Letter to the Editor: Preoperative-pathologic discordance as a structural weakness in early gastric cancer management

Abstract

Preoperative-pathologic discordance is a structural weakness in early gastric cancer management, with recent data from Jize et al, published in the recent issue of the World Journal of Gastrointestinal Oncology, showing rates exceeding 25%. Such discordance - often driven by proximal location, undifferentiated or mixed histology, lesion heterogeneity, and the limitations of standard biopsy and endoscopic imaging - can lead to non-curative endoscopic resection or to potentially avoidable gastrectomy. These patterns underscore the inadequacy of uniform guideline-based decision-making in capturing true disease biology. Improving accuracy requires explicitly integrating discordance risk into pretreatment pathways. High-risk lesions should undergo intensified assessment, incorporating virtual chromoendoscopy-guided targeted biopsies for histologic heterogeneity and selective endoscopic ultrasonography when deep submucosal invasion is suspected. Regionally calibrated treatment thresholds, second-line pathology review, and emerging artificial-intelligence tools for depth and nodal-risk prediction can further refine therapeutic selection. Recognizing discordance as an anticipated outcome of complex tumor biology - and managing it proactively rather than reactively - is essential for safer, more reliable, and organ-preserving early gastric cancer care.

Key Words: Early gastric cancer; Endoscopic submucosal dissection; Preoperative-pathologic discordance; Diagnostic accuracy; Artificial intelligence; Treatment algorithms

Core Tip: Preoperative-pathologic discordance is a common and clinically important issue in early gastric cancer management. Clinicians should recognize high-risk features early and escalate diagnostic evaluation when appropriate to preserve organ-sparing treatment options. The use of enhanced diagnostic techniques, such as targeted biopsies, virtual chromoendoscopy, and endoscopic ultrasonography, can significantly improve staging accuracy and help guide better treatment decisions, ultimately reducing unnecessary surgeries.

TO THE EDITOR

Preoperative-pathologic discordance has emerged as a consequential yet often underappreciated weakness in the contemporary management of early gastric cancer (EGC). Despite substantial advances in endoscopic detection and organ-preserving treatments, the real-world precision of clinical staging remains limited. Current treatment algorithms - particularly those informed by the Japanese Gastric Cancer Treatment Guidelines and consensus criteria for endoscopic submucosal dissection (ESD) - assume that preoperative assessment can reliably identify lesions with negligible risk of lymph node metastasis, thereby justifying endoscopic resection as a curative option in carefully selected patients[1]. However, experience from high-volume East Asian centers and newer programs elsewhere shows that these assumptions do not consistently hold in practice. Even in prospective multicenter cohorts, non-curative resection after ESD remains a meaningful concern[2]. These observations highlight a fundamental constraint of current diagnostic paradigms: Tools used to evaluate depth of invasion, histologic subtype, and ulceration do not always capture the biological heterogeneity of EGC.

The clinical manifestations of discordance are increasingly well documented across diverse healthcare settings. In a recent Western cohort of 270 superficial gastric lesions, histologic concordance between forceps biopsy and ESD specimens was only 54.1%, with underestimation in 41.1% and “severe underestimation” in 8.9% of cases - underscoring the risk that biopsy-based staging will misclassify lesions selected for organ-preserving therapy[3]. Large Asian series similarly report that discrepancies between biopsy and final resection, seen in roughly 10% of EGC cases, are associated with non-curative endoscopic treatment or missed opportunities for endoscopic therapy in surgically treated patients[4-6]. Differentiation discordance, in particular, has been linked to higher rates of submucosal invasion and lymph node metastasis, emphasizing that such mismatch is not a benign phenomenon[7]. This vulnerability also extends beyond invasive cancer. In a Korean cohort of 1643 low-grade dysplasia lesions, more than one quarter were pathologically upgraded after endoscopic resection, including 6.7% upgraded to carcinoma, with irregular mucosal patterns and larger size predicting malignant progression[8]. Together, these data show that discordance is a global issue driven by biological variability, sampling limitations, and incomplete visualization of tumor architecture.

