Licensing Opportunity: Explainable Breast-Cancer Prognosis with Coherent Voting Networks

After breast-cancer surgery, clinicians select an adjuvant strategy - hormonal therapy, chemotherapy, radiotherapy or a combination - using established protocols. Yet the expected outcome remains uncertain for the individual patient. Coherent Voting Networks (CVNs) provide an additional patient-level prognostic assessment of the selected regimen, complementing rather than replacing standard-of-care guidelines.

CVNs are a patented, explainable machine-learning technology developed at the Italian National Research Council (CNR). They transform molecular and clinical measurements into communities linking patients with prognostic biomarkers. Each community contributes evidence, providing a traceable patient-specific rationale rather than only a black-box score.

The lead licensing opportunity is five-year breast-cancer prognosis conditional on the selected post-surgical therapy profile. In a peer-reviewed study, CVNs were developed and tested using the approximately 2,000-patient METABRIC cohort and assessed on independent Gene Expression Omnibus cohorts totalling 601 patients. The method generated therapy-specific fingerprints of only 5-17 genes, approximately ten on average, compared with the 70-gene MammaPrint signature. Reported test AUCs ranged from 0.63 to 0.87 across treatment groups.

We propose prioritising the chemotherapy-based fingerprints, with triple-negative breast cancer as a high-need population for investigation. An exploratory TNBC subgroup produced encouraging results, but its small size requires confirmation. Development should therefore freeze the selected fingerprint and thresholds and validate them in larger, independent multicentre cohorts treated under contemporary protocols.

Existing evidence is retrospective and derives from microarray and RNA-sequencing data. Compact panels make transfer to lower-cost RT-qPCR or targeted RNA sequencing a credible hypothesis, but it remains untested experimentally. Value-creating milestones are assay transfer, analytical validation, locked-model multicentre validation, comparison with clinicopathological assessment and prospective clinical-utility evaluation.

The intellectual-property position is ready for technology transfer. CNR owns Italian patent IT201900019571B1 and European patent EP3813071B1, granted in August 2024 and covering breast-cancer survival prediction from gene-expression data using the CVN method. Patent protection is available across 38 European jurisdictions, including the United Kingdom, Switzerland, Germany, France, Italy and Spain. The technology is available for direct licensing from CNR, whose technology-transfer representative will participate in MIT4LS meetings.

Additional research demonstrates platform potential. In localized prostate cancer, the published fp160 fingerprint combines primary Gleason score, tumour stage and PSA with expression levels of two proteins, CDKN1B and NF2. Its best reported AUC was 0.87 for five-year disease-free survival across discovery and independent datasets. It is outside the current breast-cancer patent scope.

Further confidential research supports extension of the platform to haematological malignancies. Results can be discussed with qualified partners under appropriate confidentiality arrangements after CNR has assessed potential intellectual-property protection.

The commercial category is established: breast-cancer genomic prognostic tests have served more than one million patients, while Decipher and Prolaris demonstrate demand in prostate cancer. CVNs combine compact fingerprints, clinical and molecular information, explainable predictions, cross-indication potential and granted European IP.

We seek a molecular-diagnostics, laboratory-medicine or precision-oncology company to license the breast-cancer technology directly from CNR and lead its translation. The partner would provide cohorts, assay-development capability, IVDR expertise and commercial infrastructure. The inventor is available to transfer the methodology and software know-how and provide continuing scientific consultancy.

The opportunity is to transform protected technology and promising retrospective evidence into a clinically trusted decision-support product that gives clinicians an additional, individualised check on the expected outcome of the adjuvant strategy selected for each patient.