CLINICAL PERFORMANCE
Outperforming TMB and PD-L1 on what matters most
Negative predictive value, the ability to identify non-responders, is the most clinically important metric when a therapy is this costly and carries this much risk. IFP substantially outperforms the incumbents.
GMDx Group Ltd • Patent Portfolio Overview
AUC 0.86
versus TMB 0.71 and PD-L1 0.60 to 0.66
AID and APOBEC3G
the deaminase signatures identified as the primary differentiators between responders and non-responders across the validation studies.
Negative predictive value by cohort. Published exomes: combined Fang, Frigola, Hellmann, Miao and Rizvi (n=228; p<0.001; HR 0.25). AID and APOBEC3G were the primary deaminase signatures distinguishing responders from non-responders.
MULTI-CENTRE VALIDATION
Three independent response studies, plus leading centres
MOST ROBUST
Published exomes
n=228 validation · NPV 87% · HR 0.25 · p<0.001. Primary signatures AID and APOBEC3G.
LARGER COHORT
Tempus xT panel
n=263 validation · NPV 81% · independent of TMB and PD-L1. Primary signature AID.
BEST NPV
Tempus exomes
n=30 validation · NPV 91% · HR 0.33. Best-in-cohort; larger study planned.
COLLABORATING CENTRES
Alfred Hospital
University of Kansas Medical Center
Foundational performance: Lindley 2016 (Cancer Medicine), 95% sensitivity and 90% specificity in HGSOC. Supporting data: Mamrot 2021 (Oncotarget); Lindley 2025 (IJMS).
Monash University
Melanoma Research Victoria
CGFL · France
City of Hope · USA
Garvan Institute of Medical Research
Gustave Roussy · France
Omico
Samsung Medical Center · South Korea