Global landscape and trends of antibody-drug conjugates in oncology: A machine learning–driven decade-long informatics analysis.

Abstract

e15038 Background: Antibody-Drug Conjugates (ADCs) are reshaping the treatment standard for a wide range of tumors. However, with the explosive growth of research in this field, it is difficult for researchers to fully grasp the current scientific landscape and future trends. Therefore, this study aims to provide a panoramic view of the global scientific landscape of ADCs in oncology over the past decade. Methods: This study retrospectively collected all relevant studies regarding the application of ADCs in oncology in the last decade. Through unsupervised clustering, all the studies were classified into five research clusters. Further, the emerging research cluster and the most impactful research cluster in the whole field were identified through time series analysis and impact analysis. Finally, through Walktrap algorithm, this study identified research directions that are highly relevant to the field but not yet deeply explored. Results: Quantitative statistical analysis revealed that the application of ADCs in oncology has a favorable development trend (Annual Growth Rate = 18.57%) and global collaboration (International Co-authorship = 28.32%) in the last decade. Through unsupervised clustering algorithm and semantic integration analysis, the application of ADCs in oncology was clustered into five research clusters: Cluster 1 (Basic and Translational Research of ADCs); Cluster 2 (Clinical Trials of ADCs in Hematologic Tumors); Cluster 3 (Clinical Trials of ADCs in Her2+ Breast Cancer); Cluster 4 (Clinical Trials of ADCs in Multiple Solid Tumors); Cluster 5 ( Balance of Clinical Efficacy and Safety of ADCs and its Biomarker Discovery). Further time-series analysis and impact analysis revealed that among these five clusters, Cluster 4 was the emerging cluster (Temporal Central Tendency = 2022.87, Research Effort Dispersion = 0.96) while Cluster 3 was the most impactful cluster (Average Citation = 28.97±26.16). More importantly, the Walktrap algorithm further revealed that “Breast Cancer, Her2, T-DM1, Prognosis” (Relevance Percentage [RP] = 90.9%, Development Percentage [DP] = 36.4%), “Antibody-Drug Conjugates, Cancer, Toxicity, Tumor Microenvironment” (RP = 81.8%, DP = 63.6%), “Brentuximab Vedotin, Hodgkin Lymphoma, CD30, Relapse” (RP = 72.7%, DP = 45.5%) are highly relevant to the field but still underdeveloped, suggesting that they hold noteworthy future research prospects. Conclusions: In conclusion, this study objectively and comprehensively presents the global scientific landscape, concerns, development trends, and future research directions of ADCs in oncology through machine learning-driven informatics analysis methods, providing an important reference for subsequent clinical trial design, decision-making, and in-depth research in this field.