Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 14
Encontrada(s) 1 página(s)

Docta Complutense
Publicación Artículo científico (JournalArticle)

Mapping intratumoral heterogeneity through PET-derived washout and deep learning after proton therapy

  • Cabrales, Pablo
  • Izquierdo-García, David
  • V. Onecha, Víctor
  • Pérez Liva, Mailyn
  • Fraile Prieto, Luis Mario
  • Udías Moinelo, José Manuel
  • López Herraiz, Joaquín
The distribution of produced isotopes during proton therapy can be imaged with Positron Emission Tomography (PET) to verify dose delivery. However, biological washout, driven by tissue-dependent processes such as perfusion and cellular metabolism, reduces PET signal-to-noise ratio (SNR) and limits quantitative analysis. In this work, we propose an uncertainty-aware deep learning framework to improve the estimation of washout parameters in post-proton therapy PET, not only enabling accurate correction for washout effects, but also mapping intratumoral heterogeneity as a surrogate marker of tumor status and treatment response. We trained the models on Monte Carlo-simulated data from eight head-and-neck cancer patients, and tested them on four additional head-and-neck and one liver patient. Each patient was represented by 75 digital twins with distinct tumoral washout dynamics and imaged 15 min after treatment, when slow washout components dominate. We also introduced washed-out maps, quantifying the contribution of medium and fast washout components to the loss in activity between the end of treatment and the start of PET imaging. Trained models significantly improved resolution and accuracy, reducing average absolute errors by 60% and 28% for washout rate and washed-out maps, respectively. For intratumoral regions as small as 5.1 mL, errors predominantly fell below thresholds for differentiating vascular status, and the models generalized across anatomical areas and acquisition delays. This study shows the potential of deep learning in post-proton therapy PET to non-invasively map washout kinetics and reveal intratumoral heterogeneity, supporting dose verification, tumor characterization, and treatment personalization. The framework is available at https://github.com/pcabrales/ppw.git .
Docta Complutense
Publicación Artículo científico (JournalArticle)

Stress-induced takotsubo syndrome: dynamic changes in regional cerebral metabolism revealed by quantitative PET imaging

  • Ariza Carrasco, Alejandro
  • Yoganathan, Thulaciga
  • Leciñana, María Alonso de
  • Viel, Thomas
  • Mikail, Nidaa
  • López Herraiz, Joaquín
  • Udías Moinelo, José Manuel
  • Ibáñez García, Paula Beatriz
  • Tavitian, Bertrand
  • Pérez-Liva, Mailyn
Stress significantly contributes to cardiovascular diseases such as Takotsubo syndrome (TTS), which mimics an acute coronary syndrome without coronary obstruction. TTS is triggered by surgery, trauma, and emergency treatments in patients, and is reproduced in animal models by a catecholamine surge that impacts cardiac sympathetic innervation. The action of catecholamines on energy metabolism is well documented in the heart, less so in the brain. We investigated the effects of acute catecholaminergic stress on regional cerebral glucose metabolism and interregional metabolic organization in a TTS rat model using FDG-PET and quantitative two- tissue compartment modeling. Adult female rats received a single intraperitoneal injection of isoprenaline (ISO) (50 mg/kg). Dynamic FDG-PET imaging was performed at baseline, 2 hours (acute phase), and 7 days (recovery phase) post-injection. Kinetic parameters, namely glucose inflow (K1) and glucose phosphorylation (k3), were quantified in 58 brain regions. Interregional metabolic coordination, defined as statistically significant linear correlations between regional kinetic parameters, was assessed across functional brain areas. During the acute phase, the catecholaminergic surge induced widespread reductions in glucose inflow and regional decreases in phosphorylation, particularly in the limbic and sensorimotor areas. During the recovery phase, most regions remained below baseline. Metabolic coordination increased for glucose inflow in both phases but declined for phosphorylation, especially during recovery, indicating a disruption of metabolic synchronization. Persistent changes in brain metabolism imply that mid-to-long-term changes in regional cerebral metabolism may contribute to long-term TTS consequences.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Analytical positron range model for PET with cross-code Monte Carlo benchmarking

