4.1.1 Risk factors

Multiple complex risk factors, including environmental, genetic, and cultural, interact in the multi-stage development of EC, processing from localized damage, inflammation, cell proliferation, mutation, and eventually carcinogenesis. Smoking is a common risk factor for both ESCC and EAC, particularly for ESCC [56]. Other factors, however, pose risks specific to different pathological subtypes. While the carcinogenic risks of some factors, such as tobacco smoking, alcohol consumption, areca nut, achalasia, gastroesophageal reflux disease, BE, and helicobacter pylori infection, are relatively well-established and offer potential for prevention, others remain inconclusive. Table 1 outlines the major reported risk factors for ESCC and EAC.

The etiology of ESCC, while not fully understood, is associated with several lifestyle and environmental factors, such as tobacco smoking, alcohol consumption, and the chewing of betel nuts [50, 57, 61]. Combined risk factors like smoking and alcohol significantly elevate the risk [50]. These risk factors act directly on the esophagus and cause chronic irritation and inflammation of the esophageal mucosa, leading to DNA damage through oncogene activation and tumor suppressor genes inhibition, ultimately resulting in mutagenesis [81]. Diet-related factors constitute the majority of ESCC risk factors, such as fruit, citrus fruit, vegetables, pickled vegetables, maté tea, hot beverages and foods, hot tea, salt, folate, and vitamin B6 [57, 71, 77, 70, 82]. However, not all cases can be explained by diet alone. Chronic esophageal irritation can also occur when food is retained and broken down by bacteria, releasing various chemical irritants. Physiological conditions like achalasia significantly increase ESCC risk through food retention, chronic inflammation, and bacterial overgrowth [62, 82]. Biological infections like Human papillomavirus (HPV) promote malignant transformation of esophageal epithelium, but epidemiological and mechanistic findings remain inconsistent, even in high-incidence regions [72, 73]. Geographic variations in ESCC incidence, prevalent in China, Africa, and the Middle East, suggest a link to nutritional deficiencies, such as selenium [83].

EAC is influenced by a different combination of genetic, environmental, and lifestyle factors, characterized by prolonged esophageal exposure to acid reflux, leading to inflammation-induced hyperplasia and epithelial metaplasia [84]. Symptomatic GERD, achalasia, and the resulting BE are well-established high-risk factors and/or precancerous conditions for EAC [62-64]. Obesity links not only to GERD and its complications but also to distinct metabolic and immune features in the tumor and systemic microenvironment, promoting the development of EAC through pro-inflammatory adipokine release and insulin resistance [85]. The inverse relationship between Helicobacter pylori infection and EAC may result from reduced stomach acidity caused by chronic inflammation-induced glandular atrophy and chief cell loss, which decreases the likelihood of reflux esophagitis, BE, and EAC [66, 86]. However, epidemiological and mechanistic evidence of HPV infection for EAC risk require further clarification [74-76]. Neither alcohol consumption nor nitrosamine exposure has shown a significant risk for EAC [60, 69]. The association between the intake of vegetables, fruits and processed meats with EAC risk remains inconclusive [67, 68, 78, 79].

4.1.2 Preventive measures

Consequently, the primary prevention recommendations in ESCC are drawn from available public etiological data, emphasizing risk reduction through lifestyle modifications, such as alcohol and betel nut cessation, and increased fruit and vegetable intake [87]. Reducing or eliminating tobacco and alcohol use are the most effective strategies, with benefits increasing over time [51, 58, 59], although betel quid without tobacco (BQ-T) cessation did not show risk reversal [88]. However, in patients with achalasia, risk reduction with surgical intervention, such as surgical myotomy with or without a fundoplication, does not completely prevent tumor development [82]. Chemoprevention strategies, such as cyclooxygenase 2 (COX2) inhibitors, statins, metformin, and proton pump inhibitors (PPI), have demonstrated potential for preventing ESCC [89]. However, more robust and conclusive research is required to support formal guideline recommendations.

Risk reduction, such as smoking cessation, weight management, and reducing red meat intake, are also the basic prevention strategies for EAC. It is worth noting that bariatric surgery lowers EAC risk, but it may exacerbate GERD or cause new onset reflux, esophagitis, and BE [65, 80]. Anti-reflux therapies have the potential to curtail dysplasia in BE, subsequently reducing the EAC risk, but surgery with fundoplication did not show superiority over medication [90]. American College of Gastroenterology (ACG) current practice guidelines advise against using anti-reflux surgery for prevention in BE patients [91]. Anti-reflux medications are controversial because most EAC cases do not originate from either GERD or BE, indicating only a small proportion of the population benefits from anti-reflux therapies [92-94]. And genome-wide association studies (GWAS) data from the IEU Open GWAS project (mrcieu.ac.uk) suggested anti-reflux therapies primarily lower the risk of EAC by preventing BE formation rather than acting on existing BE [95]. While preclinical and case-control studies indicated aspirin/nonsteroidal anti-inflammatory drugs (NSAIDs) may reduce tumor progression (from metaplasia to dysplasia and cancer) more effectively than BE initiation [96-99]. Hence, the use of anti-reflux therapy might go beyond BE patients. The AspECT trial (EudraCT 2004-003836-77) provided strong evidence for EAC chemoprevention, showing that high-dose PPI (40 mg twice daily) was superior to low-dose PPI (20 mg once daily) in reducing all-cause mortality, EAC development, or high-grade dysplasia [time ratio (TR) = 1.27; 95% confidence interval (CI) = 1.01-1.58; p = 0.038]. High-dose PPIs combined with aspirin performed best (TR = 1.59; 95% CI = 1.14-2.23), though primarily for all-cause mortality reduction rather than cancer-related outcomes [100]. Thus, the ACG guidelines only endorse at least once-a-day PPI therapy in BE patients because of the result of an undouble-blind trial [91].

