Physics in medicine and biology最新文献

{"title":"Improved pharmacokinetic parameter estimation from DCE-MRI via spatial-temporal information-driven unsupervised learning.","authors":"Xinyi He, Lu Wang, Qizhi Yang, Jiechao Wang, Zhen Xing, Dairong Cao, Congbo Cai, Shuhui Cai","doi":"10.1088/1361-6560/ae0aaf","DOIUrl":"10.1088/1361-6560/ae0aaf","url":null,"abstract":"<p><p><i>Objective.</i>Pharmacokinetic (PK) parameters derived from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provide quantitative characterization of tissue perfusion and permeability. However, existing deep learning methods for PK parameter estimation rely on either temporal or spatial features alone, overlooking the integrated spatial-temporal characteristics of DCE-MRI data. This study aims to remove this barrier by fully leveraging the spatial and temporal information to improve parameter estimation.<i>Approach.</i>A spatial-temporal information-driven unsupervised deep learning method (STUDE) was proposed. STUDE combines convolutional neural networks (CNNs) and a customized Vision Transformer to separately capture spatial and temporal features, enabling comprehensive modeling of contrast agent dynamics and tissue heterogeneity. Besides, a spatial-temporal attention feature fusion module was proposed to enable adaptive focus on both dimensions for more effective feature fusion. Moreover, the extended Tofts model imposed physical constraints on PK parameter estimation, enabling unsupervised training of STUDE. The accuracy and diagnostic value of STUDE was compared with the orthodox non-linear least squares (NLLS) and representative deep learning-based methods (i.e. gated recurrent unit, convolutional neural network, U-Net, and VTDCE-Net) on a numerical brain phantom and 87 glioma patients, respectively.<i>Main results.</i>On the numerical brain phantom, STUDE produced PK parameter maps with the lowest systematic and random errors even under low signal-to-noise ratio (SNR) conditions (SNR = 10 dB). On glioma data, STUDE generated parameter maps with reduced noise compared to NLLS and demonstrated superior structural clarity compared to other methods. Furthermore, STUDE outshined all other methods in the identification of glioma isocitrate dehydrogenase mutation status, achieving the area under the curve (AUC) values at 0.840 and 0.908 for the receiver operating characteristic curves of<i>K</i>^transand V_e, respectively. A combination of all PK parameters improved AUC to 0.926.<i>Significance.</i>STUDE advances spatial-temporal information-driven and physics-informed learning for precise PK parameter estimation, demonstrating its potential clinical significance.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145131731","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance evaluation of a multiplexing circuit combined with ASIC readout for cost-effective brain PET imaging.","authors":"Fiammetta Pagano, Francis Loignon-Houle, David Sanchez, Julio Barberá, Jorge Alamo, Ezzat Elmoujarkach, Nicolas A Karakatsanis, Sadek A Nehmeh, Antonio J Gonzalez","doi":"10.1088/1361-6560/ae05ad","DOIUrl":"10.1088/1361-6560/ae05ad","url":null,"abstract":"<p><p><i>Objective.</i>A key challenge in PET systems is collecting large amounts of data with the most accurate information-time, energy, and position-to produce high-resolution images while limiting the number of channels to reduce costs and improve data collection efficiency. The new ultra-high-performance brain (UHB) scanner under development aims to tackle this issue, using a semi-monolithic detector that combines pixelated arrays and monolithic designs, along with signal multiplexing techniques.<i>Approach.</i>We assessed the time, energy, and positioning performance of the multiplexing circuit (summing signals along rows and columns) and compared it to the standard readout, both using TOFPET2 ASIC.<i>Main Results.</i>While time resolution worsens by about 15%, energy and positioning resolution-more crucial in small diameter scanners-are unaffected by signal summation. Overall, a pair of detector modules (2 × 2 arrays each) features an energy resolution of 16.9 ± 1.3% and 405 ± 29 ps coincidence time resolution. Positioning accuracy-estimated using multilayer perceptron neural network-is 1.9 ±0.4 mm and 3.0 ±0.7 mm along the monolithic and depth-of-interaction direction, respectively.<i>Significance.</i>This study demonstrates that this channel reduction readout effectively maintains high performance while allowing for reduced costs and enhanced scalability.