  • Paneque-Yunta, Robert J.
  • Encina-Baranda, Nerea
  • Carter, Lukas M.
  • Galve, Pablo
  • Ibáñez, Paula
  • Abushab, Khaled M.
  • Udías, José M Udías
  • Herraiz, Joaquín L.
22 páginas, 8 figuras, 5 tablas, Introduction. The positron range (PR) effect is a significant factor limiting spatial resolution in positron emission tomography (PET), particularly for high-resolution systems and non-standard isotopes. Objective. This study introduces a novel analytical model to accurately and rapidly describe PR distributions (PRd) for various PET radioisotopes to better include its effect in PET reconstruction algorithms. Approach. The proposed model explicitly incorporates the Coulomb repulsion effect, the multi-branch nature of certain β+ emitters, and the scaling of PR with electronic density. To minimise bias, we used a histogram-free statistical method to derive the cumulative PRd from Monte Carlo (MC) simulated annihilation datasets, avoiding arbitrary histogram binning. A comparative analysis of PR estimates was conducted across three major MC radiation transport algorithm packages: PENELOPE (via PenEasy/PeneloPET), GEANT4 (via GATE), and EGS5 (via PHITS), revealing notable discrepancies between codes, versions, and input configurations, especially at short distances from the source. Main results. The new analytical model demonstrated an excellent reproduction of the simulated data for isotopes including 11C , 13N , 15O , 18F , 64Cu , 68Ga , 82Rb and 124I , achieving in general coefficients of determination (R2) greater than 0.995 and mean absolute percentage errors ≲20%. Compared to previous methods, our model provides a more accurate description of PRd at low distances and offers improved R2 values. Significance. This work provides a robust framework for generating accurate annihilation point spread function kernels, facilitating improved PR correction in quantitative Nuclear Medical Imaging and supporting research with diverse radioisotopes., We acknowledge support from the following grants: FASCINA (PID2021-126998OB-I00), Prototwin Project (TED2021-349130592B-I00), INVENTOR (PID2022-137114OA-I00) and 3PET (PDC2022-133057-I00), funded by the Spanish Ministry of Science, Innovation and Universities (MCIU) through AEI/10.13039/501100011033 and the European Union’s NextGenerationEU/PRT programme and the Spanish Government. P. G. also acknowledges support from the grant JDC2023-051754-I, funded by MCIU/AEI/10.13039/501100011033 and ESF+., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Stress-induced takotsubo syndrome: dynamic changes in regional cerebral metabolism revealed by quantitative PET imaging

