Metachronous multiple esophageal squamous cell carcinomas and Lugol-voiding lesions after endoscopic mucosal resection

Background and study aims: Patients and methods: Results: P =  Conclusions:

Introduction ! Esophageal cancer is the seventh most common cancer worldwide, and its incidence has increased rapidly over the past three decades [1]. Marked geographic variation exists in the incidence of this cancer, and regional differences exist in the prevalence of adenocarcinoma and squamous cell carcinoma (SCC). Although most tumors are adenocarcinomas in Western societies, SCCs constitute 80 % of esophageal cancers in the world. This tumor type is especially prevalent in Asia, in countries such as China, India, and Japan [2]. Chromoendoscopy with Lugol solution is highly sensitive for identifying dysplasia and superficial SCC [3,4]. According to the Lugol staining pattern, areas completely "void" of stain were found in approximately 90 % of high-grade dysplasia and carcinoma, whereas approximately 90 % of areas that were stained less intensely than normal epithelium were nondysplastic lesions and the remaining 10 % were dysplastic [5]. Therefore, Lugolvoiding lesions (LVLs) are detectable not only in dysplasia and carcinomas but also in nondysplastic areas, such as those in esophagitis.
Superficial esophageal SCC is defined as a lesion that is confined to the esophageal mucosa or submucosa, regardless of the presence or absence of regional lymph node metastases [6]. The rate of diagnosis of superficial esophageal SCC has increased due to improved diagnostic procedures, such as endoscopic examination including chromoendoscopy with Lugol's solution [7,8]. Recently, the minimally invasive technique of endoscopic mucosal resection (EMR) has been applied to the treatment of superficial esophageal SCC [9,10]. The disease can recur in the esophagus in two patterns: local recurrence and metachronous multiple SCCs. Because EMR spares a larger area of the esophageal mucosa, metachronous multiple SCCs may occur more frequently after EMR than after surgical treatment. One of the advantages of EMR is the high quality of life resulting from preservation of the entire esophagus; however, the preserved esophagus might still require surgical resection later if metachronous multiple SCCs develop. The incidence and characteristics of superficial metachronous esophageal SCCs and the association of LVLs with metachronous esophageal SCCs are still unclear, Background and study aims: Endoscopic mucosal resection (EMR) has been applied to the treatment of superficial esophageal squamous cell carcinoma (SCC). The incidence and characteristics of metachronous multiple esophageal SCCs and Lugol-voiding lesions (LVLs) were investigated in a retrospective study in patients who had undergone EMR for superficial esophageal SCC. Patients and methods: 96 patients with esophageal SCC who had been treated by EMR were followed up by endoscopy for 12 months or longer. Clinicopathologic parameters such as tumor size and location and presence of LVLs were examined. Results: 10 patients (10 %) had synchronous multiple SCCs, and 12 (13 %) developed metachro-nous multiple SCCs. The mean annual incidence of newly diagnosed tumor was 4.4 %. The incidence of a speckled pattern of LVLs was 20/74 (27 %) in patients with solitary SCC, 5/10 (50 %) in synchronous multiple SCC, and 10/12 (83 %) in metachronous multiple SCC. The incidence of the presence of speckled pattern of LVLs was significantly higher in patients with multiple SCCs than in those with solitary SCC (68 % vs. 27 %, P = 0.0004). Conclusions: Patients who have undergone EMR for esophageal SCC, especially those with metachronous multiple LVLs in the background mucosa, should undergo follow-up with close endoscopic observation using Lugol staining. and analysis of these factors would be useful for deciding how best to follow the preserved whole esophagus following EMR. We therefore carried out a retrospective investigation to clarify the incidence and characteristics of metachronous multiple esophageal SCCs, including the presence of LVLs, following EMR treatment in patients with superficial esophageal SCC.

!
From December 1990 to March 2007, 140 patients with esophageal SCC who had been treated with EMR at Hiroshima University Hospital (Hiroshima, Japan) were enrolled in this study. Patients with lesions meeting all of the following criteria were considered suitable candidates for EMR at the institution: clinical findings of SCC confined within intraepithelial areas and lamina propria muscularis (LPM) without lymph node metastases and distant metastases. SCCs confined within these areas are in early stages, and are known to have no lymph node and distant metastases. EMR was performed using the strip biopsy method. The description of the SCCs and histologic evaluation of the resected specimens were in accordance with the Japanese Classification of Esophageal Carcinoma [11]. For tumor location, the esophagus was divided into five regions: the cervical esophagus, upper thoracic esophagus, middle thoracic esophagus, lower thoracic esophagus, and the abdominal esophagus. Periodic endoscopic examinations were recommended at least annually to look for recurrences of esophageal SCC. Synchronous multiple esophageal SCCs were diagnosed in accordance with the criteria of Warren and Gates [12]. Lesions detected within 1 year of the initial EMR were regarded as synchronous multiple lesions because microcancers might have been missed at the time of EMR. Metachronous multiple lesions were defined as new esophageal SCCs in different areas from the initial cancer and occurring at least 1 year after the initial EMR. Patients who had both synchronous and metachronous SCC were assigned to the metachronous SCC group. For patients with synchronous and metachronous multiple lesions, data for the largest or initial lesion were analyzed. The endoscopic examinations included spraying Lugol's solution via a catheter in order to detect recurrent lesions. Clinicopathologic characteristics, including patient age at diagnosis, sex, macroscopic type, location of lesion, tumor size, and presence of LVLs in background esophageal mucosae, were evaluated. A speckled pattern of LVLs was defined as cases with many more than 10 small LVLs or numerous irregularly shaped multiform LVLs in the esophageal mucosae (• " Fig. 1) [13].
For statistical analysis, the chi-squared test and Fisher's exact probability test were used. P < 0.05 indicated statistical significance.