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145034103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cross-modality transformer model leveraging DCE-MRI and pathological images for predicting pathological complete response and lymph node metastasis in breast cancer.","authors":"Ming Fan, Zhiwei Zhu, Zhou Yu, Jiaojiao Du, Sangma Xie, Xiang Pan, Shujun Chen, Lihua Li","doi":"10.1088/1361-6560/ae077c","DOIUrl":"10.1088/1361-6560/ae077c","url":null,"abstract":"<p><p><i>Objective.</i>Pathological diagnosis remains the gold standard for diagnosing breast cancer and is highly accurate and sensitive, which is crucial for assessing pathological complete response (pCR) and lymph node metastasis (LNM) following neoadjuvant chemotherapy (NACT). Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a noninvasive technique that provides detailed morphological and functional insights into tumors. The optimal complementarity of these two modalities, particularly in situations where one is unavailable, and their integration to enhance therapeutic predictions have not been fully explored.<i>Approach.</i>To this end, we propose a cross-modality image transformer (CMIT) model designed for feature synthesis and fusion to predict pCR and LNM in breast cancer. This model enables interaction and integration between the two modalities via a transformer's CA module. A modality information transfer module is developed to produce synthetic pathological image features (sPIFs) from DCE-MRI data and synthetic DCE-MRI features (sMRIs) from pathological images. During training, the model leverages both real and synthetic imaging features to increase the predictive performance. In the prediction phase, the synthetic imaging features are fused with the corresponding real imaging feature to make predictions.<i>Main results.</i>The experimental results demonstrate that the proposed CMIT model, which integrates DCE-MRI with sPIFs or histopathological images with sMRI, outperforms (with AUCs of 0.809 and 0.852, respectively) the use of MRI or pathological images alone in predicting the pCR to NACT. Similar improvements were observed in LNM prediction. For LNM prediction, the DCE-MRI model's performance improved from an area under the curve (AUC) of 0.637-0.712, while the DCE-MRI-guided histopathological model achieved an AUC of 0.792.<i>Significance.</i>Notably, our proposed model can predict treatment response effectively via DCE-MRI, regardless of the availability of actual histopathological images.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145075951","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Metal artifact reduction and contrast agent performance as a function of source spectral shape.","authors":"Jacob D Aubrey, Emmett P Perkinson, Ge Wang, Peter J Bonitatibus","doi":"10.1088/1361-6560/ae0be9","DOIUrl":"10.1088/1361-6560/ae0be9","url":null,"abstract":"<p><p><i>Objective.</i>Empirically map the x-ray source of a Medipix All Resolution Scanner (MARS) photon counting CT (PCCT) with Cu and Sn pre-filters, assess metal artifact reduction (MAR) capabilities of these pre-filters, and measure pre-filtration impact on contrast performance (CP) of FDA approved iopromide and experimental tantalum oxide nanoparticles (TaO<i>_x</i>NPs).<i>Approach.</i>The x-ray source of a MARS-PCCT system was empirically mapped with no pre-filtration, seven Cu filters (0.3-2.1 mm), and seven Sn filters (0.15-1.05 mm). A phantom with inserts containing water, lipid, iopromide, TaO<i>_x</i>NPs, and metal was scanned with no pre-filtration and Cu and Sn pre-filters. Insert noise, signal, contrast-to-pooled-noise ratio (CPNR), and a fast Fourier transform artifact metric (FFTAM) were calculated for each filter to quantify MAR and CP.<i>Main results.</i>Thick filters for Cu and Sn shifted mean energy of the unfiltered x-ray source (47.9 keV) by 19.2 and 23.4 keV, respectively. Thick filtration, 2.1 mm Cu and 0.6 mm Sn, greatly reduced noise (up to 74%) and FFTAM (up to 71%) for all inserts and energy bins. Thin filtration, 0.3 mm Cu and 0.15 mm Sn, also reduced noise (up to 47%) and FFTAM (up to 41%). In most cases, iopromide lost significant contrast (up to 50%). TaO<i>_x</i>NPs also lost contrast, though to a lesser extent (up to 38%). Pre-filtration improved image efficacy (i.e. CPNR), especially for TaO<i>_x</i>(up to 61%).<i>Significance.