Endoscopic and histopathological phenotypes of early gastric neoplasia: toward an integrative host-response framework

Early gastric neoplasia represents a clinically decisive but biologically heterogeneous interval in gastric carcinogenesis. Contemporary endoscopy has improved lesion detection and characterization through white-light endoscopy, linked color imaging, blue laser imaging, magnifying narrow-band imaging, and computer-aided systems. Histopathology remains essential for defining dysplasia, invasion depth, differentiation, mucin phenotype, and endoscopic curability. However, most current diagnostic frameworks remain predominantly lesion-centered and incompletely account for the injured mucosal field and host-response context in which early neoplastic lesions arise.

This review aims to synthesize endoscopic, histopathological, microenvironmental, microbial, metabolic, and integrative medicine evidence to propose a lesion-field-host framework for interpreting early gastric neoplasia.

We reviewed key evidence from endoscopic imaging studies, gastric premalignant lesion guidelines, Helicobacter pylori prevention literature, pathology-continuum studies, metaplasia and SPEM biology, OLGA/OLGIM and Kyoto-classification research, single-cell and spatial profiling, microbiome and metabolomics studies, digital pathology, and syndrome-related clinical research.

The proposed framework comprises three interrelated layers. The lesion layer captures visible and microscopic features of superficial neoplastic disease, including morphology, demarcation line, microvascular and microsurface patterns, biopsy and endoscopic submucosal dissection pathology, differentiation, invasion depth, and curability. The field layer captures background mucosal risk, including H. pylori status, eradication history, atrophy, intestinal metaplasia, SPEM-like metaplastic change, OLGA/OLGIM stage, Kyoto-classification features, microbiome dysbiosis, and metabolomic remodeling. The host-response layer captures inflammatory, immune, metabolic, nutritional, symptom-based, tongue-image, and syndrome-based variables. Within this layer, traditional Chinese medicine syndrome differentiation is framed as a candidate latent host-response phenotype rather than a substitute for endoscopy or histopathology.

Future progress in early gastric neoplasia will likely depend less on isolated biomarkers than on disciplined integration of optical, histological, field-mucosal, and host-response phenotypes. Prospective multicenter validation should incorporate standardized image acquisition, mapping biopsies, OLGA/OLGIM staging, digital pathology, mucosal microbiome and metabolomic profiling, blinded syndrome assessment, and prediction-model reporting aligned with TRIPOD+AI and PROBAST+AI. This framework may support more rational surveillance, prevention, and integrative risk stratification while avoiding overstatement of currently exploratory syndrome-pathology associations.