Pedro Miguel Martinez-Girones

Mr. Pedro Miguel Martinez-Girones

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 December 2021 in Electronics
Verified
Journal Article
Open access
PDF file
EAEduardo Alcaín
Eduardo Alcaín
PFPedro Fernández
Pedro Fernández
RNRubén Nieto
Rubén Nieto
AMAntonio S. Montemayor
Antonio S. Montemayor
JVJaime Vilas
Jaime Vilas
ATAngel Torrado-Carvajal
Angel Torrado-Carvajal
Medical imaging is considered one of the most important advances in the history of medicine and has become an essential part of the diagnosis and treatment of patients. Earlier prediction and treatment have been driving the acquisition of higher image resolutions as well as the fusion of different modalities, raising the need for sophisticated hardware and software systems for medical image registration, storage, analysis, and processing. In this scenario and given the new clinical pipelines and the huge clinical burden of hospitals, these systems are often required to provide both highly accurate and real-time processing of large amounts of imaging data. Additionally, lowering the prices of each part of imaging equipment, as well as its development and implementation, and increasing their lifespan is crucial to minimize the cost and lead to more accessible healthcare. This paper focuses on the evolution and the application of different hardware architectures (namely, CPU, GPU, DSP, FPGA, and ASIC) in medical imaging through various specific examples and discussing different options depending on the specific application. The main purpose is to provide a general introduction to hardware acceleration techniques for medical imaging researchers and developers who need to accelerate their implementations.

14 April 2021 in Applied Sciences
Verified
Journal Article
Open access
PDF file
PMPedro Miguel Martinez-Girones
Pedro Miguel Martinez-Girones
JVJavier Vera-Olmos
Javier Vera-Olmos
MGMario Gil-Correa
Mario Gil-Correa
ARAna Ramos
Ana Ramos
LGLina Garcia-Cañamaque
Lina Garcia-Cañamaque
ATAngel Torrado-Carvajal
Angel Torrado-Carvajal
Typically, pseudo-Computerized Tomography (CT) synthesis schemes proposed in the literature rely on complete atlases acquired with the same field of view (FOV) as the input volume. However, clinical CTs are usually acquired in a reduced FOV to decrease patient ionization. In this work, we present the Franken-CT approach, showing how the use of a non-parametric atlas composed of diverse anatomical overlapping Magnetic Resonance (MR)-CT scans and deep learning methods based on the U-net architecture enable synthesizing extended head and neck pseudo-CTs. Visual inspection of the results shows the high quality of the pseudo-CT and the robustness of the method, which is able to capture the details of the bone contours despite synthesizing the resulting image from knowledge obtained from images acquired with a completely different FOV. The experimental Zero-Normalized Cross-Correlation (ZNCC) reports 0.9367 ± 0.0138 (mean ± SD) and 95% confidence interval (0.9221, 0.9512); the experimental Mean Absolute Error (MAE) reports 73.9149 ± 9.2101 HU and 95% confidence interval (66.3383, 81.4915); the Structural Similarity Index Measure (SSIM) reports 0.9943 ± 0.0009 and 95% confidence interval (0.9935, 0.9951); and the experimental Dice coefficient for bone tissue reports 0.7051 ± 0.1126 and 95% confidence interval (0.6125, 0.7977). The voxel-by-voxel correlation plot shows an excellent correlation between pseudo-CT and ground-truth CT Hounsfield Units (m = 0.87; adjusted R2 = 0.91; p < 0.001). The Bland–Altman plot shows that the average of the differences is low (−38.6471 ± 199.6100; 95% CI (−429.8827, 352.5884)). This work serves as a proof of concept to demonstrate the great potential of deep learning methods for pseudo-CT synthesis and their great potential using real clinical datasets.

14 July 2024 in International Journal of Medical Informatics
Verified
Journal Article
Closed access
IJIgnacio Gómez-Rico Junquero
Ignacio Gómez-Rico Junquero
PMPedro Miguel Martinez-Girones
Pedro Miguel Martinez-Girones
AGAdrian Galiana-Bordera
Adrian Galiana-Bordera
AGAntonio Orduña Galán
Antonio Orduña Galán
SGSonia Ginés
Sonia Ginés
LMLuis Martí-Bonmatí
Luis Martí-Bonmatí

