Is Narrow Band Imaging Better Than White Light For Head And Neck Cancer?
Narrow band imaging brings new advancements to cancer screening, enhancing visibility of abnormalities often missed by conventional imaging techniques.
Author:Suleman ShahReviewer:Han JuOct 23, 202490.5K Shares1.2M Views
Different from the conventional imaging techniques which use white light imaging (WLI), the narrow band imaging(NBI) tries to bring in some improvements.
As the diagnostic value of NBI varies depending on the anatomical locations, we performed a meta-analysis on the clinical utility of NBI in head and neck cancer screening.
As for the methods used in this study, we calculated the pooled sensitivity and specificity of NBI and WLI.
The positive likelihood ratio (PLR), negative likelihood ratio (NLR) and the summary estimates of diagnostic odds ratio (DOR) were used to estimate the diagnostic performance.
Summary receiver operating curve (SROC)was used for the comparison of clinical utility.
Preliminary Discussion
Head and neck cancers refer to cancer arising in the upper aerodigestive tract, including:
- paranasal sinuses
- nasal cavity
- nasopharynx
- oral cavity
- salivary gland
- oropharynx
- pharynx
- hypopharynx
- larynx
- lymph node
The cancer could develop from the mucosal space synchronously or metachronously with poor prognosis.
Histologically, squamous cell carcinoma originating from the epithelial mucosal lining is the predominant form.
Some quick facts about head and neck squamous cell carcinoma (HNSCC):
1. The incidence rate of HNSCC is increasingly high (the sixth most common cancer worldwide) with poor prognosis.
2. HNSCC is the fourth most common cancer among men in the European Union.
3. In the United States, over 12,000 patients die of HNSCC every year.
4. Globally, there are approximately 650,000 new cases of HNSCC annually.
Patients with HNSCC usually present late and have a diagnosis at an advanced stage. Patients usually visit a clinic when they have already developed local-regional disease with cervical lymph node involvement.
One major reason is that early disease might not have any symptoms and is developed in the silent areas. Furthermore, the inconspicuous location of some HNSCC makes identification difficult at early stages.
HNSCC might also develop from long-standing lesions such as leukoplakia and erythroplakia, which are not recognized easily. It is estimated that the overall survival rate of HNSCC patients is <50%.
Despite the advances of cancer treatment in the last several decades, the overall survival rate of HNSCC has not considerably improved.
Early HNSCC is curable. However, at present, there is no effective tool or molecular biomarkers for screening asymptomatic cancer or in surveillance of recurrent disease.
Diagnosis of precancerous and inconspicuous head and neck cancers in the mucosa or submucosa is done via direct visualization and biopsy with the aid of image-enhanced endoscopy (IEE) along with high-resolution imaging system.
Narrow band imaging (NBI) is an emerging imaging technique for use in detecting dysplastic lesions by endoscope.
In comparison with the normal mucosa, the capillaries and microvessels in a cancerous lesion look enlarged with tortuous configuration and high density.
In contrast to conventional imaging using white light imaging (WLI), NBI is designed to enhance the viewing of superficial capillaries and neo-angiogenesis in the mucosal surface, based on the principle that light with different bandwidths has different penetration depths.
An endoscope is supplied with optical filters, which allow a selective bandwidth in the white light to pass through and illuminate the viewing sites.
The filters absorb the most of the transmitted white light (mostly red light), but they allow two major bands, 415 nanometers (wavelength ranged: 400 and 430 nm.) and 540 nanometers (wavelength ranged: 525 and 555 nm.), to pass through the endoscope.
The former short wavelength blue light (415 nm.) penetrates the mucosa superficially and highlights the superficial vasculature in brown against the blue-green mucosa background.
Furthermore, the blue light is later absorbed by the hemoglobin in the blood vessels as it covers the peak absorption spectrum of hemoglobin (410 nm.).
The larger wavelength green band (540 nm.) passes through to the submucosal layer and identifies prominent vessels.
In contrast to the conventional white light system using broadband RGB technology, the NBI endoscopy system provides images with better contrast and higher resolution on the superficial structures.
Originally, NBI was designed and used to screen and examine the mucosa of the gastrointestinal tract to characterize the intra-epithelial cancers by observing irregular microvascular patterns and neo-angiogenetic vasculature, which is usually missed by WLI. It was first used in the detection of superficial, precancerous mucosal lesions over the esophagus.
In view of the promising results of using NBI as a screening aid in cancer screening, we explore here the potential clinical utility of NBI in screening head and neck cancers.
