Esophageal Cancer

Esophageal cancer is a malignancy arising from the esophageal epithelium, comprising two principal histologic subtypes: squamous cell carcinoma (ESCC) and adenocarcinoma (EAC). The global burden is substantial: in 2022, at least 511,000 new cases were recorded worldwide, with 43,571 from Europe and 22,370 new cases in the United States in 2024 ^[1]A1c.

Mortality trends show a narrowing but persistent sex gap. Among European Union men, the esophageal cancer mortality rate declined from 5.34 to 4.99 per 100,000 between 2000-2004 and 2005-2009 (a 7% reduction), and is predicted to fall further to 4.5 per 100,000 by 2015 (approximately 22,300 deaths). For EU women, rates have remained low and stable, moving from 1.12 to 1.09 per 100,000, with a predicted 2015 rate of 1.1 per 100,000 (about 7400 deaths) ^[7]D5.

Geography and histology are tightly linked. In northern Europe, EAC is now the predominant histologic type among men, whereas ESCC remains more common in women and is still increasing in several countries ^[7]D5. Eastern populations have higher rates of squamous cell carcinoma, while adenocarcinoma predominates in the Western world; both subtypes show a male preponderance ^[1]A1c.

The disease typically afflicts older adults: the mean age across studies of esophagectomy cohorts ranges from 60.3 to 73.9 years ^[11]B2a. Age is a strong independent risk factor for postoperative pneumonia ^[11]B2a.

Several epidemiological patterns merit attention. Among individuals with a positive immunochemical fecal occult blood test (iFOBT) in a bowel cancer screening program, the incidence of esophageal cancer was 114.42 per 100,000, approximately 7.5 times higher than the background population rate of 36.08 per 100,000 ^[20]B3b. Brain metastases , though uncommon overall (pooled incidence 2.84%), occur significantly more often in adenocarcinoma (5.34%) than in squamous cell carcinoma (1.45%) ^[12]B2a. Among patients with metastatic disease at presentation, 24% have de-novo oligometastatic disease (≤5 metastases), and their median overall survival is 16 months (95% CI 13-21) ^[10]A1b.

These geographic and temporal patterns are driven by the distribution of key risk factors, which are discussed in the next section.

Pearl: A patient with a positive iFOBT has a 7.5-fold higher rate of esophageal cancer than the general population ^[20]B3b, consider upper endoscopy in this group even when colonoscopy is negative.

The incidence of esophageal cancer varies dramatically by geography and histology, reflecting distinct risk factor profiles for squamous cell carcinoma (SCC) and adenocarcinoma (EAC). The table below summarizes the key independent risk factors established in prospective studies.

Prevention

Prevention strategies target these modifiable risk factors. Tobacco cessation and alcohol avoidance remain the most powerful interventions for SCC, particularly in high-incidence regions like East Asia and sub-Saharan Africa. For EAC, weight reduction and management of metabolic syndrome are key. The Mediterranean diet, particularly adherence to the alternate Mediterranean diet score without alcohol (aMEDr), was associated with a 43% reduction in ESCC risk per 2-point increment in men ^[25]B2b. GLP-1 receptor agonists, increasingly used for diabetes and obesity, did not increase esophageal cancer risk (RR 0.46, 95% CI 0.13-1.59) ^[4]A1a; retrospective data suggest a possible protective effect, but this remains unconfirmed in randomized trials.

Screening

Screening for esophageal cancer is recommended in high-risk populations. The China guideline for the screening, early detection and early treatment of esophageal cancer (2022) provides evidence-based recommendations for screening population, technology, and procedure management ^[6]A1c. In Japan, screening endoscopy for ESCC in males over 50 with smoking and drinking history has shown increasing detection rates, with 95.2% of detected ESCCs at early stages ^[32]B3b. Minimally invasive screening tools, such as the capsule-sponge device paired with p53 immunohistochemistry, have shown high accuracy for ESCC (sensitivity 83.3%, specificity 95.4%) ^[24]C4. Public acceptance of screening is influenced by disease awareness, fear, belief in benefit, and practicalities ^[22]B2a. Notably, the rate of interval cancers after negative screening endoscopy is substantial (0.01% in a Korean cohort of 28 million), and endoscopist workload impacts detection rates: detection rates in the fourth quarter were half those in the first quarter ^[29]B3b.

These risk factors and screening strategies form the foundation for identifying high-risk individuals, which is further refined by genetic predisposition discussed in the next section.

Pearl: The sharp histologic dichotomy in risk factors means that screening strategies must be tailored: for SCC, target older men with smoking and alcohol history; for EAC, target those with obesity, metabolic syndrome, and GERD.

The risk conferred by established environmental exposures is modified by common genetic variants, most notably the ALDH2 Glu504Lys (rs671 G>A) polymorphism, which regulates acetaldehyde metabolism. This variant is almost exclusively present and highly prevalent among East Asians ^[35]C4. In carriers of the Lys/Lys and Glu/Lys genotypes, ALDH2 enzyme activity is reduced to nearly 0% and 17-38% of normal, respectively, leading to acetaldehyde accumulation after alcohol consumption ^[35]C4.

ALDH2 Polymorphism and Alcohol Consumption

In the China Kadoorie Biobank prospective cohort of 0.5 million adults aged 30-79 years, low-activity ALDH2 was not associated with increased esophageal cancer risk in the absence of alcohol consumption ^[35]C4. Among male weekly alcohol consumers, however, both self-reported flushing "soon after drinking" and the rs671 GA genotype were associated with significantly elevated risk. The adjusted hazard ratio for EC with rs671 GA versus GG was 3.31 (95% CI 1.94-5.67) ^[35]C4. The effect was confined to men consuming ≥30 g of alcohol per day; no increased risk was seen among light-to-moderate drinkers ^[35]C4. The HR for EC associated with each 15 g/d increment of alcohol was 1.28 (1.15-1.44) among men with a "soon" flushing response, compared with 1.12 (1.09-1.15) among those without (p_interaction = 0.047) ^[35]C4. For rs671 GA versus GG, the corresponding HRs were 1.41 (1.08-1.82) and 1.16 (1.06-1.27) (p_interaction = 0.493) ^[35]C4.

Self-reported facial flushing as a surrogate for inactive ALDH2 had low sensitivity (56.8%) but high specificity (88.4%) among male weekly alcohol consumers ^[35]C4. Adding past flushing response did not appreciably improve detection ^[35]C4. The population attributable fraction for EC that would be eliminated if men with rs671 GA who consumed ≥30 g/d changed to light-to-moderate consumption or abstinence was 7% ^[35]C4.

