Elsevier

Lung Cancer

Volume 73, Issue 2, August 2011, Pages 183-188
Lung Cancer

A comparison of autofluorescence bronchoscopy and white light bronchoscopy in detection of lung cancer and preneoplastic lesions: A meta-analysis

https://doi.org/10.1016/j.lungcan.2010.12.002Get rights and content

Abstract

Introduction

It is known that autofluorescence bronchoscopy (AFB) has limited value in detection of lung cancer and preneoplastic lesions. Though a substantial number of studies have evaluated the diagnostic yield of AFB, the variable estimates limited the ability to accurately assess its test performance and future role in clinical practice. The clinical utility of AFB has never been supported by a meta-analysis due to the inconsistent characteristics in some of studies. A meta-analysis was performed to re-examine the diagnostic efficiency of AFB compared with white light bronchoscopy (WLB).

Methods

Search of both MEDLINE and EMBASE database up to June 2009 was conducted and hand search was performed against the extracted reference list for relevancy. Included studies had to have a conclusive histology as diagnostic standard, and provided sufficient data to construct a 2X2 table for assessing the diagnostic yield of AFB for detection of lung cancer and preneoplastic lesions. After examining the source of variation, pooled sensitivity and specificity of AFB were estimated using a bivariate random-effects regressing model and compared with that of WLB.

Results

Of 439 publications, 14 studies, providing 15 sets of data, were suitable for analysis. The pooled sensitivity and specificity of AFB and WLB were 0.90 (95% CI 0.84–0.93) and 0.56 (95% CI 0.45–0.66), 0.66 (95% CI 0.58–0.73) and 0.69 (95% CI 0.57–0.79). The contribution of differences in excitation light source, histological criteria and biopsy strategy was not counted as a covariate.

Conclusions

The result indicated that AFB was superior to conventional WLB in detecting lung cancer and preneoplastic lesions.

Introduction

The prognosis of lung cancer strongly depends on the stage of disease at diagnosis. According to the theory of stepwise progression of carcinogenesis, early detection of preinvasive lesions, such as moderate or severe dysplasia, carcinoma in situ, and subsequently prompt surgical resection or endobronchial treatment will provide the patients with the best chance of survival and conservation of lung function. It was reported that five-year survival rate for patients treated for preinvasive (stage 0) lung cancer was greater than 90% [1]. Even for patients with invasive form of the cancer, the detection of synchronous lesions and identification of the extension of the tumor may lead to important changes in management plan [2], [3], [4]. This undoubtedly indicates the needs for early diagnosis of lung cancer in a preclinical stage. Unfortunately, the preinvasive lesions are superficial and minute in size [1], [5]. Therefore, the early detection of lung cancer presents a great challenge in clinical practice even for experienced bronchoscopist.

The difference in the intensity of autofluorescence between normal and neoplastic lesions has been reported as early as 1965 [6]. The first AFB was approved by US FDA in 1996, employing a helium-cadmium laser as excitation light source [7]. And a pseudo-color image based on autofluorescence and reflected light was used to delineate the abnormal area. As the technology improved, different systems have been developed, most using less expensive Xenon light source and a light filter [8], [9], [10]. Nowadays, AFB has become increasingly available and widely used for detection of preneoplastic lesions and lung cancer, resulting in improved appropriate sampling.

Up to recently, a number of studies have investigated the diagnostic yield of AFB and compared it with that of WLB. However, the estimates of sensitivity and specificity varied between reports, which inevitably limited the ability to accurately evaluate its test performance and future role in clinical practice [8], [9], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]. And to the current authors’ knowledge, the utility of AFB has never been supported by a meta-analysis based on all quantitative evidence. The primary endpoint of this study was to assess the diagnostic yield of AFB for detection of lung cancer and preneoplastic lesions. The second endpoint was to compare the test performance of AFB and WLB.

Section snippets

Data source and searches

We performed an evaluating search of two online databases to identify studies on comparing the diagnostic yields of AFB and WLB for identification of lung cancer and preneoplastic lesions. Sources were articles in MEDLINE and EMBASE published up to June 2009, no start date limit was applied. The search strategy included (“fluorescence” OR “fluorescence imaging”) combined with “bronchoscopy”. Search filters that often returned irrelevant studies and unlikely to decrease the number of articles to

Identification of studies and study quality

The details of study identification along with the inclusion and exclusion were shown in Fig. 1. Of the 439 references identified, 117 potentially relevant citations were selected. In 31 studies relevant to diagnostic yield of AFB, 19 were excluded, mainly due to trial design or failure to report outcome variables of interest, and 2 additional studies identified by hand search (see text documents, Supplemental Digital Content 1, which contains the studies excluded from the meta-analysis). A

Discussion

In this meta-analysis of 14 studies, the pooled sensitivity of AFB was 0.9, which is compared favorably with WLB. The pooled specificity was 0.56, lower than that of WLB. The high sensitivity and low specificity are commensurate with other imaging modalities, such as CT in the diagnosis of small malignant lung nodule [26]. It is worthwhile considering the possibility of making a preinvasive cancer diagnosis without additional risk.

As a routine study design, WLB was first carried out, and areas

Conflict of interest statement

All authors have none declared.

References (35)

  • M.M. Leeflang et al.

    Use of methodological search filters to identify diagnostic accuracy studies can lead to the omission of relevant studies

    J Clin Epidemiol

    (2006)
  • J.B. Reitsma et al.

    Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews

    J Clin Epidemiol

    (2005)
  • C.I. Henschke et al.

    Early Lung Cancer Action Project: overall design and findings from baseline screening

    Lancet

    (1999)
  • S. Lam et al.

    Localization of bronchial intraepithelial neoplastic lesions by fluorescence bronchoscopy

    Chest

    (1998)
  • P. Lee et al.

    Dual digital video-autofluorescence imaging for detection of pre-neoplastic lesions

    Lung Cancer

    (2007)
  • S. Lam et al.

    Detection and localization of early lung cancer by fluorescence bronchoscopy

    Cancer

    (2000)
  • B. Zaric et al.

    Autofluorescence videobronchoscopy (AFI) for the assessment of tumor extension in lung cancer

    Technol Cancer Res Treat

    (2009)
  • Cited by (58)

    • Endobronchial Therapies for Diagnosis, Staging, and Treatment of Lung Cancer

      2022, Surgical Clinics of North America
      Citation Excerpt :

      A meta-analysis that reviewed 14 studies showed that the pooled sensitivity of AFB was higher than WLB for the detection of lung cancer and preneoplastic lesions (0.90 vs 0.66). However, the specificity of WLB was higher than that of AFB (0.69 vs 0.56).11 The higher sensitivity and lower specificity of AFB (with WLB) compared with WLB alone was again demonstrated in a meta-analysis by Sun and colleagues.12

    • Intergroupe francophone de cancérologie thoracique, Société de pneumologie de langue française, and Société d'imagerie thoracique statement paper on lung cancer screening

      2021, Diagnostic and Interventional Imaging
      Citation Excerpt :

      However, when diagnosed at stage I and with surgical resection possible, the ten-year survival rate of lung cancer exceeds 80% [5,6]. Chest X-ray, sputum cytology, bronchoscopy and 18F-Fluoro-desoxy-glucose (18F-FDG) positron emission tomography-computed tomography (PET-CT) are ineffective for lung cancer screening [7–10]. By contrast, low-dose computed tomography (LDCT) has been or is being studied in ten randomized trials [11] (Fig. 1).

    View all citing articles on Scopus
    View full text