In the golden autumn season, Jinan, the "Spring City," hosted the annual grand event in the field of laboratory medicine – the 19th National Conference on Laboratory Medicine of the Chinese Medical Association. During the conference, the seminar "From Markers to Intelligence: Achieving Maximum Efficiency in Liver Cancer Early Screening with Optimal Resources" hosted by Dunwill Medical, was successfully held on November 1st. The event attracted over 150 experts and scholars who gathered to discuss the deep integration of intelligent technology and laboratory markers, aiming to build a more efficient and equitable new pathway for early liver cancer screening, thereby achieving a synergistic win-win situation of "minimum cost" and "maximum early diagnostic power."

In her opening speech, Professor Guo Wei, Director of the Department of Laboratory Medicine at Zhongshan Hospital affiliated with Fudan University and Chair of the session, pointed out that the rapid development of laboratory medicine over the past decade has significantly promoted the evolution of liver cancer screening models. From traditional biochemical markers to innovative combined marker detection, from imaging assistance to molecular diagnosis, and now to the deep empowerment by artificial intelligence and big data, we are witnessing a qualitative shift from "marker-driven" to "intelligent decision-making." She emphasized that this is not merely a technological upgrade but signifies the ability to extend early screening to a broader population in a more scientific, economical, and equitable manner. This special session, by convening expertise from multiple fields, holds the potential to advance liver cancer early screening from "technically feasible" to "systematically efficient," and from "localized innovation" to "comprehensive accessibility."

During the presentation session, Professor Chu Xinmin from the First Affiliated Hospital of USTC (Anhui Provincial Hospital) delivered an insightful sharing on the new paradigm for early liver cancer screening. He noted that China bears the heaviest global burden of liver cancer, accounting for 42.4% of cases worldwide. Research indicates that at least 60% of liver cancer cases can be effectively prevented through controlling risk factors. To achieve the "Healthy China 2030" goals, it is imperative to further shift the focus of cancer prevention and control forward, strengthening comprehensive prevention, active screening, and risk stratification management. He introduced Japan's successful experience with a nationwide high-risk population monitoring system, which enables over half of new liver cancer cases to be detected at an early stage. Professor Chu highlighted that, given China's vast screening population base, constructing more efficient and cost-effective screening strategies is particularly crucial.

Professor Chu further indicated that integrating multi-modal data with artificial intelligence technology can build new liver cancer screening strategies with greater health economic benefits. For example, the GALAD model, constructed based on indicators such as sex, age, AFP, and PIVKA-II, has already demonstrated good performance in early liver cancer diagnosis. Multi-marker combined diagnosis not only improves detection efficiency but also offers cost advantages. Introducing the liver cancer-related molecular marker miRNA7 on the basis of existing models can further significantly enhance the diagnostic capability for liver cancer, especially early-stage cases. Both the "GALAD+M" and "AFP+M" combinations show superior early identification performance. Leveraging "big data + AI" methods allows for the continuous optimization of the comprehensive performance of multi-modal models.

At the algorithm level, compared to the classical GALAD model, the Random Forest algorithm demonstrates more significant distinction in score distribution between liver cancer and non-liver cancer groups. Even when using a simplified feature set comprising only AFP and miRNA7, the scoring based on the Random Forest algorithm maintains excellent diagnostic performance, with an AUC value reaching 0.905. The introduction of the molecular marker miRNA7 further enhances the model's early identification ability. This marker covers multiple key pathways in the hepatitis B-related inflammation-cancer transformation process and can signal cancer risk at a very early stage. Even in patients with small liver nodules where both AFP and DCP are negative, miRNA7's AUC for identifying very early liver cancer still reaches 0.728, demonstrating significant auxiliary diagnostic value.

Finally, Professor Chu emphasized that the promotion of multi-modal models still faces challenges. Core issues include inconsistencies in the units used for serum markers across different testing platforms and the lack of systematic inter-laboratory quality assessment, which hinder the widespread application of models like GAAD/GALAD. To address this bottleneck, a standardization project led by the Shanghai Medical Association Specialty Committee of Laboratory Medicine, and jointly organized by the National Institutes for Food and Drug Control, Shanghai Municipal Center for Disease Control and Prevention, Shanghai Center for Clinical Laboratory, and Zhongshan Hospital affiliated with Fudan University, is dedicated to improving the universality of multi-modal integrated liver cancer diagnostic models in clinical promotion through the homogenized assessment of liver cancer serum protein markers. Additionally, it is necessary to promote precise risk stratification, develop differentiated monitoring pathways, and dynamically adjust testing prices based on population incidence rates to achieve optimal resource allocation. With the deepening integration of multi-omics data and the application of advanced algorithms, multi-modal fusion models are expected to play a greater role in early liver cancer screening, ultimately achieving the goal of "minimum cost, maximum early diagnostic power."

With the rapid development of testing technologies and AI algorithms, early liver cancer screening has moved from single-marker detection to a new stage of multi-dimensional integrated analysis. However, its clinical implementation still faces multiple challenges, including cost, resource allocation, and standardization. In the final panel discussion, moderated by Professor Gu Bing, Director of the Department of Laboratory Medicine at Guangdong Provincial People's Hospital, and featuring Professor Chu Xinmin, Director of the Health Management Center at the First Affiliated Hospital of USTC, Professor Situ Bo, Director of the Department of Laboratory Medicine at Nanfang Hospital of Southern Medical University, and Professor Wu Anzhao, Director of the Department of Integrated Traditional Chinese and Western Medicine at Shandong Provincial Public Health Clinical Center, experts engaged in in-depth exchanges on the topic "How to achieve precise, equitable, and accessible early liver cancer screening." The three experts, from the three perspectives of clinical practice, technology, and public health, outlined a multi-dimensional vision for early liver cancer screening.

The experts unanimously agreed that early liver cancer screening should shift from "passive screening" to "active management," with the core being "early management" rather than "waiting for disease to screen." For patients with chronic liver disease, a long-term dynamic management system should be established, utilizing AI to integrate multi-dimensional information such as test indicators, liver function data, and imaging characteristics to accurately identify risk trends, combined with the Traditional Chinese Medicine concept of "preventing disease before it occurs" for intervention. Secondly, it is essential to strengthen multi-disciplinary collaboration among laboratory medicine, imaging, clinical practice, artificial intelligence, and public health, making AI a "smart assistant" for primary care physicians and building a closed-loop system of "precise screening - dynamic monitoring - whole-process management." Finally, accessibility should be enhanced through health education and primary-level empowerment, supported by intelligent tools and regional testing centers, to lower the threshold for screening and improve affordability and availability. Only through the simultaneous advancement of technological empowerment, multi-party collaboration, and policy guidance can we achieve a precise, equitable, and sustainable early liver cancer screening system.

Dunwill Medical has always been guided by clinical needs and driven by technological innovation to advance laboratory medicine. This gathering in Jinan serves both as a concentrated display of phased outcomes and as a new starting point for progressing hand-in-hand with industry peers. Looking ahead, Dunwill Medical will continue to uphold its mission of "Creating More Possibilities for Modern Medicine," empower laboratory medicine with new quality productive forces, and contribute wisdom and strength to the building of "Healthy China!