Androgen Pathway Gene Variants

Polymorphisms in genes within the androgen pathway have been investigated for Barrett's esophagus and esophageal adenocarcinoma. In a combined analysis of 14 studies (1,508 EAC, 2,383 BE, 2,170 controls), SNPs within CYP17A1 were associated with risk of BE in the sexes combined (p = 0.002) and in males (p = 0.003), but not in females separately ^[39]D5. SNPs within JMJD1C were associated with risk of EAC in females (p = 0.001) ^[39]D5. However, none of these associations replicated in a subsequent sample, and 14 other androgen-pathway genes showed no significant association after correction for multiple testing ^[39]D5.

Emerging miRNA Disease Associations

Computational models integrating miRNA functional similarity, disease semantic similarity, and known miRNA-disease associations have predicted numerous miRNAs linked to esophageal neoplasms. The HFHLMDA model achieved 90% confirmation of top 50 predictions, EGBMMDA confirmed 98% (49 of 50), and HLPMDA confirmed 47 of 50 candidate miRNAs in case studies ^{[36]D5[37]D5[38]D5}. These findings suggest a growing role for miRNAs as biomarkers, though they are not germline variants and are not yet incorporated into clinical hereditary risk assessment.

Pearl: The ALDH2 rs671 A allele increases esophageal cancer risk only in combination with alcohol consumption ≥30 g/day; self-reported flushing has low sensitivity (56.8%) for identifying this genotype, so genotyping should be considered for risk stratification in East Asian populations who drink heavily.

The genetic alterations described in the preceding section give rise to two main histologic subtypes, esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC), that differ fundamentally in epidemiology, molecular pathogenesis, and clinical behavior. Histologic subtype is the single most important determinant of prognosis and guides the choice of systemic therapy, with emerging molecular markers further refining treatment decisions.

Histologic Subtypes

Esophageal squamous cell carcinoma (ESCC) arises from the squamous epithelium lining the upper and middle esophagus. It predominates in Eastern populations and is strongly linked to tobacco and alcohol use ^[1]A1c. In Western cohorts, ESCC accounts for approximately 14% of cases ^[47]B3b. Esophageal adenocarcinoma (EAC) develops from Barrett metaplasia in the distal esophagus and gastroesophageal junction, is more common in Western countries, and constitutes 71-85% of cases in North American series ^{[1]A1c[47]B3b[53]D5}. Both subtypes show a male preponderance ^[1]A1c.

Rare histologies include liposarcoma, which most frequently presents as a well-differentiated tumor in the hypopharynx or cervical esophagus and carries a low risk of nodal or distant metastasis ^[44]B2a.

Molecular Drivers

Molecular profiling has identified distinct driver alterations in each subtype, though the evidence base for specific mutations is limited in the current literature. Key features include:

• ESCC: Frequent TP53 mutations, CDKN2A loss, and amplification of CCND1 or EGFR. These alterations drive cell cycle dysregulation and resistance to apoptosis.
• EAC: TP53 mutations are nearly universal, with frequent CDKN2A inactivation, HER2 amplification (in ~15-20%), and KRAS or SMAD4 mutations. HER2 overexpression defines a targetable subset.
• Mismatch repair deficiency (dMMR) is an emerging biomarker that may influence treatment choice, though definitive recommendations are not yet established ^[1]A1c.
• MicroRNA dysregulation has been implicated in malignant transformation: miR-374 and miR-574 are deregulated in esophageal mucosa after caustic injury and are linked to esophageal tumorigenesis ^[54]D5.

Prognostic and Therapeutic Implications

Histologic subtype independently affects prognosis and treatment selection:

• Recurrence risk: SCC histology is associated with a higher risk of recurrence after multimodality therapy (adjusted HR not reported in source) ^[51]B3b. Patients with SCC are also more likely to have surgery canceled after neoadjuvant treatment, reflecting a more complex oncologic and functional profile ^[49]C4.
• Therapy selection: For locally advanced disease, neoadjuvant chemoradiation (e.g., CROSS regimen) is effective for both subtypes, but the optimal neoadjuvant approach may differ. The ISDE guideline panel made no separate recommendation by histology due to insufficient evidence ^[1]A1c. For EAC, perioperative chemotherapy (e.g., FLOT) is often preferred, while SCC may be more radiosensitive.
• Molecular targets: HER2 testing is standard for advanced EAC to identify candidates for trastuzumab . PD-L1 expression guides immunotherapy use. dMMR status, though rare, predicts response to immune checkpoint inhibitors ^[1]A1c.

Table: Comparison of ESCC and EAC

Histologic subtype and molecular profile directly influence the pattern of metastasis and clinical behavior, SCC more often involves the airway and brain, while EAC preferentially spreads to liver and peritoneum. These distinctions will be further explored in the next section on clinical presentation.

Pearl: In patients with esophageal cancer, histologic subtype is not merely a pathologic label, it dictates the likelihood of nodal upstaging, the choice between neoadjuvant chemotherapy and chemoradiation, and the priority of molecular testing for HER2, PD-L1, and dMMR.

The histologic subtypes described above, squamous cell carcinoma and adenocarcinoma, share a common challenge: the majority of early-stage tumors produce no symptoms, and the diagnosis is often made at an advanced stage ^[29]B3b. The clinical presentation therefore reflects the extent of local invasion, lymph node involvement, and distant metastasis.

Symptoms

• Dysphagia - progressive, first to solids then liquids; the most common presenting symptom.
• Odynophagia - pain on swallowing, often with advanced or ulcerated lesions.
• Unintentional weight loss - occurs in >50% of patients at presentation.
• Regurgitation and vomiting of undigested food.
• Chest pain or retrosternal discomfort.
• Hoarseness - due to recurrent laryngeal nerve invasion.
• Cough, choking, or aspiration - may indicate tracheoesophageal fistula. Early-stage disease is typically asymptomatic, and lesions are frequently discovered incidentally during screening endoscopy ^[29]B3b.

Signs

• Cachexia - low cachexia index (CXI) and sarcopenia are common in advanced disease and independently predict poor survival ^{[17]B3b[61]B2a}.
• Anemia - from chronic occult blood loss.
• Supraclavicular lymphadenopathy (Virchow's node).
• Hepatomegaly or ascites - from liver or peritoneal metastases.
• Pleural effusion - with metastatic spread.
• Positive fecal occult blood test - may be the first clue; in one series, esophageal cancer was found at a rate of 114.4 per 100,000 among iFOBT-positive patients, 7.5 times the general population rate ^[20]B3b.

