PTRS Partnership

Patient-level association of IFN-γ pathway activation with response and survival on first-line IO therapy in NSCLC.

Trial design impact:
Made the mechanistic case for IFN-γ signaling as a stratification biomarker in a published head-to-head against standard IO. Addresses a potential "black box" objection.

Single-cell spatial transcriptomics of NSCLC pre-treatment samples to quantify checkpoint and co-receptor populations within and outside the tumor area.

Trial design impact:
Provides the critical mechanistic foundation required to accelerate asset progression and clear clinical development milestones—directly validating dual-receptor engagement biology to justify pipeline advancement, while establishing a robust, data-driven biological rationale that de-risks internal portfolio review and regulatory filings.

Uncovered the underlying biological characterization and molecular features driving the predictive utility of an immune-inflammatory blood marker across target real-world patient cohorts.

Trial design impact:
Provides the necessary mechanistic and clinical evidence to drive high-level regulatory and protocol design discussions—validating the biological mechanism of the inflammatory ratio as a robust stratification factor to control for baseline immune variability and maximize the study's overall therapeutic signal.

Profiled target expression across more than 40 cancer indications and within disease subtypes, evaluating primary-vs-metastatic differences and tumor-purity relationships.

Trial design impact
Across more than 40 indications and their subtypes, the data ranked which secondary indications carry enough target expression to support biomarker-defined Phase 1/2 expansion, and which don't.

Characterized target expression across large HER2+ and ER/PR+ postmenopausal metastatic breast cancer cohorts, stratified by pre- and post-treatment, biopsy site, stage, ECOG, and line of therapy.

Trial design impact
Cleared the partner to write second- and third-line metastatic breast cancer trials that biopsy any metastatic site except one. Target expression holds across prior-therapy lines.

Compared molecular and ctDNA profiles in matched pre- and post-treatment metastatic prostate cancer biopsies following a prior approved radioligand therapy.

Trial design impact
Liver metastases are out. Target expression downregulates after prior radioligand exposure at that site, so the next-line trial enrolls only patients without. A single exclusion criterion, derived from matched pre- and post-treatment biopsies.

Confirmed prioritized indications for a novel antibody by characterizing real-world target prevalence and co-occurring biomarkers across the solid tumor space.

Trial design impact
Validated the partner's priority indications for a novel antibody by checking real-world target prevalence and co-occurring biomarkers across the solid tumor space. Portfolio decision held.

Assessed target copy number and RNA expression in a large pre-1L advanced solid tumor cohort to support patient stratification and control-arm planning for an ADC program.

Trial design impact
Shifted ADC eligibility from copy-number to expression. Copy-number alone would have left too few patients to power the biomarker-defined cohort.

Real-world cohort analysis evaluating the prevalence of a target gene mutation within a niche breast cancer sub-population characterized by ER-negative and PR-positive status at diagnosis.

Trial design impact
Killed a label-expansion path. The ER-negative, PR-positive subgroup with the target gene mutation is too rare in the real world to support a separate filing. Capital redeployed.

Quantified single-, pairwise-, and triple-biomarker positivity for three priority targets in a multi-modal PDAC cohort stratified by stage and line of therapy.

Trial design impact:
Three concurrent PDAC programs reset their addressable-population assumptions after the data showed how often the three priority targets co-occur, in pairs, in triples, and how often they don't.

Mapped prevalence and mutual-exclusivity of multiple actionable mutations across sidedness, metastatic status, and lines of therapy in a cohort of more than 30,000 patients.

Trial design impact:
Across more than 30,000 patients, actionable mutation prevalence in MSS/PMMR CRC was the same whether the metastasis was hepatic or not. The protocol stopped excluding liver-met patients.

Combined dual-biomarker prevalence with US treatment-landscape data across six advanced and metastatic indications to size the eligible population per indication per year.

Trial design impact:
Underpinned the Phase 2 enrollment feasibility argument at an FDA Type C meeting. Six advanced and metastatic indications, sized year-by-year, with the dual-biomarker addressable population per indication.

Reconstructed published landmark IO trial eligibility in real-world data, stratified by KRAS variant and PD-L1, with analysis of TMB, neoantigen burden, and TME composition.