The most direct prevention is endoscopic eradication for the esophageal neoplasia. The SURF (Surveillance vs. Radiofrequency Ablation) trial (NTR1198) revealed radiofrequency ablation (RFA) in low-grade dysplasia (LGD) BE patients reduced the risks of progression to high-grade dysplasia (HGD) and to EAC by 25% and 7.4%, respectively [101]. Similarly, endoscopic mucosal resection (EMR) was proven to prevent or delay the progression of esophageal squamous from LGD to HGD (dysplasia downgrading of 82.5% in EMR group vs. 49.2% in the group without EMR, p < 0.001) by a randomized study (Chi-CTR-TRC-1200201) [102].

4.2.1 Endoscopic screening

The current gold standard for EC screening is endoscopy with biopsies, allowing for the detection of precursor lesions.

Data from the Dutch National Histopathological Diagnostic Registry (PALGA) indicated that patients with LGD had a higher risk of progressing to EAC (2.51/100 person-years; 95% CI = 1.46-3.99/100) compared to those without LGD (1.01/100 person-years; 95% CI = 0.41-2.10/100), with a median progression interval of 2.93 years, emphasizing the importance of endoscopic surveillance within the first 2 to 3 years [105]. In ESCC, the only randomized controlled study (NCC1788) on endoscopic screening found that patients with mild-to-moderate dysplasia had a 1.04% overall annual risk of progression to advanced neoplasia, with critical surveillance needed within the first 2 to 3 years [106].

Endoscopic screening is recommended for high-risk EC populations. The definitions of people at risk for BE and EAC vary slightly in different guidelines [91, 107–109]. The ACG recommends screening should be performed for patients with chronic GERD symptoms and with at least three risk factors for BE (e.g., male, Caucasian, age over 50, obesity, smoking, and first-degree relatives with a family history of BE or EAC) [91]. While the latest American Gastroenterological Association (AGA) Clinical Practice Update considers GERD not an essential risk factor [108]. For ESCC, there is no uniform definition of high-risk. To maximize cost-effectiveness, the Chinese Expert Consensus on Early Diagnosis and Treatment of Esophageal Cancer recommends opportunistic screening for EC in high-incidence areas, targeting individuals ≥40 years of age with any high-risk conditions (e.g., a pre-cancerous disease in the upper gastrointestinal tract; first-degree relatives with a history of EC; squamous cell carcinoma of the head, neck, and/or respiratory tract; and other high-risk factors) [110]. In comparison, the European Society of Gastrointestinal Endoscopy (ESGE) defined high-risk individuals as those with a history of head/neck cancer, achalasia, or previous caustic injury [107].

White light endoscopy (WLE) with targeted biopsies is the primary technique for EC screening but has limitations in identifying early cancers and in adhering to the BE biopsy protocol, which requires four-quadrant biopsy at 1 cm to 2 cm intervals plus biopsy from any suspicious visual lesions [111-113]. High-resolution endoscopes (HRE) with advanced charge-coupled devices (CCD) and adjustable focus improve mucosal visualization and dysplasia detection [114]. Chromoendoscopy enhances lesion visualization using stains. Lugol chromoendoscopy (LCE) plus indicative biopsy has become the most practical and effective method to identify squamous dysplasia or cancer, with up to 98% sensitivity. The sensitivity for the detection of severe dysplasia could be 87.0% (NCT01688908) [115]. Methylene blue or acetic acid chromoendoscopy improves the sensitivity of BE and EAC diagnosis [116, 117], though the accuracy reported varies [118]. Digital chromoendoscopy, or virtual endoscopy, uses varied light wavelengths to highlight mucosal structures without dye-related risks. Techniques include narrow-band imaging (NBI), blue light imaging, autofluorescence imaging, optical enhancement, spectral fusion imaging, and photoelectric composite staining imaging. NBI is the most widely used technique, offering comparable accuracy to LCE for high-grade intraepithelial neoplasia (HGIN) detection but with reduced sensitivity for low-grade lesions (LGIN) (NCT04224896) [119]. The main advantage of NBI is its ability to enhance the visualization of intraepithelial papillary endocapillary loops (IPCL), facilitating lesion assessment and cancer infiltration depth evaluation using Japanese Esophageal Society (JES) or Inoue staging [120], and providing biopsy guidance for EAC screening. A randomized trial (NCT00576498) showed NBI-targeted biopsies matched the BE detection rate of Seattle biopsies with fewer biopsies per patient needed (3.6 in the NBI group vs. 7.6 in the high-definition-WLE group, p < 0.0001) and higher dysplasia detection (30% in the NBI group vs. 21% in the high-definition-WLE group, p = 0.01) [121]. Confocal laser microendoscope (CLE) provides a real-time, highly magnified cellular level view for histological differentiation without the need for biopsy (virtual histology). It demonstrates high sensitivity, specificity, and accuracy in detecting early ESCC (NCT01909518) [122]. A meta-analysis found that CLE-based targeted biopsies' sensitivity and specificity in detecting HGD/EAC are 68% and 88%, respectively [123]. However, CLE has not fully replaced traditional histopathology due to limited evidence of direct comparison with traditional histopathology. Optical coherence tomography (OCT)-based volumetric laser endomicroscopy (VLE) allows scanning up to 3 mm deep into mucosa and submucosa, enhancing early diagnosis of BE and EAC. A multicenter registry (NCT02215291) has shown that VLE-guided biopsy increased tumor detection rates seven-fold over random biopsy [124]. Furthermore, VLE showed higher tumor detection rates and diagnostic accuracy for superficial ESCC confined to the epithelium (EP) or lamina propria mucosa (LPM) compared to endoscopic ultrasound (EUS) [125]. However, more precise computer-aided assessments and additional prospective studies are necessary to fully validate VLE's value.

