Emerging Targets and Therapeutics in Immuno-Oncology Landscape: Insights from Natural Language Processing Analysis

Rapid and sustained growth in the field of immuno-oncology has resulted in expansion of available scientific literature. Gaining valuable insights and establishing deep and often hidden meaningful connections in such a large body of work is the need of the hour. In this report we summarize our findings from a novel Natural Language Programming (NLP)-based approach on a large dataset of >350K scientific publications in immuno-oncology research spanning across two decades (2000-2022) retrieved from the CAS Content Collection. Our analysis led to identification of >300 emerging concepts across major categories such as therapeutic targets, biomarkers, therapies, and types of cancer. We present a “Trend Landscape Map” of emerging concepts in immuno-oncology possessing layers of intricacies – at first glance providing information for the >300 identified concepts arranged hierarchically across 8 major categories and at a deeper level providing detailed quantitative metrics of growth over the last three years (2020-2022). While concepts such as immune checkpoint inhibitors (ICIs), antibody-drug conjugates (ADCs) and chimeric antigenic receptors (CARs) continue to be important in immuno-oncology, their growth over the last three years have been modest. On the other hand, concepts including protein targets such as TROP2, nectin-4, and gasdermins display rapid increase in scientific publications over 2020-2022 while their absolute number of publications remain low potentially indicative of early emergence. Finally, guided by our trend landscape analysis, we performed substance data analysis leveraging data from >3.2 million substances from the CAS Registry and identified potential higher commercial interest in protein/peptide sequences rather than small molecules in cancer immunotherapy as seen with respect to patent publications. It is our hope that our subject matter experts' knowledge-guided big data analysis approach based on the corpus of robustly CAS indexed data provides a comprehensive picture of immuno-oncology as it stands today with the trend landscape map serving as a valuable resource to researchers in this field.