007). For AFV detection of LGLs there was a statistically significant difference when comparing the oral tongue (56%) to the gingiva (88%, p = 0.01) and the oropharynx (81%, p = 0.03). The gingiva (88%) was significantly different from the buccal mucosa (62%) on AFV detection of LGLs (p = 0.04). For the sensitivity of WL+AFV detection of LGLs, the oral tongue (63%) was significantly different from the gingiva (88%, p = 0.04) and the oropharynx (95%, p = 0.001). The buccal mucosa (66%) was different from the oropharynx (95%, p = 0.004). The oral tongue (80% HGLs, 88% cancers) differed significantly from the buccal mucosa (100%, p = 0.02) and the oropharynx (100%, p = 0.03) when using AFV alone. Increased Ki-67 expression was associated with a worse histologic grade when adjusting for relevant covariates (OR = 1.36, [1.06-1.68], p = 0.014). Based on ROC analysis, Ki-67 was a significant predictor of histologic grade both when evaluated alone and in combination with other biomarkers. Potential Significance: AFV may provide enhanced detection of OPMDs and cancers that may have been missed on WLE alone. The oral tongue appears to have the lowest sensitivity for WL, AFV, and WL+AFV detection of LGLs and the lowest sensitivity for the AFV detection of HGLs and cancers when compared to other oral subsites. This may have important implications in the utility of AFV screening for the oral tongue. Ki-67 may serve as a prognosticator for increasing histologic grade. Potential biomarker profiles may be identified to predict histologic progression and identify chemopreventive targets.