</i>By empirically mapping the source spectrum of a MARS-PCCT system with pre-filters, valuable information was gathered about photon flux distribution and detector artifacts; these findings will prove insightful for applications such as energy binning for effective material decomposition. Furthermore, this information will potentially guide clinical MAR development, most notably for the MARS Extremity 5 × 120 recently deployed for first-in-human trials. Lastly, TaO<i>_x</i>NPs were shown to be more compatible than iopromide with spectral shaping.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145150440","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessment of the impact of the nuclear properties of<i>β</i>^--emitting radionuclides on the dosimetry of two radiopharmaceuticals with distinct pharmacokinetics.","authors":"Laura De Nardo, Samuele De Zan, Kevin J H Allen, Gulliermina Ferro-Flores, Ekaterina Dadachova, Laura Meléndez-Alafort","doi":"10.1088/1361-6560/ae07a2","DOIUrl":"10.1088/1361-6560/ae07a2","url":null,"abstract":"<p><p><i>Objective.</i>The aim of this study was to evaluate the impact of the nuclear properties of six<i>β</i>^--emitting radionuclides (⁴⁷Sc,⁶⁷Cu,¹¹¹Ag,¹⁶¹Tb,¹⁷⁷Lu, and¹⁸⁸Re) on the dosimetric outcomes of two tumour-targeting radiopharmaceuticals (RPs), with distinct pharmacokinetics: the peptide DOTA-folate conjugate cm09 and the monoclonal antibody HuM195. The study specifically focused on assessing the radiation-absorbed doses in organs and tumours, as well as comparing the efficacy and safety of the twelve RPs for targeted radionuclide therapy (TRT).<i>Approach.</i>Murine biodistribution data for both RPs were scaled to adult human models to determine biological residence times and the number of disintegrations in source organs and tumours. Dosimetric estimations were performed using OLINDA and MIRDCell software, considering different tumour sizes and organ-specific radiation exposure for both male and female phantoms.<i>Main results.</i>Significant differences in organ and tumour dosimetry were found across the considered radionuclides and tumour-targeting agents, attributable to the nuclear properties of the radionuclides and the RP pharmacokinetics. PFP-HuM195 labelled with¹⁶¹Tb and¹¹¹Ag demonstrated efficient dose delivery to tumour from 1-10 mm, but also higher organ-absorbed doses per unit of injected activity than other labelling radionuclides. Cm09 exhibited less variability in tumour absorbed dose as the labelling radionuclide varied, but also produced much higher kidney absorbed doses than PFP-HuM195. Normalising to the same tumour absorbed dose showed that¹⁷⁷Lu and¹⁶¹Tb are the safer options for treating small tumours (2.7-12.4 mm) with both RPs. These results demonstrate that the choice of radionuclide has a significant impact on both therapeutic efficacy and organ safety.<i>Significance.</i>This research demonstrates that selecting the appropriate radionuclide for TRT can optimise therapeutic outcomes while minimising radiation exposure to healthy tissues. The findings contribute to advancing personalised TRT approaches by considering RP-specific pharmacokinetics and radionuclide characteristics, paving the way for more effective cancer treatments.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145075972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Brain networks involved in cancer treatment response: insights from¹⁸F-FDG PET scans.","authors":"Mauro Namías, Matej Perovnik, Daniel Huff, Carolina Tinetti, María Eugenia Azar, Katja Strašek, Nežka Hribernik, Martina Reberšek, Andrej Studen, Gabriel Bruno, Robert Jeraj","doi":"10.1088/1361-6560/ae0beb","DOIUrl":"https://doi.org/10.1088/1361-6560/ae0beb","url":null,"abstract":"<p><p><i>Objective.</i>To determine whether pre-treatment brain metabolic network patterns measured with¹⁸F-FDG PET are associated with treatment response and survival in cancer patients.<i>Approach.</i>Exploratory retrospective study of two independent cohorts: stage III breast cancer patients treated with neoadjuvant chemotherapy and stage IV melanoma patients treated with anti-PD-1 immunotherapy. Metabolic brain network scores were derived from pre-treatment¹⁸F-FDG PET scans and evaluated for their ability to stratify good versus poor responders using ROC analysis (AUC). Longitudinal changes in network scores were assessed across follow-up, and progression-free survival (PFS) and overall survival (OS) analyses were performed in the melanoma cohort.<i>Main results.