  • Ariza-Carrasco, Alejandro
  • Yoganathan, Thulaciga
  • Alonso de Leciñana, María
  • Viel, Thomas
  • Mikail, Nidaa
  • Herraiz, Joaquin L.
  • Udias, Jose M.
  • Ibáñez, Paula
  • Tavitian, Bertrand
  • Pérez-Liva, Mailyn
10 páginas, 3 figuras, 1 tabla, Stress significantly contributes to cardiovascular diseases such as Takotsubo syndrome (TTS), which mimics an acute coronary syndrome without coronary obstruction. TTS is triggered by surgery, trauma, and emergency treatments in patients, and is reproduced in animal models by a catecholamine surge that impacts cardiac sympathetic innervation. The action of catecholamines on energy metabolism is well documented in the heart, less so in the brain. We investigated the effects of acute catecholaminergic stress on regional cerebral glucose metabolism and interregional metabolic organization in a TTS rat model using FDG-PET and quantitative two-tissue compartment modeling. Adult female rats received a single intraperitoneal injection of isoprenaline (ISO) (50 mg/kg). Dynamic FDG-PET imaging was performed at baseline, 2 hours (acute phase), and 7 days (recovery phase) post-injection. Kinetic parameters, namely glucose inflow (K1) and glucose phosphorylation (k3), were quantified in 58 brain regions. Interregional metabolic coordination, defined as statistically significant linear correlations between regional kinetic parameters, was assessed across functional brain areas. During the acute phase, the catecholaminergic surge induced widespread reductions in glucose inflow and regional decreases in phosphorylation, particularly in the limbic and sensorimotor areas. During the recovery phase, most regions remained below baseline. Metabolic coordination increased for glucose inflow in both phases but declined for phosphorylation, especially during recovery, indicating a disruption of metabolic synchronization. Persistent changes in brain metabolism imply that mid-to-long-term changes in regional cerebral metabolism may contribute to long-term TTS consequences., MP-L acknowledges support from the Programme Ram´on y Cajal RYC2021-032739-I, funded by MCIN/AEI/10.13039/501100011033 and by the European Union "NextGenerationEU"/PRTR". JLH acknowl- edges support from grants FASCINA (PID2021-126998OB-I00/), 3PET (PDC2022-133057-I00), Prototwin Project (TED2021-130592B-I00) from the Spanish Ministry of Science and Innovation (MCIN) AEI/ 10.13039/501100011033/Uni´on Europea Next GenerationEU/PRTR, and “Tecnologías Avanzadas para la Exploraci´on del Universo y sus Componentes” (PR47/21 TAU-CM TAU-PRTR), funded by EU Resilient and NextGeneration funds. This work has been funded with €1.026 million by the Comunidad de Madrid through the LUNABRAIN-CM R&D activities program (TEC-2024/TEC-43), granted by Order 5696/2024, and has also received funding from the Cancer Research for Personalized Medicine – CARPEM project (Site de Recherche Int´egr´e sur le Cancer SIRIC), from the Plan Cancer Physicancer [grant number C16025KS], and from the R´egion Ile-de-France SESAME funds. This work was sup- ported by Knowledge Generation Projects 2022 Type A Oriented Research, Government of Spain PID2022-137114OA-I00. In vivo imag- ing was performed at the Life Imaging Facility of Universit´e Paris Cit´e (Plateforme Imageries du Vivant - PIV), supported by France Life Im- aging [grant number ANR-11-INBS-0006] and Infrastructures Biologie- Sant´e (IBiSa). BT, MP-L, PIG, JLH and JMU acknowledge support from the European Union as part of the European Innovation Council’s Pathfinder Open Programme: RETIMAGER, 101099096., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Numerical Framework for Deep Learning Photoacoustics Image Reconstruction

