1 Introduction

In 2020, the number of esophageal cancer (EC) patients increased by 600,000, ranking 8th among all cancers, while the number of deaths increased by 540,000, ranking 6th among all cancers; In China, the incidence of EC increased by 320,000 in 2020, ranking the 6th, while the death rate increased by 300,000, ranking the 4th [1]. Recently, the nationwide statistics for cancer incidence and mortality in 2016 reported by National Cancer Center (NCC) of China revealed that the incidence of EC ranks 5th in male and 8th in female, and the mortality ranks 4th in male and 6th in female [2]. Thus, EC is still a major malignant tumor affecting the health of Chinese residents.

It included two main pathologic types, esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC). ESCC is the most common subtype, accounts for 90% of EC cases worldwide, with the highest incidence in China, Central Asia, East and South Africa, while EAC mainly occurs in South America and Europe [3]. In China, EC accounts for 70% of the world’s EC cases [4].

It still remains unknown that how EC occurs. Current studies have shown that EC is a multi-stage process that gradually develops from repeated inflammatory stimulation to precancerous lesions and then to invasive cancer. This process is very complex and involves changes in esophageal epithelial and non-epithelial cells cytogenetics in the microenvironment, and its molecular mechanisms are still being explored [5]. Some studies have shown that high frequency mutations in some driver genes are associated with the development of EC, such as p53, p16, PTEN, RB1, ZNF750 and so on [6,7,8,9,10,11,12,13].

Early esophageal cancer is defined as the lesion only involving the mucosa without suspected lymph node metastasis [14]. About 20% patients are diagnosed with early EC, and the 5-year survival rate is about 85% after lesion resection [15, 16]. The commonly used diagnostic methods, such as Lugol's iodine staining guided biopsy under white light endoscopy (WLE) or narrow-band imaging (NBI), can well diagnose esophageal precancerous lesions and early EC. However, due to the fact that patients with esophageal precancerous lesions and early EC are mostly asymptomatic, endoscopic screening is costly and painful, so endoscopic screening is not suitable for large-scale screening, which lead to a low rate of early diagnosis, resulting in a high incidence and mortality. Due to the high cost of treatment, the large surgical trauma and large amount of complications, the quality of life of patients with middle and advanced stage of EC is poor. However, endoscopic treatments for early EC such as endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) not only have less trauma, long survival time, but also preserve esophageal structure and function, which maintain a high quality of life. Therefore, the early diagnosis and treatment of EC is crucial.

Traditionally, esophageal precancerous lesion was defined as atypical hyperplasia, and it was classified as mild, moderate, and severe. In recent years, WHO classification of esophageal tumors defined esophageal precancerous lesion as intraepithelial neoplasia/dysplasia, and low grade and high grade were classified based on the depth of invasion. Low grade intraepithelial neoplasia/dysplasia invades the lower 1/2 of epithelium, while high grade intraepithelial neoplasia/dysplasia invades the upper 1/2 of epithelium. The tumor who invades the whole epithelium is called carcinoma in situ (CIS) [17].

This article reviews the recent advances in the pathogenesis, biomarkers, and diagnostic methods of esophageal precancerous lesions and early EC.