Against this backdrop, the study by Jize et al[9], published in the recent issue of the World Journal of Gastrointestinal Oncology, provides timely evidence that preoperative-pathologic discordance remains substantial in contemporary practice. In their cohort of 796 EGC lesions managed with ESD or surgery under Japanese guideline-based indications, the overall discrepancy rate between preoperative therapeutic assessment and final pathological criteria was 25.6%, meaning that nearly one in four lesions labeled as “early” ultimately fell outside preoperative expectations. Upper-third tumor location and undifferentiated histology also emerged as key predictors of under-estimation and over-estimation, respectively, underscoring that discordance systematically clusters in anatomically and biologically challenging tumors[9]. Rather than an isolated deviation, this pattern affirms the need to regard discordance as an inherent and predictable risk in EGC management and sets the stage for re-examining how guideline-based indications, diagnostic strategies, and organ-preserving algorithms should be adapted in light of real-world data. These editorial aims to encourage the field to recognize discordance as an anticipated consequence of tumor biology and to manage it proactively rather than reactively.

MAJOR RISK FACTORS AND MECHANISMS UNDERLYING STAGING DISCORDANCE

The root causes of preoperative-pathologic discordance in EGC are multifactorial, involving biologic, anatomic, and technical factors. Proximal tumors in the upper-third stomach pose significant anatomical challenges - thinner walls and proximity to vasculature make accurate depth assessment with endoscopic ultrasound (EUS) or narrow-band imaging more difficult. In Jize et al’s study[9], upper-third location was associated with two-fold higher odds of underestimation. Similar findings from 2025 nomogram-based studies identified submucosal invasion and larger tumor size as predictors of non-curative resection[10].

Undifferentiated or mixed histology also increases discordance risk. Biopsy sampling may miss heterogeneous components, leading to overestimation and avoidance of ESD in potentially curable cases. A 2020 systematic review on mixed-type EGC demonstrated that histologic heterogeneity increases non-curative resections by 10%-15%, underscoring the need for multi-site sampling[11]. More recent 2024 analyses further confirmed lymphatic and vascular invasion as key predictors of lymph node metastasis in non-curative scenarios[12].

Lesion heterogeneity and sampling errors represent additional mechanisms contributing to discordance. Comparative studies report discrepancies in tumor diameter, invasion depth, or histologic subtype in 13%-25% of cases[13], particularly in lesions with ulceration, depressed morphology, or complex surface patterns that predispose to deeper invasion. Standard modalities such as white-light endoscopy and EUS have inherent limitations. A 2025 meta-analysis showed persistent underestimation of submucosal invasion despite 70%-85% accuracy with enhanced computed tomography and double-contrast endoscopy[14].

These diagnostic challenges also appear more pronounced outside Japan, where baseline lymph node metastasis risk differs across populations. Discordance rates have been reported to be up to 20% higher in Western than East Asian cohorts[15]. Additional clinical factors - including advanced age, comorbidities, background mucosal atrophy, and multiplicity of lesions - further compound staging variability. Mismatched cases demonstrate recurrence rates as high as 12%-21%[16].

Collectively, these observations show that high-risk lesions extend beyond proximal tumors and undifferentiated histology. They encompass a broader constellation of size, morphology, heterogeneity, and biologic aggressiveness. Several predictive tools have attempted to formalize these features into structured risk-stratification systems. Recent nomograms integrate lesion diameter, ulceration, surface morphology, and suspected submucosal invasion to predict the likelihood of non-curative ESD[10], while earlier scoring systems incorporate differentiation, lymphovascular invasion, and depth of invasion to estimate nodal metastasis risk after non-curative resection[12]. Although not yet incorporated into routine guidelines, these models reflect a growing shift toward quantitatively defining “high-risk lesions” and highlight the need for explicit risk stratification within pretreatment decision pathways.

STRATEGIES TO MITIGATE DISCORDANCE IN EGC MANAGEMENT

Given the multifactorial drivers of discordance outlined above, mitigating this structural weakness requires re-engineering pretreatment pathways to explicitly integrate risk stratification. Pretreatment algorithms - defined as the sequential assessments combining high-quality endoscopy, virtual chromoendoscopy, forceps biopsy, and selective EUS to determine candidacy for endoscopic resection - must evolve from uniform evaluation to risk-adapted decision making. Lesions with features associated with greater staging variability should therefore be diverted from routine workflows into an intensified diagnostic pathway.