  • G. Muñoz, Javier
  • Ariza-Carrasco, Alejandro
  • Cruza, Jorge F.
  • Elvira Segura, Luis
  • Camacho Sosa-Días, Jorge
  • Udías, José M.
  • Alonso de Laciñana, María
  • Rodriguez-Pardo, Jorge
  • Ibáñez García, Paula
  • Pérez-Liva, Mailyn
Forum Acusticum / Euronoise 2025, the 11th EAA Annual European Conference on Acoustics and Noise Control Engineering, joint with the XLVI Spanish Acoustic Conference TECNIACUSTICA 2025, June 23rd to 26th 2025, Málaga (Spain) 4 páginas, 2 figuras, Photoacoustic Imaging (PAI) combines optical excitation with ultrasound detection to produce high-quality molec- ular images of intact biological tissues. Conventional sys- tems rely on Class IV [1] pulsed lasers, which limits clin- ical translation. This work explores the use of pulsed semiconductor diodes for their portability, versatility, and low-cost. However, their lower peak power decreases PAI signal amplitude. Implementing deep learning techniques trained with realistic synthetic data is a promising tech- nique to increase PAI image quality [2]. Here we create a framework which simulates real scenarios taking into account acoustics and optical properties, ultrasound trans- ducer characteristics and fiber bundles features, crucial to optimize reconstruction performance with a Neural Net- work. The simulated experimental setup included an ar- ray of optical fibers connected to semiconductor diodes parallel to a 128-element transducer array which registers the signal generated by the absorbed light. Optical proper- ties were simulated in MCX Monte Carlo software using GPU parallelization. k-Wave software was used for acous- tic simulations and Time-Reversal (TR) image reconstruc- tion. Finally, a Neural Network model implemented using Tensorflow enhanced the reconstruction, recovering up to 3 cm depth and 200 μm of resolution, delivering excep- tional vascular imaging quality., This work received support from the European Union’s Horizon 2020 research and Innovation Pro- gramme under the Marie Sklodowska-Curie Grant Agreement no.101030046, and the Programme Ram´on y Cajal RYC2021-032739-I, funded by MCIN/AEI/10.13039/501100011033 and by the Eu- ropean Union ”NextGenerationEU”/PRTR”. Also, we acknowledge support from grants FASCINA (PID2021- 126998OB-I00/), 3PET (PDC2022-133057-I00) from the Spanish Ministry of Science and Innovation (MCIN) AEI/10.13039/501100011033/Uni´on Europea Next GenerationEU/PRTR, Prototwin Project (TED2021-349 130592B-I00) and Tecnolog´ıas Avanzadas para la Ex- ploraci´on del Universo y sus Componentes” (PR47/21 TAU-CM TAU- PRTR), funded by EU Resilient and NextGeneration funds. This work was supported by Knowledge Generation Projects 2022 Type A Ori- ented Research, Government of Spain INVENTOR (PID2022-137114OA-I00)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Non-Invasive and Low-Cost Photoacoustic/Ultrasound Dual-Modal Imaging System for Carotid Plaque Imaging

  • Ariza-Carrasco, Alejandro
  • G. Muñoz, Javier
  • Cruza, Jorge F.
  • Elvira Segura, Luis
  • L. Herraíz, Joaquín
  • Udías, José M.
  • Alonso de Laciñana, María
  • Rodríguez-Pardo, Jorge
  • Ibáñez García, Paula
  • Pérez-Liva, Mailyn
Forum Acusticum / Euronoise 2025, the 11th EAA Annual European Conference on Acoustics and Noise Control Engineering, joint with the XLVI Spanish Acoustic Conference TECNIACUSTICA 2025, June 23rd to 26th 2025, Málaga (Spain) 4 páginas, 4 páginas, Carotid plaque rupture causes most ischemic strokes, yet current diagnostics focus on arterial stenosis, overlooking 80–90% of vulnerable plaques. This highlights the need for affordable, precise, and non-invasive imaging to as- sess plaque vulnerability. To address this, we developed a low-cost, dual Ultrasound/Photoacoustic system integrat- ing a Vermon L7.5-MHz ultrasound probe, a 128-channel programmable phased array Ultrasound scanner (Dasel, Spain), and pannels of pulsed semiconductor lasers diodes of 5W emitting 100 ns pulses at 447 nm. Synchroniza- tion was managed via external pulse generators, with laser light delivered through optical fibers. Validation utilised phantoms of black nylon threads. GPU-accelerated delay- and-sum beamforming was used for image recontruction. The system achieved 530 μm lateral and 360 μm ax- ial resolution. This system offers a cost-effective, non-invasive approach for high-resolution multiparametric tis- sue characterization. Future aims will be to improve in vivo assessment of plaque vulnerability., This work received support from the European Union’s Horizon 2020 research and Innovation Pro- gramme under the Marie Sklodowska-Curie Grant Agreement no.101030046, and the Programme Ramónn y Cajal RYC2021-032739-I, funded by MCIN/AEI/10.13039/501100011033 and by the Eu- ropean Union ”NextGenerationEU”/PRTR”. Also, we acknowledge support from grants FASCINA (PID2021- 126998OB-I00/), 3PET (PDC2022-133057-I00) from the Spanish Ministry of Science and Innovation (MCIN) AEI/10.13039/501100011033/Uni´on Europea Next GenerationEU/PRTR, Prototwin Project (TED2021-349 130592B-I00) and Tecnologías Avanzadas para la Ex- ploración del Universo y sus Componentes” (PR47/21 TAU-CM TAU- PRTR), funded by EU Resilient and NextGeneration funds. This work was supported by Knowledge Generation Projects 2022 Type A Ori- ented Research, Government of Spain INVENTOR (PID2022-137114OA-I00)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Mapping intratumoral heterogeneity through PET-derived washout and deep learning after proton therapy

