Biomedical Technology · global
From Drug Discovery to Patient Selection: LG Chem Extends AI into Cancer Clinical Trials
LG Chem plans to use tumor biomarkers to help select trial participants and has set 2029 as its target for entering clinical development with personalized mRNA cancer vaccines. The challenge is whether its models can move from identifying tumor characteristics to reliably predicting treatment response.
Whether an anticancer drug demonstrates efficacy also depends on whether a trial can identify patients who may benefit. LG Chem is extending its use of artificial intelligence from discovering drug candidates to analyzing tumor characteristics and identifying predictors of efficacy, hoping to provide a more detailed biological basis for clinical trial enrollment and treatment strategies.
The company announced this direction on October 21, 2026. The previous day, LG Chem, LG AI Research, and Flagship Pioneering jointly held the inaugural LG AI Bio Conference at LG Science Park in Seoul, attended by about 140 industry participants. The announcement focused on research and development plans in precision oncology; it was not a presentation of results showing that AI-based patient selection had improved clinical outcomes.
Specifically, LG Chem intends to use AI to analyze the complex biological characteristics of tumors, identify biomarkers that can predict treatment response, and use them to help select suitable patients for clinical trials. The question this strategy seeks to answer is: Which measurable tumor signals can indicate how likely a particular patient is to respond to a specific drug? The company also plans to use these analyses to support decisions on treatment strategies.
Technologies presented at the conference illustrated different scales of analysis. The Korea Herald reported that Inocras shared its approach to selecting patients using cancer whole-genome data, while Portrai introduced how spatial biology can be combined with AI to identify biomarkers, incorporating the distribution and interactions of cells within tissue into the analysis. However, these were technology presentations by participating companies and do not establish that the technologies have become clinically validated patient selection tools for LG Chem.
Personalized mRNA cancer vaccines are another application. LG Chem is refining AI models to predict neoantigens unique to each patient's tumor as candidate targets for eliciting an anticancer immune response. ChosunBiz also reported on this approach to identifying targets based on individual tumor characteristics. According to the company announcement and The Korea Herald, preclinical studies are currently planned for 2027, with a target of entering clinical development in 2029. This remains a research and development timeline and does not mean that authorization for human trials has been obtained.
Currently available public information does not specify the size of the models' training datasets, the cancer types involved, or their performance in independent validation, nor does it provide trial outcomes following AI-based patient selection. From the perspective of research and development validation, identifying signals associated with efficacy must be followed by confirmation that they can consistently predict responses across different patient populations. For vaccines, whether model-selected neoantigens can elicit an effective immune response likewise requires experimental and human studies to answer. LG Chem has laid out a path for bringing AI into clinical development, but the validation data supporting that path have yet to be released.