In alignment with the 2025 European Society of Gastrointestinal Endoscopy the guideline of management of epithelial precancerous conditions and early neoplasia of the stomach[17], this intensified pathway prioritizes meticulous virtual chromoendoscopy (e.g., narrow-band imaging/blue laser imaging) to delineate suspicious surface patterns and guide targeted biopsies. When histologic uncertainty stems from heterogeneity or mixed-type architecture - common causes of biopsy under-sampling - multi-site or focused biopsies can reveal concealed undifferentiated or lymphovascular components not captured by routine sampling. Conversely, when the dominant concern is invasion depth, particularly in lesions with ulceration, fold convergence, or larger size, selectively adding EUS becomes critical to determine whether deep submucosal involvement would preclude curative ESD. This modality-specific escalation provides a structured, biologically informed approach to narrowing diagnostic blind spots that frequently drive preoperative-pathologic discordance.

Such integrated assessment can meaningfully reduce discordance, while multidisciplinary collaboration - including second-line pathology review - further mitigates the risk of overestimation in undifferentiated or heterogeneous EGC, preventing unnecessary gastrectomy. Emerging predictive tools, including contemporary risk nomograms, offer an early framework for formalizing discordance-relevant variables into quantitative stratification systems and underscore the value of embedding risk assessment directly within pretreatment decision processes.

Beyond diagnostic refinement, treatment thresholds should be recalibrated using regional real-world data. For example, validating expanded ESD indications across multinational registries may help contextualize Japanese criteria to more diverse patient populations[18]. Equally important is shared decision-making that acknowledges staging uncertainty and aligns therapeutic intensity with individual preferences, particularly for elderly or comorbid patients in whom ESD may remain acceptable despite borderline indications.

Advances in artificial intelligence (AI) further complement these strategies. AI-assisted endoscopy, validated in recent trials, improves depth estimation and histologic prediction with accuracies exceeding 85%-90%, reducing operator-dependent variability and sampling errors[19]. Models designed for stage discrimination or risk stratification[20] hold promise for future integration into pretreatment algorithms, particularly when combined with virtual chromoendoscopy and high-quality imaging.

Ultimately, mitigating discordance demands a multifaceted approach that blends enhanced diagnostics, risk-adapted workflows, real-world evidence, and technological innovation. By addressing discordance explicitly rather than reactively, clinicians can move toward more precise, equitable, and organ-preserving management of EGC.

In practical terms, discordance is most likely to occur in lesions located in the upper third of the stomach, those with undifferentiated or mixed histology, and morphologically heterogeneous or large lesions. When such features are present, diagnostic evaluation should be selectively escalated beyond standard guideline-based pathways - such as incorporating virtual chromoendoscopy-guided targeted biopsies, selective EUS, or second-line pathology review - to better characterize histology and depth. This risk-adapted approach differs from conventional workflows that assume staging accuracy and uniform applicability of ESD indications across patients.

CONCLUSION

Preoperative-pathologic discordance in EGC is not a rare complication but a structural reality shaped by biologic complexity, anatomic constraints, and interpretive limitations. Recognizing this vulnerability is the first step toward improving current practice. By embedding explicit risk stratification into pretreatment pathways and adopting modality-specific escalation strategies - including selective EUS, virtual chromoendoscopy-guided biopsies, and second-line pathology review - clinicians can more reliably differentiate lesions appropriate for endoscopic resection from those requiring surgery. Future progress will depend on validating regional data, refining treatment thresholds, and integrating emerging tools such as AI-assisted endoscopy and predictive modeling into clinical workflows. These efforts must be coupled with transparent, patient-centered discussions that acknowledge uncertainty and balance oncologic safety with organ preservation. Discordance is not an error; it is a predictable outcome of a complex disease system. Managing it explicitly, rather than reactively, is essential for delivering more precise, equitable, and organ-preserving care in EGC.