2 Advances in the mechanisms of esophageal precancerous lesions and early EC

The progression of EC from precancerous lesions to invasive carcinoma is very complex. With the development of sequencing techniques, researchers have gradually revealed some key genes and pathways affecting the occurrence and development of EC through sequencing. Academician Qimin Zhan et al. found significant mutations in 8 genes through genomic analysis of 158 cases of ESCC (TP53, RB1, CDKN2A, PIK3CA, NOTCH1, NFE2L2, ADAM29, FAM135B), among which FAM135B, which was never reported to be related to ESCC, was proved to enhance the malignancy of tumor cells. At the same time, mutations in ESCC were mainly involved in Wnt, cell cycle, Notch, RTK-Ras and AKT pathways. PSMD2, RARRES1, SRC, GSK3B and SGK3 could be potential novel therapeutic targets [6]. By sequencing exons of 113 EC patients, Academician Jie He et al. found that genes related to cell cycle and apoptosis pathway were mutated in 99% of the samples, mainly including TP53, CCND1, CDKN2A, NFE2L2 and RB1. The frequency of histone modification related gene mutations was also high, including KMT2D, KMT2C, KDM6A, EP300 and CREBBP, among which EP300 mutation was associated with poor prognosis. In addition, mutations in FAT1, FAT2, FAT3 or FAT4 or AJUBA and NOTCH1, NOTCH2 or NOTCH3 or FBXW7 were found to cause dysregulation of Hippo and Notch pathways, respectively [18]. Although some studies have found some tumor biomarkers for ESCC, new biomarkers for diagnosis are still insufficient [9, 19, 20]. Recent genomic studies, including The Cancer Genome Atlas (TCGA) project, have identified many genomic mutations in ESCC by using whole-exome or whole-genome sequencing of clinical tissue samples [21]. The team of Academician Dongxin Lin found two functional single-nucleotide polymorphisms (SNPs), rs671 in ALDH2 on 4q23 and rs1229984 in ADH1B on 12q24, that were significantly associated with the risk of ESCC in a manner of interactions with alcohol drinking and tobacco smoking status [22, 23]. They further sequenced the whole genome and identified six ESCC mutational signature (E1-E6) among which E4 was significantly associated with smoking and drinking status. In addition, the frequency of Signature E4 in drinkers with ALDH2 risk genotype (rs671-AG/-AA) was significantly higher than that in drinkers without such risk genotype. It was also significantly higher than non-drinkers with this risk genotype or non-drinkers with rs671-GG genotype. The analysis results of AGH1B were similar to those of ALDH2 [21]. Although these studies revealed the pathogenesis of EC from some aspects, they were only limited to the invasive stage, rather than the process from precancerous lesions to invasive cancer, so they could not fully reveal the mechanism of the occurrence and development of EC.

Previous studies have shown that chronic inflammation of the cellular microenvironment is a strong risk factor for digestive system tumors [24, 25]. Reactive oxygen and nitrogen species (RONS) induced by chronic inflammation can damage important cellular components (such as DNA, proteins and lipids), thereby leading to malignant cell transformation [26]. Lin et al. showed that the malignant severity of esophageal mucosa histology increased with the degree of inflammation [27]. They conducted whole exon and whole genome sequencing on 227 tissues including normal epithelial tissues, simple hyperplasia tissues, intraepithelial neoplasia tissues and EC tissues, revealing the change process from normal mucosal tissues to intraepithelial neoplasia and then to invasive carcinoma [27]. Their research showed that the damage status and degree of inflammation of DNA are positively correlated with the degree of atypia of the mucosal tissue. The positive rate of phosphorylated H2AX (γH2AX) in the intraepithelial neoplasia tissue is significantly higher than that in the normal tissue and simple hyperplasia tissue. The positive rate of γH2AX was different in non-inflammatory tissues and inflammatory tissues, and the expression of γH2AX increased with the increase of the degree of inflammation and the degree of cellular atypia. In addition, they found that the increase in genomic ploidy began at the stage of intraepithelial neoplasia, for example, polyploidy was present in 68% of intraepithelial neoplasia and 55.6% of squamous cell carcinoma samples, but not in simple hyperplasia samples. Finally, they also identified several mutated genes that are common in both intraepithelial neoplasia and squamous cell carcinoma, such as CDKN2A, ASCL3, FEV, CCND1, NFE2L2, and SOX2. Although there was no difference in the size of the copy number variation (CNA) between intraepithelial neoplasia and squamous cell carcinoma, squamous cell carcinoma showed a wide increase in the CNA status. Although there are many common genetic changes in squamous cell carcinoma and intraepithelial neoplasia, there are still unique mutations between them, indicating genetic heterogeneity [28].

Only the study of genomics is far from enough, the study of proteomics is a supplement to the study of genomics. Zhang et al. performed proteomic analysis of high-grade intraepithelial neoplasia and normal mucosal tissues and compared the difference in protein expression between the two groups [29]. 4006 proteins were identified, and compared with normal mucosa tissues, 236 proteins were differential expressed, among which 138 were up-regulated and 98 were down-regulated. What’s more, 18 proteins were proved to be differential expressed between ESCC and normal tissues in previous studies. Functional enrichment analysis showed that these differentially expressed proteins were enriched in multiple functional and signaling pathways, such as regulation of actin cytoskeleton, PPAR signaling pathway, ubiquitin-mediated proteolysis, etc.

