Heterogeneity-oriented multidisciplinary and super-minimally invasive precision treatment framework for locally advanced gastric cancer

Abstract

Locally advanced gastric cancer, defined here as cT3-T4a disease with or without regional lymph node involvement and without distant metastasis, is characterized by substantial spatial and clonal heterogeneity. Distinct hypoxic, proliferative, invasive, immune-evasive, and treatment-resistant subpopulations may coexist within the primary tumor, the serosal-peritoneal interface, and regional lymph nodes. Standard perioperative systemic therapy and gastrectomy with D2 lymphadenectomy remain the established curative-intent treatment for resectable locally advanced gastric cancer; however, a uniform stage-based strategy may not fully represent the biological complexity of these tumors. To this end, we refined the Heterogeneity-oriented multidisciplinary team planning and super minimally invasive precision resection concept as a hypothesis-driven treatment framework rather than a clinically validated alternative to standard surgery. The framework comprises preoperative probabilistic heterogeneity mapping, heterogeneity-focused multidisciplinary decision-making, anatomically appropriate local treatment using endoscopic, laparoscopic or combined approaches, explicit management of the serosal-peritoneal and regional lymph node compartments, and postoperative biomarker-informed and risk-informed therapy. The proposed functional zones are not fixed anatomical layers, and endoscopic super minimally invasive surgery alone is insufficient for cT3-T4a disease. In particular, cT4a serosal involvement belongs to a serosal-peritoneal interface zone that requires laparoscopic evaluation and oncological control, while regional nodal disease requires guideline-concordant lymph node management. The model currently lacks primary clinical data, and claims of lower morbidity or superior long-term outcomes are not supported. Future validation should entail radiologic and pathologic mapping studies, followed by prospective feasibility studies measuring technical success, R0 resection, serosal and nodal control, perioperative complications, functional preservation, recurrence and survival. Heterogeneity-oriented multidisciplinary team planning and super minimally invasive precision resection should therefore be viewed as a testable framework for selected patients and clinical trials, with oncological adequacy taking priority over organ preservation.

Key Words: Locally advanced gastric cancer; Intratumoral heterogeneity; Multidisciplinary treatment; Super minimally invasive surgery; Precision oncology

Core Tip: Heterogeneity-oriented multidisciplinary team planning and super minimally invasive precision resection is an investigational heterogeneity-oriented framework, not a substitute for standard D2 gastrectomy. The central requirement of this concept is that any function-preserving local intervention must also achieve adequate control of the full gastric wall, the serosal-peritoneal interface, and regional lymph nodes. Endoscopy alone cannot meet these requirements in cT3-T4a disease; therefore, implementation would require a carefully selected, combined endoscopic-laparoscopic pathway and prospective validation.

INTRODUCTION

Locally advanced gastric cancer (LAGC) remains a major therapeutic challenge due to its biological complexity, which cannot be fully described by the anatomical stage and tumor size. Multi-region genomic, transcriptomic and proteomic studies have shown that spatially separated regions of the same gastric tumor may differ in clonal composition, proliferative activity, stromal interaction, immune state and treatment sensitivity[1-3]. Furthermore, heterogeneous subclones may also be distributed between the primary tumor and regional lymph nodes, thereby contributing to discordant responses, residual disease and recurrence. Gastrectomy with guideline-defined lymphadenectomy, combined with perioperative or postoperative systemic therapy when indicated, remains the standard curative-intent approach for resectable LAGC[4,5]. Heterogeneity-oriented multidisciplinary team (MDT) planning and super minimally invasive precision resection (HOM-SMIPR) is a novel approach that aims to convert the spatial heterogeneity of tumors into a reproducible decision framework that improves tumor characterization, local treatment planning, and postoperative stratification without compromising oncological completeness. This distinction is significant given that recurrence can occur after apparently curative D2 resection, and intensified systemic therapy has not been able to uniformly eliminate resistant or spatially protected subpopulations[6-8]. Super minimally invasive surgery (SMIS) emphasizes precise lesion targeting and maximal preservation of organ structure and function[9]. Nevertheless, most evidence supporting endoscopic or SMIS-based resection is derived from early or localized gastrointestinal neoplasia, and cannot be directly extrapolated to cT3-T4a gastric tumors. The invasive front of these tumors may extend through the muscularis propria to the subserosa or serosa, the regional lymph node metastases may contain biologically distinct subclones, and free cancer cells may be present in the peritoneal cavity. Therefore, intraluminal procedures alone cannot account for the diverse pathological components. Accordingly, the HOM-SMIPR framework involves radiology, pathology, gastroenterology, gastric surgery and medical oncology. The term “precision resection” refers to treatment planning within explicit anatomical and oncological boundaries, and does not imply that every functional subregion can be selectively ablated or that endoscopic therapy can independently provide radical treatment for LAGC.

