Advanced Endoscopic Techniques to Detect Prevalence of Synchronous Oesophageal Cancers in Patients with Head and Neck Cancers

Abstract

Objective

To study the prevalence of synchronous oesophageal cancer in patients with head and neck cancers using Narrow Band Imaging and Lugol’s chromoendoscopy.

Materials and methods

Study design: Prospective cross sectional diagnostic study. Method: 63 recruited patients with head and neck cancers, underwent haematologic evaluation, histological confirmation, imaging which included contrast enhanced computerised tomography(CECT) of the Neck and when indicated an additional Magnetic Resonance Imaging(MRI) scan followed by UGI endoscopy using white light followed by Narrow Band Imaging(NBI) and Lugol’s chromoendoscopy(LCE).

Results

Oesophageal examination picked up a gastric inlet patch in 5 subjects and epithelial hyperplasia in one while the remaining 57 were normal on using White Light Endoscopy (WLE). NBI identified an abnormal pattern in 4 patients in whom WLE was normal (3 – Type II Intra Papillary Capillary Loop (IPCL) pattern, 1- Type III IPCL pattern). LCE showed unstained mucosa of < 10 mm in six patients, the histopathology reported were normal tissue in two patients, mild chronic oesophagitis in two and, gastric metaplasia and high grade dysplasia in one each respectively. Unstained mucosa of > 10 mm in one subject showed epithelial hyperplasia. Two patients had an abnormal NBI and LCE pattern although the WLE were normal.

Conclusion

By utilizing NBI and LCE, although oesophageal cancer was not detected, premalignant pathology like high grade dysplasia, as well as gastric metaplasia, epithelial hyperplasia, esophagitis were detected. Using both NBI and LCE along with WLE would be complimentary to detect early pathological lesions. Capturing the lesions in the window period that exists between the transformation of epithelial dysplasia to carcinoma and treating them early will reduce the morbidity and also improve the outcome in oesophageal malignancy.

Introduction

Patients with head and neck cancers including laryngeal and hypopharyngeal ones are at a risk of developing second primaries in the lung, oesophageal etc. because of field cancerization. Detecting or diagnosing malignant tumours in head and neck region is difficult and identification of synchronous tumours pose an even greater challenge. With an increased incidence of synchronous cancers, detection is often delayed, adversely affecting the prognosis [1]. These patients have to undergo regular screening to diagnose second primary tumours at the earliest.

The prevalence of both high grade intraepithelial neoplasia or invasive carcinoma of the oesophagus in high risk population is around 3.2 to 28% by image enhanced endoscopy [2]. Upper gastrointestinal endoscopy with white light alone was not found useful in detecting pre-cancerous lesions in patients with suspicion of oesophageal involvement [3]. According to Hujala et al., many a time synchronous oesophageal cancer was undetected with routine endoscopy alone [4]. Of late endoscopic surveillance of oesophagus in patients that present with head and neck cancers is not a routine practice [1].

As early oesophageal cancer causes no symptoms and signs most of them present in an advanced stage. Oesophageal cancers have a poor prognosis due to detection at a late stage. Conventional white light imaging fails to identify the early mucosal changes of superficial oesophageal squamous cell cancers (ESCC).Therefore use of image enhanced endoscopy for detection of high grade dysplasia and early cancer in the oesophagus using Chromoendoscopy, Narrow Band Imaging (NBI) etc. have been advocated to identify such lesions [5]. Capturing the window period that exists between the transformation of epithelial dysplasia to carcinoma and treating these premalignant or early cancers with Endoscopic Mucosal Resection (EMR) has improved the prognosis of oesophageal cancers [6].

This prospective study was conducted with the objective to study the prevalence of synchronous oesophageal lesions in patients with head and neck cancers with the help of NBI and Lugol’s chromoendoscopy (LCE).

Materials and Methods

Approval of the institutional review board was obtained (IRB Min. Number 11355), following which this prospective observational study was conducted in the Departments of Otorhinolaryngology and Gastroenterology of a tertiary care teaching hospital in South India.