It is generally accepted that NBI has a high sensitivity and specificity in identifying premalignant lesions based on the capillary patterns on the surface mucosa.
However, whether NBI is a better alternative or supplement of WLI to routinely examine superficial cancers in the head and neck regions remains unresolved.
We performed here a systematic review on the potential use of NBI and the magnitude of benefit in detecting cancerous lesions in the head and neck regions.
Read Also: Why Does Cancer Recur? What You Need To Know
Methods And Materials
Search Strategy
A systematic literature search (publications from 1980 to February 2013) was performed independently by two of the authors (Z. H. Li and Wei Gao) in the following databases:
- Embase
- PubMed
- PubMed’s MEDLINE
- ‘head and neck/oral/tongue/pharynx/pharyngeal/nasopharynx/nasopharyngeal/oropharynx/oropharyngeal/ hypopharynx/hypopharyngeal/laryngopharynx/ laryngopharyngeal/larynx/laryngeal/glottic/ supraglottic/subglottic’
- ‘cancer/carcinoma/tumor/malignancy’
- ‘narrow band imaging/NBI’
After duplicated results were removed, the included articles were further selected by excluding the study if it was:
(1) an irrelevant article
(2) a publication not in English or Chinese
(3) a review
(4) a paper without control group; and/or
(5) an article lacking detection frequency data or without exclusive detection frequency in head and neck carcinoma
Data Extraction And Quality Assessment
Two authors (Z. H. Li and Wei Gao) reviewed and extracted the data from the selected articles independently, which included:
- first author
- publication year
- study area
- study type
- case population
- diagnostic standard and the true positive (TP)
- false positive (FP)
- false negative (FN)
- true negative (TN) results
The methodological quality of each study was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool.
Data Analysis
The diagnostic accuracy of NBI and WLI was summarized using Meta-DiSc version 1.4 (XI Cochrane Colloquium, Barcelona, Spain).
The summary estimates of pooled sensitivity and specificity were determined using the DerSimonian-Laird random-effects model.
Homogeneity among the primary studies was tested using chi-squared, and I2 statistics. I2 ≥ 50% and P < 0.05 were considered to represent substantial heterogeneity.
The diagnostic efficacy was evaluated with reference to the:
- positive likelihood ratio (PLR)
- negative likelihood ratio (NLR)
- diagnostic odds ratio (DOR)
Pooled estimates of sensitivity and specificity were used to construct the hierarchical summary receiver operating characteristic (SROC) curves.
The area under the curve (AUC)was used to compare the statistical difference between the diagnostic value of NBI and WLI. Difference over 5% was considered statistically significant.
Deeks’ plot (log of DOR vs. the standard error of the log of DOR) was constructed using STATA version 11 to evaluate potential publication bias.
P value below 0.1 for the slope coefficient was considered as significant asymmetry, which indicated potential publication bias.
Results
Here is a quick recap of the results:
Twenty-one studies including 4,880 cases (885 malignancy and 3,995 controls) were analyzed.
When using NBI for overall diagnosis in head and neck region, a pooled sensitivity of 0.90 (95% CI: 0.88 to 0.92, range: 0.47 to 1.00) and a pooled specificity of 0.97 (95% CI: 0.96 to 0.97, range: 0.79 to 1.00) were obtained.
The summary point was in the left upper quadrant of the likelihood matrix (PLR > 10 and NLR < 0.1), indicating that NBI has substantial clinical benefit in head and neck cancer screening.
In comparison with conventional WLI, the use of NBI could significantly increase the diagnostic efficacy of:
- nasopharyngeal
- laryngeal
- oral cavity; and/or
- oropharyngeal carcinoma detection
Study Characteristics And Quality Of Studies
Many studies on the use of NBI examine the sensitivity only without including any normal control in the evaluation. Hence, FP and TN values cannot be elucidated from those reports.
In order to compare the diagnostic value of NBI with WLI, only studies with appropriate NBI controls were included in subsequent analysis.
A total of 201 potentially relevant papers were initially identified from the PubMed, Embase, and MEDLINE databases.
After reading the titles and abstracts, 91 and 34 articles were excluded because of duplicated identification and irrelevance, respectively. One paper published in Japanese and three reviews were also removed.
In the remaining 75 papers, 18 articles only focused on the positive detection rate in tumors without a comparison of normal counterparts and 33 articles were without detection frequency data of NBI or WLI; these articles were also excluded.
Finally, 21 articles with detection frequency data in both cancerous and noncancerous control groups were available for the meta-analysis.
