| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Internal Medicine, Division of Pulmonary and Critical Care, Allergy and Sleep Medicine (Drs. Vincent and Silvestri), and Department of Cellular and Molecular Pathology (Dr. Fraig), Medical University of South Carolina, Charleston, SC.
Correspondence to: Gerard A. Silvestri, MD, FCCP, 96 Jonathan Lucas St, Suite 812 CSB, Charleston, SC 29425; e-mail: silvestri{at}musc.edu
Abstract
Background: The objectives of this study were to characterize the appearance of normal, dysplastic, and frankly malignant airway lesion appearance under narrow-band imaging (NBI), and to determine if NBI, when used in conjunction with white light (WL) bronchoscopy, could improve detection of dysplasia and malignancy.
Patients and methods: This was a prospective, partially blinded study at a university teaching hospital. Bronchoscopy was performed on 22 patients with known or suspected bronchial dysplasia or malignancy. Full airway examination was performed first under WL bronchoscopy and then under NBI. Directed endobronchial biopsies of likely dysplastic, malignant, and normal (control) areas were then performed and sent for examination by a pathologist blinded to the gross description of the lesion. Pathology interpretations were then compared to the corresponding WL and NBI images.
Results: There were one malignant and four dysplastic lesions in 22 patients detected by NBI when findings by WL imaging were considered normal. In cases when the WL appearance was abnormal, NBI did not improve the diagnostic yield. The increased rate of detection of dysplasia and malignancy by NBI was statistically significant (p = 0.005).
Conclusion: NBI identified dysplasia or malignancy that was not detected by WL inspection in 23% of subjects. Further studies are needed to determine the efficacy of NBI in detection of premalignant airways lesions in an at-risk population.
Key Words: bronchial dysplasia • carcinoma in situ • interventional bronchoscopy • lung cancer • malignancy • narrow-band imaging • white-light bronchoscopy
Bronchogenic lung cancer remains the most common cause of cancer death in the United States for both male and female patients.1 With the 30-year decline in smoking rates among young Americans beginning to plateau, the search for a useful lung cancer screening tool becomes ever important. While early studies2345 showed increased rates of early stage lung cancer detection, the critical end point of decreased lung cancer mortality was not realized. More recently, low-dose CT imaging has been evaluated for lung cancer screening in several large, National Cancer Institute-sponsored observational trials.6789 Cancers of the trachea and bronchi typically present at a late stage with symptoms caused by airway obstruction such as cough, hemoptysis, and postobstructive pneumonia, and early stage lesions are frequently missed on CT. Only 10% of these cancers present in a potentially surgically resectable stage.10 Since the early 1990s, studies using autofluorescence (AF) bronchoscopy to detect premalignant lesions of the central airways have, in general, shown that behavior of precancerous lesions of the central airways is unpredictable. Lesions as advanced as carcinoma in situ (CIS) have the ability to spontaneously regress, and it is not clear that lesions detected by AF will significantly shorten a patient’s life even if left untreated.1112 Furthermore, the largest randomized study13 to date concluded that screening for preinvasive lesions with AF is not recommended outside of controlled clinical trials.
Narrow-band imaging (NBI) is a new, alternative light-wavelength capture system that takes advantage of altered blood vessel morphology of bronchial dysplasia. Wavelengths of light in the visible spectrum are filtered from the illumination source, with the exception of narrow bands in the blue and green spectrum centered at 415 nm and 540 nm, coinciding with the peak absorption spectrum of oxyhemoglobin, making blood vessels more pronounced when viewed in NBI mode. It is well established that dysplastic airway lesions such as angiogenic squamous dysplasia, for example, frequently display abnormal capillary formation both grossly and histopathologically.1415 Observational studies have shown capillary loop patterns to be potentially useful in early diagnosis of both early esophageal16 and gastric17 carcinomas. Shibuya et al18 demonstrated that NBI in combination with high-magnification bronchoscopy was useful in characterizing capillary loop patterns of dysplastic airway lesions. They also characterized morphologically abnormal vessels as appearing "dotted," "tortuous," as well as abrupt-ending vessels with large caliber. The goals of this study were to characterize the appearance of different airway mucosal conditions ranging from normal to malignant under NBI mode, and to determine whether or not NBI could improve detection of dysplasia and malignancy when compared to white light (WL) alone.
Materials and Methods
This was a prospective, partially blinded study conducted at an academic medical center in the United States over a 6-month period. The study was approved by our investigational review board, and informed consent was obtained from each participant prior to the study.
