« Back

Mataj-Lokaj Belinda

Collaboratrice scientifique HES

Main contract

Collaboratrice scientifique HES

Haute école de santé - Genève
Avenue de Champel 47, 1206 Genève, CH

Faculty
Santé

Main Degree Programme
Technique en radiologie médicale

No data to display for this section

2025

Multimodal deep learning fusion of ultrafast-DCE MRI and clinical information for breast lesion classification

Scientific paper ArODES

Belinda Lokaj, Valentin Durand de Gevigney, Amal-Dahila Djema, Jamil Zaghir, Jean-Philippe Goldman, Mina Bjelogrlic, Hugues Turbé, Karen Kinkel, Christian Lovis, Jerome Schmid

Computers in biology and medicine, 2025, 188, 109721

Link to the publication

Summary:

Background : Breast cancer is the most common cancer worldwide, and magnetic resonance imaging (MRI) constitutes a very sensitive technique for invasive cancer detection. When reviewing breast MRI examination, clinical radiologists rely on multimodal information, composed of imaging data but also information not present in the images such as clinical information. Most machine learning (ML) approaches are not well suited for multimodal data. However, attention-based architectures, such as Transformers, are flexible and therefore good candidates for integrating multimodal data. Purpose : The aim of this study was to develop and evaluate a novel multimodal deep learning (DL) model combining ultrafast dynamic contrast-enhanced (UF-DCE) MRI images, lesion characteristics and clinical information for breast lesion classification. Materials and methods : From 2019 to 2023, UF-DCE breast images and radiology reports of 240 patients were retrospectively collected from a single clinical center and annotated. Imaging data were constituted of volumes of interest (VOI) extracted around segmented lesions. Non-imaging data were constituted of both clinical (categorical) and geometrical (scalar) data. Clinical data were extracted from annotated reports and were associated to their corresponding lesions. We compared the diagnostic performances of traditional ML methods for non-imaging data, an image model based on the DL architecture, and a novel Transformer-based architecture, the Multimodal Sieve Transformer with Vision Transformer encoder (MMST-V). Results : The final dataset included 987 lesions (280 benign, 121 malignant lesions, and 586 benign lymph nodes) and 1081 reports. For classification with non-imaging data, scalar data had a greater influence on performances of lesion classification (Area under the receiver operating characteristic curve (AUROC) = 0.875 ± 0.042) than categorical data (AUROC = 0.680 ± 0.060). MMST-V achieved better performances (AUROC = 0.928 ± 0.027) than classification based on non-imaging data (AUROC = 0.900 ± 0.045), and imaging data only (AUROC = 0.863 ± 0.025). Conclusion : The proposed MMST-V is an adaptative approach that can consider redundant information provided by multimodal information. It demonstrated better performances than unimodal methods. Results highlight that the combination of clinical patient data and detailed lesion information as additional clinical knowledge enhances the diagnostic performances of UF-DCE breast MRI.

Learning MRI with ImmeRgaMe :

Scientific paper ArODES

exploring the pedagogical potential of an innovative serious game for radiographer training

Marie-Anaïs Petit, Joël Piguet, Belinda Lokaj, Jerome Schmid, Céline Gaignot

Radiography, 2025, 31, 3, 102921

Link to the publication

Summary:

Introduction : This study aimed to assess the potential of ImmeRgaMe, an innovative 360° immersive serious game, to enhance motivation and engagement in MRI education for first-year radiographer students. Methods : The platform was developed as part of an educational innovation initiative and designed to complement traditional teaching methods. It integrates storytelling, progression mechanisms, and interactive tools to teach MRI safety protocols, physical principles, and clinical methods. Beta testing involved 42 participants (students and educators) who completed standardized surveys to evaluate gameplay, usability, and learning outcomes. The impact of microlearning videos embedded in the game was also assessed through pre- and post-test quizzes with 63 first-year students. Results : The beta version received positive feedback, with over 90 % of participants rating the gameplay and integrated tools as satisfactory. Students demonstrated improved performance in quizzes after using the microlearning videos, with an average score increase of 19 % across tests. Survey results indicated that 97 % of respondents found the videos helpful for understanding course content, and 86 % believed the game could support their progress during MRI internships. Conclusion : ImmeRgaMe effectively fosters student motivation and engagement, bridging the gap between theoretical and practical knowledge in MRI training. While promising, further refinement and broader testing are needed to evaluate its impact on knowledge retention and skills development, as well as its applicability to other imaging modalities. Implications for practice : The implementation of serious games like ImmeRgaMe in radiography education could reshape and modernize traditional teaching methods. By adapting this approach to other imaging modalities, educators could offer immersive and interactive learning experiences, fostering self-regulated learning and aligning training with the complex demands of clinical practice.