!
Of the 140 patients with superficial esophageal SCC who were initially treated with EMR, 96 were followed up by endoscopy for 12 months or longer and were included in this analysis. The characteristics of these 96 patients are listed in• " Table 1.
The median age at diagnosis was 64.5 years (range, 47 -88 years), the median follow-up period by endoscopy was 62.7 months (range, 12 -154 months), and the median follow-up number of endoscopic examination was 9.74 times (range, 2 -31). None of the patients had local recurrence or distant metastasis or died of esophageal SCC, and there were no treatment-related deaths. Six patients died of malignant diseases in other organs (two lung cancers, one gastric cancer, one tongue cancer, one hepatic cancer, and one malignant lymphoma). Four patients died of pneumonia, two died of hepatic failure, two died of renal failure, and three died suddenly of an unknown cause. None of the patients had serious adverse reactions associated with Lugol staining. Ten (10 %) of the 96 patients had synchronous multiple superficial esophageal SCCs (nine with double, two with triple, and one with quadruple lesions) (• " Fig. 2). Two of these patients subsequently developed metachronous SCCs. In total, 12 patients (13 %) developed metachronous multiple esophageal SCCs, and two of these patients subsequently developed a third SCC. The median interval between the diagnosis of the primary esophageal SCC and the diagnosis of the first metachronous multiple SCC was 27.3 months (range, 13 -45 months). All of these patients had intramucosal SCCs. The mean annual incidence of metachronous cancer was 4.4 % (• " Fig. 3). All of these patients with metachronous multiple SCC were treated with EMR. The characteristics of the patients with solitary, synchronous multiple, and metachronous multiple SCCs are shown in • " Table 1. There were no significant differences between solitary, synchronous, and metachronous SCCs with regard to patient sex, macroscopic type, or tumor size. The age at diagnosis of initial SCC in multiple metachronous SCC patients tended to be younger than the age at diagnosis in solitary and multiple synchronous SCC patients. The incidence of the speckled pattern of LVLs was 20/74 (27 %) in solitary SCC, 5/10 (50 %) in synchronous multiple SCCs, and 10/12 (83 %) in metachronous multiple SCCs. The incidence of the presence of the speckled pattern of LVL was signifi- Fig. 1 Representative examples of the speckled pattern of Lugol-voiding lesions (LVLs). a Many more than 10 small LVLs. b Numerous irregularly shaped multiform LVLs.
cantly higher in synchronous and metachronous multiple SCCs than in solitary SCC (P = 0.0004 by Fisher's exact probability test). The pattern of primary and secondary lesions in multiple SCCs is shown in• " Table 2.
Interestingly, 36 % of secondary lesions were located in the same segment of esophagus as the primary lesion.

