Precision Medicine
in Oncology.

50% of patients are missing the best therapy due to insufficient molecular testing*.
We have a game changer !

The issue

Precision medicine is evolving very fast, 40% of drug approvals in oncology of the 21st century are for biomarker-defined populations, however :

In NSCLC, 50% of patients are missing the best therapy due to insufficient molecular testing (Gondos et al., 2023).

Only 7% of patients do enroll in clinical trials while 70% are willing to, and 83% of oncologists agree that it would have been beneficial for patients to be enrolled

Access to molecular testing and precision medicine is highly unequal both within and across developed countries due to high cost and infrastructure limitations (Mateo et al., 2023).

Emerging trends in drug development

Source

The number of products under development in oncology has grown significantly over the last decade, with more than 2,000 products currently under development.

The demand for precisely characterized populations is increasing, mostly based on molecular biology techniques.

Oncology research and development has seen an increasing focus on targeted drugs, with innovative mechanisms of action for treatment of cancers over the last decade.

A digital pathology algorithm for early tumor characterization by virtual molecular biology.

Can predict:
+1’000 DNA and RNA alterations
+hotspots variants in
+21 cancer types

Selection of patients most at risk for emergency molecular testing: reduction of time.

Selection of the right molecular tests: division by 2 of the number of tests by removing unnecessary tests.

Enlargement of the recruitment pool for the clinical trials: Identification of the most relevant clinical studies on D0 for patients based on its eligibility criteria

Distributed & Targeted
Recruitment Solution

CROs register their study into the Ummon Crawler CRO interface

The study is added to our Clinical Trial Matching System and dispatched across our partner labs

Ummon Crawler will predict molecular biomarkers, then hits are generated when patient characteristics match your clinical study.

A report is sent to the clinician, patient consent is then collected and clinician contact is sent to the CRO

Distributed & Targeted
Recruitment Solution

CROs register their study into the Ummon Crawler CRO interface

The study is added to our Clinical Trial Matching System and dispatched across our partner labs

Ummon Crawler will predict molecular biomarkers, then hits are generated when patient characteristics match your clinical study.

A report is sent to the clinician, patient consent is then collected and clinician contact is sent to the CRO

A Predictive Patient Matching System

The tissue sample of the patient is sent to the pathologist.

chara-prAIdict predicts molecular biology from the tissue and Ummon Crawler matches with a large dataset of clinical trials.

A report with best clinical
trial opportunities is sent to
the clinician.

The clinician with the patient
decide if they accept further
characterization to fill
eligibility criteria, leading to
gold standard healthcare.

The technology

Revolutionary Dual-Model Analysis

chara-prAIdict features an innovative scoring system blending cutting-edge deep learning with contextual molecular pathways, ensuring performances and fine-grained predictions for high precision diagnostics (Morel et al., 2023).

Advanced Calibration for Consistency

Our unique calibration module overcomes interlaboratory and interscanner variations. Using a calibration process aligned with our extensive databases, we ensure consistent, accurate results, enhancing reliability (Dumas et al., 2022).

Intuitive User Experience

chara-prAIdict’s ergonomic design allows even first-time users to confidently manage the system, enabling quick and efficient analysis without compromising accuracy. Ideal for seamless integration into medical professionals’ diagnostic workflow.

History of our technology

2018 - Coudray et al.

Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning.
Nat. Med.

2020 - Fu et al.

Pan-cancer computational histopathology reveals mutations, tumor composition and prognosis.
nat. Cancer

2021 - Howard et al.

The impact of site-specific digital histology signatures on deep learning model accuracy and bias.
nat. commun

2021 - Van der Laak et al.

Deep learning in histopathology: the path to the clinic.
Nat. Med.

2024 - Valderrama et al.

Breast-NEOprAIdict: a deep learning solution for predicting pathological complete response on biopsies of breast cancer patients treated with neoadjuvant chemotherapy.
In review

2023 - Filiot et al.

Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling.
MedRxiv

2023 - Morel et al.

Preliminary evaluation of deep learning for first-line diagnostic prediction of tumor mutational status.
sci. rep.

2022 - Dumas et al.

Inter-Semantic Domain Adversarial in Histopathological Images.
ArXiv

2024 - Morel et al.

MultiVarNet: A Deep Learning and Label Engineering Approach for Predicting Tumour Mutational Status at the Protein Level.
MICCAI 2024

2018 - Coudray et al.

Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning.
Nat. Med.

2020 - Fu et al.

Pan-cancer computational histopathology reveals mutations, tumor composition and prognosis.
nat. Cancer

2021 - Howard et al.

The impact of site-specific digital histology signatures on deep learning model accuracy and bias.
nat. commun

2021 - Van der Laak et al.

Deep learning in histopathology: the path to the clinic.
Nat. Med.

2022 - Dumas et al.

Inter-Semantic Domain Adversarial in Histopathological Images.
ArXiv

2023 - Morel et al.

Preliminary evaluation of deep learning for first-line diagnostic prediction of tumor mutational status.
sci. rep.

2023 - Filiot et al.

Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling.
MedRxiv

2024 - Valderrama et al.

Breast-NEOprAIdict: a deep learning solution for predicting pathological complete response on biopsies of breast cancer patients treated with neoadjuvant chemotherapy.
In review

2024 - Morel et al.

MultiVarNet: A Deep Learning and Label Engineering Approach for Predicting Tumour Mutational Status at the Protein Level.
MICCAI 2024

A unique positioning

For healthcare

FREE: Unlike most digital pathology solutions (~25€/click on average), slowing down their deployment.

ADDITIONAL PATIENT TESTING: Increase in the number of useful tests with the indication of rare panels currently not carried out and the molecular testing by manufacturers wishing to recruit patients.

OPTIMIZE: social security costs.

With PCR: we save €152/patient with NSCLCs.
With NGS: Our solution makes it possible to screen twice as many people for the same number of tests (Morel et al., 2023).

For Clinical Research

Saves more than €300/patient in their screening strategy when recruiting patients, leading to a wider variety of clinical trials

Recruits from a much larger pool of patients: makes innovation in drug development faster, with more new treatments being developed

Game changing moment

TECHNOLOGY VALIDATED

Our AI pipelines have been validated in high quality peer reviewed journals and beta testing.

CONSORTIUM CONSTRUCTED

A consortium with 20 pathology labs, 5 partners, both in private and public sectors with European counterparts, has been created.

TAILORED PROPOSAL

We look for strategic partners that will help us tailor our process to enhance their recruitments.

The first step will be to work with our partnered labs, define the process implement it by hand,
then plug-in our solution for your offers.

Ummon Crawler

Precision Medicine
in Oncology.

The issue

Emerging trends in drug development

Distributed & Targeted
Recruitment Solution

Distributed & Targeted
Recruitment Solution

A Predictive Patient Matching System