Trial design impact
KRAS G12C patients don't share baseline IO survival with other KRAS variants. IO and KRASi combination trials that pool them lose signal.

Analyzed PFS and OS in a PIK3CA-tested CDK4/6i-treated 1L mBC cohort to evaluate PIK3CA's value as a stratification factor in trial design.

Trial design impact
Dropped PIK3CA from the stratification spec. In this 1L HR+/HER2- mBC cohort on CDK4/6i, PIK3CA status didn't track with rwPFS or rwOS. Simpler trial, same power.

Outcomes analysis stratifying mHSPC by HRR mutational status to inform the stratification approach for a Phase 3 trial of a selective PARP inhibitor.

Trial design impact
BRCA and non-BRCA HRR mutations don't behave the same way prognostically in mHSPC. The Phase 3 PARP-inhibitor stratification spec now treats them as separate subgroups, with cleaner signal in the BRCA-defined subset.

Real-world stratification of NSCLC by HER2 expression vs. mutation status to inform stratification and patient-selection strategy.

Trial design impact
HER2-mutant and HER2-overexpressing NSCLC are not the same patients prognostically. Stratification approaches differ accordingly.

Characterized real-world unmet need in a subset of PSA-detectable prostate cancer patients and evaluated prognostic outcomes across standard treatment regimens.

Trial design impact
Opened up protocol eligibility for a Phase 3 mHSPC trial after the real-world data surfaced a sizable unmet-need subgroup in PSA-detectable patients. The refined molecular stratification spec held prognostic variability under control.

Real-world survival analysis in a 2L advanced endometrial cancer population, stratified by histologic subtype with Cox adjustment for age, race, BMI, ECOG, stage, and grade; cross-validated against an independent dataset.

Trial design impact:
Histology is now in the Phase 3 stratification spec for advanced endometrial cancer. Cross-validated against an independent dataset, with Cox adjustment for the usual baseline covariates.

Real-world demonstration that a widely-available blood-test-derived ratio is predictive of survival outcomes in a late-stage oncology setting.

Trial design impact
A widely-available blood-test-derived ratio became a primary stratification variable in a late-phase oncology protocol. Smaller sample sizes carry the same power.

Analyzed longitudinal real-world outcomes in a target female cohort to evaluate the independent prognostic impact of baseline obesity and concurrent metabolic variances on response to frontline systemic therapy.

Trial design impact
Baseline obesity and concurrent metabolic variation are independent prognostic factors in advanced endometrial cancer on frontline systemic therapy. Stratifying for them isolates the therapeutic effect that randomization alone wouldn't.

Analyzed longitudinal real-world data to evaluate the comparative prognostic impact of a routinely measured metabolic blood marker versus historical smoking status on overall survival within a small cell lung cancer cohort receiving immunotherapy.

Trial design impact
A routinely measured pre-treatment serum marker beat historical smoking status as the prognostic stratification factor for OS in SCLC on IO. Frontline protocols now stratify on it, with the elevated-marker subgroup identified and screened up front.

Built classifiers leveraging multimodal data to predict the likelihood that a patient develops a progression-associated resistance mutation at the end of first-line therapy.

Trial design impact
Compressed time-to-enrollment by enriching for patients most likely to develop the progression-associated resistance mutation at end of 1L. The classifier replaces "wait and see" selection.

Real-world analysis in an NSCLC cohort to support RNA-based MET expression as a patient-selection biomarker, with attention to PD-L1 co-expression.

Trial design impact
RNA-based MET expression is now an eligibility criterion for MET-targeted and IO combination trials in NSCLC. PD-L1 co-expression is accounted for.

Characterized treatment switching patterns in KRAS-mutant patients with advanced CRC to identify clinically relevant subpopulations for targeted-therapy enrollment.

Trial design impact
Treatment-switching patterns split KRAS-mutant CRC into clinically distinct subpopulations. Eligibility criteria now line up with real-world treatment dynamics, not protocol-only assumptions about prior-line exposure.

Applied full inclusion and exclusion criteria to live multimodal data, assessing biospecimen, NGS, H&E, and Ki67 readiness across the patient funnel.