4.2.2 Non-endoscopic screening

Non-endoscopic methods are less invasive, potentially more cost-effective, and easier to administer compared to endoscopic screening in non-specialist settings such as primary care.

Cytology is an effective and more acceptable non-endoscopic screening method compared to endoscopic screening. Various tethered cell sampling devices have been developed, such as the Cytosponge. It consists of a capsule that dissolves in the stomach, releasing a spherical sponge that collects cells from the esophageal wall as it is pulled back through the mouth [126, 127]. The collected samples can be analyzed for specific immunomarkers to enhance diagnostic efficiency. The most well-established biomarker used by Cytosponge is trilobal factor family 3 (TFF3) performed with immunohistochemistry, which is a specific marker for BE [128]. The sensitivity and specificity of Cytosponge-TFF3 for BE diagnosis can reach 80% and >90%, respectively [129].

Current screening and surveillance for ESCC or EAC are recommended only for individuals with squamous epithelial hyperplasia or BE. However, 40%-50% of cases without hyperplasia at baseline eventually progress to ESCC [130]. Genomic studies show that Lugol unstained lesions already harbor genetic variants (including mutations, copy number variants, and clonal expansions) and are associated with an increased risk of developing ESCC [14]. Similarly, over 80% of EAC cases have no prior diagnosis of BE [131, 132]. This raises the question of how to screen for oncogenic initiation without dysplasia signs.

Whole-genome sequencing of BE biopsies demonstrated that genomic signals can predict the risk of progression up to 10 years before histopathologic changes [133]. Biomarkers, including multi-gene next-generation sequencing sets, differentially methylated genes, and microRNAs, were explored to identify Cytosponge cytological dysplasia [134-136]. TP53 mutations, key events in EC progression [17, 18], as well as other biomarkers, could help identify BE patients at high risk of malignant progression requiring endoscopic surveillance [137]. Preliminary results also suggested that combining Cytosponge with biomarkers, such as p53 protein staining in this study, could improve the accuracy of ESCC screening, as evidenced by a pilot study in Iran showing 100% sensitivity and 97% specificity for detecting HGIN [138].

With the advent of liquid biopsy technology and precision medicine, novel biomarkers show potential for EC screening. Genetic alterations in cell-free DNA (cfDNA) can detect the risk of BE progression [139]. Circulating microRNA (miRNA)-based signatures have been validated for early ESCC detection [140]. A Chinese study identified miRNA panels that distinguish ESCC from squamous dysplasia [141]. Circulating extracellular vehicles (EVs)-14 also displayed a high accuracy in ESCC screening [142]. Beyond genetic markers, circulating systemic inflammation biomarkers and serum metabolomics provide insights into the risk and early diagnosis of EC [143-146]. However, it remains challenging to identify malignant transformation risk early through one signature.

Integrating endoscopic and non-endoscopic modalities with a deep learning-driven artificial intelligence (AI) system can significantly enhance the identification of high-risk esophageal lesions, offering a promising adjunctive tool for EC and precancerous lesion screening (ChiCTR2100044126) [147].

5.1.1 EAC

In mucosal EAC (T1a), lymph node metastasis (LNM) risk is accorded with clinicopathologic features such as lesion size ≥2 cm, poorly differentiated, and lymphovascular invasion [148, 149]. Low-risk T1a patients rarely develop LNMs, with rates ranging from 0% to 3% [150]. The National Comprehensive Cancer Network (NCCN) guideline recommends endoscopic resection (ER) for these patients, using techniques like EMR and endoscopic submucosal dissection (ESD), without additional interventions [151]. ESD is favored for larger lesions due to higher complete resection rates and lower recurrence rates [152]. A randomized controlled study showed that ESD had a higher R0 resection rate for BE with HGIN or for early EAC with ≤3 cm focal lesions compared to EMR. Both ESD and EMR were effective in terms of surgery necessity, tumor remission, and recurrence, while ESD was more time-consuming and prone to severe stenosis (NCT01871636) [153]. There is little literature on optimal tumor size cutoffs for the selection of endoscopic resection techniques. ASGE recommends ESD for patients with early-stage, highly differentiated, non-ulcerative lesions larger than 20 mm, while either ESD or EMR can be used for similar lesions ≤20 mm [154]. In addition, eradication of residual Barrett's mucosa post-curative resection to prevent metachronous neoplasia is standard practice [154-156], but for patients unable to undergo ablation, endoscopic surveillance is an alternative [157]. High-risk T1a patients with ≥10% risk of delayed extraesophageal and LNMs require complete staging and additional treatments such as esophagectomy [158].