</i>Specific brain networks were associated with treatment outcome; the cognition/language network was the strongest predictor (AUC > 0.84 for distinguishing good vs. poor responders in both cohorts). Good responders showed lower cognition/language scores than poor responders and healthy controls. Longitudinally, cognition/language scores remained stable in good responders, while poor responders exhibited a gradual convergence toward the scores observed in good responders. In the melanoma cohort, lower cognition/language scores were significantly associated with longer PFS and OS.<i>Significance.</i>These findings indicate that metabolic brain network patterns, particularly the cognition/language network, may serve as noninvasive biomarkers linked to treatment efficacy and survival in oncology. The results support a possible complex interaction between brain metabolism, immune response, and clinical outcomes. Key limitations include the retrospective design and lack of direct immune-function and psychometric measures; prospective, multimodal studies are needed to validate these observations and elucidate underlying mechanisms.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":"70 20","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145239124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"One scan, many stories: deep learning for signal separation in multi-tracer PET imaging.","authors":"Luigi Manco, Luca Urso, Luca Filippi","doi":"10.1088/1361-6560/ae02db","DOIUrl":"10.1088/1361-6560/ae02db","url":null,"abstract":"","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144993022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Insertion of hepatic lesions into clinical photon-counting-detector CT projection data.","authors":"Hao Gong, Shravani Kharat, Jarod Wellinghoff, Ahmed Omar El Sadaney O Rabie, Joel G Fletcher, Shaojie Chang, Lifeng Yu, Shuai Leng, Cynthia H McCollough","doi":"10.1088/1361-6560/ae0975","DOIUrl":"10.1088/1361-6560/ae0975","url":null,"abstract":"<p><p><i>Objective.</i>To facilitate task-driven image quality assessment of lesion detectability in clinical photon-counting-detector CT (PCD-CT), it is desired to have patient image data with known pathology and precise annotation. Standard patient case collection and reference standard establishment are time- and resource-intensive. To mitigate this challenge, we aimed to develop a projection-domain lesion insertion framework that efficiently creates realistic patient cases by digitally inserting real radiopathologic features into patient PCD-CT images.<i>Approach.</i>This framework used an artificial-intelligence-assisted semi-automatic annotation to generate digital lesion models from real lesion images. The x-ray energy for commercial beam-hardening correction in PCD-CT system was estimated and used for calculating multi-energy forward projections of these lesion models at different energy thresholds. Lesion projections were subsequently added to patient projections from PCD-CT exams. The modified projections were reconstructed to form realistic lesion-present patient images, using the CT manufacturer's offline reconstruction software. Image quality was qualitatively and quantitatively validated in phantom scans and patient cases with liver lesions, using visual inspection, CT number accuracy, structural similarity index (SSIM), and radiomic feature analysis. Statistical tests were performed using Wilcoxon signed rank test.<i>Main results.</i>No statistically significant discrepancy (<i>p</i>> 0.05) of CT numbers was observed between original and re-inserted tissue- and contrast-media-mimicking rods and hepatic lesions (mean ± standard deviation): rods 0.4 ± 2.3 HU, lesions -1.8 ± 6.4 HU. The original and inserted lesions showed similar morphological features at original and re-inserted locations: mean ± standard deviation of SSIM 0.95 ± 0.02. Additionally, the corresponding radiomic features presented highly similar feature clusters with no statistically significant differences (<i>p</i>> 0.05).<i>Significance.</i>The proposed framework can generate patient PCD-CT exams with realistic liver lesions using archived patient data and lesion images. It will facilitate systematic evaluation of PCD-CT systems and advanced reconstruction and post-processing algorithms with target pathological features.