Assessing clinical utility of the Immune Profile Score (IPS) in a real-world cohort of patients receiving immune checkpoint inhibitors at a community cancer center

Background
Predictive biomarkers for immune checkpoint inhibitor (ICI) benefit remain imperfect in routine clinical practice. The Tempus Immune Profile Score (IPS) is a multi-modal biomarker using DNA/RNA sequencing to produce a continuous 0–100 score, stratifying patients as IPS-High (IPS-H) or IPS-Low (IPS-L) for expected ICI benefit. IPS has demonstrated improved risk stratification beyond single biomarkers such as PD-L1 or MSI in retrospective analyses. Here, we validated IPS in an internal, pan-cancer cohort.

Methods
We retrospectively identified 161 patients who initiated ICIs between 2020–2025 and had prior Tempus tumor profiling. IPS was calculated and patients stratified into IPS-high (n = 59) and IPS-low (n = 102). Demographics, tumor histology, PD-L1 status, TMB, MSI, metastatic disease, and cancer history were extracted. Overall survival (OS) and progression-free survival (PFS) were estimated by Kaplan–Meier and evaluated with Cox proportional hazards adjusted for age, sex, tumor type, PD-L1, TMB, MSI, and stage. Each ICI course was counted separately for PFS (69 IPS-high vs. 110 IPS-low cases).

Results
Median age was 69 vs. 68 years (p = 0.5). Primary sites differed, with lung (34%) and skin (15%) comprising about half of the cohort across 20 disease types; 19% had >1 diagnosis. IPS-high tumors were more frequently TMB-high (41% vs. 12%; p < 0.001). PD-L1 positivity (67% vs. 59%; p = 0.40), MSI-high (4.5% vs. 5.5%, p > 0.9), and metastatic disease (81% vs. 82%, p = 0.9) were similar. At median follow-up of 22.1 months (OS) and 13.7 months (PFS), IPS-high patients had superior OS (median 34.2 vs. 12.8 mo; HR = 2.12; 95% CI: 1.27–3.54; p = 0.004) (figure 1) and PFS (median 9.3 vs. 5.3 mo; HR = 2.04; 95% CI: 1.41–2.97; p < 0.001). Multivariate analysis confirmed IPS as an independent predictor of OS (p = 0.03) and PFS (p = 0.003). IPS-low status identified lack of benefit among patients with TMB-high or PD-L1–positive tumors. No OS benefit was observed for IPS-high tumors in PD-L1–negative or TMB-low subgroups; however, IPS-high was associated with improved PFS in TMB-low patients (HR 0.64; 95% CI: 0.42–0.99, p = 0.05) (figure 2). Exploratory findings were limited by small subgroup sizes (n = 11).

Conclusions
In this heterogeneous, real-world cohort, IPS stratifies patients for ICI benefit, outperforming PD-L1 or TMB alone. Prospective integration of IPS into clinical decision-making could refine patient selection and improve efficiency across tumor types. Additional studies are needed to further validate this multi-modal biomarker.

Overall survival by immune profile score

Overall survival by immune profile score and tumor mutational burden