Participants

A cohort of 63 consecutive head and neck cancer patients attending the ENT outpatient clinic were recruited at our institution during the period June 2018 to June 2019. A total of 63 biopsy proven newly diagnosed head and neck cancer patients above the age of 18 years were included. Patients with sino nasal and salivary gland malignancies as well as those in whom a flexible UGI endoscopy could not be performed were excluded from the study.

Methodology (Fig. 1)

Fig. 1.

Study flow chart

After obtaining an informed consent, the demographic details of each patient were documented in the proforma. This included information regarding smoking, betel nut intake and alcohol consumption as well as site and type of cancer. All underwent haematological evaluation, histological confirmation, imaging CECT Neck and when indicated an additional MRI scan and upper gastrointestinal endoscopy (UGI endoscopy) with WLE, NBI and LCE.

UGI Endoscopy

An experienced gastroenterologist performed UGI endoscopy using the new generation HD endoscopes white light endoscopy along with optical and dye based techniques as they provide better visualization of mucosal surface details [7–9]. Endoscopy was done on all patients within 6 months of the diagnosis of head and neck cancer and in a majority of them it was performed within a month of the primary diagnosis and suspicious lesions biopsied.

White Light Endoscopy

Initially the oesophagus, stomach and duodenum were examined using white light.

Narrow Band Imaging

Following white light endoscopic examination of the oesophagus, NBI imaging using the green blue filter was used to observe the intrapapillary capillary loop pattern (IPCL). In this study the IPCL findings observed during NBI enhanced endoscopy was categorized Type I to V patterns according to the Inoue et al. classification [10].

Type I-normal IPCLs (appear as brown tennis racket shaped dots on NBI seen in normal squamous epithelium.

Type II-elongation and /or dilatational changes in the IPCLs seen in esophagitis.

Type III-have minimal colour changes from Type I IPCLs and seen in chronic esophagitis and low-grade intraepithelial neoplasia (LGIN).

Type IV-Two to three of the four morphological changes corresponds with High grade intraepithelial neoplasia (HGIN) or carcinoma in situ.

Type V-IPCLs shows all the four morphological changes and corresponds to carcinoma.

A brownish area with clear, well-defined boundary was considered as suspicious lesion. This suspicious area was biopsied and sent for histopathological examination.

Lugol’s Chromoendoscopy

After the examination using NBI, LCE was done. During LCE, about 20 ml Lugol’s solution (1%) which consists of iodine and potassium iodide, is applied with a spray catheter to stain squamous epithelium of the oesophagus. The iodine in Lugol’s solution binds to intracellular glycogen and stains normal squamous epithelium as dark green to brown or black [11].

Lugol voiding lesions appear as a white or pink in colour [12]. These suspicious areas were biopsied and sent for HPE.

Institutional Review Board (IRB) and Ethics Committee Approval

Institutional Review Board (IRB) and ethics committee approval was obtained (IRB Min. Number [Observe] 11,355).

Sample Size Calculation

Based on a study by Chen et al [13], with a precision at 85% and a desired confidence interval at 95%, using the cross-sectional (one group) proportion formula a sample size of 63 was calculated.

Statistical Analysis

The statistical analysis was done using Stata/IC v.13.1. Categorical variables were summarized using counts and percentages. Continuous variables were presented as mean and SD for normally distributed variables and with median and range for skewed variables.

Prevalence of various malignancy was reported using frequency and percentage.

Gender, hoarseness, difficulty in swallowing, haematemesis, addictive habits and site of cancer was reported using frequency and percentage. Continuous variables between two independent groups were compared using either Student t test or Mann-Whitney U test depending on whether they were normally distributed or not.

Results

Demographic Data (Table 1)

Table 1.

Baseline characteristics and demographic profile of patients with Head Neck cancers

Parameter	Number	%
Total subjects	63
Sex
Male	54	86
Female	9	14
Age median(range), years	59(25–74)
Risk factors:
Betel nut chewing
Yes	28	44.4
No	35	55.6
Tobacco use
Yes	36	57.1
No	27	42.9
Smoking Pack Years (median, range)	7(0–90)
Alcohol consumption
Yes	15	23.8
No	48	76.2
Site of Primary cancer
Oral cavity	6	9.5
Oropharynx	11	17.5
Hypopharynx	13	20.6
Larynx	33	52.4

A total of 63 patients who fulfilled the inclusion criteria were recruited to the study.

