Clinical Medicine, Oncology and Diagnostics
16:05 - 17:25 | Meeting Room 103 (Level 1)
Chair: Dr Christian Zammit
Dr Oriana Mazzitelli
Department of Applied Biomedical Sciences, Faculty of Health Sciences
Lung cancer is the leading cause of cancer mortality worldwide. Conventional small-molecule anti-cancer drugs still encounter important challenges, including low response rates and drug resistance amongst patients. Therapeutic antibodies supersede the efficacy of small molecules due to their higher specificity and stronger affinity for their molecular targets. These antibodies have the potential to be used for diagnostic purposes and as drug-delivery vehicles when combined with cytotoxic drugs.
The aim of Project LuCSIB (Lung Cancer Specific Immuno-Biologics) is to discover and validate high-affinity single-chain antibodies that specifically target known tumour-associated antigens (TAAs) and immune checkpoint proteins present in Non-Small Cell Lung Cancer (NSCLC) tumour cells. This is being achieved using the 2018 Nobel Prize–winning antibody phage display technique, in which genetically modified bacteriophages display single-chain antibody fragments on their surface, enabling the identification of target-specific binders through iterative selection.
Additionally, we are also adopting a non-biased approach to ensure the antibodies recognise target antigens as presented on the cancer cells by performing cell-based phage biopanning on commercially available NSCLC cell lines and control (non-tumour) immortalised lung cells.
We have uncovered binders to the tumour-associated antigens ROR1 and TROP2, as well as to the immune checkpoint protein CD47, validated by ELISA. Additionally, flow cytometry confirmed binders to the human lung adenocarcinoma cell line A549, with no binding to immortalised human alveolar epithelial cells (huAEC), highlighting the potential of phage display to discover antibodies targeting novel cancer-specific molecules.
Future work will determine the nucleotide sequences of candidate antibodies and employ computational modelling to predict and refine the 3D structures of their CDRs, with the aim of characterising and enhancing antigen-binding affinity.
Ms Deborah Mizzi
Department of Radiography, Faculty of Health Sciences
Purpose: To develop evidence-informed policy recommendations to improve supplementary breast cancer screening for women with dense breasts across Europe, by addressing current gaps in guidelines, access to imaging, and professional preparedness.
Methods: This work draws on a four-phase mixed-methods investigation comprising:
(1) a systematic review and meta-analysis of supplementary screening modalities;
(2) qualitative interviews with breast screening professionals across Europe;
(3) a cross-sectional online survey assessing knowledge, attitudes, and practices; and
(4) integration of findings to formulate actionable policy recommendations. Triangulated evidence informed both a proposed policy change document and a targeted dissemination framework.
Results: Evidence supports the high sensitivity of modalities such as Magnetic Resonance Imaging, Digital Breast Tomosynthesis, and Contrast Enhanced Mammography for dense-breast screening, yet access to imaging and the implementation of supplementary screening strategies vary widely across Europe. Survey data revealed that 91.6% of professionals would adopt new practices with adequate training and resources, and 90.4% recognised the need for further education. Barriers include the absence of harmonised guidelines, workforce shortages, and unequal access to advanced imaging. Eight policy recommendations are proposed: harmonised European guidelines; mandatory breast density reporting; risk-based screening pathways; enhanced professional training; investment in infrastructure; improved patient communication; research registry development; and cross-border collaboration. The dissemination plan targets policymakers, healthcare professionals, academics, and the public through policy briefs, academic publications, engagement with professional society, and public awareness campaigns.
Conclusion: Implementing these recommendations could harmonise screening practices, reduce inequities, and improve outcomes for women with dense breasts.
Dr Nicoletta Riva
Department of Pathology, Faculty of Medicine and Surgery
Background: Apixaban is a direct oral anticoagulant which inhibits coagulation factor Xa. This study aimed to assess the effect of apixaban on a global coagulation assay, the thrombin generation assay (TGA).