  • Cabrales, Pablo
  • Izaquierdo-García, David
  • Onechea, Víctor V.
  • Pérez-Liva, Mailyn
  • Fraile, Luis Mario
  • Udías, José Manuel
  • Herráiz, Joaquín L.
13 páginas, 7 figuras, 4 tablas, The distribution of produced isotopes during proton therapy can be imaged with Positron Emission Tomography (PET) to verify dose delivery. However, biological washout, driven by tissue-dependent processes such as per fusion and cellular metabolism, reduces PET signal-to-noise ratio (SNR) and limits quantitative analysis. In this work, we propose an uncertainty-aware deep learning framework to improve the estimation of washout param eters in post-proton therapy PET, not only enabling accurate correction for washout effects, but also mapping intratumoral heterogeneity as a surrogate marker of tumor status and treatment response. We trained the models on Monte Carlo-simulated data from eight head-and-neck cancer patients, and tested them on four additional head-and-neck and one liver patient. Each patient was represented by 75 digital twins with distinct tumoral washout dynamics and imaged 15 min after treatment, when slow washout components dominate. We also in troduced washed-out maps, quantifying the contribution of medium and fast washout components to the loss in activity between the end of treatment and the start of PET imaging. Trained models significantly improved resolution and accuracy, reducing average absolute errors by 60% and 28% for washout rate and washed-out maps, respectively. For intratumoral regions as small as 5.1 mL, errors predominantly fell below thresholds for differentiating vascular status, and the models generalized across anatomical areas and acquisition delays. This tudy shows the potential of deep learning in post-proton therapy PET to non-invasively map washout kinet ics and reveal intratumoral heterogeneity, supporting dose verification, tumor characterization, and treatment personalization. The framework is available at https://github.com/pcabrales/ppw.git., This work was supported by Spanish MCIN/AEI/10.13039/ 501100011033 and the European Union NextGenerationEU/PRTR un der grants TED2021-130592B-I00 PROTOTWIN, PID2021-126998OB- I00 FASCINA, and PDC2022-133057-I00 3PET; by Comunidad de Madrid (Spain) under projects ASAP-CM P2022/BMD-7434 and TAU- CM PR47/21 (PRTR), and by Grupo de Física Nuclear (910059) at Universidad Complutense de Madrid. Additionally, Pablo Cabrales acknowledges support from the Complutense University Predoctoral Fellowship (CT15-23), funded by the Universidad Complutense de Madrid and Banco Santander, as well as from the Ayuda para estancias breves fellowship (EB33/24), funded by the Universidad Complutense de Madrid. M.Pérez-Liva was supported by the Program Ramón y Cajal RYC2021-032739-I, MCIN/AEI/10.13039/501100011033, by the European Union NextGenerationEU/PRTR, by INVENTOR PID2022- 137114OA-I00, and by LUNABRAIN (PR37/24 TEC-2024/TEC-43), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

