Multiple foci of dilated vessels as a new predictor of metachronous esophageal cancer

Prediction of the risk of esophageal squamous cell carcinoma (SCC) by endoscopic findings without iodine staining, which is irritating to the esophagus, would be beneficial. In a previous retrospective study, we found that multiple foci of dilated vascular areas (MDV) of the esophageal mucosa, seen in narrow‐band imaging (NBI)/blue laser imaging (BLI), are associated with iodine‐unstained lesions and, thus, may be a predictor of esophageal SCC. This prospective study aimed to investigate the association between MDV and metachronous esophageal SCC.


INTRODUCTION
T HE PRESENCE OF multiple iodine-unstained lesions in the esophageal mucosa is a risk factor for the development of metachronous esophageal squamous cell carcinoma (SCC) after endoscopic resection. 1,2Metachronous cancer occurs at an annual frequency of about 10% in patients with this condition.In contrast, iodine is highly irritating to the esophagus, and the use of iodine staining in clinical practice is decreasing. 3Recently, the usefulness of equipment-based image-enhanced endoscopy (IEE), such as narrow-band imaging (NBI) and blue laser imaging (BLI), in the diagnosis of esophageal cancer has been reported; [4][5][6][7] NBI or BLI provide diagnostic accuracy comparable to that of iodine, without the use of staining. 8,9Further research is required into whether equipment-based IEE alone, without iodine staining, could predict the risk of multiple heterochronic cancer.
During endoscopy, horizontal wrinkles or low vascular permeability in white light observations have been associated with multiple iodine-unstained lesions. 10In addition, low vascular permeability, nonuniform color tone, and dilated vessels 11 in BLI have been associated with metachronous esophageal SCC.We also found that multiple foci of dilated vascular areas (MDV) were associated with iodine-unstained lesions (Fig. 1a-c), and background coloration (BC) was slightly, but not notably, associated with iodine-unstained lesions 12 (Fig. 1d).However, all of these studies were retrospective.Considering that the various endoscopic findings reported here are very subtle changes compared to the endoscopic findings of cancer, the accuracy of evaluating such minor findings in a retrospective study is limited.Therefore, we conducted a prospective study to investigate the association between metachronous esophageal SCC and endoscopic findings by evaluating MDV and BC in real time during endoscopy.

Patients
T HIS SINGLE-CENTER, PROSPECTIVE, observa- tional study was conducted at the Osaka International Cancer Institute.This study was approved by the Ethical Review Committee of the Osaka International Cancer Institute (No. 1802069366) on February 6, 2018, and conducted in accordance with the Declaration of Helsinki.Informed consent to perform endoscopy was obtained from all patients and participation in the study was obtained on an opt-out basis.The inclusion criteria were as follows: (i) age 20 years or older; (ii) history of endoscopic resection for esophageal SCC (one or two lesions); and (iii) no residual or recurrent esophageal SCC.The exclusion criteria were as follows: (i) patients who had undergone esophagectomy; (ii) patients with a history of prior radiotherapy to the esophagus; and (iii) patients with reflux esophagitis of Los Angeles classification grade A or higher, or other endoscopically proven esophagitis.
High-resolution endoscopes (EG-L590ZW, EG-L600ZW7; Fujifilm, Tokyo, Japan) and a video endoscopy system (LASEREO; Fujifilm) were used for the BLI evaluation.A black soft hood (MB162 or MB46; Olympus) was attached to the tip of the endoscope to maintain a sufficient distance for observation.The initial examination was performed by two endoscopists using NBI/BLI (endoscopists R.I. and K.M., with endoscopy experience of 27 and 8 years, respectively).The endoscopist first evaluated the number of MDV and BC on endoscopy after enrollment and noted the number on the monitor before iodine staining.After the evaluation of MDV and BC, iodine staining and evaluation of the grade of iodineunstained lesions were conducted. 9The patients were monitored by surveillance endoscopy with a follow-up interval of 6-12 months for the development of metachronous esophageal SCC.Surveillance of the esophagus was mainly conducted by NBI/BLI.Iodine staining was conducted only when NBI/BLI was not sufficient for the evaluation of the esophagus.Biopsy was taken only from suspected lesions for cancer.Alcohol consumption and smoking were assessed based on patients' declaration.

