Radiomics
Gaining the advantage in patient diagnosis and treatment using Radiomics.
Radiomics is a method that extracts numerous quantitative features from medical images that are unrecognizable to the naked eye.
This approach enables the creation of new imaging biomarkers and access to additional information at a microstructural level.
Why radiomics
- Provides new imaging biomarkers for very precise and detailed exploration of pathologies.
- Minimizes interobserver variability by leveraging AI-enabled segmentation and analysis.
- Expounds medical images to numbers which helps clinicians to see the lesions in multiple dimensions and compound them with structural and textural characteristics.
- Enables more accurate treatment planning by exposing data and patterns within a lesion that were hidden beforehand.
- Better patient selection and stratification on the very early stage in drug development trajectory.
- Data derived from images combined with texture analysis fuels diagnosis, therapy and prognosis.
- Delivers value to development of personalized medicine – the combination of imaging and phenotypic data may reveal the key to facilitating better clinical decision making and treatment.
- Provides early indicators of treatment effectiveness, which makes it a great aid in patient response assessment and clinical trials..
How we work on radiomics based on MRI/CT
- Image data acquisition and reconstruction – for reliable results outcomes, important is homogenize data during the acquisition or pre-processing step. Data preparation should be not only pixel or voxel oriented, but we also need to check that the data is unbiased to avoid modelling problems.
- Segmentation – we develop precise and automated segmentation algorithms. We fully understand that this step is the most critical component of radiomics – following feature data and quantification are generated from the segmented ROI (region of interest).
- Features extraction and selection – we develop radiomic features in accordance with the IBSI – Image Biomarker Standards Initiative. We work closely with clinicians worldwide to develop additional handcrafted features specific to various fields of specialty, particular organs and diseases. Our tailored made data-characterisation algorithms analyse images with selected biomarkers. Numbers are huge: 5467 features in total were extracted in the liver cirrhosis project.
- Data analysis – on this step we build a model. To build a model, we use the extracted, selected and grouped features. Modelling allows us to respond to the asked question – for example to determine which patients are likely to benefit from therapy or to predict the risk of developing recurrence of a disease.
- Validation – in medical imaging analysis a great deal depends on the radiologist, but on the other hand: clinical variability is an issue, that needs to be addressed when it comes to proper validation. We have successfully conducted numerous validations compliant with ISO 13485 and CE/FDA regulations.
- Report and interpretation – this is the last phase of the process of our work related to radiomics. Here, a report is generated and the data provided by the algorithm is interpreted.
Why us
- We are able to prepare properly data for learning. We can also verify whether the data is unbiased. We are aware of that the model results depend on data quality.
- We involve medical professionals in the radiomic project – if it is necessary. Thanks to their involvement, we are able to create additional functions adapted to various specialties, organs and diseases.
- Numerous validations in accordance with ISO 13485 and CE/FDA regulations have been completed with success by our company – taking into account of clinical variability.
- We have experience in the extraction, selection and grouping of radiomic features and data analysis in a radiomic project. Thanks to prepared functions, we create a model answering the question asked.
- We create automatic, advanced and accurate machine learning algorithms – as a basic element of the radiomic project. According to your current need, we are able to design an algorithm for your specific needs.
- We provide specialized algorithms for data characterization that analyse medical images using particular biomarkers. With the help of the algorithms we’ve created, we can extract a plethora of characteristics from medical photos.
Learn more about our competences
Radiomics: a great potential to impact the future landscape of digital healthcare.
The road to perfection in medical image segmentation.
Medical image segmentation as an advancement in medical imaging.