Submucosal EAC (T1b) carries up to a 30% risk of LNMs [159]. Esophagectomy is its standard care, showing improved overall survival (OS) compared to ER [160], with five-year survival rates of 89% versus 59% in a multicenter study [161]. However, T1b patients with low-risk features (e.g., pT1b sm1, good to moderately differentiated, and no lymphovascular invasion [LVI]) have a 2%-4% risk of LNMs [149, 162]. For these low-risk T1b EAC, endoscopic eradication and recurrence rates are comparable to T1a patients [163]. Therefore, current AGA guidelines recommend endoscopic treatment as a reasonable alternative to esophagectomy despite limited direct comparative evidence [164]. Chemoradiotherapy (CRT) is another organ-preserving option, though it is less studied in T1b EAC. A small two-center retrospective study demonstrated CRT is a safe and effective curative treatment for selected T1 EAC, with a 3-year OS of 78% [165]. However, a Japanese retrospective study compared ESD followed by CRT with ESD followed by esophagectomy for clinical T1bN0M0 EAC patients, and found no significant difference in 2-year disease-free survival (DFS) rates between two groups (88% in the CRT group vs. 100% in the esophagectomy group, p  =  0.43) [166].

The treatment of T1 EAC is evolving towards minimally invasive procedures. Therefore, histologic evaluation is essential to identify high-risk individuals and select appropriate treatment. However, significant inconsistencies exist in the histologic evaluation of ER specimens. The Amsterdam University Medical Center has proposed improvements in histopathological assessment, though further clinical validation and updates are imperative [167].

5.1.2 ESCC

ER is the primary treatment for superficial ESCC in various guidelines. cT1a- epithelium (EP, M1)/ lamina propria mucosae (LPM, M2) ESCC, evaluated according to the JES classification [168], is an absolute indication for ER [158], featuring a LNM rate of 0%-5% [169, 170], and a 5-year DFS of nearly 100% [171]. ESD is preferred for its superior curative resection and reduced local recurrence rate [172], especially for lesions >15 mm [154].

The LNM rate increases to 18% when the tumor invades the mucosal muscle (MM, M3) and exceeds 50% when it extends into the submucosa [173]. Preoperatively differentiating between cT1a-MM and cT1b-submucosa ≤ 200 µm (SM1) using endoscopic techniques is challenging. Consequently, most studies and guidelines categorize these as cMM/SM1 cancers. ESD without additional treatment is an appropriate option for T1a-MM/T1b-SM1 ESCCs [174, 175].

The current debate centers on the optimal treatment for high-risk T1b patients. It is widely accepted that ER alone is inadequate for patients with deep submucosal invasion (T1bSM2-3), lymphovascular invasion, or poor differentiation [176]. Esophagectomy remains the standard treatment of T1b ESCC [151]. However, assessing tumor depth before ER is challenging; thus, ESD serves as an initial diagnostic tool and potential cure [177]. Noncurative ER followed by esophagectomy significantly improves OS and recurrence-free survival (RFS) of cT1N0M0 ESCC patients compared to esophagectomy alone [178]. Definitive CRT (dCRT) offers OS comparable to radical surgery while preserving the organ [179, 180], making it a guideline-recommended option [181]. However, it carries a higher risk of recurrence. Additional CRT post-ER improved local control and reduced radiation dose to minimize toxicity compared to dCRT [182, 183]. Compared to ESD alone, a multicenter, cross-sectional study demonstrated adjuvant radiotherapy after ESD could improve survival in pT1b ESCC with LVI+, with a 5-year OS of 91.7% versus 59.5% [184]. Therefore, the first question is what the best adjuvant therapy after ER is. Reported 5-year OS rates ranged from 90% to 100% for esophagectomy and from 75% to 85% for CRT after ER, though the CRT dose was low in most studies [176]. Esophagectomy is preferred for extensive submucosal tumor invasion due to the lower risk of recurrence [185]. The ongoing phase III Ad-ESD trial (NCT04135664), which compared esophagectomy and dCRT for patients with cN0-pT1b ESCC after ESD will answer the question [186]. Secondly, JCOG 0508 (UMIN000000553) showed that the 5-year OS of pT1b or of high-risk pT1a (LVI) ESCC patients receiving additional CRT post-ER are comparable to those receiving surgery alone [187, 188]. Further exploration is needed to answer the question of the value and optimal candidates for the organ-preserving ER-CRT approach as an alternative to surgery.