CHARACTERISTICS OF LAGC HETEROGENEITY AND LIMITATIONS OF UNIFORM TREATMENT

Functional zoning and its anatomical boundaries

For conceptual purposes, LAGC may contain several overlapping functional regions. A central hypoxic or poorly perfused region usually harbors quiescent, apoptosis-resistant, and metabolically adapted cells, while the marginal invasive region may show epithelial-mesenchymal transition, stromal interaction, lymphovascular access, and enhanced migratory capacity. On the other hand, the paravascular regions may contain highly proliferative cells that are relatively accessible to systemic therapy, although biomarker expression remains spatially heterogeneous. Furthermore, additional subclones may be present in the lymphatic channels, regional lymph nodes, and the peritoneal cavity[1-3,10]. Notably, these regions are probability-based biological constructs inferred from imaging, pathological examination, and molecular testing, and therefore should not be interpreted as fixed histological compartments with universally visible borders that can be removed with a different endoscopic device. Their purpose is to guide targeted sampling, risk assessment, and treatment selection. Nevertheless, the accuracy, reproducibility, and interobserver agreement of such zoning have not been established so far. The primary safety boundary or surgical margin for tumor resection is the anatomical depth. A cT3 tumor extends through the muscularis propria into the subserosa, whereas a cT4a tumor directly involves the serosa or visceral peritoneum. Therefore, in the HOM-SMIPR framework, the involved serosal area is classified as the serosal-peritoneal interface zone that lies outside the reliable reach of purely intraluminal treatment, and requires laparoscopic inspection, assessment of the external gastric wall, and evaluation for peritoneal dissemination. The extent of serosal change is clinically relevant because serosal invasion is associated with peritoneal spread and intraperitoneal free cancer cells[11,12]. Altogether, the possible indications for organ-preserving HOM-SMIPR, generally after standard preoperative treatment, are as follows: (1) Major response or tumor downstaging; (2) No distant or peritoneal metastasis; (3) Negative or acceptably managed peritoneal cytology; and (4) A primary lesion for which complete gastric-wall, serosal and nodal control can be achieved through a planned hybrid approach. If these conditions are not met, or if intraoperative findings exceed the predefined boundary, standard gastrectomy with D2 lymphadenectomy should remain the default treatment.

Mechanisms by which heterogeneity may limit treatment

The spatial heterogeneity of LAGC limits treatment efficacy through several non-exclusive mechanisms. First, systemic therapy may preferentially eliminate proliferating sensitive clones while leaving quiescent, stem-like or microenvironment-protected cells. Second, fibrosis, hypoperfusion and altered extracellular matrix may reduce drug delivery to selected regions. Third, biomarker expression, including human epidermal growth factor receptor 2 (HER2), programmed cell death ligand 1, mismatch-repair proteins and other emerging targets, may vary within the same tumor or between the primary lesion and metastases[10]. Fourth, regional lymph nodes may undergo independent clonal selection and may respond differently from the primary tumor. These mechanisms provide a biological rationale for improved mapping and stratification, but they do not establish that local zonal treatment improves survival. D2 gastrectomy provides full-thickness removal of the primary tumor and systematic regional lymph node control and remains the reference oncological operation[4,5]. It does not, however, directly individualize the extent of treatment according to intratumoral subclonal variation. Conversely, an organ-preserving local strategy may reduce anatomical loss but creates a greater risk of undertreating deep-wall, nodal or peritoneal disease. HOM-SMIPR is proposed to study this trade-off rather than to assume that precision local treatment is superior.