Of the 63 study subjects the median age was 59 years, the youngest was 25 and the oldest was 74 years old, and most (85.7%) of them were men. The geographical distribution of the study subjects showed the majority of them were from eastern India. Geographically 23 patients were from West Bengal, 16 from Bangladesh, 8 from Tamil Nadu, 7 from Assam, 4 from Jharkhand, 2 from Odisha, and 1 each from Bihar, Karnataka and Andhra Pradesh. While two-thirds of the subjects were tobacco users, the habit of betel nut chewing and alcohol consumption was absent in the majority. The median pack years was 7 years (range 0–90).

Over half of the patients (52.4%) had a primary cancer of the larynx. The others had hypopharyngeal (20.6%), oropharyngeal (19%) and oral cavity (9.5%) cancers. The distribution of the histological grading showed that nearly 50% (n = 31) patients had moderately differentiated cancers. Among 12 patients with oropharyngeal cancers, 7 patients were p16 HPV seropositive, 4 patients were p16 seronegative and one was equivocal. In the 63 patients, the most common presenting symptom was hoarseness (n = 39, 61.9%) followed by dysphagia (n = 24, 38.1%).

White Light Endoscopy

Endoscopy using white light to examine the oesophagus was useful in detecting gastric inlet patch in five subjects and epithelial hyperplasia in one subject. However in these subjects the NBI and LCE were otherwise normal.

Narrow Band Imaging

Oesophageal examination though normal in 57 subjects on using WLE, abnormal vascular pattern were identified in 4 of these patients using NBI (Fig. 2). Among these 4 patients, three had IPCL pattern type II and one type III. All lesions were biopsied. Among those with type II pattern, histopathology was reported as reflux oesophagitis in one, mild chronic oesophagitis with gastric heterotopia in the other and epithelial hyperplasia in the third patient and normal architecture in the patient with IPCL pattern type III (Table 2).

Fig. 2.

NBI showing Type 2 IPCL pattern for a patient with squamous cell carcinoma of the glottis

Table 2.

HPE of tissue with abnormal narrow Band Imaging pattern compared with LCE and WLE

Nature / Pattern of lesion	Number	Biopsy report - NBI	White light endoscopy report	Lugol’s chromoendoscopy report
Type II	3	Reflux oesophagitis	Normal	Normal
		Mild chronic oesophagitis with gastric heterotopia	Normal	Less than 10 mm unstained area Biopsy-mild chronic oesophagitis
		Epithelial hyperplasia	Normal	Normal
Type III	1	Normal	Normal	Less than 10 mm unstained area Biopsy-normal

Lugol’s Chromoendoscopy

Lugol’s Chromoendoscopy that was also done for all 63 patients showed unstained mucosa of < 10 mm in six and > 10 mm in one (Fig. 3). The histopathology of the six with < 10 mm unstained mucosa was reported as normal tissue in two, mild chronic oesophagitis in two, gastric metaplasia and high grade dysplasia in one each respectively. The one patient that showed unstained mucosa > 10 mm on Lugol’s chromoendoscopy, tissue biopsy taken was found to be epithelial hyperplasia (Table 3). The prevalence of the premalignant lesion –high grade dysplasia noted in the cohort was 1.5%. From Tables 2 and 3, it can be observed that WLE was normal in those with an abnormality either on NBI or LCE. The HPE report matched only in one patient with abnormal NBI and LCE. Therefore using both NBI and LCE along with WLE would be complimentary to detect early pathological lesions.

Fig. 3.

Lugol’s Chromoendoscopy showing (A) unstained area less than 10 mm in a patient with Carcinoma right tonsil and (B) more than 10 mm in a patient with Supraglottic carcinoma

Table 3.