5.2.1 Surgical approaches

Esophagectomy remains the cornerstone of curative treatment for localized EC. Common surgical procedures encompass left thoracic (Sweet), transhiatal esophagectomy (THE), and right thoracic esophagectomy [three-field (McKeown or modified McKeown) and two-field (Ivor Lewis)]. The right thoracic approach is preferred for better lymphadenectomy as supported by studies such as NST1501 (NCT02448979) [189-191]. Despite a higher complication rate than the left thoracic approach, it offers better safety than THE [189-191], regardless of neoadjuvant therapy, as evidenced by the SAKK trial 75/08 (NCT01107639) [192]. Procedure choice depends on the individual circumstances, such as tumor location, the patient's overall health, surgical risks, and the expected quality of life post-surgery. McKeown or Ivor Lewis esophagectomies are suitable for middle and lower esophageal tumors, while tumors at or above the carina benefit from three-field lymphadenectomy for sufficient proximal resection margins [193, 194]. Compared to Ivor-Lewis esophagectomy, McKeown has a comparable 5-year OS and perioperative mortality [195], but with a higher complication rate [196]. This also pertains to another debate, namely the question of two-field or three-field lymph node (LN) dissection. A national ESCC study suggested approximately one-third of patients had undetected laryngeal and/or cervical lymph node metastases, underscoring the benefit of three-field LN dissection in enhancing 5-year OS [197]. High cervical and recurrent laryngeal cervicothoracic LNM rates were also discovered in EAC [198]. However, a randomized clinical trial (NCT01807936) demonstrated no improvement in OS or DFS with three-field versus two-field lymphadenectomy [199]. It is important to note that both groups underwent dissection of the laryngeal recurrent nerve lymph nodes, introducing a confounding factor.

The trend toward improved overall mortality from EC is attributed to early detection and the development of minimally invasive esophagectomy (MIE) [200]. MIE has a low postoperative mortality rate (1.7% at 30 days) and fewer pulmonary complications compared to open surgery [201]. Randomized controlled trials (such as the TIME Trial (NTR TC 2452) and NCT00937456) suggested comparable or superior 3-year OS and progression-free survival (PFS) for MIE compared to open esophagectomy (OE) [202-204]. However, MIE had a longer operative time, a higher 30-day reoperation rate, and an increased risk of esophageal hiatal hernia [205, 206].

Despite challenges like limited visualization, restricted instrument movement, and a steep learning curve, MIE is recommended in institutions with the necessary expertise. Robotic-assisted minimally invasive esophagectomy (RAMIE) addresses these challenges with better three-dimensional visualization and flexible instruments. A randomized controlled trial (NCT01544790) has shown that RAMIE offers outcomes comparable to OE but with fewer surgery-related and cardiopulmonary complications, reduced postoperative pain, improved short-term quality of life, and faster short-term functional recovery [207]. Though no breakthroughs in perioperative outcomes and OS than conventional MIE [208], RAMIE improves LN clearance, especially along the left recurrent laryngeal nerve, enhancing PFS [208]. The RAMIE trial (NCT03094351) found shorter operative times and better lymph node clearance with similar complication rates compared to MIE, even after neoadjuvant therapy [209].

5.2.2 Neoadjuvant treatment

Neoadjuvant ICRT vs. CRT

The PALACE-1 (NCT04435197) and the SCALE-1 (ChiCTR2100045104) studies incorporated immunotherapy into nCRT (nICRT) for ESCC, achieving the highest pCR rates in history [238, 239]. A network meta-analysis confirmed that the highest pCR and MPR rates remain in radiotherapy groups [neoadjuvant immunotherapy combined with CRT (ICRT) and nCRT] [240]. However, the long-term survival impact of pCR achieved through immunotherapy requires further follow-up. Several phase III trials in ESCC, including iCROSS (NCT04973306) and NEOCRTEC2101 (NCT05357846) based on the standard nCRT regimen, are currently underway and expected to yield practice-changing data [241].

However, a pooled analysis of 17 non-randomized prospective trials revealed that adding immunotherapy to nCRT does not improve pCR in resectable esophageal or gastroesophageal junction (GEJ) cancers but significantly increases severe adverse events (AEs) [242]. The recent ECOG-ACRIN EA 2174 study (NCT03604991) presented at the 2024 ASCO meeting is poised to provide critical insights into neoadjuvant therapy for EAC [243]. This phase II/III randomized trial evaluated immunotherapy in a neoadjuvant setting in combination with nCRT, and in an adjuvant setting with either nivolumab or ipilimumab/nivolumab. This report spotlights the outcomes of the neoadjuvant treatment phase. Unfortunately, the addition of nivolumab to nCRT has not improved the pCR rate in E/GEJ adenocarcinoma (24.8% in nICRT vs. 21.0% in nCRT). Though the adjuvant element is still pending, the neoadjuvant approach for EAC does not endorse immuno-enhancement with nCRT.

To date, neoadjuvant or perioperative immunotherapy is only recommended by the NCCN guidelines for resectable, locally advanced dMMR/MSI-H G/GEJ patients, based on the GERCOR NEONIPIGA (NCT04006262) and INFINITY (NCT04817826) [151, 244, 245]. In these ICIs-sensitive tumors, the pCR reached 60%.

The quest for the most effective neoadjuvant treatment strategies in the era of immunotherapy continues, with ongoing phase III clinical trials depicted in Figure 5. Limited data from retrospective comparisons indicated that nICT may be more beneficial for metastatic lymph nodes [246], while nCRT is more valuable for the primary tumor and metastatic lymph nodes at levels 1 and 2R [247]. Thus, personalizing treatment strategies for different patient groups is likely the future direction in oncology care.