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12492340/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145092285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated computation of detectability index and generation of contrast-detail curves for CT protocol optimization.","authors":"Choirul Anam, Ariij Naufal, Heri Sutanto, Kusworo Adi, Chai Hong Yeong, Geoff Dougherty","doi":"10.1088/1361-6560/ae0ab0","DOIUrl":"10.1088/1361-6560/ae0ab0","url":null,"abstract":"<p><p><i>Objective.</i>The aim of this study was to develop an automatic method for generating a detectability index (<i>d'</i>)-based contrast-detail (<i>C</i>-<i>D</i>) curve across multiple object sizes and contrasts, and to evaluate its performance under varying tube current settings and reconstruction filter types.<i>Approach.</i>To compute<i>d'</i>for a given object size and contrast, the task-transfer function and noise power spectrum were obtained from ACR 464 computed tomography (CT) phantom images acquired at tube currents of 80, 120, 160 and 200 mA, using Edge, Lung, and Soft filter types. The task objects were varied in size (1-15 mm) and contrast levels (1-15 HU) with both flat and Gaussian signal types. For each defined task object,<i>d'</i>was calculated using a non-prewhitening model observer. This process was iterated for every predefined task function across multiple object sizes and contrasts, resulting in a<i>d'</i>map corresponding to the synthetic low-contrast images. A<i>C</i>-<i>D</i>curve was then generated using a<i>d'</i>cut-off value defined by the user. For comparison, a separate<i>C</i>-<i>D</i>curve was generated based on visual assessment by five human observers (HOs).<i>Main results.</i>The automated method successfully computed<i>d'</i>values and arranged synthetic low-contrast images into a grid according to object size and contrast.<i>C</i>-<i>D</i>curves using<i>d'</i>cut-off values of 3 or 4 most closely reflected HOs performance. For tube current variations, increasing the current led to higher detectability. For filter type variations, the Lung filter resulted in relatively lower detectability compared to the Edge and Soft filters.<i>Significance</i>. An automated method to calculate<i>d'</i>across a wide range of object sizes and contrasts, and to generate a<i>d'</i>-based<i>C</i>-<i>D</i>curve for CT protocol optimization was developed. The results were consistent with HO trends and effectively captured detectability changes across different imaging parameters.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145131700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning-enhanced 3D real-time photoacoustic imaging using experimental ground truths obtained from fluctuation imaging.","authors":"Ivana Falco, Godefroy Guillaume, Maxime Henry, Véronique Josserand, Emmanuel Bossy, Bastien Arnal","doi":"10.1088/1361-6560/ae0f70","DOIUrl":"https://doi.org/10.1088/1361-6560/ae0f70","url":null,"abstract":"<p><p>3D conventional photoacoustic (PA) imaging often suffers from visibility artifacts caused by the limited bandwidth and constrained viewing angles of ultrasound transducers, as well as the use of sparse arrays. PA fluctuation imaging (PAFI), which leverages signal variations due to blood flow, compensates for these visibility artifacts at the cost of temporal resolution. Deep learning (DL)--based photoacoustic image enhancement has previously demonstrated strong potential for improved reconstruction at a high temporal resolution. However, generating an experimental training dataset remains problematic. &#xD;Herein, we propose creating an experimental training dataset based on single-shot 3D PA images (input) and corresponding PAFI images (ground truth) of chicken embryo vasculature, which is used to train a 3D ResU-Net neural network.&#xD;The trained DL-PAFI network predictions on new experimental test images reveal effective improvement in visibility and contrast. We observe, however, that the output image resolution is lower than that of PAFI. Importantly, incorporating only experimental data into training already yields a good performance, while pre-training with simulated examples improves the overall accuracy. Additionally, we demonstrate the feasibility of real-time rendering and present preliminary in vivo predictions in mice, generated by the network trained exclusively on chicken embryo vasculature. These findings suggest the potential for achieving real-time, artifact-free 3D PA imaging with sparse arrays, adaptable to various in vivo applications.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145225684","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}