Cancer Management Solutions Powered By Deep Learning Technologies

About The Company

Curatio.DL provides cancer management solutions by delivering accurate diagnosis & prognosis, allowing personalized treatment for cancer patient based on Deep Learning-powered pathology analysis. This innovative pre-processing segmentation technology will allow pathologists, and in particular, pathologists who are not necessarily experts in a specific field, to obtain expert-level decision support diagnosis and prognosis.

The company has an exclusive groundbreaking patented technology developed at Harvard University aiming to take the development of DL-powered cancer diagnostic platform (AI-CDP) to the next level of commercialization. The Technology was tested successfully on 10 types of cancer malignancies and showed high validation in diagnosing types and subtypes of cancers.

Our Vision

To provide personalized medicine for each cancer patient by delivering accurate diagnosis & prognosis, based on DL-powered technology analysis

Our platform and services

CuratioDL’s solution is to provide Accurate diagnosis and treatment matching based on Deep Learning analysis of histopathologic biopsies. We are using neural networks combined with an innovative pre-processing segmentation technology to process dozens of thousands of images of histopathology slides. Providing us with all the information needed for the decision :

Cancer subtypes
Genetic profiling
Biomarkers
Survival period estimation
Treatment suggestion

The advantages are clear

The process relies on big data – enabling analysis of much bigger areas of samples at once providing accurate and fast diagnosis and in low cost.

It can discover novel clinically relevant patterns objectively, as the system teaches itself to identify new patterns. And therefore, the procedure can be copy paste to other tumor types or diseases.

On top the features just mentioned – the system can also highlight regions of interest for the pathologists with unique coloring. As the examination samples are huge, this feature can save time for the pathologists and draw his attention to the areas which have clinical significance, by than save a lot of diagnostic errors.

On top of diagnosis and sub type classification our models were able to also provide estimation of the patients’ survival period. This has a huge impact on the disease management both for physicians and for patients.

Technology

DL Model is a new pathology diagnostic technology developed at the Harvard Medical School to analyze histopathology slide images from large dataset within a very short period of time with over 90% detection accuracy

Fully-automated DL – Histopathology slide images are fed into an image processing neural network

Fit into the category of clinical decision support software (CDSS) for FDA approval, providing Automated diagnostic classification and Survival Outcomes Prediction

To be offered as a cloud-based, Software as a Service (SaaS) solution

Automated regions of interest selection by DL - A key advantages to be able and save a lot of diagnostic errors

The Unmet Needs

Manual evaluation by expert pathologist is laborious and subjective
Large net deficit of pathologists results in increasing histopathology evaluation by general pathologists – not field-specific / disease-oriented pathologists
Tumor has great genetic heterogeneity, which can resist treatment causing low treatment response rate
“Low” detection accuracy (only around 60%) of existing products in the market
Hundreds cancer subtypes, results in more diagnostic & prognostic mistakes and wrong treatments

The Cancer Management Solution

More accurate diagnosis & prognosis

Shows regions in image of most significance to the resulting interpretations of histopathology slides
Allow general pathologists, even not experts pathologists, to obtain expert-level decision support for cancer diagnosis and treatment recommendations
Further differentiates subtypes where a diagnosis of cancer has already been determined

Personalized treatment for patient

Predicts mutation burden for more suitable selection of immuno-therapeutic treatment
Contains previously unrecognized signals indicative of disease progression and treatment response
Match between patient and his cancer subtype to the right drug

Competitive Advantages

First to develop digital pathology on kidney cancer or RCC
First to classify subtypes of lung cancer
First to demonstrated results on 10 different types of cancer with >90% accuracy using the DL Model, including kidney, lung, leukemia, brain, breast, ovarian, colon, rectal, stomach cancer and liver hepatocellular carcinoma, while competitors in the market which focus only on 1-2 types of cancer
Able to successfully test the technology on a new type of cancer with a small number of samples without the need to change algorithm, while competitors in the market need numerous samples to test their technology on a new type of cancer