Numerical Framework for Deep Learning Photoacoustics Image Reconstruction

  • G. Muñoz, Javier
  • Ariza-Carrasco, Alejandro
  • Díaz Cañadas, Bárbara
  • Cruza, Jorge F.
  • Elvira Segura, Luis
  • Camacho Sosa-Días, Jorge
  • L. Herraiz, Joaquín
  • Udías, José M.
  • Bertolet Reina, Alejandro
  • Alonso de Laciñana Cases, María
  • Rodriguez-Pardo, Jorge
  • Mariño Trillo, Eduardo
  • Laso García, Fernando
  • Ibáñez García, Paula
  • Pérez-Liva, Mailyn
23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026, 08-11 April 2026, London, UK. 4 páginas, 3 figuras., Photoacoustic imaging (PAI) combines pulsed optical excitation with ultrasound detection to produce high-contrast molecular and structural images of biological tissues. Conventional PAI systems rely on high-energy pulsed lasers that, despite their excellent performance, remain bulky, costly, and constrained by strict safety requirements, limiting their clinical translation. This work explores the use of pulsed semiconductor diodes as a compact, low-cost, and portable alternative. Although these sources exhibit lower peak power and reduced fluence, advanced computational approaches can compensate for these limitations. We present a numerical framework that integrates realistic optical and acoustic simulations with deep learning reconstruction to enhance image quality under low-fluence conditions. The framework accounts for light propagation, transducer response, and acoustic field effects to generate training datasets that accurately represent experimental conditions. The proposed hybrid reconstruction recovers fine vascular structures up to 3 cm in depth with 200μ m spatial resolution, demonstrating the potential of diode-based PAI systems for affordable and viable PAI., This work received support from: Programme Ramón y Cajal RYC2021-032739-I, funded by MCIN/AEI/10.13039/ 501100011033 and by the European Union “NextGenerationEU”/PRTR”; FASCINA (PID2021-126998OB-I00/), 3PET (PDC2022-133057-I00) from the Spanish Ministry of Science and Innovation (MCIN) AEI/10.13039/501100011033/ Unión Europea Next GenerationEU/PRTR, Prototwin Project (TED2021-349 130592B-I00) and Tecnologías Avanzadas para la Exploración del Universo y sus Componentes” (PR47/21 TAU-CM TAU- PRTR), funded by EU Resilient and NextGeneration funds; INVENTOR (PID2022-137114OA-I00);project LUNABRAIN-CM (TEC-2024/TEC-43), funded with € 1.026 million by the Comunidad de Madrid through the RD activities program in technologies, granted by Order 5696/2024., Peer reviewed
Docta Complutense
Publicación Artículo científico (JournalArticle)

Robustness of single- and dual-energy deep-learning-based scatter correction models on simulated and real chest X-rays

  • Freijo, Clara
  • López Herraiz, Joaquín
  • Arias Valcayo, Fernando
  • Ibáñez García, Paula Beatriz
  • Moreno, Gabriela
  • Villa Abaunza, Amaia
  • Udías Moinelo, José Manuel
Chest X-rays (CXRs) represent the first tool globally employed to detect cardiopulmonary pathologies. These acquisitions are highly affected by scattered photons due to the large field of view required. Scatter in CXRs introduces background in the images, which reduces their contrast. We developed three deep-learning-based models to estimate and correct scatter contribution to CXRs. We used a Monte Carlo (MC) ray-tracing model to simulate CXRs from human models obtained from CT scans using different configurations (depending on the availability of dual-energy acquisitions). The simulated CXRs contained the separated contribution of direct and scattered X-rays in the detector. These simulated datasets were then used as the reference for the supervised training of several NNs. Three NN models (single and dual energy) were trained with the MultiResUNet architecture. The performance of the NN models was evaluated on CXRs obtained, with an MC code, from chest CT scans of patients affected by COVID-19. The results show that the NN models were able to estimate and correct the scatter contribution to CXRs with an error of <5%, being robust to variations in the simulation setup and improving contrast in soft tissue. The single-energy model was tested on real CXRs, providing robust estimations of the scatter-corrected CXRs.
Docta Complutense
Publicación Artículo científico (JournalArticle)

Optimizing point source tracking in awake rat PET imaging: a comprehensive study of motion detection and best correction conditions