Definition of MDV, BC, and metachronous cancer
MDV was defined as multiple foci of dilated vascular area (Fig. 1a-c).The dilated vascular area is a cluster of dilated intrapapillary capillary loops sized 3-10 mm but is usually ≤5 mm.Each dilated vascular area was distinguished from cancer by small size and the absence of typical findings of cancer, such as dot-like vessels. 13,14BC was defined as a brownish area without dilated vessels (Fig. 1d).Each BC was distinguished from cancer by the absence of dot-like vessels, 14 which is a typical finding of cancer.Metachronous cancer was defined as pathologically diagnosed SCC of the esophagus that developed during follow-up.However, local recurrent cancer that developed adjacent to an endoscopic resection scar was excluded from metachronous cancer.For the analysis, patients were divided into two groups based on the number of MDV and BC.The cutline (number 5) was selected for MDV because it stratified the risk of metachronous cancer in the preliminary investigation, and we believed that this cutline (number 5) would be memorable to endoscopists.The same cutline was applied for the analysis of BC.For the analysis of iodine-unstained lesions, the severity was evaluated according to the grading proposed by Katada et al., 2 which is widely accepted in Japan and several other countries (grade A, 0 iodine-unstained lesions per endoscopic view; grade B, 1-9 iodine-unstained lesions per endoscopic view; grade C, ≥10 iodine-unstained lesions per endoscopic view).

Statistical analysis
The endpoint was the rate of esophageal metachronous SCC after 2 years in the subgroup of patients stratified based on the number of MDV and BC.A two-sided P-value of less than 0.05 was considered statistically significant.A preliminary investigation showed that MDV ≥5 was detected in 25% of patients, and the 2-year metachronous SCC rate was 20% in patients with MDV ≥5 and 6% in patients with MDV ≤4.The study sample size was calculated at ~200 patients to achieve 80% power with a two-sided level of 0.05 to detect any differences between the two groups.Therefore, we estimated that 205 patients would be required, considering some deviations and dropouts.The incidence of metachronous esophageal SCC in patients with MDV and BC was estimated using the Kaplan-Meier method and compared using a log-rank test.Logistic regression analysis was conducted using four variables that had a relatively low Pvalue in univariate analysis, because of the limited number of patients with metachronous cancer.Statistical analyses were performed using R version 4.0.3(http://www.r-project.org).

Characteristics of patients and cancers
B ETWEEN FEBRUARY 2018 andMay 2019, 206 patients were enrolled in the study.After excluding four patients due to endoscopic detection of esophageal SCC at the time of enrollment and one patient because of prior radiotherapy, 201 patients were included in the analysis.Patients with reflux esophagitis were not evaluated for MDV and BC and were not enrolled in the study.The median patient age was 71 years; 164 patients were male and 37 were female (Table 1).A total of 172 patients had a history of alcohol consumption, 161 had a history of smoking, and   1.The association between the number of MDV and the number of patients with metachronous cancers is shown in Table 2.

Predictors of metachronous cancer
The association between MDV and the risk of metachronous esophageal SCC was examined.Patients were divided into two groups based on the number of MDV: MDV 0-4 (159 patients); and MDV ≥5 (42 patients).Kaplan-Meier curves were plotted for these two groups, and the incidence of metachronous esophageal SCC was significantly different between the two groups (P < 0.01) (Fig. 2a).The incidence of metachronous esophageal cancer at 2 years was 7.1% in patients with MDV ≤4 and 13.9% in patients with MDV ≥5.The association between BC and metachronous esophageal SCC was also examined.Patients were divided into two groups based on the number of BC: BC ≤4 (194 patients); and BC ≥5 (seven patients).The Kaplan-Meier curve was plotted, and the incidence of metachronous esophageal cancer was not statistically different between the two groups (P = 0.42).The incidence of metachronous esophageal cancer at 2 years was 9.0% in patients with BC 0-4 and 0% in patients with BC ≥5 (Fig. 2b).Thus, there was no association between the number of BC cases and metachronous esophageal SCC.
The association between iodine-unstained lesions and metachronous esophageal SCC was examined.Patients were divided into three groups based on the number of iodineunstained lesions in the endoscopic view: grade A, 0 (25 patients); grade B, 1-9 (148 patients); and grade C, ≥10 (28 patients).The incidence of metachronous esophageal cancer was significantly different among the three groups (P = 0.02).Concordance between MDV and iodineunstained lesions is shown in Table 3. Fifty percent (21/42 patients) of MDV ≥5 corresponded to grade C, while 95.6% (152/159 patients) of MDV ≤4 corresponded to grade A or B.
In the univariate analysis, MDV and iodine-unstained lesions were significantly associated with the development of metachronous esophageal SCC (Table 4).In the multivariate analysis, MDV and smoking were significantly associated with the risk of metachronous esophageal SCC.The odds ratio (95% confidence interval) of MDV and smoking were 2.37 (1.06-5.31)and 3.38 (1.1-10.8),respectively (Table 4).

