AI&BM for Leukemia

The “AI&BM for Leukemia” project – Artificial Intelligence and Biophotonic Microscopy for rapid and early tests on leukaemia cells

A new biophotonic microscopy-based technology to accelerate and simplify the diagnosis and monitoring of patients with leukaemia, with significant time- and cost-savings to benefit individual health and the operation of the healthcare system

The challenge

Many of the most recent advances in biomedical research are associated with the identification of new methods for studying cells. The combined application of modern technologies results in considerable reductions in the time required to analyse and classify certain cells, thereby revolutionising the systems used to diagnose a number of diseases. The “AI&BM for Leukemia” project – proposed by the University of Ferrara and Plastic Jumper and funded by the IFAB Foundation – intends to develop a new system for the early diagnosis of leukaemia and its monitoring in the therapy and post-therapy phases. The new system aims to overcome some of the limits of current diagnostic practices based on specialised laboratory testing, in order to simplify and expedite analysis and study the effectiveness of treatments in a non-invasive and highly accurate manner.

The solution

As its name implies, the “AI&BM for Leukemia” project combines two techniques: hyperspectral biophotonic microscopy (BM) and artificial intelligence (AI) on microscopically-obtained images.

Biophotonic microscopy is a technology that exploits the interaction between light and biological material in order to study its characteristics. This technology makes it possible to view cell characteristics without having to apply stains that could cause changes in the biological sample, by generating digital images representative of the light reflection spectrum emitted by biological material.

The aim of the project is to use biophotonic microscopy on a blood cell line of human acute promyelocytic leukaemia (APL) (HL60), for which there is already a therapy that does not require use of chemotherapy drugs. It will thus be possible to reproduce in vitro on these cells the therapeutic process (obtained by differentiation with substances such as retinoic acid) and monitor the evolution of the blood cells at various stages of treatment.

By applying quantitative analyses using Artificial Intelligence techniques, the digital images acquired at the various stages can be used to obtain information on cell membrane composition.

Again using Artificial Intelligence, it will then be possible to automate the identification of these characteristics that correlate with the pathological status of a cell at specific therapy timepoints. Essentially, the system will be able to profile the health/disease of the blood cells and response to therapy, thereby greatly increasing the speed and accuracy not only of diagnosis, but also and above all of monitoring during therapy (response to treatment) and after recovery (relapse).

The Project consists of three phases:

Phase 1 – The data acquired by biophotonic imaging are compared with conventional staining techniques to map how the cells react to light at the various stages of differentiation as a result of the pharmacological treatment.
Phase 2 – The sample acquired is profiled using algorithms that recognise certain imaging characteristics (shape, pixel clusters, etc.), to provide an analysis of the characteristics of the blood cells.
Phase 3 – The statistical analyses acquired are also compared with the reference literature.

Benefits

Having completed the experimental phase, the project aspires to develop a new diagnostic technology, based on biophotonic microscopy and Artificial Intelligence techniques, that can be applied to ex vivo blood cells, therefore taken directly from patients, such as, for example, a single drop of blood sampled by fingerstick. This would be a very quick and simple test, that does not generate waste and that can be carried out even in non-hospital settings (e.g. pharmacies), and that therefore helps to promote near-patient medicine. Such a rapid and inexpensive new type of test would bring undeniable advantages for both patients and the community, by permitting faster access to treatment and avoiding the need for other more invasive tests and longer than necessary precautionary therapy for patients who are already cured.

Partners

University of Ferrara
Plastic Jumper Srl