Onset and Progression

Symptoms appear insidiously over weeks to months. Dysphagia initially to solids then progresses to liquids; weight loss accelerates as obstruction worsens. The median time from first symptom to diagnosis is often several months, contributing to the advanced stage at presentation.

Red Flags

• Complete dysphagia (inability to swallow saliva)
• Hematemesis or melena
• Severe malnutrition (e.g., >10% body weight loss in 6 months)
• Stridor or respiratory compromise (airway invasion)

Atypical Presentations

• Brain metastases - occur in 2.84% of patients (pooled incidence), with a higher rate in adenocarcinoma (5.34%) than squamous cell carcinoma (1.45%); may present with headache, seizures, or focal deficits ^[12]B2a.
• Positive iFOBT - up to 75% of patients with UGI cancer in one study had an unexplained positive fecal occult blood test, suggesting that a gastroscopy may be warranted in this setting ^[20]B3b.

Pearl: Esophageal cancer is most often asymptomatic in its early, curable stage; any new, progressive dysphagia in a patient over 50, especially with risk factors, should prompt immediate upper endoscopy regardless of the absence of other alarm signs ^[29]B3b.

After clinical suspicion of esophageal cancer is raised, tissue acquisition for histologic confirmation is the critical next step. Endoscopic forceps biopsy remains the gold standard for initial diagnosis, but its accuracy is limited by sampling error and lesion heterogeneity. In superficial esophageal squamous neoplasms, the histopathologic discrepancy rate between forceps biopsy and endoscopic resection specimens is 34.5%, with upgrades (from HGIN to SCC or LGIN to SCC) accounting for 27 of 29 discordant lesions ^[72]C4. Similarly, in a large multicenter ESD registry, 48.7% of biopsies disagreed with the resection specimen, and 40.4% were underestimated; carcinoma was unsuspected on biopsy in 54.8% of cases ^[74]B3b. A larger lesion (≥22 mm) was a significant predictor of histologic upstaging ^[74]B3b.

Minimally Invasive Sampling Alternatives

The capsule-sponge device offers a safe, office-based alternative for detecting esophageal squamous neoplasia. In the EDEN trial, capsule-sponge sampling with cytologic assessment for intraepithelial neoplasia achieved accuracy 94.2% (95% CI 89.6-96.8%), sensitivity 83.3% (95% CI 60.8-94.2%), and specificity 95.4% ^[24]C4. Adding p53 immunohistochemistry to atypical cytology increased the odds ratio for neoplasia to 161.1 (95% CI 28.7-903.2) for a dual-positive test ^[24]C4. No adverse events occurred, and median acceptability was neutral-to-favorable ^[24]C4.

Adjunctive Optical Diagnostic Techniques

Optical coherence tomography (OCT) provides cross-sectional imaging of esophageal wall layers. In prospective studies, OCT identified intestinal metaplasia with sensitivity 81-97% and specificity 57-92%, and dysplasia/early cancer with sensitivity 68-83% and specificity 75-82% ^[70]A1a. For staging early squamous cell carcinoma, OCT accuracy exceeded 90% when compared to histologic resection specimens ^[70]A1a. However, current performance does not meet recommended thresholds to replace 4-quadrant biopsies in clinical practice ^[70]A1a.

Histopathologic and Immunohistochemical Assessment

Histologic classification of Barrett's-related neoplasia can be challenging. Immunohistochemical markers help distinguish between lesions: MUC2 and LI-cadherin expression differs significantly between normal glandular mucosa, Barrett's mucosa, intraepithelial neoplasia, and invasive carcinoma, but neither marker distinguishes low-grade from high-grade intraepithelial neoplasia ^[84]D5. Mutational profiling of biopsy tissue is increasingly feasible: CDKN2A/B and FHIT somatic copy number alterations are detected in 88% of progressors versus 24% of non-progressors, suggesting a role for early risk stratification ^[79]D5. In esophageal squamous cell carcinoma, MDM2 copy number increases occur in 9% of cases and are associated with poor prognosis ^[80]D5.

After neoadjuvant chemoradiotherapy, post-treatment biopsies have sensitivity of only 11.1% for detecting residual disease; an algorithm combining endoscopic category with biopsy results improves sensitivity to 64.8% and specificity to 95.9% for predicting pathological complete response ^[73]C4.

Pearl: When a biopsy shows low-grade or high-grade intraepithelial neoplasia, anticipate a 34-49% chance of histologic upgrade on endoscopic resection; a larger lesion (≥22 mm) and upper esophageal location are independent predictors of upstaging ^{[72]C4[74]B3b}.

With histologic confirmation established, imaging defines the anatomic extent of disease and drives treatment decisions. The choice and sequencing of modalities follow guideline recommendations and the clinical question being asked.

Staging Modalities

CT of the chest and abdomen is the initial imaging study for locoregional and distant staging, though its accuracy for T stage is only 40-50% ^[1]A1c. CT-based T staging relies on indirect signs such as fat plane obliteration, organ deformity, and tumor contact angle; equivocal findings (cT3br) are common ^[96]C4.

EUS provides the best T and N staging among available modalities, but its accuracy is suboptimal. For esophageal adenocarcinoma, overall T staging accuracy is 65.5%; sensitivity and specificity vary by T category (Table). Nodal staging accuracy is 77.9%, with sensitivity 77.3% and specificity 67.4% ^[93]B2a. Despite these limitations, EUS remains essential for locoregional assessment, particularly when combined with fine-needle aspiration.

PET/CT is recommended for detecting distant metastases and alters management in 13.2% of patients (major change in treatment intent), while reducing the need for additional imaging from 36.6% to 3.1% ^[95]B2b. Pretreatment metabolic parameters carry prognostic value: higher SUVmax is associated with worse PFS (HR 1.06, p=0.011) and RFS/DFS (HR 1.09, p=0.019); higher MTV with worse OS (HR 1.02, p=0.049); higher TLG with worse RFS/DFS (HR 2.02, p=0.022) ^[87]B2a. For cT2N0 disease, the ISDE guideline panel recommends both FDG-PET and EUS before treatment ^[1]A1c.

MRI offers superior soft-tissue contrast for assessing T4 invasion. A standardized MRI scoring system (T-score for tracheal invasion, V-score for vascular invasion) achieves AUCs of 0.94-0.99 for tracheal invasion and 0.88 for vascular invasion, compared to ≤0.71 for CT ^[96]C4. Inter-reader agreement is substantial to almost perfect (κ=0.77-1.00) vs moderate for CT (κ=0.37-0.49) ^[96]C4. MRI should be considered when CT findings are equivocal for T4 disease.