Although these studies have revealed the important role of genomic alterations in ESCC, it is not clear how normal epithelial cells progress from precancerous lesions to invasive cancers by mutation because most of these studies are based on cross-sectional designs rather than sequential progression or involve fewer stages of disease. In addition, a few somatic mutations may not be sufficient to initiate the occurrence and development of EC, as such mutations also occur in pathologically normal human esophageal tissue [30].

The complex background of tumor microenvironment (TME) also plays an important role in the occurrence and development of tumors. Therefore, it is of great significance to elucidate the dynamic transcriptome changes of cells in TME during tumorigenesis to reveal the development of ESCC. Single cell transcriptome analysis can analyze complex cell composition and decipher cell state transitions in tissue samples [31]. However, it is difficult to analyze continuously progressing tumors because it is almost impossible to obtain these lesions from patients, so the solution has to be found in animal models. The chemical carcinogen 4-nitroquinoline 1-oxide (4NQO) can induce the development of ESCC in mice in a manner similar to the tumorigenic process of human ESCC [32, 33].

Therefore, the study on the development process of ESCC by constructing a mouse model is helpful to understand the development process of human ESCC. The analysis of each stage of the progression of ESCC in model mice by single cell sequencing can show the status of various cells in the esophageal epithelium at different stages of development, thus describing the panoramic picture of the progression of esophageal squamous cell carcinoma [5]. Studies have shown that during the development of ESCC, epithelial cells exhibited six different status (EpiC 1–6), which were respectively related to division and proliferation, detoxification, extracellular carcinogen stimulating response and pro-inflammatory response, keratinization, epithelial-mesenchymal transformation and angiogenesis, and control of tumor invasion and metastasis. EpiC 1–3 existed in different stages, but in different proportions, indicating that these three types of cells were basic components of esophageal epithelium, while EpiC 4 and 5 existed in inflammatory and proliferative stages, and EpiC 6 existed only in invasive tumors. In addition, these cells showed two pathways of change, both starting from EpiC 1, some cells transformed from proliferative EpiC 1 to normal differentiated EpiC 4, and some cells developed from EpiC 2 to EpiC 5, and finally to malignant EpiC 6. These cell subtypes well demonstrate the dynamic transformation of esophageal epithelial cells during tumor development.

Previous research on human esophageal genome mutations showed that although the exons mutation burden in normal human esophageal epithelial increased with age, but no cancer related lesions occurred, suggesting that transcriptome change caused by carcinogens and immunosuppressive factors might play an important role in ESCC [5]. Generally speaking, the occurrence of ESCC is caused by the combined effects of external carcinogenic stimulation, inflammatory response, gene mutation and immunosuppression.

3 Advances in early diagnosis of EC

At present, early detection, diagnosis and treatment are the best strategies to improve the prognosis of EC. If early EC is detected and treated with endoscopy or surgery, the 5-year overall survival rate can be greatly improved to more than 90% [34, 35]. However, for the screening of EC and precancerous lesions, endoscopic screening is not suitable for large-scale screening due to its high price and low efficiency. At present, it is mainly applied to the screening of EC in high-incidence areas in China. Therefore, to find effective non-invasive biomarkers for screening, reduce the cost of screening, improve the efficiency and accuracy of early diagnosis of EC will help to reduce the mortality [36].

4 Advances in biomarkers for screening

In recent years, some biomarkers for the diagnosis of EC have been developed based on esophageal cytology samples, blood samples, respiratory samples, oral microbiome, etc. [36, 37], for example, blood antibodies such as p53 antibodies, circulating tumor cells, circulating miRNAs and lncRNAs [38,39,40,41]. In addition, there are diagnostic methods that combine biopsies with biomarkers, such as DNA methylation markers, mutated genes and SNPs [11, 42].