Data on NBI along with data on WLI were present in 16 articles.
Comparative analysis for NBI in cancerous and control groups was found in the remaining five articles (see picture below). The picture shows the literature search strategy and review process.
Study Quality And Publication Bias
According to QUADAS, publication scores of between 7 and 14 are required for selecting articles for meta-analysis. All the selected studies had a QUADAS score of >8. Of the 21 studies, three articles got a score of 9 and 19 studies got a score of ≥10.
Deeks’ funnel plot asymmetry test was applied for publication bias assessment. P value for NBI and WLI was 0.95 and 0.46, respectively, which indicated no potential publication bias (see picture below). The picture shows Deeks’ funnel plot with a superimposed regression line.
Overall Diagnostic Accuracy Of NBI And WLI In Head And Neck Carcinoma
In total, 4880 cases (885 malignancy and 3995 controls) in 21 studies were included in the evaluation for NBI.
NBI had a pooled sensitivity of 0.90 (95% CI: 0.88 to 0.92, range: 0.47 to 1.00) and a pooled specificity of 0.97 (95% CI: 0.96 to 0.97, range: 0.79 to 1.00) (Figure 3; see picture below).
The picture shows Forest plot of the pooled overall sensitivity and specificity for head and neck carcinoma detection by NBI and WLI, respectively.
Heterogeneity between malignancy and control groups was observed in both sensitivity (P < 0.01, I2 = 79.6%) and specificity (P < 0.01, I2 = 86.4%).
The pooled PLR, NLR and DOR were 18.30 (95% CI: 11.07 to 30.26), 0.09 (95% CI: 0.06 to 0.15), and 220.33 (95% CI: 105.57 to 459.83), respectively.
The AUC was 0.98 (Figure 4; see picture below). The picture shows the summary receiver operating characteristic (SORC) curve for head and neck carcinoma detection by NBI and WLI.
For WLI, 16 studies including 4287 cases (701 malignancy and 3586 controls) were analyzed.
The conventional WLI had a pooled sensitivity of 0.62 (95% CI: 0.59 to 0.66, range: 0.41 to 0.90) and a pooled specificity of 0.85 (95% CI: 0.83 to 0.86, range: 0.55 to 1.00). Again, heterogeneity in sensitivity (P < 0.01, I2 = 84.5%) and specificity (P < 0.01, I2 = 96.2%) were observed (Figure 3).
The pooled PLR, NLR, and DOR were 9.12 (95% CI: 5.29 to 15.73), 0.38 (95% CI: 0.28 to 0.52) and 29.88 (95% CI: 17.24 to 51.81) respectively.
The picture below (Figure 5) shows the:
- SROC curve of NBI in head and neck cancer diagnosis
- forest plot of the pooled sensitivity and specificity for nasopharyngeal carcinoma (NPC) detection by NBI and WLI
In the picture below, the NPC also shows the SROC curve.
The AUC was 0.89 (Figure 4).
In comparing the AUC of NBI and WLI, the difference between the two AUCs was 10%, indicating a statistically significant difference between AUCs of NBI and WLI.
Diagnostic Accuracy Of NBI And WLI For Nasopharyngeal Carcinoma (NPC)
Further analyses on different anatomical locations were performed to examine the region-specific variations in diagnostic value of NBI and WLI in the head and neck region.
Comparative analysis for cancer detection in the nasal cavity was found in 10 studies.
High sensitivity (0.93, 95% CI: 0.89 to 0.96) and specificity (0.99, 95% CI: 0.98 to 0.99) were observed in the NBI group (Figure 5).
In comparison, WLI had a relatively lower sensitivity (0.82, 95% CI: 0.75 to 0.88) and specificity (0.78, 95% CI: 0.76 to 0.80) for NPC detection (Figure 5).
Between these two methods, heterogeneity was observed in both sensitivity (NBI: P < 0.01, I2 = 76.3%; WLI: P = 0.03, I2 = 67.4%) and specificity (NBI: P < 0.01, I2 = 84.3%; WLI: P < 0.01, I2 = 95.6%).
The pooled PLR and pooled NLR were 32.79 (95% CI: 9.70 to 110.82) and 0.06 (95% CI: 0.02 to 0.26), respectively, in NBI, with a summary point in the left upper quadrant (PLR > 10 and NLR < 0.1) of the likelihood matrix.