Restaging After Neoadjuvant Therapy

After neoadjuvant chemoradiotherapy, accurate restaging is critical to identify patients who may avoid surgery. DCE-MRI using a qualitative five-grade system shows high performance: grade 1 (compatible with complete response) yields 100% PPV and specificity; combining grades 1-2 increases sensitivity to 73.5% with specificity 88.9% and accuracy 78.8% ^[97]B3b. Adding DCE-MRI to PET/CT improves accuracy from 65.4% to 80.8% (p=0.03), and to endoscopy from 55.8% to 76.9% (p=0.005) ^[97]B3b. DWI sequences also predict response: responders show a significant percent increase in ADC during treatment (mean difference 21.06%, p<0.01) ^[88]B2a.

Surveillance for Recurrence

The pooled incidence of brain metastases from esophageal carcinoma is 2.84% (95% CI 1.92-4.19%), with higher rates in adenocarcinoma (5.34%) vs squamous cell carcinoma (1.45%) ^[12]B2a. This supports regular brain imaging during follow-up, particularly for adenocarcinoma. CT-based radiomics models integrating clinical features predict early recurrence (within 1 year) after trimodal therapy, with AUC 0.809 in the validation cohort ^[92]A1b.

Emerging Techniques

Radiomics and deep learning applied to CT improve lymph node metastasis detection, with pooled AUC 87% in training sets ^[60]B2a. PET/CT radiomics combines PET and CT features for T stage, node, and pathological stage prediction, achieving AUCs of 0.815-0.837 in validation cohorts ^[99]B3b. A spectral CT-based nomogram incorporating neutrophil-to-lymphocyte ratio, clinical stage, effective atomic number, and normalized iodine concentration predicts neoadjuvant chemotherapy response with AUC 0.825 in the primary set ^[98]B3b. MRI-based radiomics models combining clinical factors predict DFS and OS with concordance indices of 0.714-0.730 ^[90]A1b.

Pearl: For suspected T4 disease, an MRI-based T/V score provides substantially better accuracy and inter-reader reliability than CT, and should be integrated into the staging algorithm when CT findings are equivocal ^[96]C4.

Following imaging-based staging, molecular characterization of the tumor guides the selection of targeted and immunotherapeutic agents and provides prognostic information. The clinical utility of biomarkers in esophageal cancer depends on histologic subtype, stage, and the availability of matched therapies.

HER2 Testing

HER2 overexpression or amplification is the most established actionable biomarker in esophageal adenocarcinoma (EAC). In a nationwide Dutch cohort of 2846 patients with synchronous metastatic gastroesophageal adenocarcinoma treated with systemic therapy, HER2 testing increased from 18% in 2010 to 88% in 2016 ^[101]B3b. The overall HER2 positivity rate was 19% ^[101]B3b. Among HER2-positive patients, 77% received trastuzumab , and those treated with a trastuzumab-containing regimen had a median overall survival of 11.6 months versus 6.6 months with chemotherapy alone ^[101]B3b. The College of American Pathologists, American Society for Clinical Pathology, and American Society of Clinical Oncology recommend HER2 testing in all patients with metastatic gastroesophageal adenocarcinoma eligible for targeted therapy, using a validated algorithm: immunohistochemistry (IHC) 3+ is positive, 0 or 1+ is negative, and equivocal IHC 2+ requires confirmatory in situ hybridization (ISH) with a HER2:chromosome 17 ratio of ≥2 ^[101]B3b.

Microsatellite Instability and Genomic Biomarkers

Microsatellite instability-high (MSI-H) status is a rare but therapeutically relevant phenotype in esophageal carcinoma. In a multi-center real-world study of 1485 patients, the incidence of MSI-H was 0.49%-1.76% across centers ^[106]D5. All MSI-H tumors showed loss of PMS2 expression, with concurrent MLH1 loss in 75%; whole-exome sequencing revealed high tumor mutational burden (19.12-32.68 mutations/Mb) ^[106]D5. MSI-H is a biomarker for immune checkpoint inhibitor response, irrespective of histology. Other genomic alterations with prognostic significance include:

• MDM2 copy number increase - detected in 9% of ESCC cases, associated with shorter overall survival (p = 0.033) ^[80]D5.
• CDKN2A/B and FHIT deletions - found in 88% of nondysplastic Barrett's esophagus tissues from patients who progressed to dysplasia/adenocarcinoma, versus 24% in nonprogressors (p = 0.0004), suggesting a role as early progression biomarkers ^[79]D5.
• p16 overexpression - present in 9% of high-grade esophageal squamous cell carcinomas and 100% of small-cell carcinomas, but driven by RB1 pathway inactivation rather than HPV infection, and not predictive of favorable outcome ^[112]D5.

Circulating Biomarkers

Liquid biopsies offer a non-invasive approach to diagnosis, prognosis, and treatment monitoring.

Circulating tumor DNA (ctDNA). In 86 patients with previously treated metastatic gastroesophageal adenocarcinoma, the highest baseline ctDNA levels (top 20%) were associated with significantly shorter progression-free survival (median 1.8 vs. 4.6 months; HR 2.2, p = 0.005) and overall survival (median 5.5 vs. 9.9 months; HR 2.2, p = 0.004) ^[100]B2b. In resectable EAC (stage II/III), combining amplicon sequencing and shallow whole-genome copy number analysis detected ctDNA in 50.5% of patients at baseline; non-clearance of a baseline variant or ichorCNA >3% in pre-surgery samples was associated with a high risk of early progression (HR 4.58, 95% CI 2.22-9.46, p < 0.001) ^[107]D5.

MicroRNAs (miRNAs). Plasma miR-21 is significantly elevated in ESCC patients, while miR-31 and miR-375 are decreased. For differentiating ESCC from healthy controls, miR-31 and miR-375 each showed 87.5% sensitivity and 98.4%-100% specificity ^[103]B3b. Serum miR-18a and miR-21 have area under the curve (AUC) values of 0.948 and 0.914, respectively, outperforming conventional tumor markers ^[114]B3b.

Circulating tumor cells (CTCs). CTCs are detected in 84.6% of ESCC patients; TWIST-positive CTCs (75% of patients) are associated with poor histologic differentiation ^[108]D5.

