+
PLATFORM
/
We are much more than a product portfolio. Our technology platform is the engine that drives the development, integration and continuous improvement of every QUANTUS solution: from initial clinical research to deployment in real hospital environments.
&
How to integrate the solutions?
>
Our solutions integrate easily into hospital environments without impacting other existing systems
technical_dossier.pdf
Custom solution development
Our team can carry out new research proposed by the client to cover new use cases where AI provides differential value
On-premise installations
- The standard reports of our products can be customized to the client’s preference.
- A product’s positivity detection thresholds can be adapted to the specific administrative and population needs of the service, while maintaining the quality and regulatory safety inherent to our products.
+
Research platform
/
Every QUANTUS product is born from a rigorous research process carried out in collaboration with leading clinical centers. From use-case design to certification as a medical device, we follow a proven methodology that guarantees clinical validity and safety in every solution.
Use case study
- Applicability study within the current protocol
- Ethics committee (CEIC) approval and collaboration agreements
- Identification of the gold standard (clinical observation, clinical trial, etc.)
Dataset creation
- Definition of the image acquisition method
- Definition and assignment of the clinical outcome to studies/images
- Delineation of regions of interest by clinical experts
- Traceability and comparison of assignments across different experts
Research – Proof of concept
- Definition of the sample size (N) required for the proof of concept
- Running experiments to achieve an acceptable model
Research – Final model development
- Incorporation of all studies/images into the experiment
- Running experiments to achieve the final model
- Comparison of validation results with clinical support for evaluation
Validation of results with independent data
- Definition of the sample size (N) required for validation
- Selection of new images (prospective or retrospective) and assignment of the clinical outcome.
- Testing of the resulting model with the images and evaluation together with clinical experts.
The process repeats as many times as necessary until the desired goal is achieved.
If the chosen model does not perform adequately with an external dataset, the process returns to the research step to reinforce the model.
Support tools
GUI/BIOBANK
Allows organizing image datasets, reviewing them, delineating regions of interest, labeling images for binary or multi-class classification, and assigning variables for regression research.
PROTO
Web-based platform that enables the development and validation of AI models without the need for local software installation.
#
AI Training
We provide specialized training in medical imaging AI. We train professionals in AI applied to medical imaging so they can use our systems and tools as efficiently as possible.
;
