Detection of Bronchial Neoplasia in Uranium Miners by Autofluorescence Endoscopy (SAFE-1000)

The increase in the detection rate for premalignant changes of bronchial epithelium was studied in 56 symptom-free volunteers from the risk group of Czech uranium miners (mean age 50.69 years, mean WLM 21.06 (1 Working Level Month is equal to the absorption of latent energy of 2.08 × 10–5 J/m3 in one month, i.e. 170 working hours)) by the additional employment of the System of Autofluorescence Endoscopy (SAFE-1000 Pentax) to conventional white-light bronchoscopy, comparing results with those of bronchial biopsy histopathology examination. Histopathology using hematoxylin and eosin staining confirmed intraepithelial neoplasias in 15 areas in 10 persons. White-light bronchoscopy sensitivity was 21.05%, and specificity 93.7% which an autofluorescence bronchoscopy sensitivity was 78.95% and specificity 81.89%.


INTRODUCTION
Lung cancer is the final stage of a multistep carcinogenic process [1]. The intraepithelial stage of neoplastic development typically lasts for a number of years before invasion occurs [2,3]. In uranium miners 68% of lung cancers are located in the central area of the bronchial tree [4].
Mining uranium ore is a high risk factor for lung cancer which can act after many years leaving the job in the mines. When the symptoms occur, the disease is usually advanced [5].
Lung cancer is recognized as an occupational disease in about 75 Czech uranium miners annually [6], but the real incidence is much higher and reaches some hundred cases per year. Epidemiologic studies PS-AFB98: Pilot study Autofluorescence bronchoscopy 1998, SAFE-1000.

MATERIALS AND METHODS
In the prototype of the System of Autofluorescence Endoscopy (SAFE-1000, PENTAX, Tokyo, Japan) infrared light from a Xenon light source Pentax LX-750P is eliminated by a special infrared filter, and only blue light of 420-480 nm is delivered through an excitation filter and transmitted via a light guide. The emitted autofluorescence is transmitted via an image guide to a TV camera with a Hamamatsu ICCD Camera Controller C 3510, fluorescence filter Subjects Seventy five subjects were examined, 56 persons of whom satisfied the conditions of the study: symptom-free volunteers from risk group of Czech uranium miners employed ten or more years in uranium mines with WLM evaluated cummulative radon exposure. They included smokers, ex-smokers and non-smokers.

Clinical Protocol
Conventional fiberoptic bronchoscopy was performed under local anesthesia. First WLB examination was done, then autofluorescence bronchoscopy (AFB) by the SAFE-1000 device. Areas suspected of premalignant changes were studied with alternating white and blue light and photographs were taken. Both investigations were recorded on videotape. Biopsy specimens for histopathology were taken from all suspected areas on WLB and AFB. Then random biopsies were carried out in visually normal areas using both techniques from right carina 2 or (RC2 and RC1) and left carina or 2 (LC1 and LC2), and other places as needed. Biopsy specimens underwent standard staining hematoxylin and eosin (HE). They were examined by an experienced pathologist.
All subjects gave informed consent. The approval for the study was given by the Ministry of Health Care of the Czech Republic and Czech Medical Chamber.

Data Analyses
The aim of the study was to evaluate the presence of premalignant changes ofbronchial mucosa in Czech AUTOFLUORESCENCE BRONCHOSCOPY 93 uranium miners and the sensitivity of the SAFE-1000 in conjunction with conventional WLB using bronchial biopsy histopathology for objective of comparison of visual findings of both endoscopy methods. The 157 biopsies performed in 56 subjects were divided into the 3 groups: Group consisted 108 random biopsies from apparently normal areas on autofluorescence and WLB, of this 104 specimens showed normal or mild inflammatory mucosal finding, and 4 specimens were found to show mild intraepithelial premalignant changes of bronchial mucosa (1 basal cell hyperplasia, mature squamous metaplasia, 2 mild dysplasias). Group 2 included 11 biopsies with alternating normal respiratory epithelium and squamous metaplasia without atypia. They were not included in AFB sensitivity estimation.
Group 3 consisted 38 specimens from areas with decreased autofluorescence intensity. Histopathology examinations confirmed 15 areas with intraepithelial premalignant changes (i.e. true positive AFB findings), and 23 areas exhibited no premalignancy (i.e. false positive AFB findings). False positive AFB findings from the histopathologic point of view consisted 18 specimens which were found to be normal and 5 biopsies which were found to show inflammation.
Significant decrease of autofluorescence intensity was seen in 38 areas (Table I). Histopathology confirmed premalignant changes in 15 areas, 12 were found to be dysplasia, including 5 with moderate dysplasia (Fig. 2) and 7 with mild dysplasia (Fig. 3), 2 were found to have metaplasia and there was basal cell hyperplasia. Another 23 areas had no evidence of premalignancy (Table II). WLB showed pathological findings in 12 areas in 10 persons (Table III).
The group of 13 persons (19 areas) with evidence of intraepithelial neoplasia (true positive and false negative findings) included 3 non-smokers, exsmoker and 9 smokers.

DISCUSSION
In order to control and reduce mortality prevention and early detectionare necessary. Autofluorescence imaging facilitates the study ofthe natural history of    This pilot study proved the sensitivity of the prototype autofluorescence system for lower grade atypia of bronchial mucosa. The difficulties at the beginning of the learning curve have gradually disappeared. The AFB has 3.75-times higher sensitivity than WLB but about 12% lower specificity (Table V).
From the analysis of causes of false positivity ( Fig. 4 and Table VI) the groups I, II and IV with normal histopathology, group III with pathological findings and group V revealed various causes such as anatomically related shadows (SAC e.g. Fig. 5), by motility related abnormality of decreased autofluorescence intensity (MOD), adherent mucus  Accurate evaluation was difficult in 11 specimens.
They were classified as checkered findings according to histopathology, with alternating normal respiratory epithelium and squamous metaplasia without atypia. Ten specimens were located in areas of normal autofluorescence intensity i.e. visually negative AFB findings. The other specimen of this type came from an area with decreased autofluorescence intensity i.e. visually positive AFB findings. The 11 specimens were not included in the evaluation of AFB sensitivity. Further examinations of specimens and follow-up investigation ofthe involved areas are necessary.
All false negative biopsies (Fig. 7) were found to be areas of low grade intraepithelial premalignancy. Thin epithelial layers could explain the false negativity of 3 specimens, with basal cell hyperplasia, with mature squamous metaplasia, and with mild dysplasia. False negativity of the second biopsy with mild dysplasia could have been caused by low grade atypia, by the biological properties of the tissue, or by technical reasons.
The detection of basal cell hyperplasia based on autofluorescence requires consideration of some of the causes of false positivity and optically nonsignificant presence of basal cell hyperplasia in the FIGURE 7 False negativity schema. epithelium, the biological properties of the bronchial tissue and also the technical characteristics of the equipment.
The development of methods for detection and treatment of intraepithelial lesions within the bronchial tree is promising both from surgical and medical points of view [11][12][13][14][15][16]. Autofluorescence technology, endobronchial sonography, ultrathin bronchoscopy, chemoprevention, photodynamic therapy and electrocautery offer challenges to work together.