Emerging Biomarkers

• Supersulfides. In a study of 50 esophageal cancer patients and 30 controls, plasma homocysteine hydropersulfide (HomoCysSSH) distinguished cancer from healthy subjects with AUC 0.93, sensitivity 89%, specificity 96% ^[102]B3b. Cysteine hydropersulfide (CysSSH) in exhaled breath condensate (EBC) had AUC 0.71, offering a non-invasive diagnostic alternative ^[102]B3b.
• Oncostatin M (OSM). OSM expression is elevated in ESCA tissues and correlates with advanced stage, lymph node metastasis, and poor prognosis; in vitro knockdown reduces proliferation and migration ^[109]D5.
• Stemness gene signatures. A prognostic model incorporating stemness genes, with RBBP7 as the most influential factor, accurately predicts 1-, 2-, and 3-year survival, outperforming TNM staging ^[113]D5.

Integration with Staging

Molecular biomarkers complement anatomic staging by identifying patients most likely to benefit from targeted therapy (HER2, MSI-H) or immunotherapy (MSI-H, high TMB), and by providing early readouts of treatment response or resistance (ctDNA). The choice of which biomarkers to test is guided by histology, stage, and available therapeutic options, and results should be incorporated into the multidisciplinary treatment planning discussed in the next section.

Pearl: In metastatic gastroesophageal adenocarcinoma, HER2 testing is mandatory, the nationwide increase from 18% to 88% over 6 years paralleled a survival improvement from 6.9 to 7.9 months, highlighting the impact of biomarker-guided therapy ^[101]B3b.

The molecular and histologic characterization of the tumor must be integrated with anatomic staging to guide treatment decisions. Accurate staging is the cornerstone of esophageal cancer management, determining whether a patient is a candidate for endoscopic resection, neoadjuvant therapy, or palliative treatment, and every staging modality has limitations that clinicians must understand.

TNM Classification and Stage Groups

Staging follows the AJCC 8th edition, which incorporates histologic type and grade into the stage grouping. For example, T2N0M0 adenocarcinoma is upstaged from Stage IC to IIA in the presence of poor differentiation, and squamous cell cancer is upstaged from Stage IB to IIA unless the cells are well-differentiated ^[1]A1c. After neoadjuvant therapy, the ypStage (pathologic stage after therapy) is used, but it may not fully capture prognosis: patients with ypT0N0 (pathologic complete response, pCR) have a 5-year overall survival (OS) of 75.1%, whereas those with ypT+N0 have 42.4%, ypT0N+ 54.9%, and ypT+N+ 26.1% ^[120]B3b. Even within ypStage I, survival differs between ypT0N0 and ypTis-2N0 ^[120]B3b.

Staging Modalities and Accuracy

No single modality provides perfect staging, and the accuracy of clinical staging (cTNM) compared with pathologic staging (pTNM) is often poor. CT has an accuracy of only 40-50% for T stage ^[1]A1c. Endoscopic ultrasound (EUS) performs better but still under 90% for nodal status, and occult nodal metastasis rates of up to 50% have been reported ^[1]A1c. Low-dose hydro-CT with iterative reconstruction offers a sensitivity of 93%, specificity of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 88% for esophageal cancer diagnosis, with near-perfect inter-reader agreement for T, N, and M stages (κ > 0.90) ^[64]B3b.

PET/CT has a significant impact on management: in a prospective registry, major changes (e.g., curative to palliative or vice versa) occurred in 13.2% of patients, and minor changes in 24.1% ^[95]B2b. Curative-intent therapy based on PET/CT was associated with a median OS of 3.5 years versus 0.9 years for palliative intent ^[95]B2b.

MRI outperforms CT for detecting tracheal and vascular invasion in suspected T4 disease. The area under the ROC curve (AUC) for tracheal invasion was 0.943-0.990 for MRI versus 0.529-0.706 for CT; for vascular invasion, MRI AUC was 0.878 versus ≤0.706 for CT ^[96]C4. Inter-reader agreement was substantial to almost perfect for MRI (κ = 0.771-1.000) but only moderate for CT (κ = 0.369-0.487) ^[96]C4. A combined MRI score (T-score or V-score) achieved specificity ≥ 0.88 without loss of sensitivity, supporting its use when CT findings are equivocal ^[96]C4.

Optical coherence tomography (OCT) may identify intestinal metaplasia and dysplasia, but its accuracy (sensitivity 68-83%, specificity 75-82% for dysplasia and early cancer) may not meet thresholds to replace four-quadrant biopsies ^[70]A1a. OCT may be more accurate than EUS for staging early esophageal cancer, but randomized trials are lacking ^[70]A1a.

The Challenge of Nodal Staging

Occult lymph node metastasis (LNM) is the most critical staging pitfall. In patients with clinical T1-T2 N0 esophageal squamous cell carcinoma (ESCC) undergoing upfront surgery, occult LNM was found in 21.8% ^[91]A1b. A risk model identified four independent predictors: tumor maximum standardized uptake value (SUVmax) > 3.8, higher histological differentiation grade (poorer differentiation), tumor length > 25 mm, and advanced clinical T stage ^[91]A1b. The risk scoring system showed high accuracy (AUC 0.81 in the training set, 0.743 in the testing set) ^[91]A1b.

For cT2N0 disease, the correlation between clinical and pathologic staging is poor. Accurate staging was reported in only 6-42.8% of cases, and upstaging after histopathologic review occurred in 21.4-63.4%, with nodal upstaging in up to 48% of patients ^[1]A1c. In early esophageal adenocarcinoma, the overall metastatic rate (including LNM and distant metastases) increases with depth: sm1 11.9%, sm2 27.3%, sm3 32.1% ^[124]C4. Lymph node metastases were absent in m1 and m2 tumors and rare in m3 (1/18) and m4 (5/21), but more frequent in sm2 (11/44) and sm3 (18/78) ^[124]C4.

Risk Stratification for Occult Disease and Treatment Selection

Given the high rate of understaging, the ISDE guideline panel recommends neoadjuvant therapy followed by surgery for most patients with cT2N0 esophageal cancer (conditional recommendation, low certainty of evidence) ^[1]A1c. The panel identified low-risk features that may allow primary surgery alone: small (< 2 cm), well-differentiated tumor, no lymphovascular invasion (LVI), and high confidence in staging workup ^[1]A1c. High-risk features favoring neoadjuvant therapy include poor differentiation, LVI, tumor length ≥ 2 cm, and higher SUVmax ^{[1]A1c[91]A1b}.