CONSTRUCTION OF THE HOM-SMIPR FRAMEWORK

The HOM-SMIPR framework consists of three linked stages: (1) Preoperative heterogeneity decoding and MDT decision-making; (2) Intraoperative management of the primary tumor together with the serosal-peritoneal and regional nodal compartments; and (3) Postoperative biomarker-informed and risk-informed treatment. The conceptual workflow of the HOM-SMIPR framework for LAGC is shown in Figure 1. The Figure 1 represents a hypothesis-generating pathway and does not imply established clinical efficacy.

Figure 1 Conceptual workflow of the heterogeneity-oriented multidisciplinary team planning and super minimally invasive precision resection framework for locally advanced gastric cancer. HOM-SMIPR: Heterogeneity-oriented multidisciplinary team planning and super minimally invasive precision resection; MDT: Multidisciplinary team; SMIS: Super minimally invasive surgery.

Preoperative probabilistic mapping and heterogeneity-focused MDT

The preoperative map should integrate conventional staging with spatially registered information rather than depend on a single radiomic output. Contrast-enhanced computed tomography, endoscopy, endoscopic ultrasonography and, where clinically indicated, staging laparoscopy provide the anatomical foundation, while radiomics can estimate intratumoral and peritumoral variation[13,14], and targeted multi-region biopsy can be used to examine differentiation, proliferation, immune markers and actionable alterations. As superficial biopsy cannot reliably sample deep subserosal or serosal disease, the map should explicitly record areas of uncertainty instead of treating unobserved regions as negative. At present, the accuracy of central, marginal, and paravascular zoning is unknown. Therefore, it is necessary to compare preoperative maps with spatially registered gastrectomy specimens and determine the sensitivity, specificity, spatial overlap, calibration, and interobserver agreement. Multi-point biopsy also carries sampling error; therefore, the number and location of samples should be prespecified, documented through photographs or by a coordinate system, and updated when intraoperative or final pathology identifies discordant regions. The distinctive feature of HOM-SMIPR is a heterogeneity-focused, question-driven output. The radiologist defines anatomical depth, radiomic uncertainty and suspected extramural disease, the pathologist assesses spatial variation and specimen orientation, the endoscopist evaluates intraluminal boundaries and technical feasibility, the gastric surgeon determines serosal, peritoneal and nodal control, and the medical oncologist interprets systemic-treatment response and biomarker implications. The conclusion of the MDT should codify treatment boundaries, acceptable uncertainties, conversion triggers and postoperative requirements before intervention. Nevertheless, the transition from a static preoperative map to intraoperative guidance remains experimental, and may warrant fluorescence imaging, repeat endoscopic ultrasound, optical or molecular imaging, image-to-patient registration and artificial intelligence-assisted recognition. These technologies remain candidates for future validation and are not yet established as real-time navigation tools for LAGC.

Primary-tumor and serosal-peritoneal management

LAGC treatment involves zone-specific management because not every functional zone has a validated selective treatment. For instance, superficial residual components confined to the mucosa or submucosa can be treated by endoscopic resection in selected investigational settings, while deeper intramural disease may require endoscopic full-thickness resection, laparoscopic local resection, or laparoscopic-endoscopic cooperative surgery. Non-exposure techniques are usually employed for full-thickness resection as they reduce direct communication between the tumor surface and the peritoneal cavity[15]. However, these techniques have been established mainly for subepithelial tumors and selected early cancers, and not for the routine treatment of cT3-T4a LAGC. There is currently no validated method that can identify a dormant hypoxic subclone with sufficient spatial precision to permit selective ablation while reliably preserving the normal gastric wall. Radiofrequency ablation, cryoablation or other energy-based treatments should therefore not be presented as standard components of HOM-SMIPR. Their use would require a clinical protocol with a validated biological target, a clearly defined treatment boundary, and mandatory pathological or imaging assessment. Similarly, occult submucosal micrometastatic foci may not always be detected with fluorescence or intraluminal ultrasound. Practical tumor control depends on en bloc resection when feasible, adequate lateral and deep margins, systematic specimen mapping, and a low threshold for conversion to standard surgery. Therefore, complete intraoperative resection should be verified on the basis of negative resection margins, adequate full-thickness control when the deep wall is involved, absence of untreated serosal or peritoneal disease, and a complete nodal strategy, rather than visual disappearance alone. For suspected cT4a disease, the serosal-peritoneal interface must be examined laparoscopically, and staging laparoscopy and peritoneal lavage cytology should be considered according to current guidelines and institutional practice. Gross serosal extension, positive cytology, peritoneal deposits, inability to obtain a safe full-thickness margin, or concern for tumor spillage should preclude an endoscopy-dominant organ-preserving procedure and prompt standard oncological management.

