Is random biopsy necessary for normal esophageal mucosa during chromoendoscopy? Evidence from a population-based cohort study

Abstract

Background

Esophageal squamous intraepithelial neoplasia could be detected in normally stained mucosa under Lugol’s chromoendoscopy. We aim to determine whether an active biopsy should be performed on such mucosa.

Methods

This study was based on a population-based screening cohort where participants underwent Lugol’s chromoendoscopy with biopsies taken from abnormal-unstaining areas and normally stained standard site. A total of 641 participants with esophageal squamous intraepithelial neoplasia or more severe lesions were included in the analysis. The cumulative incidence of high-grade intraepithelial neoplasia/esophageal squamous cell carcinoma (HGIN/ESCC) and ESCC-specific mortality were compared between participants biopsied from iodine unstained and stained areas. Additionally, paired eligible tissues were utilized for comparative genomic analysis.

Results

For participants diagnosed with low-grade intraepithelial neoplasia (LGIN), HGIN, and ESCC, 292 (54.8%), 11 (14.9%), and 1 (2.9%) cases, respectively, were biopsied from normal-staining mucosa. Over a median 9.5-year follow-up, no incident HGIN/ESCC cases were identified among individuals with LGIN diagnosed from normal-staining esophageal mucosa. In contrast, LGIN cases detected in unstained lesions exhibited a cumulative incidence of HGIN/ESCC of 15.8 (95% CI, 11.4–21.0) per 100 persons. No ESCC-related deaths occurred in patients having normal-staining lesions, irrespective of pathological grade. Contrastingly, cumulative ESCC-specific mortality per 100 persons for LGIN, HGIN, and ESCC patients of unstained lesions were 2.1 (95% CI, 0.7–4.8), 14.3 (95% CI, 6.7–25.4), and 36.4 (95% CI, 20.4–54.9), respectively. Genomic analysis revealed minimal copy number variants in normally stained lesions compared to substantial alterations in unstained lesions.

Conclusions

Random biopsies of normally stained esophageal mucosa under Lugol’s chromoendoscopy should be unnecessary for population-level ESCC screening.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12916-026-04759-2.

Background

Esophageal cancer (EC) ranks the eleventh and the seventh globally in terms of incidence and mortality, respectively [1]. Over half of new cases of and deaths from EC worldwide occur in China yearly. Of the two histological subtypes of EC, esophageal adenocarcinoma is the predominant histologic subtype in western countries, whereas 86% of cases in China are esophageal squamous cell carcinoma (ESCC) [2, 3]. To date, well-accepted etiologic factors for ESCC have not yet been identified worldwide, which poses a barrier to the implementation of primary prevention. Secondary prevention, also known as early diagnosis and treatment, including population-level screening, has thus been adopted as the main strategy for ESCC prevention.

Esophageal squamous intraepithelial neoplasia, also referred to as dysplasia, has been identified as the detectable precancerous lesion of ESCC, which could serve as an important target for early intervention and management [4]. It is commonly classified using a two-tier system, namely low-grade and high-grade intraepithelial neoplasia, corresponding to the previous categories of mild and moderate dysplasia (low-grade) and severe dysplasia and carcinoma in situ (high-grade) [5, 6]. According to the current national guideline for EC screening, patients with low-grade intraepithelial neoplasia (LGIN) should undergo endoscopic surveillance examination every 1 to 3 years, and those with high-grade intraepithelial neoplasia (HGIN) or more severe lesions should receive early clinical treatment [7]. The intraepithelial neoplasia could be visualized using Lugol’s chromoendoscopy through the reaction between iodine and intracellular glycogen, which has been proved to be a sensitive method for detecting early-stage malignancies in the squamous epithelium of the esophagus [8–10]. Lugol’s chromoendoscopy has thus been utilized in most population-level screening programs for ESCC [11–13].