Trial design impact:
Thousands of protocol-eligible patients, screened against the full inclusion and exclusion spec on live multimodal data with biospecimen, NGS, H&E, and Ki67 readiness baked into the funnel. Phase 3 recruitment risk priced down.

Constructed a well-characterized Stage III NSCLC cohort with linked multimodal data to enable predictor analyses for an IO and anti-angiogenic combination program.

Trial design impact:
A well-characterized Stage III NSCLC cohort with linked multimodal data, ready to feed predictor analyses for an IO and anti-angiogenic combination program. Biomarker analyses won't run underpowered.

Benchmarked OS and PFS in US patients receiving adjuvant standard-of-care therapies, using risk-set adjustment and stratification across multiple cohort definitions.

Trial design impact:
Comparator-arm expectations calibrated against actual US standard adjuvant practice. Risk-set adjusted, stratified, multi-cohort. The investigational arm now competes against the real benchmark, not an aspirational one.

Built a large HNSCC cohort anchored on surgery date, stratified by PD-L1 status and anatomical site, with visual mapping of neoadjuvant, resection, and adjuvant flows.

Trial design impact:
Set the SoC denominators the Phase 2 platform study has to clear, with anatomical site and PD-L1 stratification. Plus an IO-exposed subgroup the study can use as an external control arm.

Replicated the inclusion criteria of a published landmark Phase 3 IO trial in real-world data, and compared real-world median OS to trial-reported median OS across PD-L1 strata.

Trial design impact:
Real-world median OS, computed on the inclusion criteria of a published landmark Phase 3 IO trial, matched against the trial's reported OS across PD-L1 strata. Phase 3 sizing assumptions now rest on real-world replication.

Real-world data analysis of more than 500 patients treated with prior radioligand therapy in mCRPC, with outcomes matched against published landmark trial benchmarks.

Trial design impact:
More than 500 patients in mCRPC with prior radioligand therapy. A benchmark in a setting where there isn't supposed to be one. SoC comparator-arm assumptions about prior-therapy prevalence got rewritten.

Multiple standardized ctDNA timepoints across early treatment cycles in an IO-treated solid tumor cohort, modeled against progression-free survival.

Trial design impact:
Cut the wait for radiographic confirmation. Early ctDNA dynamics across multiple standardized timepoints predict PFS, and the interim endpoint becomes ctDNA-based.

Longitudinal real-world data validating a routinely-measured circulating-marker trajectory as an independent prognostic biomarker in late-line standard-of-care metastatic colorectal cancer.

Trial design impact:
An investigator-initiated cellular therapy trial in advanced CRC now uses a routinely measured circulating-marker trajectory as its molecular response endpoint. Faster, cheaper, and less prone to delayed efficacy signals than radiographic response.

Demonstrated correlation between IHC and RNA-seq quantification of clinical biomarkers across solid tumor indications, validated against established companion diagnostic tests where available.

Trial design impact:
RNA-seq stands in for IHC when companion diagnostic testing is unavailable or inconsistent, across solid tumor indications, validated against established CDx tests. Eligibility criteria stop being gated by which site can run which assay.

Built a modular technical platform to run Tempus pathology algorithms, including TIL quantification, directly on sponsor clinical trial H&E images.

Trial design impact:
A modular platform that drops Tempus pathology algorithms onto sponsor clinical trial H&E images, including TIL quantification. The integration friction that usually delays novel-AI-assay decisions is gone.

Established correlation between IHC and RNA-seq quantification of GPC3 expression in HCC, enabling RNA-based identification of GPC3-positive patients for biomarker-driven trial enrollment.

Trial design impact:
RNA-seq picks GPC3-positive HCC patients out of the funnel even when IHC isn't available. The site list a trial can run through expanded immediately.

Correlation analysis of FOLR1 RNA expression and FRα IHC staining across ovarian and endometrial cancer to enable RNA-based identification of FOLR1-targetable patients.

Trial design impact:
RNA-based FOLR1 calls catch the FRα-positive patients IHC misses across ovarian and endometrial cancer. The eligible pool for FOLR1-targeted therapies got bigger.