Study Population
Patients were eligible for enrollment in the study if they were 18 years old and scheduled for flexible bronchoscopy for the following reasons: radiologic airway obstruction from known or suspected cancer with or without associated pneumonia and/or mediastinal and/or hilar lymphadenopathy; hemoptysis from known or suspected airway malignancy or dysplasia; abnormal sputum cytology findings; and radiologic airway obstruction from known or suspected pulmonary metastases from primary cancer other than lung. Twenty-two patients met enrollment criteria. Exclusion criteria included inability to safely tolerate bronchoscopy with endobronchial biopsies, such as unstable angina pectoris, bleeding diathesis, uncontrolled hypertension, and sensitivity to midazolam, fentanyl, or xylocaine, and inability to understand and sign informed consent.
Study Design
Bronchoscopy was performed in a dedicated bronchoscopy center. All standard monitoring and alert systems were used during each procedure. Videobronchoscopy was performed by the investigators (B.V. and G.S.) using a flexible videobronchoscopy system (EVIS EXERA II model BF-Q180 bronchoscope, CLV 180 light source, CV-180 video processor, and OEV-191H LCD monitor; Olympus; Center Valley, PA). Examination of the airways was performed starting with the trachea, followed by the right and left bronchial trees. Each separate area was first examined under WL and then with NBI mode. Abnormal mucosa were first identified under WL and then with NBI mode. Congruent and incongruent areas under WL and NBI were confirmed by the second bronchoscopist. An abnormal appearance under NBI mode was defined as any area discordant in appearance vs WL appearance by either blood vessel concentration or appearance (ie, dotted, tortuous, or large-caliber, abrupt-ending vessels). Once consensus was attained between the two bronchoscopists regarding the location and appearance of discordant areas, these sites were targeted for endobronchial biopsy. In 20 cases, no more than three and no less than two sites were sampled for biopsy. In two cases, only one site was sampled due to only one area being abnormal by either WL, NBI, or both. A randomly selected area of the trachea appearing normal under both WL and NBI was selected for control biopsy. Three specimens were obtained at each target site, including the control.
NBI and WL mode images of all areas sampled were captured using a digital-image storage device. Radial serrated-jaw forceps were then used to obtain samples, starting with the most distal targets and ending with the trachea (control). New biopsy forceps were used at each biopsy site to avoid contamination of subsequent biopsy sites. Images were then printed using a high-quality dye sublimation color printer (OEP-3; Olympus) with photo-quality paper (UPC-510; Sony; Tokyo, Japan). The procedure was also stored via a real-time videotape recorder (SVO-9500MD; Sony). Biopsy specimens were stored in 10% formalin solution and processed for pathology interpretation. Control specimens were stored in our research tissue bank. All specimens were interpreted by a dedicated lung pathologist (M.F.) blinded to all clinical information except the anatomic location of the specimen. Biopsy specimens were considered evaluable and included in the final analysis if the bronchial epithelium was intact for interpretation and matching WL and NBI images of the biopsy site were available for review. Dysplastic and cancerous lesions were classified according to the World Health Organization classification system. The preinvasive lesion classification contains seven categories, including normal histology; reserve-cell hyperplasia; squamous metaplasia; mild, moderate, and severe dysplasia; and CIS.
Statistical Analysis
Statistical comparisons were performed using a Fisher two-sided exact test; p values < 0.05 were considered significant. Ninety-five percent confidence intervals (CIs) about the mean were calculated using a paired t test. Statistical software (SPSS 14.0; SPSS; Chicago, IL) was used for statistical analysis.
Results
Patient Characteristics
There were 8 men and 14 women (mean age, 59 years; range, 28 to 77 years). All participants were either current or former smokers (mean, 38 pack-years [packs per day times the number of years smoked]; range, 5 to 100 pack-years). Fifty percent of the participants had a diagnosis of lung cancer, and all of these had received treatment for lung cancer. Three patients underwent surgical resection only, and eight patients underwent chemotherapy and radiation. Patient characteristics are shown in Table 1
. Indications for procedures in this study were the following: malignant airway obstruction from known lung cancer (50%), radiologic abnormality of the central airway (32%), undiagnosed parenchymal lung mass (36%), hemoptysis (23%), and mediastinal adenopathy (23%). Eighteen subjects had more than one indication. There were no adverse events reported during the study.
