Autofluorescence bronchoscopy (AFB) has been shown to be sensitive to detect preneoplastic lesions in central airways. Apart from bronchial mucosa thickness, tissue autofluorescence is also related to the biochemical properties of the target cells. Genetic studies have shown molecular abnormalities to be present in histologically normal mucosal specimens. Forty-six high-risk individuals, free of micro-invasive cancer at the initiation of the study, were included in this analysis and have been subjected to repeat bronchoscopic examinations every 4-6 months. They had previous curatively treated lung cancer (n=18), ENT tumor (n=11) or were at risk to acquire lung cancer primaries (n=17). Baseline AFB is scored for each suspicious lesion, thus the total score represents the number of AFB suspicious lesions present in each individual at risk. Baseline AFB score was correlated to outcome, i.e. the development of squamous-cell cancer (SCC) in each individual. So far, 11/46 (24%) of the individuals acquired SCC. Follow up has been 12-80 months. All five individuals with >/=3 lesions (100%, 12-36 months), five of the ten (50%, 12-48 months) individuals who had two lesions and one among the 12 (8%, 36 months) individuals with one suspicious AFB lesion, developed SCC. Up till now (12-80 months), the remaining 19 individuals without any suspicious AFB lesion have not acquired SCC. The average AFB score for the group of individuals which developed SCC was significantly different (P<0.001) from the remaining individuals who did not acquire SCC (2.64+/-1.1 vs. 0.6+/-0.7 S.D.). The number of suspicious lesions at baseline AFB is a good predictor for the development of SCC in the individuals at risk in our study population. This finding is compatible with field carcinogenesis and warrants a more upfront use of AFB in a lung cancer screening to sift the different risk-cohorts in a population mainly at risk for developing metachronous lung cancer.