However, in a population-based screening trial conducted among ~ 16,000 participants in a high-risk area for ESCC in northern China, we proposed that approximately half of histopathological confirmed squamous intraepithelial neoplasia cases (45.9% for mild dysplasia, 55.3% for moderate dysplasia) originated from normally stained esophageal mucosa, indicating that the sensitivity of Lugol’s chromoendoscopy in detecting LGIN is only 40% to 50% [14]. Cross-sectional studies in other high-risk ESCC regions, employing biopsies from normally stained mid-esophageal mucosa, also reported that Lugol’s chromoendoscopy demonstrated comparable sensitivities of 50% to 60% for detecting LGIN [15, 16]. Nevertheless, current screening guidelines in both high-risk areas such as China and non-high-risk Western countries do not require biopsies of normally stained mucosa. These raise a critical question regarding whether population-level screening practice should incorporate extensive “random biopsies” alongside targeted ones to secure comprehensive detection of precursor lesions [17]. To address this question, a prospective follow-up is needed for these histologically confirmed neoplastic lesions with different Lugol’s iodine staining statuses, to determine their risk of progression and death. Only if the risk of progression and death is comparable for intraepithelial neoplasia and early malignancy detected in normally stained esophageal mucosa versus unstained lesions should active “random biopsies” be recommended in screening practice. However, no such evidence has been reported to date.

In this study, based on a median follow-up of 9.5 years from a large-scale population-based screening trial, we compared the incidence of malignant lesions among participants with LGIN, and ESCC-specific mortality among those with LGIN, HGIN, and ESCC, stratified by whether diagnostic biopsies were obtained from normal-staining or abnormal-unstaining mucosa under Lugol’s chromoendoscopy. Additionally, we performed comparative genomic analyses to characterize molecular differences between lesions with and without Lugol’s unstaining.

Methods

Study participants

Participants in this study were obtained from the screening arm of the Endoscopic Screening for Esophageal Cancer in China (ESECC) randomized controlled trial (ClinicalTrials.gov: NCT01688908). The study design and interim results of the ESECC trial have been reported previously [13, 18]. Briefly, 668 villages were randomly selected from Hua County, a well-known high-risk area for ESCC in the Taihang Mountain region of China, and were assigned to either the screening arm or the control arm in a 1:1 ratio (334 villages and ~ 16,000 participants in each arm). A computer-aided one-on-one questionnaire investigation was conducted for all participants at the baseline enrollment to collect information on characteristics and potential risk factors for ESCC. The recruitment and upper gastrointestinal endoscopic screening were conducted from January 2012 to September 2016.

The inclusion criteria of the current study included (1) participants in the screening arm of the ESECC trial who were diagnosed with esophageal squamous LGIN, HGIN, or ESCC at baseline endoscopic screening; and (2) having available endoscopic images to confirm the iodine staining status.

Endoscopic examination and pathological diagnosis

In the screening arm, standard upper gastrointestinal endoscopy was performed by experienced endoscopists, including chromoendoscopy with 1.2% Lugol’s iodine solution [8] to spray the entire esophagus. All endoscopically detectable lesions, whether identified under white light or after Lugol’s chromoendoscopy, were biopsied. Given that squamous cell carcinoma demonstrates a predilection for the middle to lower esophageal thirds [19], biopsies were additionally obtained from predefined standard sites (28 and 33 cm distal to the incisors in the 6-o’clock position) in the absence of visual abnormalities. Following the biopsy, the site was rinsed with saline to confirm the absence of active bleeding and to ensure that the biopsied area had been adequately sampled. The number of biopsy specimens collected at each Lugol’s unstained areas depended on the size of the unstained areas, ranged from 1 to 6. For each biopsy site submitted for pathological evaluation, a paired fresh tissue sample was collected from the corresponding Lugol’s unstained areas or the standard 28 cm, and stored at − 80 ℃ for subsequent research. Endoscopic images of each participant were routinely captured every 5 cm along the entire esophagus, and close-up pictures were taken for each area where biopsy tissues were sampled.

All biopsy specimens obtained during endoscopy were processed following a standardized procedure, including fixation in 10% formaldehyde, paraffin embedding, sectioning at 5 μm, and hematoxylin and eosin staining. All specimens were evaluated by a uniform panel of expert pathologists, with disclosure of biopsy site status (normal or abnormal staining). Two senior pathologists (each with over 10 years’ experience and more than 3000 esophageal lesion diagnoses) from Anyang Cancer Hospital, a regional referral hospital in this high-risk region, independently reviewed all biopsy slides. Histopathological evaluations were conducted in accordance with the standard criteria [20], with pathologists blinded to detailed endoscopic features (including size, uniformity, regularity, etc.) and the endoscopist’s assessment. Diagnostic discrepancies were resolved through consensus review. To validate the pathological diagnosis at Anyang Cancer Hospital, a back-to-back expert review was conducted during the initial phase of the ESECC trial, wherein the slides were sent to pathology facilities at Peking University and the Department of Pathology at Peking University Cancer Hospital. This evaluation ensures the quality and reliability of pathological diagnoses.