Methods: A pool of normal plasma was spiked with increasing concentrations of apixaban (range 50-750 ng/ml), corresponding to sub-therapeutic, therapeutic and supra-therapeutic levels. Concentrations were confirmed using Ultra-High Performance Liquid Chromatography coupled with tandem Mass Spectrometry (UHPLC-MS/MS). The TGA analyser provided the following parameters: Lag Time (LT), Time to Peak (ttP), Peak Thrombin (Peak), Endogenous Thrombin Potential (ETP), Velocity Index (VelIndex). Standard coagulation assays, such as prothrombin time (PT) and activated partial thromboplastin time (APTT), were also performed.
Results: Apixaban concentrations positively correlated with PT and APTT (both p<0.001); however, the magnitude of effect was modest (up to 2.2 seconds for PT, 12.5 seconds for APTT).
TGA showed that increasing concentrations of apixaban had a strong positive correlation with LT (r=0.956, p=0.011), a non-significant positive correlation with ttP (r=0.788, p=0.113), strong negative correlations with Peak (r=-0.880, p=0.049), ETP (r=-0.911, p=0.031), and VelIndex (ρ=-1.0, p<0.01).
When the effect of branded apixaban was compared to different generic apixabans spiked at the same concentrations, no significant differences were found (p=0.948 for LT; p=0.551 for ttP; p=0.836 for Peak; p=0.901 for ETP; p=0.615 for VelIndex).
Conclusion: Global coagulation assays are more sensitive than standard coagulation assays for measuring the in vitro effect of apixaban. There was no difference between branded and generic apixabans.
Ms Dora Lee Borg
Department of Systems and Control Engineering, Faculty of Engineering
Cardiovascular Disease (CVD) is the leading cause of mortality globally and poses an increased risk to individuals with diabetes, often developing silently before clinical diagnosis. Despite widespread availability of Electrocardiograms (ECGs), current CVD risk assessment tools rely on multi-parameter clinical data and are not suited for continuous or population-level screening. Advances in Artificial Intelligence (AI) and smart wearable technologies provide an opportunity to enable earlier, accessible, and scalable cardiovascular risk prediction. This study aims to develop and validate ECG-based Explainable AI (XAI) models capable of predicting CVD risk in individuals with a history of diabetes, with a specific focus on translating hospital-grade 12-lead ECG intelligence to single-lead ECGs in concordance with smart wearable devices. Deep learning models, primarily convolutional neural networks, are trained on raw ECG waveform data. Both 12-lead clinical ECGs and derived single-lead ECGs are used. Data are obtained from large, open-access repositories, including ethically approved, controlled-access clinical datasets, comprising adults with diabetes and longitudinal cardiovascular outcomes. Models are evaluated using robust cross-validation and standard performance metrics. Explainable AI techniques are applied to identify ECG waveform regions that contribute most strongly to predicted cardiovascular risk, providing visual explanations that support physiological interpretation and clinical plausibility. The ongoing work will demonstrate the feasibility of accurate CVD risk prediction using single-lead ECGs, with strong correlation to 12-lead models. By enabling explainable and scalable cardiovascular risk assessment, the project supports earlier intervention and prevention strategies, paving the way for the integration of AI-driven ECG analysis into e-health systems and smart wearable platforms.
Prof. Nikolai P. Pace
Department of Anatomy, Faculty of Medicine and Surgery
Background: Hidradenitis suppurativa (HS) is a chronic inflammatory skin disorder with complex aetiology. The Maltese population exhibits a high HS prevalence, and we identified a founder NCSTN variant, providing a unique model to dissect genotype-phenotype relationships.
Methods: We conducted a multi-phase study in an ethnically Maltese cohort, integrating deep phenotyping, targeted genotyping for the NCSTN:p.Val224_Thr227del variant, whole exome sequencing (WES) of variant-negative familial/atypical cases, serum biomarker analysis, and Mendelian randomisation (MR).
Results: The founder NCSTN:p.Val224_Thr227del variant accounted for 31.1% of familial HS, and identity-by-descent analysis confirmed a founder effect. WES revealed novel and potentially deleterious variants in NCSTN and NOD2 (p.Ala731Val). Elevated serum Immunoglobulin G (IgG) correlated strongly with disease severity and specifically with the NCSTN indel variant, establishing it as a novel biomarker. MR confirmed a causal role for body fat percentage in HS risk. Phenotypically, distinct sex-specific anatomical patterns and a high prevalence of metabolically unhealthy phenotypes were identified.