5.2.3 Adjuvant treatment

Another potential integrative treatment strategy is postoperative adjuvant therapy. Previous studies on perioperative chemotherapy for G/GECJ adenocarcinoma above advocate continuing the same regimens for adjuvant chemotherapy who responded to nCT [216, 224]. However, postoperative adjuvant chemotherapy for ESCC failed to provide additional survival benefits in JCOG studies [255, 256]. Despite the absence of prospective studies, adjuvant therapy after nCRT and surgery is not generally recommended in ESCC [257, 258]. Retrospective studies indicate that high-risk ESCC patients, such as those with poor responses to nCRT or those with advanced pathological stages, may benefit from adjuvant chemotherapy [259, 260].

Nivolumab represents a most significant advancement in adjuvant therapy for EC. In the Checkmate 577 trial (NCT02743494), it prolonged DFS regardless of histology (22.4 months vs. 11.0 months with placebo) and is approved for EC patients with residual pathological disease following CRT [227]. Figure 5 presents ongoing phase III studies investigating the integration of immunotherapies and novel targeted agents in the perioperative treatment of EC.

5.2.4 Organ preservation

Given the approximately half of pCR after nCRT, organ preservation based on active surveillance is an important direction for resectable EC, especially ESCC. The FFCD 9102 study (NCT00416858) indicated no survival benefit for surgery after response to CRT compared to continued dCRT for ESCC patients [261]. A phase II randomized study (NCT02959385) in ESCC patients achieving clinical complete response (cCR) reached the same conclusion [262]. A multicenter study showed that active surveillance for EC patients achieving cCR from nCRT yielded equivalent PFS and OS outcomes compared to immediate esophagectomy [263]. However, the ESOPRESSO study (NCT01740375) on ESCC patients suggested a trend toward better DFS, PFS, and time to progress in the surgery group and a higher local recurrence in the observation group, despite no survival differences in only 37 cCR patients due to slow enrollment [264]. The main challenge of ESCC active surveillance post-CRT is the subtlety of residual tumors, characterized by sparse viable cancer cells, low lymph node pCR rates, and varied regression patterns [265]. The PreSANO study (NTR6803), utilizing a multimodal approach including repeated endoscopic ultrasound examinations, bite-on-bite biopsies, and fine needle aspiration of suspicious lymph nodes, reported a 10% miss rate for tumor regression grades 3 and 4 tumors [266]. Innovative strategies like liquid biopsy for circulating tumor DNA and the Cytosponge™ device are emerging to enhance the detection of residual tumors [267, 268]. However, it is noteworthy that the negative predictive value is low [266]. Ongoing trials, such as SANO (NCT05953181), SANO-2 (NCT04886635), Esostrate-Prodige 32 (NCT02551458), preSINO (NCT03937362), and NEEDS (NCT04460352), are further evaluating surveillance strategies, as shown in Figure 6 [269-272]. Additionally, the PALACE3 study (NCT06339060) aims to explore the value of immunotherapy combined with chemoradiotherapy as an alternative to surgery.

To clearly outline the key clinical research landscape for locally advanced resectable EC, Table 2 provides a systematic summary.

5.3.1 cCRT: standard regimen

Concurrent chemoradiotherapy (cCRT) remains the standard treatment for locally advanced inoperative EC based on the result of RTOG 8501 [273]. Even in elderly ESCC patients, phase III studies have confirmed that cCRT with S-1 offers superior response rates and survival compared to radiotherapy alone [274, 275]. From RTOG 9405 to recent studies, including CONCORDE (PRODIGE-26, NCT01348217), ARTDECO (NCT01457677), and two Chinese randomized studies (NCT01937208, NCT02850991), have established the recommended dose regimen for dCRT in EAC and ESCC [276-280]. No significant local control benefit with radiation doses escalated to 60 Gy over 50 Gy [278-280]. However, local recurrence remains the predominant failure following cCRT, particularly in ESCC [281]. The latest phase III randomized study (ChiCTR-IPR-15007172), which compared both dose and irradiation field, suggested that a high dose of 60 Gy to the involved field was most effective in improving local control and survival [282]. The ESO-Shanghai 12 study (NCT03790553) is ongoing to explore positron emission computed tomography-guided dose escalation in ESCC [283].

About the optimal concurrent chemotherapy regimen, no differences in PFS and in 2-year OS were shown in both ESCC and EAC patients among various fluorouracil-based regimens, including capecitabine in the CRTCOESC study (NCT02025036) and PRODIGE5 study (NCT00861094) [284, 285]. Similarly, an ESCC phase III trial (NCT02459457) found no survival advantage for paclitaxel-based regimens [286]. Direct comparisons between these regimens are lacking.

5.3.2 Immunotherapy combination

Another strategy involves intensifying drug combinations. The benefits of induction or consolidation chemotherapy with dCRT are disputed, and meta-analysis showed no significant long-term survival benefits [287]. Inspired by the PACIFIC study (NCT02125461) in locally advanced non-small cell lung cancer (NSCLC) [288], the incorporation of immunotherapy offers new therapeutic insights for locally advanced EC.

TENERGY (EPOC1802) study (UMIN000034373) is the first reported single-arm phase II trial on atezolizumab post-dCRT for ESCC, reported a 40% cCR rate but disappointing median PFS (3.2 months) and 12-month PFS rate (29.6%), which fell short of expectations [289]. Conversely, a Korean phase II study (NCT03377400) of consolidation dual immunotherapy with durvalumab and tremelimumab in ESCC showed 24-month PFS and OS rates of 57.5% and 75%, markedly better than historical controls [290].