Staging After Neoadjuvant Therapy

Pathologic complete response (pCR) after neoadjuvant therapy offers the best prognosis, with a 5-year OS of 75.1% ^[120]B3b. However, residual nodal disease substantially worsens survival: patients with ypT0N+ (pCR of primary but residual nodal metastases) had a 5-year OS of 54.9%, and those with ypT+N+ had only 26.1% ^[120]B3b. The presence of a positive supraclavicular lymph node (cSCN+ or pSCN+) did not independently predict OS after three-field lymph node dissection in ESCC ^[121]B3b.

Clinical Implications

Staging accuracy directly drives treatment strategy. Understaging leads to undertreatment and higher recurrence risk, while overstaging exposes patients to unnecessary neoadjuvant therapy. MRI should be integrated when CT findings are equivocal for T4 disease ^[96]C4. PET/CT frequently changes management and should be part of the staging workup ^[95]B2b. The management section that follows builds on these staging principles to outline the multimodal treatment algorithm.

Pearl: The critical challenge in esophageal cancer staging is the high rate of occult nodal metastasis (up to 48% in cT2N0), which makes neoadjuvant therapy the preferred approach for most patients unless low-risk features, small size, well-differentiated histology, no LVI, and high-confidence staging, are all present ^[1]A1c.

Staging, histology, and patient fitness jointly determine the treatment strategy for esophageal cancer. The paradigm is multimodal: endoscopic resection for early disease, neoadjuvant or perioperative therapy for locoregionally advanced disease, definitive chemoradiotherapy for unresectable tumors, and systemic therapy, increasingly including immunotherapy, for metastatic or recurrent cancer. Dedicated sections on each modality are linked below; this overview provides the framework for selecting among them.

Treatment Algorithm by Stage and Histology

The initial decision is whether the tumor is amenable to curative-intent treatment. The table below summarizes the standard approach by stage grouping.

Neoadjuvant and Perioperative Therapy

For resectable esophageal cancer, preoperative therapy improves survival compared with surgery alone. The choice between nCRT and neoadjuvant chemotherapy (nCT) is debated. A meta-analysis of three randomized trials found that nCRT significantly increases R0 resection rate and pCR rate, but 3- and 5-year overall survival did not differ between the approaches ^[126]A1a. Thus, the addition of radiotherapy to chemotherapy does not clearly confer a survival advantage, despite improving local response. In patients with cT2N0 disease, the ISDE guideline makes a conditional recommendation for neoadjuvant therapy followed by surgery over primary surgical resection, acknowledging that staging uncertainty is high ^[1]A1c.

Definitive Chemoradiotherapy for Unresectable Disease

For patients with locally advanced, unresectable tumors (T4b, bulky nodal disease, or medically inoperable), dCRT is the standard of care. A correlation analysis of 11 randomized trials (N = 2812) demonstrated that event-free survival (EFS) is a strong surrogate for overall survival (R² = 0.80; slope 0.91; 95% CI 0.62-1.20) ^[129]A1a. Median EFS was 17.5 months and median OS was 25.1 months in these trials, supporting the use of EFS as an endpoint in future dCRT studies. Similarly, for resectable disease, recurrence-free survival or disease-free survival shows a strong correlation with OS (ρ = 0.89) ^[138]B2a, suggesting that these endpoints can accelerate trial completion.

Immunotherapy in Advanced Disease

First-line pembrolizumab plus chemotherapy has become a standard for advanced esophageal cancer, particularly in patients with PD-L1 combined positive score (CPS) ≥10. In the Japanese subgroup of KEYNOTE-590, with median follow-up of 60.6 months, the combination improved median OS from 11.7 to 17.7 months (HR 0.65; 95% CI 0.45-0.94) and raised the 60-month OS rate from 8.5% to 24.0% ^[127]A1b. Objective response rate was 56.8% vs 38.8%. However, cost-effectiveness analyses from Japan show an ICER of $176,479 per QALY in the overall population and $126,862 in the CPS ≥10 subgroup, exceeding the willingness-to-pay threshold of $50,000-$100,000, so the addition is not cost-effective in that setting ^[136]A1b.

Special Populations and Conversion Surgery

Older adults (≥70 years): In a target trial emulation of 432 patients with ESCC, concurrent chemoradiotherapy (CCRT) was associated with lower treatment completion rates (adjusted RR 0.86) and higher grade ≥3 toxicity (38.1% vs 17.8%) compared with radiotherapy alone ^[133]A1b. Effects were more pronounced in those aged ≥75, with Charlson Comorbidity Index ≥6, or frail (G8 ≤14). These data support careful patient selection and shared decision-making when considering intensification.

Initially unresectable disease: After effective conversion therapy, conversion surgery improves survival compared with continued non-surgical treatment. A meta-analysis of 10 studies (973 patients) found that conversion surgery was associated with significantly better OS at 1 year (RR 1.53), 2 years (RR 1.65), and 3 years (RR 2.03) ^[132]B2a. Postoperative complications included respiratory issues, anastomotic leakage, and recurrent laryngeal nerve injury.

Prehabilitation and Supportive Care

Multimodal prehabilitation, combining exercise, nutrition, and psychosocial support, during neoadjuvant chemotherapy is feasible and maintains cardiorespiratory fitness and functional capacity ^[9]C4. A 12-week walking and dietary education program improved rest-activity circadian rhythm and sleep quality at 6 months ^[137]A1b.

Thromboprophylaxis: Prolonged (1-month) low-molecular-weight heparin after esophagectomy did not reduce thromboembolic events compared with standard in-hospital prophylaxis in a randomized trial. However, preoperative prothrombin fragment 1+2 (F1+2) levels predicted postoperative VTE (OR 1.64 per 50 pmol/L increase) ^[8]A1b, suggesting a role for risk stratification.

Dumping syndrome occurs after esophagectomy in a pooled prevalence of 27% (95% CI 14-39%), but rises to 67% when assessed by specialized questionnaires ^[135]A1a. Awareness and dietary counseling are important to mitigate its impact on quality of life.

Controversies and Guideline Disagreement

Pearl: For resectable esophageal cancer, neoadjuvant chemoradiotherapy improves pathologic response and R0 resection but does not clearly improve overall survival compared with neoadjuvant chemotherapy alone; in older or frail patients, radiotherapy alone may be better tolerated than concurrent chemoradiotherapy.