  • Arias Valcayo, Fernando
  • Galve Lahoz, Pablo
  • Udías Moinelo, José Manuel
  • Vaquero, Juan José
  • Desco, Manuel
  • Herraiz, Joaquín L.
  • López Herraiz, Joaquín
Preclinical PET animal studies require immobilization of the animal, typically accomplished through the administration of anesthesia, which may affect the radiotracer biodistribution. The use of (Formula presented.) F point sources attached to the rat head is one of the most promising methods for motion compensation in awake rat PET studies. However, the presence of radioactive markers may degrade image quality. In this study, we aimed to investigate the most favorable conditions for preclinical PET studies using awake rats with attached point sources. Firstly, we investigate the optimal activity conditions for the markers and rat-injected tracer using Monte Carlo simulations to determine the parameters of maximum detectability without compromising image quality. Additionally, we scrutinize the impact of delayed window correction for random events on marker detectability and overall image quality within these studies. Secondly, we present a method designed to mitigate the influence of rapid rat movements, which resulted in a medium loss of events of around 30%, primarily observed during the initial phase of the data acquisition. We validated our study with PET acquisitions from an awake rat within the acceptable conditions of activity and motion compensation parameters. This acquisition revealed an 8% reduction in resolution compared to a sedated animal, along with a 6% decrease in signal-to-noise ratio (SNR). These outcomes affirm the viability of our method for conducting awake preclinical brain studies.
Docta Complutense
Publicación Artículo científico (JournalArticle)

Reconstruction of multi-animal PET acquisitions with anisotropically variant PSF

  • Arias Valcayo, Fernando
  • Galve Lahoz, Pablo
  • López Herraiz, Joaquín
  • Vaquero, J J
  • Desco, M
  • Udías Moinelo, José Manuel
Among other factors such as random, attenuation and scatter corrections, uniform spatial resolution is key to performing accurate quantitative studies in Positron emission tomography (PET). Particularly in preclinical PET studies involving simultaneous acquisition of multiple animals, the degradation of image resolution due to the depth of interaction (DOI) effect far from the center of the Field of View (FOV) becomes a significant concern. In this work, we incorporated a spatially-variant resolution model into a real time iterative reconstruction code to obtain accurate images of multi-animal acquisition. We estimated the spatially variant point spread function (SV-PSF) across the FOV using measurements and Monte Carlo (MC) simulations. The SV-PSF obtained was implemented in a GPU-based Ordered subset expectation maximization (OSEM) reconstruction code, which includes scatter, attenuation and random corrections. The method was evaluated with acquisitions from two preclinical PET/CT scanners of the SEDECAL Argus family: a Derenzo phantom placed 2 cm off center in the 4R-SuperArgus, and a multi-animal study with 4 mice in the 6R-SuperArgus. The SV-PSF reconstructions showed uniform spatial resolution without significant increase in reconstruction time, with superior image quality compared to the uniform PSF model.
Docta Complutense
Publicación Artículo científico (JournalArticle)

Analytical positron range model for PET with cross-code Monte Carlo benchmarking

  • Paneque-Yunta, Robert J
  • Encina-Baranda, Nerea
  • Carter, Lukas M
  • Galve Lahoz, Pablo
  • Ibáñez García, Paula Beatriz
  • Abushab, Khaled M
  • Udías Moinelo, José Manuel
  • López Herraiz, Joaquín
Introduction. The positron range (PR) effect is a significant factor limiting spatial resolution in positron emission tomography (PET), particularly for high-resolution systems and non-standard isotopes. Objective. This study introduces a novel analytical model to accurately and rapidly describe PR distributions (PRd) for various PET radioisotopes to better include its effect in PET reconstruction algorithms. Approach. The proposed model explicitly incorporates the Coulomb repulsion effect, the multi-branch nature of certain beta+ emitters, and the scaling of PR with electronic density. To minimise bias, we used a histogram-free statistical method to derive the cumulative PRd from Monte Carlo (MC) simulated annihilation datasets, avoiding arbitrary histogram binning. A comparative analysis of PR estimates was conducted across three major MC radiation transport algorithm packages: PENELOPE (via PenEasy/PeneloPET), GEANT4 (via GATE), and EGS5 (via PHITS), revealing notable discrepancies between codes, versions, and input configurations, especially at short distances from the source. Main results. The new analytical model demonstrated an excellent reproduction of the simulated data for isotopes including 11C, 13N, 15O, 18F, 64Cu, 68Ga, 82Rb and 124I, achieving in general coefficients of determination (R2) greater than 0.995 and mean absolute percentage errors less than or similar to 20%. Compared to previous methods, our model provides a more accurate description of PRd at low distances and offers improved R2 values. Significance. This work provides a robust framework for generating accurate annihilation point spread function kernels, facilitating improved PR correction in quantitative Nuclear Medical Imaging and supporting research with diverse radioisotopes.
Docta Complutense
Publicación Artículo científico (JournalArticle)

