A comparison of video and autofluorescence bronchoscopy in patients at high risk of lung cancer

The proportion of individuals with bronchial carcinoma in situ in whom invasive cancer will develop is likely to be >40% 1. Treatment of cervical dysplasia or carcinoma in situ has significantly reduced the incidence and mortality of invasive cervical cancer 2. By coupling sensitive diagnostic tools used in the investigation of early lung cancer with new treatment modalities, such as chemoprevention and various endobronchial therapies, it is hoped that the traditionally poor prognosis for patients with lung cancer can be altered in the near future 1. An important factor needed for the success of such programmes is the easy availability of diagnostic tools.

The effectiveness of autofluorescence bronchoscopy (e.g. lung imaging fluorescence endoscopy (LIFE)) as a guide for endobronchial biopsy has been established for the localisation of intra-epithelial neoplasia 3, 4. Endoscopic changes have been described using white light bronchoscopy that can be used to detect dysplasia and carcinoma in situ 5. Several studies have described the benefit of autofluorescence bronchoscopy over white light bronchoscopy 3, 6, 7. In a multicentre study, Lam et al. 3 reported that the addition of autofluorescence bronchoscopy to white light bronchoscopy resulted in the detection of a significant number of early intra-epithelial neoplastic lesions that would otherwise have been missed. In these studies, white light bronchoscopy was carried out using a fibreoptic bronchoscope. The white light flexible video bronchoscope (FVB) has a miniature charge couple device built into its tip, which helps to deliver a clearer image. Furthermore, the images obtained with the FVB do not have the graininess seen with images obtained from the conventional fibreoptic bronchoscope. To the current authors' knowledge, the utility of FVB and autofluorescence bronchoscopy, specifically LIFE, has not been compared previously in the investigation of individuals at a high risk of lung cancer (moderate dysplasia or worse on sputum cytology) to localise intra-epithelial dysplasia or neoplasia. Hence, this study was undertaken to correlate the findings observed with FVB and LIFE, and to assess their diagnostic yield in the investigation of individuals at a high risk of developing lung cancer.

METHODS

From September 1998–September 2002, 158 consecutive high-risk individuals, who underwent mass screening for lung cancer having sputum cytology findings of class C, D and E, as defined by the Japan Lung Cancer Society, were referred to the Dept of Thoracic Surgery, Chiba University, Chiba, Japan, for bronchoscopic investigation 8. Criteria for cytological diagnosis by sputum cytology were: A) no histiocytes in sputum (insufficient specimen); B) normal, only normal epithelial cells, proliferative basal cells, mild dysplasia and ciliated columnar cells; C) moderate dysplasia and columnar cells with changes of nuclear size or strong staining; D) severe dysplasia or cells suspicious for malignancy; and E) cells of malignant tumour 4, 8. All patients underwent chest radiography and computed tomography of the chest. Patients with peripheral lesions underwent transbronchial biopsy in addition to transbronchial needle aspiration cytology under fluoroscopy guidance as they had abnormal sputum cytology. None of the patients had previously received photosensitising agents, such as haematoporphyrin derivative, chemopreventative drugs, radiotherapy to the chest, or cytotoxic chemotherapy agents. Seven patients were excluded (four due to inadequate endobronchial biopsy specimens and three due to vocal cord malignancy). Patients with vocal cord malignancy were excluded from the analysis to maintain a uniform cohort, as none of the remaining 151 patients included in the study had a known extrapulmonary cancer.