Induction immuno-chemotherapy followed by cCRT was also evaluated in a phase II single-arm ESCC trial (NCT03985046), and it showed improved radiotherapy sensitivity and local control rates via vascular normalization and hypoxia alleviation [291]. These findings were validated by the ImpactCRT trial (ChiCTR2000034304) and a real-world study [292, 293].

The EC-CRT-001 study (NCT04005170), employing concurrent CRT and immunotherapy in ESCC patients, achieved a 95% completion for planned chemoradiotherapy, with 62% attaining cCR and a 1-year PFS of 54.5%, demonstrating encouraging activity [294]. These studies highlight the potential of combining immunotherapy with CRT in locally advanced ESCC, though few studies report on combining dCRT with immunotherapy in EAC. Ongoing phase III trials (Figure 7) like KEYNOTE-975 (NCT04210115) and KUNLUN (NCT04550260) are expected to provide further insights. Concurrently, the immunotherapy combination has introduced a cascade of additional queries regarding the optimal combination regimens, such as optimal timing, radiation dose, radiation volume, and immunotherapy course, requiring further investigation.

5.3.3 Targeted therapy combination

Despite limited success in neoadjuvant therapy, targeted therapy holds promise for locally advanced unresectable ESCC. A phase II study (NCT02375581) in elderly ESCC patients showed that combining erlotinib with radiotherapy significantly improved OS (mOS, 24.0 months vs. 16.3 months) compared to radiotherapy alone, particularly in those with EGFR-overexpression, offering options for patients intolerant to cCRT [295]. However, another phase II study (NCT02577341) found that cCRT combined with nimotuzumab failed to prolong PFS or OS, although it reduced distant metastasis, possibly due to unselected EGFR expression status of the included subjects [296]. Further phase III NXCEL1311 study (NCT02409186) demonstrated that nimotuzumab combined with CRT increased response rate, though long-term follow-up is warranted for survival outcomes [297]. Another phase III study has revealed that cCRT combined with erlotinib improved long-term survival in locally advanced ESCC, particularly the patients with EGFR-overexpression [27]. However, RTOG 0436 (NCT00655876) and SCOPE1 (ISRCTN47718479) studies demonstrated no benefit of cetuximab to cCRT for EC [298, 299]. Therefore, the role of EGFR-targeted agents requires further validation.

5.3.4 Conversion surgery

Contrary to the debate on organ preservation in resectable EC, recent advances in improved regression by adding immunotherapy to CRT and the adoption of a lower radical dose of 50 Gy have reinvigorated the discussion on conversion surgery for initially unresectable diseases, particularly in ESCC. A phase II Japanese study (UMIN000011089) on unresectable T4b and supraclavicular lymph node metastatic ESCC patients using DCF chemotherapy ± radiotherapy for conversion surgery showed a 41.7% conversion rate, 95% R0 resection rate, and 20% pCR rate. Postoperative complications occurred in 25% of patients, with a 3-year OS of 46.6% and significantly better OS in R0 resection patients (71.4% for R0 patients vs. 30.1% for patients who did not) [300, 301]. A multicenter real-world study from China employing an immunotherapy combination for induction treatment of ESCC achieved a 74.8% conversion rate, 94% R0 resection rate, and 22.3% pCR rate, surpassing chemotherapy alone with fewer postoperative complications [302]. Conversion CRT in a phase II study (NCT04278287) in China reported a pCR rate of 47.6% in 46.7% of ESCC patients undergoing surgery and a significantly improved OS compared with dCRT (2y-OS: 84.2% vs. 54.4%; HR = 3.41; 95% CI = 1.10-10.61) [303]. Furthermore, they conducted another phase II NEXUS-1 study (ChiCTR2100054327) incorporating immunotherapy into the conversion CRT and achieved an R0 resection rate of 100% and pCR rate of 61.5% in ESCC patients [304, 305]. Despite the promising prospects of conversion surgery, it is crucial to note that selecting cT4b candidates for surgery after induction therapy is a considerable challenge. Surgery isn't advisable for patients unable to achieve R0 resection, as palliative procedures have higher complication rates and offer no long-term survival benefits [306].

Similarly, Table 3 summarizes the key clinical research landscape for locally advanced unresectable EC.

5.4.1 First-line treatment for advanced ESCC and HER2 negative EAC

5.4.2 Second- or later-line treatment

Advanced EAC

HER-2 negative EAC

Chemotherapy remains the cornerstone of second-line treatment of advanced EAC. Docetaxel, irinotecan, and paclitaxel have been established as standard second-line options through phase III clinical trials [COUGAR-02 (ISRCTN13366390), WJOG 4007 (NCT01224652) and AIO (NCT00144378)], with ORRs of 15%-25%, a mPFS of 2-3 months, and a mOS of 8-9 months [361-363].