HPE of suspicious lesions detected using Lugol’s chromoendoscopy(LCE) compared with that of WLE and NBI

Nature / Pattern of lesion	Number	Biopsy report LCE	White Light Endoscopy	NBI biopsy report
Less than 10 mm unstained area	6	Gastric metaplasia	Normal	Normal
		Mild chronic oesophagitis	Normal	Normal
		High grade dysplasia	Normal	Normal
		Normal	Normal	Normal
		Mild chronic oesophagitis	Normal	Mild chronic oesophagitis with gastric heterotopia
		Normal	Normal	No biopsy(Normal pattern)
More than 10 mm unstained area	1	Epithelial hyperplasia	Normal	No biopsy(Normal pattern)

There was no significant difference between the age and number of pack years (P = 0.44& P = 0.96; Mann- Whitney U test) when subjects with abnormal findings on LCE(n = 7) were compared with those with normal findings(n = 56).

Discussion

Identification of synchronous oesophageal cancer is important as it alters the management plan for the patient. Of late endoscopic surveillance of oesophagus in patients that present with head and neck cancers is not a routine practice [1]. This prospective study was conducted to detect synchronous malignancy in head and neck cancer, although pathology like gastric metaplasia, epithelial hyperplasia, high grade dysplasia was picked up, none of the subjects had a second primary malignancy detected in the oesophagus. However, Lugol’s chromoendoscopy and narrow band imaging was found useful in detecting pre-cancerous lesions and such patients could be followed closely.

The subjects of this study belonged to South-east Asia, a region in close proximity to the Asian oesophageal cancer belt across Central Asia, one of the two regions in the world with the highest risk for oesophageal cancer [6]. A higher prevalence of synchronous oesophageal cancer is seen among the Asian population as compared to the western because of an increased exposure to risk factors like alcohol, tobacco and betel nut chewing as well as the genetic polymorphism for alcohol metabolism. The other attributable probable factors being immune deficiency, occupation exposure to carcinogens and human papilloma virus infection which predisposes them to multiple tumours [14, 15]. HPV related tumours present at a younger age with low tobacco exposure and the major site of index tumour being oropharyngeal [16].

As seen in several other studies, the subjects were predominantly male and the mean age was 56.7 years [17–20].

The recent trends show that the incidence of head and neck cancers along with second primary tumours have increased in younger patients, probably due to the increase in alcohol consumption and heavy smoking habits prevalent among this age group. Various risk factors like alcohol intake, cigarette smoking and betel nut chewing were assessed in this study and smoking was found to be the most common risk factor among these patients with head and neck cancer [21]. The incidence of synchronous oesophageal cancer in head and neck cancer patients as reported from various studies ranges between 1.7 and 28% [2].

Wang et al. reported a higher incidence of ESCC among those with a primary involving the hypopharynx and oropharynx as compared to the larynx [19]. Oesophageal cancers usually present in an advanced stage as they remain either asymptomatic for long or have mild symptoms of dyspepsia. Strategies for early detection of oesophageal cancers should be developed as 5 year survival outcomes following treatment are significantly higher (80–90%) for early disease, than advanced disease (5–20%) [22]. Endoscopic treatment of the early lesions (high grade dysplasia and CIS) picked up during screening reduces the morbidity and mortality in these high risk group individuals.

In this study, 58% (7 out of 12) of those diagnosed with oropharyngeal cancers were found to be p16 positive and 28.5% were smokers. Those with HPV seropositive oropharyngeal SCC index tumours have a lower risk of developing second primary as compared to seronegative cancers [23]. According to Peck et al., second primary tumours in HPV seronegative patients occur in the head and neck region and other conventional sites like lung and oesophagus. However, HPV seropositive patient with associated smoking develop second primary tumours in other sites like kidney, endometrium and other less common sites [24].

White light endoscopy is an important tool in diagnosing synchronous malignancy especially in a limited resource setting. Hashimoto et al. observed that WLE was comparable with LCE in detecting advanced oesophageal cancers but was inadequate for squamous dysplasia (missed 45% of the lesions) [25]. Lugol’s chromoendoscopy has been found to be useful for the detection of early stage oesophageal cancer in patients with head and neck cancer [26]. Lugol’s chromoendoscopy can be used as a screening tool with a high sensitivity (92–100%) for detecting oesophageal squamous dysplasia. The low specificity (37–82%) has been attributed to the staining pattern in esophagitis [27–30].

