Abstract A15: Autofluorescence-guided detection of oropharyngeal neoplasms and precursor lesions

Introduction : The incidence of oropharyngeal (OP) cancer has been increasing, in contrast to the overall head and neck tumors, likely related to human papillomavirus infection (HPV). Early clinical detection of oral cavity (OC) and OP tumors and precursors is essential for early intervention and chemoprevention. Since the oropharynx is less accessible than the OC, altering the current screening practice may be necessary. The addition of autofluorescence visualization (AFV) to standard examination has proven valuable by accurately detecting low and high-grade preneoploastic lesions (LGLs, HGLs) and oral cancers but its value in the OP has not been studied to date. We aim to assess if adding AFV to routine white light exam (WLE) increases the detection of OP neoplasms and precursor lesions. Study Design : High-risk patients with suspicious OC or OP lesions or recently diagnosed, untreated carcinoma underwent examination with WLE followed by AFV at 405nm from a 10mm rigid endoscope. Biopsies were obtained from areas with positive suspicion on either WLE or AFV examinations. Lesions were stratified on central pathology review as benign, low-grade (parakeratosis with atypia (PKA) or mild dysplasia), high-grade (moderate dysplasia, severe dysplasia, or carcinoma in situ (CIS)), or cancer. Sensitivity and specificity were calculated for WLE, AFV, and WLE + AFV for all patients with an OP biopsy on the first visit. Results : A series of 143 patients under active surveillance at our institute were examined over 218 visits and a total of 664 biopsies were collected. Twenty-nine patients (20%) had 41 biopsies (6%) sampled from the OP or OC/OP junction as follows: 5 base of tongue (BOT), 17 soft palate (SP), 10 tonsil, 5 SP/tonsillar junction, 3 hard palate/SP junction, and 1 BOT/posterior lateral tongue junction. Of the 29 LGLs/HGLs on the tonsil/SP/BOT, 14/29 were missed on WLE and 3/29 were missed on AFV. Out of 13 LGLs/HGLs on the SP, 5/13 were missed on WLE and 2/13 were missed on AFV. One severe dysplasia and one CIS were missed on WLE of the SP but none were missed with AFV. Of the 11 LGLs/HGLs on the tonsil or SP/tonsillar junction, 5/11 were missed on WLE and none were missed on AFV. One moderate dysplasia was missed on WLE of the tonsils but none were missed on AFV. Out of 5 LGLs/HGLs on the BOT, 4/5 were missed on WLE and 1/5 was missed on AFV. For the BOT, 4 mild dysplasias were missed on WLE and one was missed on AFV. Overall, WLE alone vs. AFV alone vs. AFV+ WLE detected 35% vs. 85% vs. 95% of LGLs respectively. Of the LGLs, 7 were PKA and 13 were mild dysplasia. WLE alone vs. AFV alone vs. the addition of AFV+ WLE detected 75% vs. 100% vs. 100% of HGLs respectively. Of the HGLs, one was moderate dysplasia, 5 severe dysplasias and 6 were CIS. All 3 modalities 100% detected the 2 cancers. Seven biopsies were benign. The addition of AFV+WLE greatly improved sensitivity in detecting LGLs when compared with WLE alone (95% vs. 35%) and HGLs (100% vs. 75%). The specificity in detecting both LGLs and HGLs decreased from 85.7% with WLE to 42.9% with AFV+WLE. Conclusions : AFV has high sensitivity in detecting OP lesions that may have been missed by WLE. Multi-modality visualization has utility in detecting pre-malignant lesions that may serve as potential targets for chemoprevention or early intervention. The addition of AFV to WLE may be useful in screening patients for chemoprevention trials. Due to its low specificity, AFV may be best limited to screening high-risk populations. Citation Information: Cancer Prev Res 2011;4(10 Suppl):A15.