Pharmacokinetics, safety, and pharmacodynamics of intravenous methoxyethyl etomidate hydrochloride (ET-26-HCl) in Chinese participants with hepatic impairment.

Gastric cancer (GC) originates from gastric epithelial cells. As of 2022, gastric cancer ranked fifth worldwide in both incidence and mortality among malignant tumors, with approximately 970,000 new cases and 660,000 deaths, posing a serious threat to human health (). In China, approximately 360,000 new cases of gastric cancer and 260,000 related deaths occur each year, making it the third leading cause of cancer-related death (). More than 60% of Chinese patients with gastric cancer are already at an advanced stage at the time of diagnosis, a situation closely associated with a relatively low 5-year survival rate of approximately 35% to 40% (). In sharp contrast, the 5-year survival rate of patients with early gastric cancer can exceed 90% after curative treatment, such as surgery or endoscopic therapy (). Therefore, early detection, early diagnosis, and early treatment constitute the core strategy for reducing gastric cancer mortality (). The current core diagnostic framework for gastric cancer comprises four major components: screening, imaging examination, endoscopic examination, and pathological diagnosis. Among these, gastroscopy combined with pathological biopsy remains the “gold standard”, as it can determine the histological type, degree of differentiation, and depth of invasion of the tumor. However, this examination requires transoral insertion of an endoscope and biopsy sampling; it is therefore an invasive procedure that must be performed by specialized physicians. It can easily cause discomfort such as nausea and vomiting, and population adherence is less than 30% (, ), making large-scale screening difficult to implement. Tumor marker testing, such as CEA and CA724, and serum pepsinogen (PG) testing are noninvasive, but their sensitivity and specificity are limited. They are suitable only for risk stratification and cannot provide a definitive diagnosis. Even for emerging liquid biopsy markers in recent years, such as exosomal lncRNAs, the results may still be affected by individual differences, disease stage, testing platform, and other factors; clinically, they are usually used only as auxiliary evidence () and cannot replace pathological diagnosis. Among imaging examinations, contrast-enhanced CT is mainly used for tumor staging. Non-contrast CT has traditionally been considered unsuitable for screening because gastric distension varies greatly, interference from gastric contents is common, and early lesions are subtle (, ). In addition, uneven endoscopic diagnostic proficiency in primary medical institutions and imbalanced allocation of medical resources further aggravate the challenges of early diagnosis. Therefore, efficient, convenient, and broadly applicable technological breakthroughs are urgently needed. Artificial intelligence, particularly the deep learning (DL) branch, can automatically learn complex and subtle feature patterns from massive datasets through multilayer neural network models. It has demonstrated performance surpassing that of human experts in fields such as image recognition () and has already played an important role in specific diseases, including pulmonary nodule detection and the diagnosis of diabetic retinopathy (). Its core advantages include: ① quantitative analysis of subtle features that are difficult for the human eye to detect, thereby improving the detection rate of early lesions (, ); ② reduction of diagnostic subjectivity and realization of standardized assessment (); ③ efficient processing of massive medical data, thereby alleviating physicians’ workload (); ④ adaptability to multimodal data integration, enabling the construction of a comprehensive disease representation system ().