All patients underwent FVB (BF-240; Olympus, Tokyo, Japan) followed by LIFE (Xillix LIFE; Xillix Technologies Corp., Richmond, BC, Canada) in the same session, which was applied using a fibreoptic bronchoscope (BF40; Olympus). Findings of FVB and LIFE were documented prospectively during the procedure and were not changed once the procedure was completed. FVB was performed under local anaesthesia and sedation (diazepam/midazolam). Visual findings at FVB were first recorded either as normal, abnormal or suspicious for high-grade dysplasia or cancer, as follows. Normal: areas without any visual abnormality; abnormal: areas with increased redness and hypervascularity, swelling or thickening of the bronchial mucosa, or focal thickening of a subcarina; suspicious for high-grade dysplasia or cancer: nodular or polypoid lesions or irregularity of the bronchial mucosa (fig. 1⇓) 3, 6, 9. FVB was followed by LIFE, and findings were also recorded as either normal, abnormal or suspicious 3. With LIFE bronchoscopy, normal areas appear as green; areas with a definite brown or brownish-red colour (loss of green fluorescence) were labelled as suspicious for high-grade dysplasia or cancer; and areas that were slightly brown with ill-defined margins were labelled as abnormal.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

a) and d) show normal appearance with flexible video bronchoscopy (FVB) and abnormal lesion with lung imaging fluorescence endoscopy (LIFE), respectively, at the same site. b) and e) show abnormal appearance at FVB and suspicious lesion at LIFE, respectively, at the same site. c) and f) show suspicious lesions at both FVB and LIFE, respectively, at the same site.

Endobronchial biopsies (EBX) were obtained using 2-mm diameter fenestrated cup forceps (FB-19C-1; Olympus) from abnormal or suspicious areas at both FVB and/or LIFE. In patients with normal FVB and LIFE, EBX was obtained from a second-generation carina. Biopsy specimens were reported and classified as either: 1) normal; 2) inflammation/bronchitis; 3) hyperplasia; 4) squamous metaplasia; 5) mild dysplasia; 6) moderate dysplasia; 7) severe dysplasia; 8) carcinoma in situ; or 9) carcinoma 4, 10.

Histological findings from 1–5 were classified as negative and 6–9 as positive 3. Relative sensitivity was calculated to evaluate the contribution of LIFE to FVB to detect intra-epithelial dysplasia or neoplasia 3. Relative sensitivity of >1 would indicate a significant improvement of LIFE over FVB. The Chi-squared test was used to assess the association between the findings of FVB and LIFE with the pathological diagnosis.

RESULTS

A total of 151 patients (142 males, nine females) with a mean age 67±8 yrs (range 44–82 yrs) underwent EBX from 343 sites. A history of smoking was present in 143 (95%) individuals with a mean of 54±24 pack-yrs (range 11–153). Sputum cytology classes C, D and E were present in 42 (28%), 58 (39%) and 50 (33%) individuals, respectively. Clear interpretation of sputum cytology was not possible in one patient with haemoptysis. The pathological and bronchoscopic characteristics of the EBX are presented in table 1⇓. Normal, abnormal and suspicious areas were observed during FVB at 162 (47%), 151 (44%) and 30 (9%) biopsy sites, respectively; and LIFE at 69 (20%), 212 (62%) and 62 (18%) biopsy sites, respectively (table 2⇓). Abnormal or suspicious findings during FVB and/or LIFE were present at 295 out of 343 (86%) biopsy sites.

Of the 83 out of 343 (24%) pathologically positive EBX, 81 out of 295 (27%) were obtained from sites that had abnormal or suspicious findings at FVB or LIFE, and two out of 48 (4%) were obtained from areas that were normal at both FVB and LIFE. The sensitivity of FVB was 72% (60 out of 83) and LIFE 96% (80 out of 83). The relative sensitivity of LIFE over FVB was 1.33 (80 over 60). The specificities of FVB and LIFE were 53% (139 out of 260) and 23% (66 out of 260), respectively. The positive and negative predictive values of FVB were 60 out of 181 (33%) and 139 out of 162 (86%), respectively; and for LIFE 80 out of 274 (29%) and 66 out of 69 (96%), respectively.