MLTiming: pairwise-constrained learning for improving timing resolution in TOF-PET

  • Avellaneda, José A.
  • Sánchez Tembleque Verbo, Víctor
  • Rodríguez, Miriam Caballero
  • Fraile Prieto, Luis Mario
  • Udías Moinelo, José Manuel
  • López Herraiz, Joaquín
Accurate timing characterization of detected radiation events in Positron Emission Tomography (PET) provides Time-of-flight (TOF) information for detected coincidences, which improves the signal-to-noise ratio of the reconstructed images. In this work, we propose a method, called MLTiming, to train machine learning (ML) models to assign accurate timestamps to events measured in radiation detectors training using just the signals detected by the individual single-channel detectors, without the need of resorting to other detector signals nor simulations. We performed a proof-of-concept study using three different neural network (NN) architectures, including the recently proposed Kolmogorov-Arnold network (KAN), trained with event pairs constructed with the measured signal and a delayed version of it. The delay is used as a label in the training process. The proposed trained model was evaluated with coincidence data acquired with a 22Na point source using a pair of LaBr3(Ce) crystals in the shape of truncated cones coupled to fast photomultiplier tubes (PMTs). The detected γ-ray pulses were sampled at 5 Gsamples/s. A 3% improvement in Coincidence Time Resolution (CTR) over an ultra optimized genetic algorithm–enhanced digital CFD (D-CFD) method (246±3 ps to 238±1 ps) was achieved. The possibility of training ML models for each individual detector using only the information contained in their own detected events may facilitate the use of ML models for large and medium-sized scanners, as well as other research areas. MLTiming is available at https://github.com/wjosefw/MLTiming.git.
Docta Complutense
Publicación Artículo científico (JournalArticle)

PROTOTWIN-PET: A deep learning and GPU-based workflow for dose verification in proton therapy with PET

  • Cabrales, Pablo
  • Onecha, Víctor V.
  • Izquierdo-García, David
  • Fraile Prieto, Luis Mario
  • Udías Moinelo, José Manuel
  • López Herraiz, Joaquín
In proton therapy (PT), accurate dose delivery verification is critical for detecting treatment plan deviations. This can be achieved by imaging activated positron emitters with a positron emission tomography (PET) acquisition and converting the data into a delivered dose image. This work presents PROTOTWIN-PET (PROTOn therapy digital TWIN models for dose verification with PET), a patient-specific, deep learning (DL) and GPU-based workflow for 3-D dose verification. The proposed workflow generates a dataset of simulated, realistic 3-D PET and dose pairs that reflect possible clinical deviations in patient positioning and physical parameters. Using this dataset, a DL model is trained to estimate the delivered dose from the PET image, incorporating a deviation-predicting branch (DPB) to estimate patient positioning deviations. PROTOTWIN-PET is demonstrated on a two-field oropharyngeal cancer treatment plan, estimating the delivered dose in milliseconds with an average mean relative error of 0.6% and near-perfect gamma passing rates (3 mm, 3%). Positioning deviations are estimated on average within a tenth of a millimeter and degree. PROTOTWIN-PET can be implemented within the one-day interval between the plan CT acquisition and the first treatment session, potentially enabling timely treatment plan adjustments and maximizing the precision of PT. PROTOTWIN-PET is available at github.com/pcabrales/prototwin-pet.git.