5 Advances in endoscopic screening

At present, endoscopic screening is still the main clinical screening and early diagnosis method. The main purpose of endoscopic screening is to detect asymptomatic precancerous and early cancerous lesions and remove them, so it is very dependent on the accuracy of diagnosis. In a multicenter cohort study, EC missed in endoscopic diagnosis accounted for 6.4% of all cases [43]. Thus, improving the accuracy is the focus of the diagnosis and treatment of EC. Narrow-band imaging (NBI) can detect more areas of dysplasia while reducing the number of biopsies required [44].

In order to more intuitively and accurately evaluate the staging and histopathological types, and realize targeted biopsy and even pathological visualization, new endoscopic imaging techniques are emerging constantly. Image-enhanced endoscopy (IEE) improves the detection rate of precancerous lesions and early cancer on the basis of white light endoscopy, and avoids unnecessary biopsies [45]. Lugol chromoendoscopy (LC) can show precancerous lesions not seen under WLE and enable more accurate biopsy by showing precise boundaries [46]. I-scan technology is also a digital comparison method that can enhance the tiny mucosal structure and subtle color changes. Compared with random biopsies, i-scan guided biopsies have a higher diagnostic rate (66%) with an accuracy rate of 96% and need fewer biopsies [47, 48].

Artificial intelligence (AI), especially deep learning, is widely used in the medical field. After learning how endoscopy experts diagnose and interpret images, AI can complete the work better and more efficiently. In addition, AI is not significantly different from clinical experts in sensitivity, specificity and accuracy, and even better than some junior or intermediate physicians [49]. In the future, AI will play an increasingly important role in the field of medicine.

6 Advances in endoscopic treatment for early EC

Endoscopic eradication therapy (EET) is currently the basic treatment method for precancerous lesions of the esophagus and early EC restricted to the mucosa and submucosa without regional lymph node metastasis. It mainly includes endoscopic mucosal resection (EMR) and endoscopic submucosal resection (ESD). EET can achieve a 5-year survival rate of 95% to 100% [50]. Radiofrequency ablation (RFA) is usually used for ablation of flat or non-visible lesions, as well as those that remain after endoscopic ablation [51]. Bleeding, perforation and esophageal stenosis are common complications after endoscopic therapy. Careful coagulation of exposed vessels at the resection site can reduce the risk of delayed bleeding; Clip closure, stenting and endoscopic suture are effective methods for the treatment of intraoperative perforation [52]. Common treatments for stricture are self-help inflatable balloon, prophylactic placement of fully covered self-expandable metal stents, biodegradable stents, oral steroid administration, and local steroid injection or topical steroid gel application [49] (Fig. 1).

Fig. 1
figure 1

Treatment strategies for early EC

Full size image

7 Advances in other treatments for early EC

Though endoscopic treatment is the basis for precancerous lesions of the esophagus and early EC, surgery and chemoradiotherapy are also needed. According to Esophageal cancer practice guidelines 2022 edited by the Japan Esophageal Society (JES), when it is diagnosed as having pT1a-muscularis mucosae (MM), patients with non-circumferential lesions that are recommend receiving endoscopic treatment, while patients with whole-circumferential lesions are recommend receiving surgery (if can tolerate) or chemoradiotherapy. As for patients diagnosed with submucosal (SM) after endoscopic treatment, additional treatment such as surgery or chemoradiotherapy should also be considered [53, 54]. What’s more, patients with lesion which length is > 5 cm are suggested to receive surgery. When patients are unable or unwilling to receive surgery, chemotherapy, radiotherapy or chemoradiotherapy is recommended (Fig. 1) [54].

8 Conclusion

In order to improve the therapeutic effect of patients with EC and improve the quality of life of patients, the screening and early diagnosis and treatment of high-risk groups of EC are very important, which should be one of the key points in the prevention and treatment of malignant tumors. Although the research on the pathogenesis of EC is gradually deepening, the evidence found from the genome-transcriptome-proteome is still insufficient, and no effective biological tumor markers have been found that can be used for the primary screening of large-scale population. Therefore, we still need to continue to study and explore the mechanism of occurrence and development of EC, the discovery and utilization of screening biomarkers, and early diagnosis and treatment.