On the other hand, the pooled PLR (8.57, 95% CI: 2.61 to 28.12) and pooled NLR (0.22, 95% CI: 0.12 to 0.41) of WLI contributed to a summary likelihood matrix point in the right lower quadrant (PLR < 10 and NLR > 0.1), indicating that WLI is relatively weaker in NPC detection.
The DOR of NBI and WLI was 484.67 (95% CI: 120.95 to 1942.11) and 36.46 (95% CI: 18.29 to 72.68), respectively.
Although the AUCs of NBI and WLI were relatively high (0.99 vs. 0.92), a significant difference (7.6%) was still observed (see picture below). The picture shows the summary receiver operating characteristic (SORC) curve for NPC detection by NBI and WLI.
Diagnostic Accuracy Of NBI And WLI For Cancer Originating In The Oral Cavity And/Or Oropharyngeal Regions
Fixed effect model was adopted for the calculation in the WLI group, as the I2 of all the diagnostic parameters was less than 25%.
In comparison with WLI, NBI was more sensitive in detecting cancers originating from the oral cavity and/or oropharyngeal regions (0.92, 95% CI: 0.87 to 0.95 vs. 0.50, 95% CI: 0.41 to 0.61).
Furthermore, PLR (25.11, 95% CI: 6.56–96.22) and NLR (0.09, 95% CI: 0.04 to 0.22) of the NBI group was better than that of the WLI group (PLR: 21.1, 95% CI: 2.92 to 152.30; NLR: 0.52, 95% CI: 0.42 to 0.62).
For NBI, the AUC was 0.97.
For WLI, however, the SROC curve was not available as data from two independent studies only were insufficient for the generation of the SROC curve.
Diagnostic Accuracy Of NBI And WLI For Laryngeal Carcinoma
In laryngeal carcinoma, heterogeneity was not observed (I2 < 50%) in sensitivity, specificity, PLR, NLR, and DOR for both NBI and WLI groups.
With the use of the fixed effect model, a significant difference in AUCs (7.6%) was observed.
Main Discussion
At present, there is no fully validated molecular marker for use in early HNSCC detection.
Diagnosis of superficial cancer in the head and neck regions relies heavily on endoscopic techniques and specific imaging modalities. However, in asymptomatic patients, detection of early mucosal and intra-epithelial lesions in a routine endoscopic examination is challenging.
Although it is generally accepted that NBI is of great benefit in detecting superficial mucosal lesions, the diagnostic efficacy changes according to the anatomical location. In colon cancers, it has been reported that NBI does not improve the detection of colorectal polyps in comparison with conventional colonoscopy.
In a randomized controlled trial, it was shown that visual inspection of the polyp by conventional colonoscopy had a better detection rate in comparison with the NBI group.
In addition, the efficacy of NBI is affected by the histological features of the mucosal surface and tumor morphology. In contrast, NBI offers a good detection rate at other anatomical sites, such as the upper aerodigestive tract.
As the head and neck region is covered by squamous epithelium presenting similar vascular architecture in esophagus, many studies therefore delineated the effectiveness and suggested that NBI is also potentially useful in HNSCC detection.
Endoscopic examination with multiple biopsies is the golden standard of NPC diagnosis. However, the use of conventional nasopharyngoscopy in screening suspicious lesions or early cancers is limited due to low sensitivity.
Development of NPC is commonly found in Rosenmüller fossa or RF (pharyngeal recess) of the nasopharynx, with the possibility of intracranial spread and erosion of the skull base.
Examination of the nasopharyngeal regions by endoscopic methods provides valuable information on the mucosal involvement and tumor extension.
With the use of rigid Hopkin endoscope and/or flexible fiberoptic endoscope, the following can be examined by visual inspection:
- the nasal floor
- roof of the nasopharynx
- the Eustachian tube
- the Rosenmüller fossa (RF) of the nasal cavity
The use of NBI is beneficial in identifying suspicious lesions located beneath the mucosal surface with high chance of unrepresentative biopsies and reduced FN rate.
NBI is good for NPC detection as the characteristic changes in the intrapapillary capillary loops of the nasopharyngeal lesions are remarkably revealed due to the high contrast image obtained.
Furthermore, NBI is particularly useful in differentiating cancerous lesions in case of a typical flat lesion and is suggested for use in guiding tissue sampling for subsequent histological evaluation.
The use of NBI is useful in post-irradiated NPC cases as the squamous pharyngeal mucosal surface is exposed from the lymphoid tissue due to radiation-induced tissue degeneration.
Early cancerous changes in oral cancer are visible with the use of good lighting. They could be detected by visual inspection of the abnormalities on the mucosal surface.