As previously described [13], tissue biopsies were originally classified using dysplasia-based terms. In this study, following the established correspondence between dysplasia and intraepithelial neoplasia, we categorized specimens into non-intraepithelial neoplasia, LGIN (mild or moderate dysplasia), HGIN (severe dysplasia or carcinoma in situ), and squamous cell carcinoma [5, 6].

Confirmation of iodine staining status

To minimize misclassification regarding the staining status of lesions, two well-trained researchers (W.Y. and Z.Q.) retrospectively reviewed endoscopic images for all included participants, blinded to pathological diagnosis. Any discrepancies that arose were resolved through discussion or consultation with a third reviewer (M.L.). As previously described [14], suspicious lesions identified during the process of image review were reclassified as Lugol-unstained lesions (LULs). In this study, biopsies obtained from areas exhibiting Lugol-unstained regions were categorized as LUL ( +). Conversely, biopsies taken from standard sites without evidence of Lugol unstaining were classified as LUL ( −). Study participants were classified based on the iodine staining status (LUL ( +) or LUL ( −)) corresponding to their highest-grade histopathological lesion.

Outcome and follow-up

There were 2 outcomes in this study: (1) incident HGIN/ESCC identified at reexamination or during follow-up among participants initially diagnosed with LGIN, and (2) ESCC-related death among participants with LGIN, HGIN, and ESCC at baseline screening.

Per the trial protocol [18], participants who presented with either LULs or were pathologically confirmed as having LGIN at baseline examination (n = 1468) were invited to undergo a surveillance Lugol’s chromoendoscopy between May 2017 and November 2018. During the reexamination, the position and diagnosis of the lesion found at baseline were disclosed to the endoscopist for a targeted inspection and biopsy.

In addition, two approaches were used to identify incident cancer cases and all-cause death events: active surveillance involving annual door-to-door interviews conducted by trained village doctors using standardized protocols, and passive surveillance through linkage with the Hua County New Rural Cooperative Medical Scheme claims database (providing cancer type, diagnosis date, and treatment) and the county Death Registry from the Centers for Disease Control and Prevention (providing date and cause of death). This follow-up framework has demonstrated high sensitivity and specificity in incident cases and deaths identification [21, 22]. The current analysis included follow-up until May 31, 2023.

Biopsy tissue sequencing and copy number variants analysis

Paired fresh-frozen tissues, collected alongside diagnostic biopsies, were subjected to whole genome sequencing (WGS) analysis. Genomic DNA was extracted from the fresh-frozen tissues using the DNeasy Blood & Tissue Kit (QIAGEN, USA) following the manufacturer’s protocol. Shallow WGS was performed on the esophageal tissue samples using the Illumina NovaSeq platform (Illumina Inc., San Diego, CA, USA) with 150-bp paired-end reads at an average sequencing depth of 10 × coverage across the genome. Raw sequencing reads were processed following the GATK4 [23] best-practice workflow, with alignment to the UCSC human reference genome (hg19). Copy number variants (CNVs) were identified with Control-FREEC [24], with genomic regions exhibiting an absolute log2 ratio > 0.1 classified as CNVs. The Wilcoxon test was used to conduct comparisons among groups.

Statistical analysis

We evaluated the risk of incident esophageal malignancy (HGIN/ESCC) among participants with LGIN, and the risk of ESCC-related death among participants with LGIN, HGIN, and ESCC, stratified by Lugol’s staining status. Cumulative incidence and mortality per 100 persons were calculated as the number of incident cases or ESCC-related deaths divided by the number of participants in each specific subgroup. The Nelson-Aalen cumulative hazard function was utilized to illustrate cumulative incidence curves, and one minus the Kaplan–Meier estimator was used to delineate cumulative mortality curves. Differences between cumulative curves were estimated using the log-rank test.

All analyses were performed using Stata (version 16.0, StataCorp., TX) and R software (version 4.3.2, R Foundation). Two-sided P-values < 0.05 were considered statistically significant.

Results

Patients’ iodine staining status

The ESECC trial screening arm initially enrolled 17,151 eligible participants. Among them, 15,299 individuals completed both baseline upper gastrointestinal endoscopy and standardized questionnaire investigation. Subsequent pathological evaluation revealed 112 participants without available pathological diagnosis, while 14,513 participants were not diagnosed with esophageal neoplastic lesions; all of these individuals were excluded from the current analysis. Additionally, 33 participants were excluded due to unavailable endoscopic images. Consequently, a total of 641 participants were included in the current study (Fig. 1). Among the remaining patients, 533 cases of LGIN, 74 cases of HGIN, and 34 cases of ESCC were identified. Of these, 292 (54.8%) LGIN cases, 11 (14.9%) HGIN cases, and 1 (2.9%) ESCC case were detected in mucosa with normal iodine staining. Treatment of HGIN and ESCC patients stratified by Lugol’s staining status is presented in Additional File 1: Table S1.