!
In the present study, we showed that metachronous multiple esophageal SCCs develop frequently in patients who have undergone EMR treatment for esophageal SCC, and that careful endoscopic observation is required for these patients, especially those with LVLs in the background esophageal mucosa. An increased rate of metachronous multiple esophageal SCCs was expected in patients treated by EMR for esophageal SCC because the largest area of esophageal mucosa remains in these patients. Several researchers have reported that the rates of metachronous multiple esophageal SCCs after EMR for superficial esophageal SCC were 7.8 % -20.0 % [7, 14 -16]. In the present study, the mean annual incidence of metachronous cancer was 4.4 %, which was comparable with previous reports. In addition, 20 % of the patients in the present study developed metachronous lesions within 3 years after initial EMR. Therefore, careful annual follow-up endoscopic examinations may be needed for at least 3 years after initial EMR for patients who have been treated with EMR for esophageal SCC. In the present study, synchronous multiple esophageal SCCs were detected relatively frequently (10 %), as were metachronous lesions (13 %). Kuwano et al. [17] detected synchronous multiple esophageal lesions in Japanese patients more frequently, in 11 of 43 patients (26 %). The difference in rates between the study by Kuwano et al. and the present study may reflect different populations. Because synchronous multiple esophageal lesions may be observed relatively frequently, other areas in the esophagus must be checked carefully to avoid missing the synchronous lesion(s). To detect synchronous or metachronous esophageal SCCs, the Lugol spraying method has been recommended. Because squamous dysplasia and SCC lack glycogen, these lesions are revealed as LVLs. Researchers have reported the usefulness of this method to diagnose esophageal SCC. For example, Shimizu et al. [18] examined 326 consecutive patients with head and neck cancer (patients who are at high risk of esophageal SCC), and reported that 45 % of high-grade intraepithelial neoplasias were missed without Lugol staining. Dubuc et al. [19] examined the advantages of systemic Lugol staining in a total of 1095 patients in 62 endoscopy centers. Of the 35 cancers detected, 20 % were detected only after Lugol staining. Lugol staining improves the diagnostic accuracy of endoscopy. In addition to aiding in the diagnosis of esophageal SCC, use of the Lugol spraying method can evaluate the risk states of developing SCC. The speckled pattern of LVLs is the endoscopic characteristic. It has been reported that metachronous multiple esophageal SCC developed in 38 % of patients treated with EMR for esophageal SCC who had the speckled pattern of LVLs in the background mucosa compared with only 9 % of those without LVLs [13]. The cumulative proportion of metachronous multiple SCCs was significantly higher in patients with the speckled pattern of LVLs than in those without LVLs. The presence of multiple LVLs was the independent risk factor for multiple cancers (relative  [20]. In the present study, the incidence of presence of the speckled pattern of LVLs was significantly higher in synchronous and metachronous multiple SCCs than in solitary SCCs. These results may truly imply that the existence of the speckled pattern of LVLs may indicate that the esophageal SCC is at high risk of progressing to metachronous multiple SCCs. Characteristics of synchronous and metachronous multiple superficial esophageal SCCs were not significantly different in the present study, and these lesions arose in an area of the esophagus similar to that of the primary lesion. It is well established that multiple cancers can occur in the esophagus. This phenomenon has been explained by the concept of "field cancerization," in which repeated exposure to carcinogens leads to an accumulation of genetic alterations, resulting ultimately in the development of multiple and independent cancers [21]. Multiple LVLs in the esophageal mucosa are significantly associated with heavy consumption of alcohol. Additionally, p53 mutations and the pat-tern of DNA methylation in intraepithelial neoplasia (IEN) and SCC occurring in the same patients have been examined [22]. The pattern of methylation in individual loci indicated a progressive accumulation of methylations from non-neoplastic epithelium to IEN to SCC in individual patients. The authors suggested a field of methylation may precede and has some causative relation with the development of IEN and SCC. However, they also found that the specific mutations in p53 differed between IEN and SCC from the same patients. Mutations in p53 may occur independently in multiple sites and may contribute to carcinogenesis as well as to DNA methylation. These findings on the p53 mutations may be contradictory to the concept of "field cancerization". Risk factors for esophageal SCC such as genetic factors, including genetic polymorphisms, and environmental factors, such as tobacco smoking and alcohol consumption, have been reported to be important [23]. As for genetic polymorphisms, the *1 and *2 genotypes of ALDH2, a gene encoding an alcohol-metabolizing   [26]. Alcohol consumption is also linked to increased risk of esophageal cancers. For persons who consume more than 4 drinks (= 47.5 g of pure ethanol) per day, the relative risk of developing such cancers is 4.2 compared with nondrinkers. A synergistic effect was observed in persons who both smoke tobacco and drink alcohol. The relative risks of developing the cancers are 12.7 among those who smoke > 30 cigarettes a day and consume > 4 drinks per day. In the present study, the age at diagnosis of initial SCC in patients with multiple metachronous SCC tended to be younger than the age at diagnosis in patients with either solitary or multiple synchronous SCC. The findings might represent an inherited genetic susceptibility to cancer and/or patients' habits of tobacco smoking and alcohol consumption.
To prevent death due to esophageal SCC, screening programs that use endoscopy or other modalities are crucial. As described previously, Lugol staining is useful to detect early and small lesions of the esophagus. However, Lugol staining often causes mucosal irritation leading to retrosternal discomfort [27]. Complications of Lugol staining, including acute esophageal and gastric mucosal damage, have also been reported [28]. Recently, the usefulness of a narrow-band imaging system with magnifying endoscopy has been reported [29 -32]. This method requires no chemical agents and is without complications. It has been applied clinically to the detection of esophageal SCC and also head and neck cancers and may overtake Lugol staining in efficacy [33]. Further research is needed to clarify the significance of the narrow-band imaging system in screening programs. In addition to screening programs, educational programs warning of the risks of tobacco smoking and alcohol consumption are crucial. People at high risk, such as heavy smokers and drinkers and those with a family history of upper aerodigestive tract cancer, have an eightfold greater risk of developing a second cancer compared with those without a family history of any cancer [34]. Such at-risk people need to acquire this knowledge and participate in screening programs.
In conclusion, esophageal SCCs are often newly detected after initial EMR for esophageal SCC. Patients who have undergone EMR for esophageal SCC, especially those with multiple LVLs in the background mucosa, should be followed with close endoscopic    Competing interests: None