Management by Modality — Dedicated Pages

Detailed, modality-specific management is covered on dedicated pages:

• Esophageal Cancer Surgical Management — operations by stage, approach, resectability, adjuvant triggers
• Esophageal Cancer Radiation Management — EBRT, brachytherapy, concurrent chemoradiation, dose/fractionation, OAR constraints
• Esophageal Cancer Systemic Therapy — perioperative / definitive chemotherapy, targeted therapy, immune checkpoint inhibitors
• Esophageal Cancer Palliative Care — symptom management, stenting/nutrition, early integration, end-of-life
• Esophageal Cancer Surveillance and Follow-up — post-treatment surveillance schedule, late toxicity, survivorship
• Esophageal Cancer Recurrent and Metastatic Disease — locoregional salvage, systemic therapy for advanced disease, oligometastatic

Building on the multimodal framework outlined above, the current standard of care for esophageal cancer has emerged through several pivotal shifts in surgical technique, systemic therapy, and endoscopic management, each grounded in landmark trials that redefined practice.

Surgery and Lymphadenectomy

The extent of lymphadenectomy has long been debated. Historically, 2-field lymphadenectomy (2FL) included mediastinal and abdominal nodes, while 3-field lymphadenectomy (3FL) added bilateral supraclavicular dissection. However, inconsistent definitions have complicated comparisons ^[143]B2b. Recent evidence, including meta-analyses and large multicenter retrospective studies, has not demonstrated a clear survival benefit of prophylactic 3FL over contemporary 2FL ^[143]B2b. The ongoing randomized phase III JCOG2013 trial is expected to clarify the therapeutic value of prophylactic supraclavicular dissection ^[143]B2b. Current Japanese guidelines weakly recommend 3FL for upper and middle thoracic esophageal cancer ^[143]B2b.

Systemic Therapy: From Chemotherapy to Immunotherapy

Early meta-analyses of cisplatin -based combination chemotherapy showed no significant survival benefit when used as adjuvant or neoadjuvant therapy (odds ratio 0.96, 95% CI 0.75-1.22) ^[139]A1a. This led to the adoption of neoadjuvant chemoradiation as the standard for locally advanced disease. The International Society for Diseases of the Esophagus (ISDE) now conditionally recommends neoadjuvant therapy followed by surgery over primary surgical resection for cT2N0M0 esophageal cancer ^[1]A1c.

The landmark KEYNOTE-590 trial established immunotherapy as a cornerstone for advanced disease. In the Japanese subgroup, first-line pembrolizumab 200 mg every 3 weeks plus chemotherapy (cisplatin 80 mg/m² and 5-fluorouracil 800 mg/m²/day) significantly improved overall survival compared with placebo plus chemotherapy (median OS 17.7 vs 11.7 months; HR 0.65, 95% CI 0.45-0.94) ^[127]A1b. The 60-month OS rates were 24.0% and 8.5%, respectively, yielding an NNT of 7 to prevent one death over 5 years ^[127]A1b. Progression-free survival also improved (HR 0.57, 95% CI 0.39-0.83), with 60-month PFS rates of 16.9% vs 0% ^[127]A1b.

Endoscopic and Minimally Invasive Approaches

Endoscopic submucosal dissection (ESD) has become the preferred treatment for early esophageal neoplasms. A post hoc analysis of a multicenter RCT comparing insulated-tip knife (ESD-IT) and needle-type knife (ESD-N) found comparable procedural times (51.0 vs 49.5 minutes), complete resection rates (90.5% vs 90.7%), and complication rates (1.8% vs 3.7%) ^[145]A1b. However, the operator handover rate due to difficulty was significantly higher with ESD-N (13.0% vs 0.6%, P=0.001), which improved with traction assistance ^[145]A1b.

For stricture prevention after extensive ESD, a randomized trial compared intralesional triamcinolone 100 mg injection with oral prednisolone (30 mg/day tapered over 8 weeks). The oral group required significantly fewer dilation sessions (5 vs 19, P=0.04), though stricture rates were similar (66.7% vs 40%, P=0.27) ^[128]A1b. For refractory benign esophageal strictures , a biodegradable stent achieved dysphagia improvement maintained at 3 months in 46.7% (95% CI 28.3-65.7) of patients, with median dysphagia-free survival of 98 days ^[141]B2b. However, caution is warranted in patients with prior chemoradiotherapy, as one fatal esophago-left atrium fistula occurred ^[141]B2b.

Supportive Care and Rehabilitation

Multimodal prehabilitation (exercise, nutrition, psychosocial support) during neoadjuvant chemotherapy is feasible, with compliance rates of 72% for supervised and 77% for home-based exercise, and both groups maintained cardiorespiratory fitness ^[9]C4. A nationwide Swedish RCT showed that a 12-week home-based exercise program 1 year after surgery significantly improved lower extremity strength (mean difference 4.48 vs 2.73, P=0.03) ^[140]A1b. A 12-week walking and dietary education program improved rest/activity circadian rhythm and total sleep time at 3 months (Cohen's d=0.86) ^[137]A1b.

Weight loss during neoadjuvant chemotherapy is an independent risk factor for postoperative infectious complications: >5% weight loss increased the odds ratio to 2.69 (95% CI 1.12-6.46, P=0.027) ^[15]B2b. Famotidine 40 mg orally daily before radiotherapy sessions reduced thrombocytopenia (P<0.0001) and preserved lymphocyte and platelet counts ^[144]A1b. A randomized trial found no benefit of prolonged (1-month) thromboprophylaxis after esophagectomy over standard in-hospital prophylaxis, though preoperative prothrombin fragment 1+2 levels predicted venous thromboembolism (odds ratio 1.64 per 50 pmol/L increase, 95% CI 1.17-2.54) ^[8]A1b. Non-placement of a cervical drainage tube after McKeown esophagectomy did not demonstrate non-inferiority to placement for anastomotic leakage (risk difference -0.100, P=0.0591) ^[142]A1b.

Special Populations: Older Adults

In patients aged ≥70 years with esophageal squamous cell carcinoma, concurrent chemoradiotherapy (CCRT) was associated with significantly lower treatment completion rates (adjusted risk ratio 0.86, 95% CI 0.79-0.93) and higher grade ≥3 toxicity (38.1% vs 17.8%; aRR 2.01) compared with radiotherapy alone ^[133]A1b. Effects were more pronounced in those aged ≥75 years, with Charlson Comorbidity Index ≥6, or frail (G8 ≤14) ^[133]A1b. These findings highlight the need for careful patient selection.

Controversies and Guideline Disagreement

These treatment advances have directly influenced prognostic outcomes, as discussed in the next section.

Pearl: Endoscopic techniques (ESD, biodegradable stents) and supportive care (prehabilitation, exercise, famotidine) have improved outcomes and quality of life, but careful patient selection is needed for older adults ^[133]A1b.

