Menu Close


Perthera is honored to have been accepted to present at numerous conferences and symposiums throughout the years.

Perthera AI, Perthera’s Artificial Intelligence system, integrates the patient’s multi-omic cell architecture and treatment history with our leading-edge Computational Engine; the output is assessed and approved by our every-patient, real-time physician, and scientific tumor board.

The collective data from Perthera Precision Medicine along with our patient outcomes data provides for statistically significant findings for research and publication.

Precision Medicine for Pancreatic Cancer Patients: Preliminary Results from the Know Your Tumor (KYT) Program
  • 932 reports created from 287 high-volume practices in 44 states
    • 152 initiated treatment with report-listed therapy (Figure 2)
  • Actionable genomic mutations were found in 48% of the 932 patients
    • Molecular profiling revealed 27% of the genomic mutations were highly actionable
    • Highly actionable: Literature supports clinical evidence of high response rate in patients with that molecular abnormality in any cancer type
    • Actionable proteomic mutations found in 5% of the 932 patients
  • Patients with highly actionable alterations who received matched therapy (n=20) had a median progression-free survival (PFS) of 5.5 months
    • Significantly longer than patients with highly actionable alterations that did NOT receive matched therapy (n=23, PFS = 1.9 months)
    • PFS 5.5/1.9 = 2.9 x’s longer PFS with matched therapy
    • This effect has not yet been observed for overall survival (OS)
  • Follow-up studies are underway for OS analyses

Outcome Driven Persona-Typing for Precision Oncology: Beyond a Genomics Centered View of Individualized Therapy

  • 10 population-based “personas” were defined by our data on 1,014 pancreatic adenocarcinoma patients
    • Personas share common outcomes and identify response predictors to any given therapy
    • Personas were created by combining multi-omic data, treatment history, clinical-epidemiological data, and outcomes data
    • All personas have at least one significantly associated genomic alteration
    • Half of the personas have one significantly associated proteomic alteration
    • Overall survival (OS) was calculated for the 10 persona types
  • Integration of multi-omic and clinical data can improve predictive biomarkers of treatment response as compared to genomics alone
  • Proteomic information significantly associated with outcomes more frequently than genomic information
  • Persona-typing can be used to define and map key characteristics that associate with outcome and specific treatment
    • The future goal is to map individual patients to pre-defined personas to more accurately describe their potential outcome and determined personalized treatment options

Multi-omic Molecular Comparison of Primary Versus Metastatic Pancreatic Tumors

  • The comparison of the molecular characteristics of
    primary vs metastatic pancreatic adenocarcinoma in
    patients who have received Perthera Reports
    reveals that the molecular characteristics are very
    similar and that actionable alterations are identified
    at the same frequency.
  • Our findings support the belief that primary
    pancreatic cancers metastasize very early and thus:
  • The role of biomarker-directed therapy for early-stage
    pancreatic cancers in lieu of, or in
    addition to standard therapy could be further
    evaluated in prospective clinical trials.
  • In addition, re-biopsy at recurrence may not be
    necessary which can decrease patient
    discomfort and anxiety, cost, and delay to next
A Computational Model for Integrating
Genomic Data with Public Datasets
For Molecular Tumor Board Recommendations
  • The method presented here can be useful in summarizing the available evidence linking mutations with drug response, and in prioritizing multiple, sometimes conflicting, biomarkers:
    • Hypothesis generation for in vitro validation
    • Interpretation of the impact of molecular profiles on observed differences in drug response across patient populations
  • Computational challenges related to scale-up of this approach:
    • Parameter estimation against cell line and other experimental data2,3
    • Simulation of networks with cycles (feedback loops)4
  • As outcomes data are collected for Perthera patients, molecular profiles of patients will be simulated to suggest potential mechanism for observed response/lack of response to targeted therapies
Preliminary Observations of Blood-based Molecular Testing in a Subset of Patients
with Pancreatic Cancer Participating in the Know Your Tumor (KYT) Initiative
  • Based on this pilot study, given that BB-cfDNA testing could not identify known KRAS mutations in a small percentage of patients with disseminated and metastatic disease, we conclude that BBcfDNA testing is not ready for clinical decision making, especially in the arena of precision therapy for PDA.
  • Tumor tissue testing should still remain the gold standard.
  • Future studies evaluating BB-cfDNA platforms in PDA cohorts
    should consider: 1) side by side testing with TT and 2) the use of KRAS mutations as a gold standard biomarker for this disease.
  • Sequencing germline (constitutional gDNA not from the tumor) should be considered in an effort to define true somatic events.

Molecular and clinical characterization of
BRAF mutations in pancreatic ductal
adenocarcinomas (PDACs)

  • BRAF mutations are significantly and inversely correlated with KRAS alterations
  • The most common BRAF alteration, V600E mutation, was
    found to be mutually exclusive with the KRAS mutation
  • Clinical trials targeting BRAF alterations in KRAS wildtype
    pancreatic cancer appears warranted