Predict
Lung Prediction Panel
Predict genomic alterations in NSCLC samples
in minutes with an AI-powered assay
Lung Prediction Panel is an AI-powered research-use assay built on the CanvOI foundation model. It predicts the presence of eight genomic alteration biomarkers in NSCLC from H&E whole-slide images and provides results in minutes. The RUO panel includes EGFR, KRAS G12C, ERBB2, and BRAF V600E mutations; ALK, RET, and ROS1 fusions; and MET exon 14 skipping.
Results can support sample prioritization across a variety of research use applications including retrospective studies, tissue stewardship, and biomarker analysis.
This RUO product is not for clinical use.
How It Works
Upload WSI and perform QC
Following an upload, a quality control step ensures tissue integrity and quantity of analyzable tissue.
Rapid multi-gene AI analysis
Rapid analysis across all eight biomarkers in a single H&E-based workflow.
Exploratory biomarker data review
Review visual patterns, histological feature summaries, and heatmaps.
Research Evaluation
Lung Prediction Panel has been evaluated in internal and external retrospective research studies, including academic and biopharma collaborations.
Download the datafile to learn more about our internal nested cross-validation test results. Contact
Imagene to learn more about our completed and ongoing research programs.
Download the datafile
Transparent & Explainable Data
In addition to summarizing prediction results, the platform provides a heatmap for each biomarker, helping researchers explore morphological patterns associated with the predicted detection of the biomarker. For added explainability, results include ranked fields of view highlighting regions associated with predictions of the presence or absence of a biomarker.
Schedule a demo to explore the supporting visualizations and see how they enable a more informed review of biomarker data.
Applications
Preserve Tissue Through Smarter Triage
To maximize the use of NSCLC samples for research use, Lung Prediction Panel helps identify samples that may warrant deeper molecular analysis, supporting more focused use of tissue and testing resources.
- Prioritize samples for deeper molecular profiling to confirm the presence of the biomarker
- Focus resources on prioritized samples
- Preserve limited tissue for downstream analyses
- Streamline sample review
Analyze
Analyze with the Lung Prediction Panel
Prioritize
Advance high priority samples to molecular confirmation
Confirm
Confirm presence of biomarker
Generate Hypotheses from Real-World Data
Apply H&E-based predictions to historical slide collections to estimate predicted alteration prevalence and conduct hypothesis-generating analyses of treatment and outcome associations when molecular data are limited or incomplete.
- Estimate predicted alteration prevalence
- Exploratory treatment and outcome analysis
- Unlock research value from historical slide archives
collections
algorithm
MutationStatus
biomarker profile
Status
Associations
Advance Retrospective Biomarker Analysis
Combine Lung Prediction Panel outputs with molecular, clinical, and outcomes data to investigate morphology-linked genomic patterns and generate biomarker hypotheses for future validation.
- Explore morphology-linked genomic patterns
- Investigate spatial features associated with outcomes
- Generate hypotheses for future biomarker development
Tissue Morphology (H&E)
Genomic pattern
| S1 | S2 | S3 | S4 | S5 | |
|---|---|---|---|---|---|
| EGFR | |||||
| ALK | |||||
| RET | |||||
| ROS1 | |||||
| BRAF | |||||
| KRAS |
Used Across Global Research
and Biopharma Workflows
Across more than 20 global sites, Lung Prediction Panel has supported more than 4,300 panel analyses in academic, biopharma, and CRO research workflows.
20+
Sites implemented globally
4,300+
Test runs completed
+6
New sites by year-end
In collaboration with leading pharma partners
Available through the OI Predict Platform designed for scalable deployment, or through seamless integration with supported third-party image management systems(IMS). Imagene AI is ISO 27001 and 27799 certified to ensure research data security.
FAQs
How has Lung Prediction Panel been evaluated?
Lung Prediction Panel has been assessed in internal and external research settings in academia and biopharma. Select classifiers early in development were additionally evaluated in npj Precision Oncology. Contact Imagene to review available study details.
Does Imagene AI develop custom models for other use cases?
Yes. Through OI Profile, Imagene AI develops custom models for research applications across multiple tissue types. The breadth and depth of the CanvOI foundation model support algorithm development for diverse research applications. Contact Imagene AI to discuss a pilot study.
Can Lung Prediction Panel be used for clinical diagnostics?
No. Lung Prediction Panel is For Research Use Only. Not for use in diagnostic procedures. Results must not be used to diagnose disease, guide treatment, make patient-management decisions, or determine clinical-trial eligibility.
For information about Imagene’s separate clinical offerings, contact us or visit the LungOI webpage.
Why Imagene AI?
Imagene AI develops AI-powered pathology tools for research and, separately, supports clinical laboratory offerings through a CAP-accredited, CLIA-certified laboratory. Lung Prediction Panel is not part of those clinical offerings and remains For Research Use Only.
For Research Use Only. Not for use in diagnostic procedures. Lung Prediction Panel results are intended for research sample prioritization and are not intended for patient diagnosis, treatment decisions, or determination of clinical-trial eligibility.