Conclusion: This study delineates a significant burden of rare deleterious variants, particularly in NCSTN, within a founder HS population. Integration of genomic, phenotypic, and biochemical profiling refines HS subclassification and supports precision medicine approaches. Ongoing work is investigating the functional impact of the NCSTN founder variant using CRISPR-Cas9 edited keratinocytes to elucidate its pathogenic mechanism.
Prof. Josephine Attard
Department of Midwifery, Faculty of Health Sciences
Background: Professional Doctorates (PDs) offer a practice-focused doctoral pathway that integrates advanced scholarship, clinical leadership, and applied research. While well established in larger systems, their relevance and feasibility within small-resource contexts remain underexplored.
Aim: To examine the development, perceived value, and strategic considerations for implementing a Professional Doctorate in midwifery and nursing within a small-state higher education system.
Methods: An exploratory, sequential, mixed-methods programme of research was undertaken. Phase one involved qualitative interviews with UK-based PD programme coordinators to identify key domains underpinning sustainable PD design. These findings informed the development and validation of the Perspectives on a Professional Doctorate in Education for Midwifery and Nursing.
Results: Across stakeholder groups, PDs were perceived as valuable for strengthening research competence, clinical leadership, and professional credibility. However, perceived behavioural control shaped by workload pressures, funding constraints, and limited supervisory capacity emerged as a key barrier. Seven Essential Domains for PD development were consistently identified: programme structure, admissions, learning outcomes, supervision, implementation enablers and constraints, quality assurance, and sustainability.
Conclusion: Professional Doctorates are both desirable and feasible in small-state contexts when grounded in flexible, interprofessional design and supported by strong institutional commitment. The Seven Essential Domains framework offers a transferable, evidence-informed guide for developing sustainable PD pathways that advance midwifery and nursing leadership, research capacity, and practice impact.
Dr Jose Guilherme Couto
Department of Radiography, Faculty of Health Sciences
Background: Persistent disparities in healthcare education across Europe create significant expertise gaps that can compromise the quality of patient care. While educational variations are documented, there is a lack of standardised, practical tools to systematically identify and address these gaps in expertise in a multi-professional context.
Objectives: This study aimed to develop and validate a comprehensive tool to identify expertise gaps across four domains (knowledge, practice, research, and teaching methods) and to assess the extent of these gaps across European healthcare settings.
Methods: A cross-sectional, anonymous survey was conducted between July and September 2024, and 230 European healthcare professionals and academics were recruited using convenience sampling. Internal consistency reliability was evaluated using Cronbach’s alpha, and construct validity was assessed using Confirmatory Factor Analysis (CFA).
Results: The tool demonstrated strong internal consistency (α>0.877) and robust construct validity (Factor Loadings between 1.18–2.05; p<0.001). Expertise gaps were greater in digital (mean=5.04/10) and green skills (mean=6.16/10) across all four domains (knowledge, practice, research and teaching methods). Participants reported specific deficiencies in AI integration, sustainable healthcare practices, and innovative pedagogies such as simulation-based learning. Notably, no significant demographic or regional correlations were found, indicating that these challenges are systemic across the European landscape.
Conclusion: This validated tool offers a practical framework for mapping expertise gaps at local and international levels. The study identified major gaps in digital and green skills, emphasising an urgent need for targeted interventions in digital technology and sustainability to align healthcare education with the UN Sustainable Development Goals.
Prof. Cristiana Sebu
Department of Mathematics, Faculty of Science
Electrical impedance tomography (EIT) is a technology for imaging the electrical conductivity distribution within an object from boundary measurements of electric currents and voltages. EIT can be used as a method of medical imaging, and one possible application is breast cancer detection. The main focus of this project is to develop novel direct algorithms to obtain conductivity reconstructions for arbitrary geometry, extremely rapidly and without relying on accurate a priori information. This is achieved by reformulating the inverse problem as a pair of coupled integral equations, one of which is of the first kind. We solve them by using variational mollifier methods – a powerful regularisation method that has not previously been implemented for EIT. Reconstructions of the conductivity distributions obtained from simulated noisy data will be presented.