Anti-angiogenic targeted drugs have a place in the second-line and subsequent treatment for EAC. Ramucirumab, a VEGFR antibody, combined with paclitaxel, improved OS compared to paclitaxel alone in the second-line treatment of advanced GEJ/G adenocarcinoma after first-line chemotherapy in the RAINBOW study (NCT01170663), and has been recommended for patients regardless of HER2 status by the NCCN guideline [151, 327]. The FRUTIGA study (NCT03223376) showed that the combination of fruquintinib and paclitaxel prolonged mPFS compared to paclitaxel monotherapy (5.6 months vs. 2.7 months; HR = 0.57; 95% CI = 0.48-0.68; p < 0.0001) [328]. Phase III trial (NCT01512745) of apatinib for subsequent antitumor therapies in patients with chemotherapy-refractory GEJ adenocarcinoma indicated a co-endpoint OS benefit [364]. The combination of camrelizumab with apatinib and chemotherapy showed promising anti-tumor activity regardless of PD-L1 expression, with a mPFS of 6.5 months in a phase II study (NCT04345783) [365]. Phase III trials combining anti-angiogenic targeted drugs with immunotherapy are ongoing (NCT04385550, NCT06341335). However, the combination of ICIs, such as cabozantinib combined with atezolizumab (COSMIC-021 study, NCT03170960), and the addition of tremelimumab to durvalumab plus chemotherapy (PRODIGE 59-FFCD 1707-DURIGAST trial, NCT03959293), has shown minimal activity preliminarily [366, 367].

These findings above highlight a significant unmet clinical need in later-line treatment of advanced EAC refractory to chemotherapy. Most importantly, the optimal approach for EAC that progresses after ICI treatment remains unknown, representing a critical gap in scientific and clinical knowledge. A retrospective analysis reveals promising response and survival with ramucirumab plus chemotherapy in patients with post-first-line checkpoint inhibitors and chemotherapy [368]. Further research is needed. Figure 8 presents the ongoing phase III studies of various lines of novel targeted agents in advanced EC.

HER2 positive EAC

ADC drugs have also achieved breakthroughs. The DESTINY-Gastric01 study (NCT03329690) led to the approval of trastuzumab deruxtecan for second-line or later treatment of HER-2-positive GEJ/G adenocarcinoma, with an ORR of 51% and a mOS of 12.5 months [329]. Disitamab vedotin, developed in China, was approved based on the RC48-C008 study (NCT03556345) results, showing an ORR of 24.8% and a mOS of 7.9 months, even for patients with HER2 immunohistochemistry 2+ without Fluorescence In Situ Hybridization (FISH) testing, achieving an ORR of 23.0% and a mOS of 7.1 months at third-line [369]. Upcoming studies are set to investigate the synergy of HER-2 ADCs with immunotherapy (NCT06221748) and to determine the potential benefits of ADCs for patients with low HER-2 expression after progression on anti-HER2 therapy [DESTINY-Gastric04(NCT04704934), NCT06123494]. Besides, trophoblastic cell-surface antigen 2 (TROP2) ADCs are investigated for third-line or later treatment regardless of HER-2 status (NCT06356311).

The optimal approaches after progression on targeted drug regimens are actively being explored. In HER2-expressing patients who progressed after first-line trastuzumab-containing chemotherapy, DESTINY-Gastric02 (NCT04014075) supports the use of trastuzumab deruxtecan as second-line therapy with an ORR of 42% [330]. Other intensified anti-HER2 regimens, including ramucirumab plus trastuzumab and paclitaxel, as well as anti-HER2 bispecific antibody KN026, have also demonstrated substantial efficacy in HER-2 positive G/GEJ cancer by multicenter studies (NCT04888663, NCT03925974) [370, 371].

5.4.3 Local treatment for advanced EC

Although systemic therapy remains the standard of care for patients with advanced disease, the integration of local treatments, such as radiotherapy, rather than concurrent chemoradiotherapy, has significantly improved the management of local swallowing symptoms [372, 373]. In phase III clinical trials, despite the administration of efficacious chemoimmunotherapy, the mPFS for these patients without local radiotherapy remains approximately seven months [319]. Radiotherapy has been demonstrated to enhance survival in patients with advanced ESCC [374, 375]. The ESO-Shanghai 10 trial (NCT03000816) assessed the efficacy of stereotactic body radiotherapy (SBRT) in treating oligometastatic ESCC, revealing an mPFS of 13.3 months and a 2-year OS of 58.0% [376]. Subsequently, The ESO-Shanghai 13 trial (NCT03904927), employing a randomized controlled design, has conclusively demonstrated that the addition of local therapy for metastatic foci markedly improves the PFS in patients with oligometastatic ESCC undergoing systemic treatments, including immunotherapy [331]. Cohort study has demonstrated that cCRT for both primary and oligometastatic ESCC lesions as a first-line treatment had a superior survival to chemotherapy alone (mPFS: 9.7 months vs. 7.6 months) [377]. The ESO-Shanghai 20 study (NCT06190782) evaluates the combination of anti-PD-1 antibodies with radiotherapy for primary and metastatic sites in oligometastatic ESCC.

The AIO-FLOT3 trial (NCT00849615) also demonstrated a combination of chemotherapy followed by surgery yields favorable OS for oligometastatic G/GEJ adenocarcinoma [332]. The Delphi consensus advocates that the management of the primary tumor in oligometastatic EC should adhere to the guidelines for locally advanced EC [378]. The phase III AIO-FLOT5 (RENAISSANCE, NCT02578368) trial evaluating perioperative FLOT therapy followed by surgery of the primary tumor and metastatic lesions against palliative chemotherapy for EAC or GAC, and the ECOG-ACRIN EA2183 study (NCT04248452), which evaluating chemotherapy with consolidative radiotherapy to all sites of disease in oligometastatic EAC, are ongoing [379].