DISCUSSION

White light flexible fibreoptic bronchoscopy is an important diagnostic tool in the histological diagnosis of lung cancer 11. However, it has major limitations in the diagnosis of dysplasia and early malignant lesions. Autofluorescence bronchoscopy has been reported to be useful in the diagnosis of dysplasia and early lung cancer 3, 4. Conventionally, white light bronchoscopy is usually performed prior to autofluorescence bronchoscopy in the same session. The order in which the bronchoscopist carries out these procedures does not appear to affect the outcome of the results 7. In this study, FVB was performed prior to LIFE. The addition of autofluorescence bronchoscopy to white light bronchoscopy has been reported to increase the detection of pre-invasive lung cancer. However, white light bronchoscopy has been commonly performed with a fibreoptic bronchoscope whilst comparing its utility with autofluorescence bronchoscopy 3, 6, 7, 12. FVB has superior optical resolution when compared with fibreoptic bronchoscopy and, therefore, has potential advantages with respect to the image resolution. To current authors' knowledge, this is the first study that has compared the utility of FVB and LIFE in the investigation of individuals at a high risk of developing lung cancer. The sensitivity of FVB in this study was 72%. For FVB, the yield of abnormal and suspicious findings was significantly higher when compared with normal findings. The present study shows that EBX obtained from abnormal or suspicious areas on FVB have a good diagnostic yield in individuals at a high risk of lung cancer.

In the current study, biopsy areas that were visualised to be normal during FVB had a significant increase in yield when EBX was guided by changes noted with LIFE when compared with the random biopsies obtained when both FVB and LIFE were normal. In areas with abnormal FVB findings, there was a significant increase in the diagnostic yield of biopsies obtained from lesions that also showed abnormal and suspicious changes at LIFE when compared with normal LIFE. These findings indicate the benefit of LIFE in identifying lesions when FVB is either normal or abnormal.

An important factor associated with the success of a screening programme is the development of a diagnostic tool that is routinely available and not expensive. Following evolutions in technology, conventional white light fibreoptic bronchoscopy is now being replaced with FVB, and subepithelial vascular patterns in bronchial dysplasias have been described using a high-magnification bronchovideoscope 13. Whether conventional FVB can be utilised for the evaluation of individuals at a high risk of lung cancer in centres that do not have the facilities for autofluorescence bronchoscopy needs to be determined. However, the good yield of FVB (72%) in this study has led the current authors to propose that it may have a potential role in the primary evaluation of individuals at a high risk of lung cancer. The FVB used in this study had a black and white video chip, and colour on the monitor was achieved through the light source. There are no data comparing the colour video chip white light flexible bronchoscopy findings with autofluorescence bronchoscopy. Such a study would be of great interest, as the video chip white light flexible bronchoscope may have the potential to further enhance the sensitivity and specificity of FVB.

In another multicentre study, abnormal findings at white light bronchoscopy and autofluorescence bronchoscopy were considered negative 3. In the present study, anatomical anomalies and bronchoscopic trauma were not included as abnormal findings. Furthermore, increased redness and hypervascularity, and focal thickening of a subcarina were included as abnormal findings. The sensitivity of FVB in the present study, if guided by suspicious lesions only, would be 23 out of 83 (28%). However, abnormal lesions were used as markers for EBX, which led to a high sensitivity of FVB (72%). Although this study does not compare the findings of fibreoptic bronchoscopy with FVB, the authors believe that the putative advantages of the FVB image assisted in the better classification of FVB findings.

The diagnostic yield of random biopsies when FVB and LIFE were normal was only 4% (two out of 48). When FVB was normal and LIFE showed either abnormal or suspicious changes, the diagnostic yield of biopsies guided by changes at LIFE was 18% (21 out of 114). When FVB was abnormal and LIFE normal, the yield of biopsies guided by changes at FVB was only 5% (one out of 21). Therefore, LIFE was able to distinguish FVB false-positive as well as false-negative biopsy sites. Hence, LIFE is useful both when FVB is normal or abnormal, and is also a good diagnostic tool in detecting pre-invasive lesions.

In conclusion, endobronchial biopsy of suspicious findings at lung imaging fluorescence endoscopy as well as flexible video brochoscopy have a good diagnostic yield in individuals at a high risk of developing lung cancer. Lung imaging fluorescence endoscopy is more useful when flexible video bronchoscopy is either normal or abnormal.