|
|
Discussion
It is estimated that squamous cell carcinoma of the central airways represents up to one third of all lung cancers diagnosed in the United States,22 yet the natural history of bronchial dysplasia is not as well understood. Previous studies232425 estimated that invasive carcinoma will develop in 40 to 83% of patients with severely dysplastic lesions. It has been demonstrated in animal studies2627 that many of these lesions can regress spontaneously, and it has also been suggested that invasive carcinoma can develop where no previous abnormality had been detected.2829 Breuer et al30 followed up 52 high-risk patients with varying degrees of dysplasia over 8 years and showed a 9 to 32% rate of malignant development for all dysplastic lesions, ranging from squamous metaplasia to severe dysplasia. They also demonstrated a regression rate of 54% of all preinvasive lesions as well as non-stepwise transformation.30
The search for imaging technologies able to reliably detect precancerous airways disease led to the development AF bronchoscopy. AF bronchoscopy takes advantage of the fact that the loss of extracellular matrix protein as well as thickening of the superficial mucosa and increase vessel concentration all lead to an alteration of the fluorescent properties of the tissue.3132 The advantage of alternative light-wavelength capture systems such as AF over WL is a color-contrasted view of surface architecture allowing subtle differences to become more apparent to the bronchoscopist. Since 1991, AF bronchoscopy has been evaluated in numerous studies,1920333435363738 and has improved the detection of moderate-to-severe dysplasia and carcinoma 1.5- to 6.3-fold when compared to WL. The largest study to date, a multicenter trial reported by Haeussinger et al,13 had an overall improvement in detection of preinvasive lesions by a factor of 1.6 in patients at high risk for lung cancer, similar to the relative sensitivity of NBI compared to WL bronchoscopy in this study. Importantly, the detection of CIS by AF was not significantly improved over WL in that study,13 and an important conclusion was that screening with AF could not be recommended based on their findings. A recent study by George et al39 combining AF bronchoscopy and yearly chest CT demonstrated development of 11 cancers in nine patients with high-grade lesions (either high-grade dysplasia or CIS) in a population at very high risk for invasive cancer. Six of these cancers occurred at sites of previously diagnosed high-grade airway lesions. Most invasive cancers were able to be treated with curative intent.
This was an initial proof of principle study that showed that NBI improves the detection of dysplasia when WL bronchoscopy findings are considered normal. NBI did not improve detection of dysplasia or malignancy in areas where WL findings were also abnormal. NBI+/WL– was, however, less specific for cellular atypia than NBI+/WL+. This was expected based on the fact that many of our patients had locally advanced lung cancer and the majority had radiographic and/or symptomatic central airway obstruction from known lung cancer, thereby increasing the chances that WL bronchoscopy results would also be abnormal. One advantage of AF vs NBI is the track record and availability of data on AF. A disadvantage of AF is the requirement of additional equipment and training. NBI capability is included in the latest videoprocessor unit and would obviously represent additional expense to those without it, albeit less costly than an AF unit.
This study had several limitations. First, since this was new technology at the time of our evaluation, we sought to establish proof of concept that NBI would indeed improve detection of dysplasia compared to WL. A larger study with patients at very high risk is the logical next step to our initial findings. Second, there is currently no established "gold standard" for detecting preinvasive lesions in the airways. AF bronchoscopy is the most studied and accepted standard and, while it would be interesting to see whether or not NBI was equivalent in detection rate compared to AF, this was simply not feasible for this study. Recent presented data from Herth et al40 suggest that NBI may, in fact, be more specific than AF for dysplasia when used as an adjunct to AF bronchoscopy to evaluate vascular patterns of AF-abnormal areas. Third, we acknowledge that there may have been observer bias in overestimating incongruity or abnormality of airway lesions where none may have existed. Some AF trials1920333435363738 identify pathologist interobserver variability as a limitation, and this has been demonstrated in clinical studies.2141 Interobserver variability was controlled for by single-pathologist interpretation for all specimens. Another possible limitation could have been that eight patients studied had systemic treatment for lung cancer in the form of chemotherapy and radiation therapy because this may have had the effect of reducing the dysplastic burden and decreasing diagnostic yield.
In some ways, both AF and NBI are "technologies in search of an indication." We feel that three advances must occur to make either technology worthwhile. First, one must attempt to define an "ultra high risk" population in whom surveillance bronchoscopy identifies patients with lesions that are likely to progress to cancer. Second, we must know more about the natural history of these lesions both with and without treatment. Finally we need viable therapeutic strategies to either prevent the progression to overt cancer or treat disease once discovered. Without these advances, both NBI and AF remain largely research tools and should not be routinely performed outside well-designed clinical trials or special clinical circumstances.
Conclusions
The addition of NBI to WL bronchoscopy significantly improved detection of bronchial dysplasia compared to WL bronchoscopy alone. NBI requires closer examination both as a stand-alone technology and in comparison with other imaging strategies including AF bronchoscopy in the setting of clinical trials. Future directions include combining NBI with molecular markers in those at high risk for lung cancer.4243
Footnotes
Abbreviations: AF = autofluorescence; CI = confidence interval; CIS = carcinoma in situ; NBI = narrow band imaging; WL = white light
This work was supported by a research grant provided by Olympus America Inc.
The authors have no conflicts of interest to disclose.
Received for publication November 17, 2006. Accepted for publication February 8, 2007.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
