Predicting Immune-Related Adverse Events: A Key Breakthrough in Cancer Treatment

Strategic Insights -

Joanne B. Weidhaas, MD, PhD, MBM, a renowned expert in the field of cancer research, has made a groundbreaking discovery in predicting immune-related adverse events (irAEs) in patients undergoing anti-PD1 therapies. Weidhaas, a professor at the David Geffen School of Medicine at UCLA Health Jonsson Comprehensive Cancer Center, has co-founded MiraDx and MiraKind, and recently shared her team's findings at the 2025 American Society of Clinical Oncology (ASCO) Annual Meeting.

The study, which included patients with various types of non-melanoma cancers, demonstrated the ability to predict late irAE toxicity in a non-heavily pretreated cohort. The microRNA-based genetic signature showed impressive results, with a sensitivity of 0.692, specificity of 0.861, and an area under the curve (AUC) of 0.776. In contrast, baseline clinical models showed significantly lower performance, with a sensitivity of 0.538, specificity of 0.788, and AUC of 0.663.

The model predicted acute toxicity with 18 microRNA-based single-nucleotide polymorphisms (mirSNPs) versus baseline clinical models with 0.700 vs 0.533 sensitivity, respectively, 0.804 vs 0.730 specificity, 0.344 vs 0.225 positive predictive value (PPV), 0.948 vs 0.914 negative predictive value (NPV), and 0.752 vs 0.632 AUC.

Key Takeaways:

  • A microRNA-based genetic signature was developed to predict immune-related adverse event (irAE) risk in patients with non-melanoma cancers.
  • The model showed a sensitivity of 0.692, specificity of 0.861, and an area under the curve (AUC) of 0.776 in predicting late irAE toxicity.
  • Baseline clinical models showed lower performance, with a sensitivity of 0.538, specificity of 0.788, and AUC of 0.663.
  • The model predicted acute toxicity with 18 microRNA-based single-nucleotide polymorphisms (mirSNPs) versus baseline clinical models.
  • Weidhaas emphasized the importance of expanding the selection of patients to those who hadn't received prior treatment and identifying patient biomarkers that predicted irAEs.
  • The study included over 250 patients and investigated the timing of toxicity, specifically irAEs.

Statistics:

  • Sensitivity: 0.692 (microRNA-based model) vs 0.538 (baseline clinical model)
  • Specificity: 0.861 (microRNA-based model) vs 0.788 (baseline clinical model)
  • Area under the curve (AUC): 0.776 (microRNA-based model) vs 0.663 (baseline clinical model)
  • Positive predictive value (PPV): 0.344 (microRNA-based model) vs 0.225 (baseline clinical model)
  • Negative predictive value (NPV): 0.948 (microRNA-based model) vs 0.914 (baseline clinical model)
  • AUC: 0.752 (microRNA-based model) vs 0.632 (baseline clinical model)

Sources:

1. Weidhaas J, Marco N, Scheffler AW, et al. Germline biomarkers predict toxicity to anti-PD1/PDL1 checkpoint therapy. J Immunother Cancer. 2022;10(2):e003625. doi:10.1136/jitc-2021-003625

2. Weidhaas JB, McGreevy K, Drakaki A, et al. MicroRNA-based signatures of early and late immune-related adverse events to anti-PD1 treatment. J Clin Oncol. 2025;43(suppl 16):2661. Doi:10.1200/JCO.2025.43.16_suppl.2661