In general, thorough and systematic examination of the mucosal lining is critical to identify early cancerous lesions.
For visualizing lesions in the lower part of the oropharynx and tongue base, a laryngeal mirror can be used.
Our data indicated that NBI had a better diagnostic value for detecting epithelial changes in the oropharyngeal epithelia. With the use of NBI, the highly elongated and dilated capillaries resulting from cancer angiogenesis could be distinguished from benign lesions.
However, in case of hyperplasia where mucosal erosion is commonly observed, the irregular vasculature is difficult to observe even with NBI.
In particular areas where oral tongue cancers are associated with habitual betel quid or areca nut chewing, the oral mucosa has a scarring, precancerous appearance due to oral submucous fibrosis resulting from long-term chewing.
At present, there is no sufficient data for use to analyze the usefulness of NBI in distinguishing:
- oral submucous fibrosis
- oral leukoplakia
- oral cancer
The current data also suggested that in comparison with WLI, NBI is advantageous in detecting laryngeal carcinoma.
Clinically, the appearance of the larynx can be viewed by a rigid telescope or laryngoscopy with WLI. However, using light-aided endoscopy has limitations in the laryngeal regions as certain areas such as the pre-epiglottic space and paraglottic space cannot be clinically assessed.
In contrast to studies that report the significant advantage of using NBI alone in cancer screening, we here explored the relative benefits of using NBI in screening the mucosal changes in the head and neck regions.
Therefore, we only included comparative studies with normal counterparts as control in this meta-analysis.
We sought to explore whether the use of NBI could be used as a supplementary tool to improve the overall detection rate. In considering the use of NBI, it should be noticed that the performance of NBI is affected by the anatomical locations.
The current data suggests that the use of NBI could improve the diagnostic performance of the conventional WLI techniques, especially in cancerous changes from the:
- nasopharynx
- larynx
- oral cavity
- oral pharynx
The potential of NBI in other head and neck regions remained to be explored.
The advantage of NBI is that it can reveal the early benign changes based on the twisted elongation of intra-epithelial papillary capillary patterns.
At present, however, studies on the detection of benign or early head and neck cancers are limited.
Additional studies are necessary to answer whether NBI could improve patient prognosis by enhancing the detection of early head and neck cancer.
Conclusion
The diagnostic value of narrow band imaging varies in different anatomical locations of head and neck.
In comparison with white light imaging, narrow band imaging offers better diagnostic value in the screening of carcinoma originating from the head and neck mucosa.
Using narrow band imaging in combination with white light imaging is recommended in head and neck cancer screening.
Suleman Shah
Author
Suleman Shah is a researcher and freelance writer. As a researcher, he has worked with MNS University of Agriculture, Multan (Pakistan) and Texas A & M University (USA). He regularly writes science articles and blogs for science news website immersse.com and open access publishers OA Publishing London and Scientific Times. He loves to keep himself updated on scientific developments and convert these developments into everyday language to update the readers about the developments in the scientific era. His primary research focus is Plant sciences, and he contributed to this field by publishing his research in scientific journals and presenting his work at many Conferences.
Shah graduated from the University of Agriculture Faisalabad (Pakistan) and started his professional carrier with Jaffer Agro Services and later with the Agriculture Department of the Government of Pakistan. His research interest compelled and attracted him to proceed with his carrier in Plant sciences research. So, he started his Ph.D. in Soil Science at MNS University of Agriculture Multan (Pakistan). Later, he started working as a visiting scholar with Texas A&M University (USA).
Shah’s experience with big Open Excess publishers like Springers, Frontiers, MDPI, etc., testified to his belief in Open Access as a barrier-removing mechanism between researchers and the readers of their research. Shah believes that Open Access is revolutionizing the publication process and benefitting research in all fields.
Han Ju
Reviewer
Hello! I'm Han Ju, the heart behind World Wide Journals. My life is a unique tapestry woven from the threads of news, spirituality, and science, enriched by melodies from my guitar. Raised amidst tales of the ancient and the arcane, I developed a keen eye for the stories that truly matter. Through my work, I seek to bridge the seen with the unseen, marrying the rigor of science with the depth of spirituality.
Each article at World Wide Journals is a piece of this ongoing quest, blending analysis with personal reflection. Whether exploring quantum frontiers or strumming chords under the stars, my aim is to inspire and provoke thought, inviting you into a world where every discovery is a note in the grand symphony of existence.
Welcome aboard this journey of insight and exploration, where curiosity leads and music guides.
Latest Articles
Popular Articles