Detection of these Lugol voiding lesions is easy and requires no special training or magnification endoscopy. Unstained area of 10 mm or more was strongly associated with squamous cell carcinoma or dysplastic epithelium [31]. Use of large volumes of the dye is avoided in upper part of oesophagus to prevent accidental aspiration [32].

Narrow band imaging not only requires a skilled endoscopist but also an experienced one for the interpretation of the vascular pattern of the IPCL. While NBI has better sensitivity, specificity and positive predictive value, but both NBI and Lugol’s chromoendoscopy and have similar diagnostic accuracy and negative predictive value. NBI is probably the best method used for screening superficial oesophageal neoplasms in the head and neck region [30]. Diagnostic procedures like PET-CT have been used for detection of synchronous oesophageal lesions thus avoiding pan endoscopy. According to Thakur et al., transnasal oesophageal endoscopic examination using NBI technology as compared to WLE and Lugol’s chromoendoscopy resulted in shorter examination time and lesser complications [20]. However, in our study the precancerous lesions were picked only on LCE which stained less than 10 mm.

Risk of progression of squamous dysplasia, a precursor of oesophageal squamous cell carcinoma depends on the degree of dysplasia [33]. The importance of early detection and treatment of oesophageal squamous dysplasia before malignant transformation has been well recognised. The endoscopist puts to good use the advances in medical technology utilising NBI enhancement and chromoendoscopy as screening tools in recognising asymptomatic oesophageal squamous dysplasia and early ESCC. Regular follow up of patients with dysplasia will lead to detection of cancer in an early stage. All visible lesions whether overt or subtle should be biopsied. The pink-silver sign with a high sensitivity (91.9%) and specificity (94%) on Lugol’s chromoendoscopy and NBI system helps in diagnosing high grade intra-epithelial neoplasia (HGIN) or invasive carcinoma [15]. High grade dysplasia and ESCC can be treated effectively with endoscopic mucosal resection (EMR)/endoscopic submucosal dissection thus improving overall cure rates in oesophageal malignancies [34].

Strength of the Study

This study is the first prospective Indian study based in a tertiary care hospital in South India that has incorporated white light upper GI endoscopy, Narrow Band Imaging and Lugol’s Endoscopy for detection of synchronous oesophageal lesions.

Limitations

One of the limitations in our study was that a genetic work up was not included among the risk factors for developing synchronous oesophageal lesions. Another limitation was that only a single screening UGI endoscopy was performed. Also in this study, the number of patients that presented with oral cavity and oropharyngeal cancers were few probably due to the referral pattern followed in our institution.

Conclusion

In the present study, synchronous oesophageal malignancy was not detected in any patients with head neck cancers at the time of diagnosis however the prevalence of the premalignant lesion –high grade dysplasia was 1.5%. Utilizing NBI and Lugol’s chromoendoscopy, pathology like high grade dysplasia, gastric metaplasia and epithelial hyperplasia as well as esophagitis were detected which escaped recognition with white light endoscopy. Detection of the premalignant condition of high grade dysplasia and treating it with endoscopic resection will improve the overall treatment outcome in this cohort. Multicenter studies utilizing advanced endoscopic techniques may further enhance the knowledge in this field.

Summary

• Identification of synchronous oesophageal cancer is important as it alters the management plan for the patients with head and neck cancers, however none were detected in this study.

• Premalignant lesions like high grade dysplasia and synchronous oesophageal cancers must be detected at the earliest for better outcome.

• Combining White light endoscopy with Narrow Band Imaging and Lugol’s Chromoendoscopy is complimentary and useful in detection of even premalignant lesion like high grade dysplasia etc.

Funding

This research was partly funded by the Fluid Research Grant from the institution.

The authors received no financial support for the authorship, and/or publication of this article.

Declarations

Ethics Approval

The institutional review board (Research and Ethics committee) reviewed and approved this study ( IRB Min. Number [Observe] 11355).

Competing Interests

The authors declare that they have no conflict of interest.