Fig. 1.

Flowchart of this study. Abbreviation: ESECC, Endoscopic Screening for Esophageal Cancer in China

Cumulative incidence of malignancy

During a median follow-up of 9.5 years (maximum follow-up: 11.5 years), among participants in the LUL ( +) group, 38 cases of LGIN progressed to HGIN/ESCC, corresponding to cumulative incidence of 15.8 (95% CI, 11.4–21.0) per 100 persons (Table 1). In contrast, no progression to HGIN/ESCC was observed in the LUL ( −) group, with cumulative incidence rates of 0.0 (95% CI, 0.0–1.3) per 100 persons. The Nelson-Aalen cumulative hazard plot consistently revealed that LUL ( −) patients had minimal to no risk of HGIN/ESCC progression during the study period, contrasting with LUL ( +) individuals (Fig. 2A; log-rank P < 0.001).

Table 1.

Cumulative incidence and mortality stratified by Lugol’s staining status and pathology over a median follow-up period of 9.5 years

Lugol’s staining status	Pathological diagnosis	No. of participants (n = 641)	No. of incident HGIN/ESCC (n = 38)	Cumulative incidence of HGIN/ESCC (per 100 persons, 95% CI)a	P valuesb	No. of ESCC-related deaths (n = 26)	Cumulative ESCC-specific mortality (per 100 persons, 95% CI)a	P valuesb
LUL ( −)	LGIN	292	0	0.0 (0.0–1.3)c		0	0.0 (0.0–1.3)c
	HGIN	11	—			0	0.0 (0.0–28.5)c
	ESCC	1	—			0	0.0 (0.0–97.5)c
LUL ( +)	LGIN	241	38	15.8 (11.4–21.0)	< 0.001	5	2.1 (0.7–4.8)	0.02
	HGIN	63	—			9	14.3 (6.7–25.4)	0.21
	ESCC	33	—			12	36.4 (20.4–54.9)	0.65

Abbreviations: ESCC esophageal squamous cell carcinoma, HGIN high-grade intraepithelial neoplasia, LGIN low-grade intraepithelial neoplasia, LUL Lugol-unstained lesion

^aThe exact confidence interval was calculated

^bOne-side Fisher’s exact test was used to compare the incidence or mortality between different LUL groups within the same pathological diagnosis

^cOne-sided 97.5% confidence interval

Fig. 2.

Cumulative incidence of HGIN/ESCC and ESCC-specific mortality. Abbreviations: ESCC, esophageal squamous cell carcinoma; HGIN, high-grade intraepithelial neoplasia; LGIN, low-grade intraepithelial neoplasia; LUL, Lugol-unstained lesion. A Cumulative incidence of HGIN/ESCC among participants diagnosed as LGIN (log-rank P < 0.001); B cumulative ESCC-specific mortality among participants diagnosed as LGIN, HGIN, and ESCC. Log-rank test: LUL ( −) LGIN vs. LUL ( +) LGIN, P = 0.02; LUL ( −) HGIN vs. LUL ( +) HGIN, P = 0.18; LUL ( −) ESCC vs. LUL ( +) ESCC, P = 0.51

Cumulative ESCC-specific mortality

Among all included participants, 26 died of ESCC. The cumulative ESCC-specific mortality per 100 persons was 2.1 (95% CI, 0.7–4.8) for LUL ( +) with LGIN, 14.3 (95% CI, 6.7–25.4) for those with HGIN, and 36.4 (95% CI, 20.4–54.9) for those with ESCC. Detailed analysis of the death cases indicated that these patients were generally older males, and most had received multimodality treatment regimens (see Additional File 1: Table S2). No ESCC-related deaths were observed in the LUL (–) group. Cumulative ESCC-specific mortality curves stratified by pathological diagnosis are presented in Fig. 2B (log-rank test: LUL (–) LGIN vs. LUL ( +) LGIN, P = 0.02; LUL (–) HGIN vs. LUL ( +) HGIN, P = 0.18; LUL (–) ESCC vs. LUL ( +) ESCC, P = 0.51).