Despite advances in multimodality therapy, the overall 5-year survival for esophageal cancer remains below 20% globally ^[135]A1a. Outcomes vary substantially by stage at diagnosis, histologic subtype, and treatment response. For early-stage disease managed with endoscopic submucosal dissection (ESD), 5-year disease-free survival (DFS) reaches 91.7% ^[119]B3b. In locally advanced disease, neoadjuvant therapy followed by surgery yields 5-year overall survival (OS) rates approaching 45-50% ^[146]B2b. For advanced or metastatic disease, first-line pembrolizumab plus chemotherapy achieved a 5-year OS of 24.0% compared with 8.5% for chemotherapy alone (HR 0.65, 95% CI 0.45-0.94) in the Japanese KEYNOTE-590 subgroup ^[127]A1b.

Survival by Stage and Treatment

Survival varies markedly by treatment strategy, particularly in patients with cervical lymph node metastasis (CLNM). In a nationwide Dutch cohort, median OS was 24 months for neoadjuvant therapy followed by surgery, 18 months for definitive chemoradiotherapy (dCRT), 15 months for chemotherapy alone, 7 months for palliative radiotherapy, and 3 months for best supportive care. The 3-year OS rates were 38%, 21%, 10%, 2%, and 1%, respectively. Neoadjuvant therapy followed by surgery was associated with longer survival than dCRT (HR 0.56, 95% CI 0.34-0.91) ^[146]B2b. For patients with unresectable locally advanced disease receiving dCRT, median event-free survival (EFS) is 17.5 months and median OS 25.1 months, with a strong EFS-OS correlation (R² 0.80) supporting EFS as a surrogate endpoint ^[129]A1a.

Prognostic Factors Beyond Stage

Several patient- and tumor-related factors refine prognosis beyond TNM stage. The table below summarizes validated prognostic factors.

Postoperative Complications and Quality of Life

Postoperative dumping syndrome is common after esophagectomy , with a pooled prevalence of 27% (95% CI 14-39%) using standard assessment, rising to 67% (95% CI 60-73%) when specialized symptom questionnaires are used ^[135]A1a. Although not directly linked to survival, dumping syndrome impairs nutritional status and quality of life, potentially affecting long-term outcomes. Venous thromboembolism occurs in 13-15% of patients within 30 days of surgery, and preoperative prothrombin fragment 1+2 (F1+2) levels predict this risk (OR 1.64 per 50 pmol/L increase) ^[8]A1b.

Surrogate Endpoints for Survival

In clinical trials, RFS and DFS show strong correlation with OS (Pearson ρ = 0.89, P < 0.001), supporting their use as surrogate endpoints in resectable esophageal cancer ^[138]B2a. Similarly, EFS is a valid surrogate in unresectable locally advanced disease treated with dCRT (R² = 0.80) ^[129]A1a. These endpoints enable earlier evaluation of treatment efficacy.

Pearl: When counseling patients, remember that the strongest prognostic stratifiers beyond stage include the cachexia index (CXI), weight loss during neoadjuvant therapy, and pretreatment PET/CT metabolic parameters, each independently predicting survival across early and advanced disease.

Prognostic factors such as cachexia index and performance status carry particular weight in older patients, who represent a growing proportion of the esophageal cancer population and for whom treatment decisions must balance oncologic benefit against physiologic reserve.

Older Adults (Geriatric Population)

Age itself is an independent predictor of poorer disease-free survival after endoscopic submucosal dissection (ESD): each additional year increases the hazard by 1.8% (HR 1.018) ^[119]B3b. In patients undergoing esophagectomy , the impact of risk factors differs by age. Among those aged ≥70 years, low body mass index (BMI) poses a significantly greater risk for severe postoperative complications (composite of mortality, reoperation, renal replacement therapy, or mechanical ventilation within 30 days) compared with younger patients, with a significant age-BMI interaction (p = 0.024) ^[18]B3b. Prolonged surgical time (OR 1.061 per 10 min, p < 0.001) and hypoalbuminemia (OR 0.312, p = 0.002) are independent predictors of adverse outcomes in the overall cohort, and these associations persist in both age strata ^[18]B3b.

The cachexia index (CXI), calculated as skeletal muscle index × serum albumin / neutrophil-to-lymphocyte ratio, is independently associated with overall survival (OS) and disease-free survival (DFS) after esophagectomy (both p = 0.002) ^[17]B3b. Low CXI is more common in older patients (p < 0.001) and predicts non-EC-related mortality, making it a useful tool for preoperative risk stratification ^[17]B3b.

Performance status (PS) is a dominant prognostic factor in advanced disease. In patients receiving second-line nivolumab monotherapy, PS ≥2 and elevated C-reactive protein were independently associated with shorter OS; median OS was 22 months for those with 0 risk factors, 15 months with 1, and 4.4 months with 2 ^[16]B2b. Similarly, in patients treated with immune checkpoint inhibitor (ICI) combination regimens, PS ≥2 was independently associated with shorter progression-free survival (PFS) (p = 0.012), while occurrence of immune-related adverse events (irAEs) was associated with longer PFS (p = 0.003) ^[147]C4. In a nationwide cohort of patients with cervical lymph node metastasis, poorer PS and higher cN stage were independently associated with worse survival ^[146]B2b.

Patients with Steatotic Liver Disease

Steatotic liver disease (SLD) subtypes show differential associations with esophageal cancer risk. In a Korean nationwide cohort, metabolic dysfunction-associated steatotic liver disease (MASLD) was not significantly associated with EC risk (HR 0.81, 95% CI 0.63-1.05), whereas alcohol-driven subtypes, MetALD (HR 1.68, 95% CI 1.17-2.42) and alcohol-related liver disease (HR 2.18, 95% CI 1.36-3.49), conferred significantly elevated risk ^[27]B3b. This suggests that alcohol exposure, rather than metabolic dysfunction, is the primary driver of EC risk in patients with SLD.

Pregnancy and Immunocompromised Populations

The provided evidence does not address the management of esophageal cancer in pregnant or immunocompromised patients. In the absence of disease-specific data, treatment decisions in these populations should follow general oncologic principles with multidisciplinary input, weighing maternal-fetal risks or infection risk against oncologic benefit.

Pearl: In older adults (≥70 years), low BMI is a stronger predictor of severe post-esophagectomy complications than in younger patients; the cachexia index (CXI) can help identify those at highest risk for non-cancer mortality and guide preoperative optimization ^{[17]B3b[18]B3b}.