Regional lymph node heterogeneity and oncological management

The curative-intent therapy for LAGC involves primary tumor removal along with adequate treatment of metastatic lymph nodes. The HOM-SMIPR framework therefore separates primary-lesion control from regional nodal control. For patients with clinically node-positive or otherwise resectable LAGC, guideline-concordant lymphadenectomy, usually D2 in an appropriate surgical setting, remains the reference approach[4,5]. A combined endoscopic-laparoscopic procedure would only be oncologically coherent if the planned operation also includes adequate lymph node dissection or another prospectively validated nodal strategy. Sentinel-node navigation demonstrates that stomach-preserving surgery can be investigated in early gastric cancer. However, the available evidence is confined to carefully selected early-stage disease[16], and should not be extrapolated to cT3-T4a tumors or used to justify omission of D2 lymphadenectomy. Although treatment response, radiomics, circulating tumor DNA or molecular profiling cannot currently replace pathological nodal assessment, their ability to improve nodal risk stratification may be investigated in future.

Postoperative biomarker-informed and risk-informed therapy

Postoperative treatment should be based on the final pathological stage, treatment response, nodal status, resection margins, molecular biomarkers, and the evidence supporting each therapy. Spatial heterogeneity can aid in the interpretation of these biomarkers, especially when pretreatment biopsy and resection specimens are discordant. Repeat testing of a resected specimen or metastatic site may be appropriate when the results alter the treatment approach. The current LAGC biomarkers include HER2 amplification or overexpression, mismatch-repair deficiency or microsatellite instability, and programmed cell death ligand 1 combined positive score, along with emerging targets such as CLDN18.2 and FGFR2b[10]. However, most target-directed evidence is derived from advanced or metastatic disease, and biomarker positivity does not automatically establish an adjuvant indication. HER2-directed therapy, immune-checkpoint blockade or emerging targeted agents should be used in accordance with applicable guidelines, regulatory indications or clinical trials. The negative or incomplete results of perioperative immunotherapy studies, including KEYNOTE-585, also indicate that systemic treatment alone should not be assumed to overcome all spatial heterogeneity[8]. In a practical workflow, suspected resistance should lead to reassessment of pathology, biomarker discordance, treatment exposure and clinical-trial eligibility rather than an unspecified reversal therapy. Invasion-associated programs such as epithelial-mesenchymal transition or transforming growth factor-β signaling may support risk stratification and translational research, but they are not established postoperative targets in LAGC. Dynamic surveillance may incorporate conventional imaging, endoscopy, nutritional and functional assessment and, where validated, circulating tumor DNA.

DIFFERENCES FROM STANDARD TREATMENT, POTENTIAL ADVANTAGES AND LIMITATIONS

The conceptual difference between standard treatment and HOM-SMIPR lies in the decision logic rather than in the simple addition of MDT or endoscopy. Standard treatment is primarily stage-based and anatomy-based and has established oncological efficacy. HOM-SMIPR adds probabilistic spatial mapping and requires each treatment decision to address the primary tumor, the full gastric wall, the serosal-peritoneal interface, regional lymph nodes, and postoperative systemic risk. The framework therefore seeks biological refinement of standard care, not replacement of its oncological principles. The potential advantage of HOM-SMIPR is functional preservation. In cases where the primary lesion can be completely controlled by limited gastric-wall resection, and the nodal and peritoneal risks are adequately managed, preservation of gastric volume and continuity could theoretically reduce postgastrectomy nutritional impairment, dumping symptoms, and quality-of-life loss. Laparoscopic-endoscopic cooperative surgery illustrates how precise local resection can minimize unnecessary gastric-wall sacrifice[15].