Copy number variants stratified by iodine staining status

A total of 403 eligible paired esophageal mucosa tissues were subjected to WGS, comprising 94 (23.3%) without LULs and 309 (76.7%) with LULs. Four tissues failed quality control and were excluded from the analysis. As shown in Fig. 3, tissues from the LUL ( −) group showed minimal CNV burden (< 2% of genomic regions affected), whereas tissues from the LUL ( +) group demonstrated significantly higher CNV involvement (P < 0.0001). Focal CNVs regions identified in LULs ( +) included amplification of 2q21.1 (LRP1B) and 8q24.21 (MYC), as well as deletion in 3p14.2 (FHIT) and 9p21.3 (CDKN2A) (Fig. 4).

Fig. 3.

Copy number variants burden in esophageal lesions with different Lugol’s staining status. Abbreviations: CNV, copy number variants; ESCC, esophageal squamous cell carcinoma; HGIN, high-grade intraepithelial neoplasia; LGIN, low-grade intraepithelial neoplasia; LUL, Lugol-unstained lesion. Differences between groups were evaluated using the Wilcoxon test. ****P < 0.0001

Fig. 4.

Average copy number variants in esophageal lesions with different Lugol’s staining status. Abbreviations: CNV, copy number variants; ESCC, esophageal squamous cell carcinoma; HGIN, high-grade intraepithelial neoplasia; LGIN, low-grade intraepithelial neoplasia; LUL, Lugol-unstained lesion

Discussions

Over the past decades, esophageal squamous intraepithelial neoplasia has been recognized as the asymptomatic precursor lesion of ESCC, which can be endoscopically visualized after staining with Lugol’s iodine solution. Lugol’s chromoendoscopy technique, followed by targeted biopsy and pathological examination, has been widely considered as the “gold standard” for population-level ESCC screening given its high sensitivity [7]. Nevertheless, the ESCC trial, which represents the first real-world assessment of the sensitivity of iodine staining in a large-scale, population-based chromoendoscopic screening for a high-risk population in China [14], has found a lower sensitivity of LGIN compared to previous smaller and less representative studies [15, 16]. These findings suggested that endoscopically normally stained mucosa may occasionally exhibit histopathological abnormalities, underscoring the potential need to incorporate active “random biopsy” into current Chinese screening guidelines to enhance the detection of precursor lesions and optimize the beneficial effect of screening. However, given the invasive nature and substantial cost of biopsies, the long-term clinical outcomes of these “histologically neoplastic but normally stained” lesions should be thoroughly assessed before recommending random biopsies for visually normal-staining esophageal mucosa.

In this study, based on a large-scale prospective screening cohort with long-term follow-up conducted in a high-risk area for ESCC in China, we found a notable disparity of the risk of progression and ESCC-related death between participants with the same pathological diagnosis but different iodine staining statuses. Specifically, individuals with LGIN whose biopsies were taken from normally stained mucosa under Lugol’s chromoendoscopy showed a remarkably low incidence of HGIN/ESCC, comparable to those with abnormalities detected during chromoendoscopic examination. Furthermore, mortality analysis revealed that pathologically confirmed but normal-stained lesions did not contribute to any ESCC-related death, regardless of the severity of pathological abnormalities. This result underscores an extremely low risk of progression to death, even for HGIN and ESCC originating from visually normal squamous mucosa under Lugol’s chromoendoscopy. Further analysis showed that none of the HGIN and ESCC patients in the normally stained group had received timely treatment (see Additional File 1: Table S1). Previous studies with a median follow-up of 8.5 years reported ESCC-related deaths emerged two years after baseline screening in untreated HGIN (severe dysplasia/carcinoma in situ) patients [25, 26]. Notably, our study, with a median 9.5-year follow-up, observed no ESCC-related deaths in these patients, which further supports the indolent nature of the normally stained lesions. Taken together, these findings highlight the critical role of Lugol’s iodine staining status in determining the long-term clinical outcomes for early esophageal lesions, supporting current screening guidelines in China and Western countries, which recommend targeted biopsies from Lugol's unstained areas while confirming that random biopsies in normally stained mucosa are unnecessary.

The heterogeneity between normally stained and abnormal-unstained neoplastic lesions was also observed in genomic analysis of paired fresh-frozen tissues. Normally stained neoplastic lesions showed an extremely low frequency of CNVs compared to unstained lesions identified by Lugol’s chromoendoscopy. Focal CNV regions, such as CDKN2A [27], previously reported in neoplastic lesions, were absent from the genomic profile of normally stained neoplastic lesions. These findings further demonstrated the molecular distinction between lesions with different iodine staining statuses, even within the same pathological diagnosis, suggesting that iodine staining may reflect molecular diversity independent of histopathology.

