AI is an important driver of innovation in haematology

Clinical information in haematology, such as real-world data, is often unstructured, noisy, predominantly longitudinal, and multimodal. Prof. Federico Alvarez (Universidad Politécnica de Madrid, Spain) explained that, despite these challenges, artificial intelligence (AI) can generate data of substantial value in haematology, for example, by accelerating clinical research [1].

AI refers to technologies that emulate human or beyond-human intelligence. Machine learning (ML) is a subset of AI that uses data to train models capable of making predictions or generating insights. A particularly transformative branch of ML, often compared in impact to electricity, is generative AI (GenAI). GenAI encompasses deep learning models that can learn from raw data to generate statistically plausible outputs when prompted. At a high level, these models encode a simplified representation of their training data and use this to produce new, similar (but not identical) outputs. In other words, they can autonomously generate content.

Generative models have long been used in statistics to analyse numerical data; however, advances in deep learning now enable the integration of complex data types such as images, speech, and multimodal clinical data, according to Prof. Alvarez. Applications such as AI-assisted radiology reporting and clinical decision support systems exemplify this capability.

GenAI has the potential to accelerate clinical trials across multiple stages, including trial design, data sharing, data augmentation, bias mitigation, model validation, digital twin development, single-arm controlled trial optimisation, health technology assessment, and regulatory processes. Prof. Alvarez highlighted that GenAI can support the setup of clinical trials by leveraging historical data, thereby reducing design time. It can also assist in defining key trial parameters (e.g. dosing, sample size, endpoints), improve patient stratification, predict patient adherence, and streamline the management of a large volume of case reports and regulatory documentation.

Importantly, GenAI may also contribute to faster clinical trial approvals. A recent study reported that the approval times in the UK decreased from an average of 91 days to 41 days following major reforms supported by new digital platforms at the Medicines and Healthcare products Regulatory Agency (MHRA) [2].

Prof. Alvarez is the coordinator of Genomed4All, a European initiative aimed at transforming the management of haematological diseases through AI. The project pools genomic and other omics data within a secure, federated learning infrastructure, he explained. Its goal is to create a large-scale, distributed repository of omics health data across Europe, enabling the integration of currently fragmented or non-standardised datasets.

In recent years, this initiative has enabled the application of AI in several haematological conditions:

Myelodysplastic syndromes: identification of at-risk individuals through genomic screening; omics-based personalised classification and prognosis; prediction of treatment response to support clinical decision-making; and identification of candidates for drug repurposing in specific myelodysplastic syndrome subgroups.
Multiple myeloma: improved understanding of the disease’s heterogeneity; characterisation of temporal disease evolution; risk stratification; and integration of radiomics and radiogenomics to predict treatment response and progression-free survival.
Sickle cell disease: identification of gene mutations associated with inflammatory pathways; AI-driven patient stratification; development of predictive models for clinical outcome; and creation of probability scores based on brain MRI analysis to predict silent cerebral infarctions in paediatric patients.

Prof. Alvarez pointed out that initiatives such as this represent a significant step toward advanced precision medicine.

Alvarez F. AI-driven innovation in hematology – transforming transplantation and cellular therapy. P01-5, EBMT congress 2026, 22–25 March, Madrid, Spain.
Manfrin A, et al. Br J Clin Pharmacol. 2026;92(3):822-829.