Nevertheless, such benefits have been demonstrated mainly in subepithelial tumors and early gastric cancer, and cannot be extrapolated to LAGC. Furthermore, the limitations are substantial; for instance, mapping may be inaccurate, the biopsy may miss resistant or deep subclones, the planned local boundary may underestimate microscopic spread, full-thickness intervention may cause bleeding, perforation, leakage, intra-abdominal infection, stenosis, delayed gastric emptying or tumor-cell spillage, lymphadenectomy may negate part of the minimally invasive advantage, and unexpected findings may require conversion to gastrectomy. No reliable complication rate is currently available for HOM-SMIPR in cT3-T4a disease, and long-term oncological evidence is also absent. Existing SMIS guidelines and isolated reports of endoscopic resection after downstaging demonstrate technical possibilities, but do not validate the complete HOM-SMIPR model or establish non-inferior survival in LAGC[9,17]. The model should therefore remain investigational, and organ preservation should never take precedence over R0 resection, nodal control, or management of peritoneal disease.

CLINICAL VALIDATION ROADMAP AND FUTURE DIRECTIONS

Stage 0: Validation of the heterogeneity map

The first step should not be an organ-preserving intervention trial but a prospective mapping study in patients undergoing standard gastrectomy. Preoperative computed tomography, endoscopic ultrasound, endoscopy, radiomics, and targeted biopsy should be spatially registered to the resected specimen. Primary outcomes should include the accuracy of invasion-depth assessment, sensitivity and specificity for predefined biological regions, spatial overlap between predicted and pathological regions, interobserver agreement, and the frequency of clinically important sampling error. This stage would determine whether the heterogeneity map is sufficiently reproducible to guide treatment.

Stage I: Feasibility and safety

If mapping is sufficiently accurate, an initial prospective feasibility cohort of approximately 20-30 highly selected patients could be considered at experienced centers. Eligibility should require response to standard preoperative therapy, absence of distant and peritoneal metastasis, predefined technical resectability, and an explicit nodal plan. Primary endpoints should include protocol completion without unplanned oncological compromise and 30-day major morbidity, and secondary endpoints should include en bloc and R0 resection, deep and serosal margin status, adequacy of lymph node management, conversion to standard gastrectomy, leakage, infection, bleeding, length of stay, gastric function, nutritional status and patient-reported quality of life. Recruitment time would depend on the rarity of eligible patients and should be assessed during protocol development rather than assumed from the present concept.

Stage II: Comparative oncological evaluation

A multicenter comparative study should be undertaken only after feasibility and safety have been adequately demonstrated. The control group should receive guideline-concordant standard treatment. Depending on equipoise and recruitment feasibility, the study could use either a randomized design or a rigorously defined prospective control cohort. Disease-free survival, local and peritoneal recurrence, overall survival and gastric-cancer-specific survival should be evaluated along with functional and quality-of-life endpoints. Sample size and any non-inferiority margin must be based on stage I event rates and clinically acceptable oncological risk. Future technical development should focus on validated image registration, non-exposure full-thickness methods, reliable deep-margin assessment, and methods that connect preoperative maps to the orientation of the final specimen. Finally, research should remain focused on gastric cancer until proof of concept is established.

CONCLUSION

HOM-SMIPR reframes spatial heterogeneity as a clinically testable source of information rather than as a justification for endoscopic treatment of all LAGC. Its essential principle is that biological precision must remain subordinate to anatomical and oncological completeness. For cT3-T4a disease, this requires explicit control of the full gastric wall, the serosal-peritoneal interface and regional lymph nodes, usually through an integrated endoscopic-laparoscopic and systemic pathway. The framework is currently conceptual, has no validated complication or survival advantage, and should only advance through stepwise prospective studies. If future evidence demonstrates that selected patients can achieve equivalent oncological control with improved function, HOM-SMIPR may become a useful extension of standard multidisciplinary care. At present, D2 gastrectomy and guideline-concordant systemic therapy remain the reference treatment.