DISCUSSION
T O THE BEST of our knowledge, this is the first prospective study to identify a high-risk group for metachronous esophageal SCC based on the esophageal mucosa observed by NBI/BLI; the number of MDV was significantly associated with the occurrence of metachronous esophageal SCC.
Studies investigating the association between NBI/BLI findings and metachronous esophageal SCC have been reported in the past. 11A single-center retrospective study reported that BLI findings of low vascular permeability, nonuniform color tone, and dilated vascular areas were associated with metachronous esophageal SCC.However, the findings associated with metachronous esophageal SCC are very subtle compared to those of SCC.The accuracy of assessing such minor findings in a retrospective study is limited.In this study the findings were assessed at the time of endoscopy, which we believe resulted in a more accurate and reliable evaluation.
Multiple foci of dilated vascular areas was strongly associated with metachronous esophageal SCC (P < 0.01) and was confirmed as an independent predictor of metachronous esophageal cancer in the multivariate analysis.The benefit of MDV over iodine-unstained lesions is that MDV can be evaluated during routine esophageal observations without staining.The disadvantage is that MDV is usually subtle and subject to interobserver variation.Further investigation to confirm the validity of the endoscopic assessment of MDV by less-experienced endoscopists is needed.Conversely, iodine staining may provide clear and objective images of iodine-unstained lesions, allowing easier evaluation than MDV.The disadvantage is that iodine staining involves an increased patient burden and longer procedure time.As described above, the two methods have different characteristics, which should be considered when applying them in daily practice.
The histological findings of MDV have not been investigated in detail.Previous studies showed that dotlike blood vessels and BC are important findings in the diagnosis of cancer. 13,14BC is reportedly caused by the thinning of the normal epithelium as the normal epithelium is replaced by cancer cells arising from the basal layer. 15Vascular changes and BC observed in MDV are weaker than those in SCC, which may represent dysplastic or inflammatory changes that were localized to the basal part of the epithelium.
We have previously reported that MDV corresponded with iodine unstained lesions. 12A histological study of iodine-unstained lesions using surgical specimens showed that dysplastic and inflammatory changes were found mainly in the basal one-third of the epithelium. 16Since these histological findings are consistent with the endoscopic images of MDV, we believe that MDV also shows dysplastic and inflammatory changes localized to the basal part of the epithelium.
Our study has some limitations.First, this was a singlecenter study, and the application of these findings need to be validated in a multicenter setting.Second, although drinking and smoking habits were not the most important points of this study, this information was obtained based on patient reports, which limits the accuracy of the information.Third, 15 patients did not receive 2-year endoscopic surveillance; however, the drop-out rate (less than 10%) was not high.Fourth, the severity of dilatation or other morphological changes in vessels may be another important determinant of the risk of metachronous cancer.However, the association between the risk of metachronous cancer and morphological change in vessels was not investigated in this study.Fifth, although each dilated vascular area is distinguished from cancer by small size and the absence of typical findings of cancer, such as dotlike vessels, the possibility that some MDV are cancer cannot be ruled out because the pathological findings of all MDV were not confirmed.However, based on the endoscopic findings of MDV, we believe that this possibility is unlikely.
In conclusion, MDV of the esophageal mucosa observed using BLI/NBI is a new way to stratify the risk of metachronous esophageal SCC without staining.

CONFLICT OF INTEREST
A UTHOR R.I. HAS received honoraria for lectures from Olympus and FUJIFILM Medical.N.U.has received honoraria for lectures from Olympus, FUJIFILM Medical, Takeda Pharmaceutical, Otsuka Pharmaceutical, EA Pharma, AstraZeneca, Boston Scientific Japan, Daiichi-Sankyo and Miyarisan Pharmaceutical Co., Ltd.K.W. has received honoraria for lectures from Takeda Pharmaceutical, Otsuka Pharmaceutical and AI Medical Service.S.S. has received honoraria for lectures from AI Medical Service, EA Pharma, AstraZeneca and Janssen Pharmaceutical.Yoji Takeuchi has received honoraria for lectures from Olympus, FUJIFILM Medical, EA Pharma, AstraZeneca, Boston Scientific Japan, Daiichi-Sankyo, Miyarisan Pharmaceutical Co., Ltd., Asuka Pharmaceutical, Zeria Pharmaceutical, Kaneka Medix, Kyorin Pharmaceutical and Japan Gastroenterological Endoscopy Society.The other authors declare no conflict of interest for this article.

Figure 2
Figure 2 (a) Multiple foci of dilated vascular areas (MDV) and incidence of metachronous esophageal squamous cell carcinoma.(b) Background coloration (BC) and incidence of metachronous esophageal squamous cell carcinoma.

Table 2
Association between number of multiple foci of dilated vascular areas (MDV) and the risk of metachronous cancer followed up until October 31, 2022.The median (interquartile range) follow-up period was 1260 (1105-1348) days.Among the 201 patients, 15 did not undergo a 2-year follow-up endoscopy.Of 159 patients with MDV ≤4, 33 metachronous cancers developed in 22 patients.Of the 42 patients with MDV ≥5, 28 metachronous cancers developed in 15 patients.Characteristics of 61 metachronous cancers are shown in Table

Table 3
Concordance of multiple foci of dilated vascular areas (MDV) and iodine-unstained lesions classification

Table 4
Univariate and multivariate analyses for predictors of esophageal metachronous cancer CI, confidence interval; ER, endoscopic resection; MDV, multiple foci of dilated vascular areas; OR, odds ratio.