2024

Efficient clinical information extraction from breast radiology reports in French

Book chapter ArODES

Jamil Zaghir, Belinda Lokaj, Karen Kinkel, Amal-Dahila Djema, Hugues Turbé, Mina Bjelogrlic, Valentin Durand de Gevigney, Jerome Schmid, Christian Lovis, Jean-Philippe Goldman

Dans Andrikopoulou, Elisavet, Arvanitis, Theodoros N., Benis, Arriel, Demiris, George, Gallos, Parisis, Hasman, Arie, Mantas, John, Marschollek, Michael, Ognjanovic, Ivana, Saranto, Kaija, Zoulias, Emmanouil, Studies in health technology and informatics (5 p.). 2024, Amsterdam : IOS Press Pays-Bas : IOS Press

Link to the publication

Summary:

Radiology reports contain crucial patient information, in addition to images, that can be automatically extracted for secondary uses such as clinical support and research for diagnosis. We tested several classifiers to classify 1,218 breast MRI reports in French from two Swiss clinical centers. Logistic regression performed better for both internal (accuracy > 0.95 and macro-F1 > 0.86) and external data (accuracy > 0.81 and macro-F1 > 0.41). Automating this task will facilitate efficient extraction of targeted clinical parameters and provide a good basis for future annotation processes through automatic pre-annotation.

Barriers and facilitators of artificial intelligence conception and implementation for breast imaging diagnosis in clinical practice :

Scientific paper ArODES

a scoping review

Belinda Lokaj, Marie-Thérèse Pugliese, Karen Kinkel, Christian Lovis, Jerome Schmid

European radiology, 2024, 34, 3, 2096-2109

Link to the publication

Summary:

Objective : Although artificial intelligence (AI) has demonstrated promise in enhancing breast cancer diagnosis, the implementation of AI algorithms in clinical practice encounters various barriers. This scoping review aims to identify these barriers and facilitators to highlight key considerations for developing and implementing AI solutions in breast cancer imaging. Method : A literature search was conducted from 2012 to 2022 in six databases (PubMed, Web of Science, CINHAL, Embase, IEEE, and ArXiv). The articles were included if some barriers and/or facilitators in the conception or implementation of AI in breast clinical imaging were described. We excluded research only focusing on performance, or with data not acquired in a clinical radiology setup and not involving real patients. Results : A total of 107 articles were included. We identified six major barriers related to data (B1), black box and trust (B2), algorithms and conception (B3), evaluation and validation (B4), legal, ethical, and economic issues (B5), and education (B6), and five major facilitators covering data (F1), clinical impact (F2), algorithms and conception (F3), evaluation and validation (F4), and education (F5). Conclusion : This scoping review highlighted the need to carefully design, deploy, and evaluate AI solutions in clinical practice, involving all stakeholders to yield improvement in healthcare.

2024

Barriers and facilitators of implementation of AI

Conference

Mataj-Lokaj Belinda, Pugliese Marie-Therese, K. Kinkel, C. Lovis, Schmid Jérôme

European Society of Breast Imaging (EUSOBI), 03.10.2024 - 04.10.2024, Lisbon, Portugal

Efficient clinical information extraction from breast radiology reports in French

Conference

J. Zaghir, Mataj-Lokaj Belinda, K. Kinkel, A. Djema, H. Turbé, M. Bjelogrlic, Durand De Gevigney Valentin, Schmid Jérôme, C. Lovis, J.-P. Goldman

Medical Informatics Europe, 25.08.2024 - 29.08.2024, Athens, Greece

Comparative analysis of breast lesion classification with deep learning in dynamic MRI using different types of regions of interest

Conference

Mataj-Lokaj Belinda, Durand De Gevigney Valentin, A. Djema, K. Kinkel, C. Lovis, Schmid Jérôme

Swiss Congress of Radiology, 20.06.2024 - 22.06.2024, Genève, Suisse

Enhancing breast lesion classification by integrating lesion characteristics and clinical data information with ultrafast MRI

Conference

Mataj-Lokaj Belinda, Durand De Gevigney Valentin, J.-P. Goldman, K. Kinkel, C. Lovis, J. Zaghir, A. Djema, Schmid Jérôme

European Congress of Radiology, 28.02.2024 - 03.03.2024, Vienna, Austria

2023

Combining ultrafast MRI sequence with artificial intelligence (AI) for breast cancer detection

Conference

Mataj-Lokaj Belinda, A. Djema, K. Kinkel, C. Lovis, Schmid Jérôme

European Congress of Radiology, 01.03.2023 - 05.03.2023, Vienna, Austria

2020

Evaluation of an ultra-fast 4D sequence for detection of breast lesions in MRI

Conference

Mataj-Lokaj Belinda, K. Kinkel, J.-N. Hyacinthe, Schmid Jérôme, Gaignot Céline

International Society for Magnetic Resonance in Medicine (ISMRM), 18.04.2020 - 23.04.2020, Sydney, Australia

Achievements

PEOPLE@HES-SO Directory and Skills inventory

Mataj-Lokaj Belinda

Collaboratrice scientifique HES

Collaboratrice scientifique HES

PEOPLE@HES-SO
Directory and Skills inventory