31 July 2026 in European Radiology Experimental
Verified
Journal Article
Open access
PDF file
AGAdrian Galiana-Bordera
Adrian Galiana-Bordera
JAJavier Aquerreta-Escribano
Javier Aquerreta-Escribano
PMPedro Miguel Martinez-Girones
Pedro Miguel Martinez-Girones
PLPau Lozano
Pau Lozano
GRGloria Ribas
Gloria Ribas
Luis Martí-BonmatíLuis Martí-Bonmatí
Luis Martí-Bonmatí
Objective: Artificial intelligence (AI) shows promise for improving cancer management, but clinical adoption is limited by the absence of standardized validation tools. As an important step toward bridging the AI chasm, the European CHAIMELEON project addresses this gap by developing a platform for in silico validation of AI models in oncology, aligned with the EUCAIM project. Materials and methods: We designed a web-based platform using Kubernetes microservices architecture, integrating imaging, clinical, and AI components. The backend combines a REST API, an ORTHANC-PACS, and a Keycloak/OAuth2 security layer. The frontend includes a customized OHIF DICOM viewer for oncological case review. Clinicians evaluated cases across three sequential stages: (1) Standard clinical assessment, (2) Assessment with AI predictions, and (3) Final review against clinical endpoint ground truth. The platform also collects clinicians’ perceptions of AI utility, trust, workflow integration, and usability through a Likert-scale questionnaire. Results: The platform provided efficient access to multimodal data for five cancer types. Clinicians completed structured validations, and survey responses indicated favorable perceptions: 93% found the platform intuitive, over 80% would recommend it, and agreement on AI utility exceeded 40% across most endpoints, indicating a positive reception of the AI as a supportive “second reader” tool that prompts clinical re-evaluation rather than simple consensus. Workflow impact was rated positively, underlining the potential to support clinical decision-making. Conclusion: This platform offers a reproducible, scalable, and user-friendly environment for clinical validation of AI models in oncology. Its modular architecture allows integration of additional AI models and is designed to ensure sustainability and interoperability, fostering evidence-based AI adoption in European radiology practice. Trial registration: ClinicalTrials.gov, NCT06950996. Key Points: How can the medical community standardize the in silico clinical validation of oncology AI models to safely bridge the gap before real-world deployment? We successfully implemented a web-based microservices platform, demonstrating high clinical usability and defining human-AI trust dynamics across five distinct cancer types. This standardized validation framework directly enhances diagnostic confidence and interdisciplinary medical collaboration, bridging technical performance with clinical trust to ensure that innovative artificial intelligence technologies safely optimize daily treatment decisions and ultimately maximize therapeutic success and safety for oncology patients.

15 July 2025 in Computer Methods and Programs in Biomedicine
Verified
Journal Article
Open access
ÁPÁngela Pérez-Benito
Ángela Pérez-Benito
AGAdrian Galiana-Bordera
Adrian Galiana-Bordera
PMPedro Miguel Martinez-Girones
Pedro Miguel Martinez-Girones
GUGemma Urbanos
Gemma Urbanos
AIAnna Nogué Infante
Anna Nogué Infante
MAMaria Angeles Perez Anson
Maria Angeles Perez Anson
Prostate cancer remains a significant global health concern, with treatment response varying among patients. Radiotherapy, often combined with hormone therapy, is a key treatment approach, but predicting individual outcomes remains challenging. Computational models have emerged as valuable tools to simulate tumour behaviour and optimise treatment strategies. This study presents a patient-specific computational model designed to predict tumour response by associating Prostate-Specific Antigen (PSA) dynamics with tumour biological behaviour under therapy. The model integrates patient-specific clinical data and imaging biomarkers from a retrospective study, including apparent diffusion coefficient values from diffusion-weighted imaging to represent tumour cellularity and perfusion parameters from dynamic contrast-enhanced MRI to characterise vascular properties. Clinical data from five patients undergoing radiotherapy, hormone therapy, or combination therapy are used for model development and validation. Due to the limited availability of patient data, PSA is the only parameter used for calibration and validation. One patient is used for calibration, while six serve for validation. Model performance is evaluated by calculating the mean absolute error (MAE) between simulated and observed PSA values post-treatment. The model also estimates tumour shrinkage, though this cannot be directly validated. To assess predictive capacity, two patients are selected for additional analysis simulating different treatment strategies and their impact on PSA dynamics and tumour shrinkage. The model successfully replicates PSA trends, with MAE values of 0.1, 0.08, 0.23, 0.14, 0.11 and 0.15 ng/mL and RMSE of 0.18, 0.15, 0.24, 0.18, 0.12 and 0.15 ng/mL for the six validation patients, with Patient C showing the closest correspondence to clinical data (MAE = 0.08). Overall, the MAE ranges from 0.08 ng/mL to 0.23 ng/mL, indicating the model's ability to approximate treatment response. In the two selected patients, simulated treatment variations result in distinct PSA dynamics and estimated tumour shrinkage, highlighting interpatient variability in treatment response. This computational model provides a predictive framework for assessing prostate cancer treatment response based on patient-specific PSA dynamics and imaging biomarkers. While tumour shrinkage estimates cannot be validated, the model offers insights into treatment-induced PSA fluctuations. The findings support the potential of in-silico tools in personalised medicine, aiding clinical decision-making by evaluating different therapeutic strategies. Further validation with larger datasets is necessary for clinical integration.