Elsevier

Lung Cancer

Volume 48, Issue 3, June 2005, Pages 307-313
Lung Cancer

Effective detection of bronchial preinvasive lesions by a new autofluorescence imaging bronchovideoscope system

https://doi.org/10.1016/j.lungcan.2004.11.023Get rights and content

Summary

Autofluorescence bronchoscopy is an important tool for the early detection of preinvasive bronchial lesions. However, autofluorescence bronchoscopy has difficulty distinguishing between preinvasive lesions and other benign epithelial changes. A new autofluorescence imaging bronchovideoscope system (AFI) comprises three signals, including an autofluorescence (460–690 nm) on excitation blue light (395–445 nm) and two different bands of reflected light: G′ (550 nm) and R′ (610 nm). We hypothesized that color analyses of these three wave lengths would improve our ability to differentiate between inflammation and preinvasive lesions. In order to prove this hypothesis and to evaluate the efficacy of AFI for detecting preinvasive lesions, we conducted a prospective study. A total of 32 patients with suspected or known lung cancer were entered into this study. Conventional white light bronchovideoscopy (WLB) and light induced fluorescence endoscopy (LIFE) were performed prior to using AFI. WLB and LIFE detected 62 lesions, including lung cancers (n = 2), squamous dysplasias (n = 30), and bronchitis (n = 30). By utilizing AFI, 24 dysplasias and 2 cancer lesions were magenta in color, while 25 bronchitis lesions were blue. The sensitivities of detecting dysplasia by LIFE and AFI were 96.7% and 80%, respectively. The specificity of AFI (83.3%) was significantly higher than that of LIFE (36.6%) (p = 0.0005). We conclude that AFI appears to represent a significant advance in distinguishing preinvasive and malignant lesions from bronchitis or hyperplasia under circumstances where LIFE would identify these all as abnormal lesions.

Introduction

Lung cancer is one of the most common causes of cancer-related deaths. Detection of tumors at an early stage followed by complete surgical resection of tumors is the only curative treatment approach currently available for the large majority of lung cancer patients. Laser-induced fluorescence endoscopy (LIFE) lung-system (Xillix Technologies Corp.; Richmond, BC, Canada) is a new modality that has facilitated early detection of preinvasive bronchial lesions [1], [2], [3], [4], [5]. However, LIFE has difficulty distinguishing between preinvasive lesions and other benign epithelial changes such as bronchitis, which frequently is present in patients whose sputum cytology is suspicious of malignancy or positive for malignancy [3], [6]. Bronchoscopic findings have been divided into three classes: class I (normal), class II (inflammation and mild dysplasia), and class III (changes suggesting moderate or severe dysplasia, carcinoma in situ and/or invasive cancer) [5], [7]. At present, it remains difficult to predict pathological diagnosis from the observed grade of autofluorescence, as studies have found that some lesions identified as class III turned out to be inflammatory lesions on biopsy [7], [8]. The specificity of fluorescence bronchoscopy is reported to be rather low with up to one-third of the areas of abnormal fluorescence representing false positives [1], [9]. Kennedy et al. note that this low specificity leads to unnecessary biopsies, resulting in greater economic cost and longer examination time [10]. Therefore, the ability to distinguish preinvasive lesions from bronchitis would be of great value to endoscopic technology. Recently, a new autofluorescence imaging bronchovideoscope (AFI) system (Olympus Optical Corporation, Tokyo, Japan) combining autofluorescence with two reflected light signals has been developed in an attempt to improve the diagnostic precision of fluorescence bronchoscopy. We have therefore conducted a prospective clinical study in order to evaluate whether, compared with LIFE, AFI can more precisely distinguish between preinvasive bronchial lesions and bronchitis.

Section snippets

Autofluorescence imaging videoscope system

AFI is a video-endoscopic system that displays a composite image integrating three signals—an autofluorescence signal plus reflected green (G′) and red (R′) light signals. The Storz D-light system has already combined autofluorescence and reflected light [11], but AFI has further developed this approach. Fig. 1 shows spectra of the three AFI signals and the absorption coefficient of HbO2 (solid line). The excitation light (395–445 nm) produces an autofluorescence signal (490–700 nm). Green

Representative case of bronchoscopic findings

Fig. 3A, B, and C show typical bronchoscopic findings of WLB, LIFE, and AFI for squamous cell carcinoma, squamous dysplasia, and bronchitis, respectively. Arrows indicate abnormal areas of bronchial epithelium. Fig. 3A shows a case of squamous cell carcinoma. Irregular mucosa and protruding nodules are identified at the bifurcation of the right middle and lower lobe bronchi by WLB, LIFE, and AFI. The AFI image clearly is magenta at the abnormal area. Fig. 3B shows a case of squamous dysplasia

Discussion

It has been reported that autofluorescence bronchoscopy is more useful for detecting preinvasive lesions than white light bronchoscopy [1]. The results of multi-center clinical trials and our own experience suggest that autofluorescence bronchoscopy increases the diagnostic accuracy for squamous dysplasia, CIS, and early hilar lung carcinoma when used simultaneously with conventional bronchoscopy [2], [3], [5], [6], [9], [14], [15]. Although several types of autofluorescence systems, such as

Conclusion

AFI could accurately and objectively distinguish preinvasive and malignant lesions from bronchitis through color tone analysis.

Acknowledgments

We wish to thank Isami Hirao, Yuichi Morizane, Sakae Takehana (Olympus Optical Corp.; Tokyo, Japan) for their technical assistance. Supported in Part by Grant-in-Aid for Scientific Research (C) 13671376 from Japan Society for the Promotion of Science.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

View full text