The mechanism underlying the low risk of progression and death observed in normally stained mucosa remains to be elucidated. Glycogen metabolism may provide a plausible explanation. It is well recognized that iodine staining depends on its reaction with intracellular glycogen [8–10]. Iodine unstained lesions indicate glycogen deficiency in squamous epithelium cells, which may be due to a high proliferation rate and poor cell differentiation associated with malignant changes [28]. By contrast, normally staining areas may represent mucosa with a more stable metabolic state and lower proliferative potential, corresponding to more benign biological behavior. Further studies are needed to investigate the cellular behavior associated with normally stained mucosa.

Pathological evaluation remains the cornerstone in the diagnosis, treatment, and management of ESCC in current population-level screening programs [7]. Our study, which incorporates epidemiological and genomic analyses, explicitly suggests that surveillance and management of individuals with esophageal precursor lesions should be guided by iodine staining status. That is, unstained lesions detected by Lugol’s chromoendoscopy indicate a significantly elevated risk of malignant progression and ESCC-related death. Previous studies also demonstrated that these patients are at risk of disease progression and death [26, 29]. Consequently, post-treatment management is also essential, particularly for those with multiple LULs, who face an increased risk of developing second primary ESCC or head and neck cancer [30, 31]. Therefore, intensive surveillance using Lugol’s chromoendoscopy or image-enhanced endoscopy should be recommended. Contrastingly, normally stained esophageal mucosa carries minimal risk of progression and death, regardless of the severity of pathological abnormalities, making endoscopic surveillance and early treatment unnecessary for such cases. Our previous studies based on the ESECC trial demonstrated that endoscopic features of LULs, including size [32], multiplicity [33], and persistence [34], are independent predictors of precursor lesions progression. Genomic and epidemiologic evidence further indicates that non-neoplastic LULs should be the initial stage in the carcinogenic process of ESCC, earlier than cytological changes [35]. These findings collectively underscore that Lugol’s iodine staining is indicative of the biological behavior and ultimate outcome of the esophageal squamous carcinogenic process [36].

The strengths of this study include the utilization of a large-scale, population-based screening cohort with up to 11-year follow-up, which ensured the representativeness of a real-world screening scenario and accuracy of outcome determination, and the Lugol’s iodine staining statuses for biopsied lesions were reconfirmed through a comprehensive review of endoscopic images. However, it should be noted that this study was conducted in a single center located in a high-risk area of rural China, and as such, our findings still need to be further validated for generalizability to other settings.

Conclusions

In summary, by leveraging prospective long-term follow-up of patients with esophageal neoplastic lesions and integrating epidemiological and genomic analyses, our study demonstrated that neoplastic lesions biopsied from normally stained mucosa exhibit an extremely low risk of malignant progression and ESCC-related death, and harbor a minimal burden of CNVs. Therefore, random biopsies and surveillance for normally stained mucosa under Lugol’s chromoendoscopy are unnecessary in ESCC screening practice.

Abbreviations

CNVs

Copy number variants

EC

Esophageal cancer

ESCC

Esophageal squamous cell carcinoma

ESECC

Endoscopic Screening for Esophageal Cancer in China

LULs

Lugol-unstained lesions

LGIN

Low-grade intraepithelial neoplasia

HGIN

High-grade intraepithelial neoplasia

WGS

Whole Genome Sequencing

Authors’ contributions

ZHH and YK conceived and designed the study. WLY, MFL, ZFQ, YH, ZL, CHG, AXL, HJY, FLL, YL, FFL, YQP, HC, ZHH, and YK acquired the data. WLY, MFL, ZFQ, and YH analyzed and interpreted the data. WLY, MFL, ZFQ, and YH drafted the article. WLY, MFL, ZHH, and YK made critical revisions to the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the National Key R&D Program of China (2021YFC2500405), the National Natural Science Foundation of China (82303952, 82073626, 81773501), Science and Technology Planning Project of Tibet Autonomous Region (XZ202501JD0021).

Data availability

The clinical data generated and/or analyzed from this study will not be publicly available due to protection of patient privacy but may be accessible upon reasonable request and subject to approval by the corresponding author. The raw sequencing data generated and/or analyzed in this study have been deposited in the Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA015647) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of the Peking University School of Oncology (approval number 2011KT27). Participants gave informed consent to participate in the study before taking part.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.