Physics in medicine and biology最新文献

{"title":"Syn2Real: synthesis of CT image ring artifacts for deep learning-based correction.","authors":"Dennis Hein, Staffan Holmin, Vladimir Prochazka, Zhye Yin, Mats Danielsson, Mats Persson, Ge Wang","doi":"10.1088/1361-6560/adad2c","DOIUrl":"10.1088/1361-6560/adad2c","url":null,"abstract":"<p><p><i>Objective</i>. We strive to overcome the challenges posed by ring artifacts in x-ray computed tomography (CT) by developing a novel approach for generating training data for deep learning-based methods. Training such networks require large, high quality, datasets that are often generated in the data domain, time-consuming and expensive. Our objective is to develop a technique for synthesizing realistic ring artifacts directly in the image domain, enabling scalable production of training data without relying on specific imaging system physics.<i>Approach</i>. We develop 'Syn2Real,' a computationally efficient pipeline that generates realistic ring artifacts directly in the image domain. To demonstrate the effectiveness of our approach, we train two versions of UNet, vanilla and a high capacity version with self-attention layers that we call UNetpp, withℓ2and perceptual losses, as well as a diffusion model, on energy-integrating CT images with and without these synthetic ring artifacts.<i>Main Results.</i>Despite being trained on conventional single-energy CT images, our models effectively correct ring artifacts across various monoenergetic images, at different energy levels and slice thicknesses, from a prototype photon-counting CT system. This generalizability validates the realism and versatility of our ring artifact generation process.<i>Significance.</i>Ring artifacts in x-ray CT pose a unique challenge to image quality and clinical utility. By focusing on data generation, our work provides a foundation for developing more robust and adaptable ring artifact correction methods for pre-clinical, clinical and other CT applications.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143024013","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":"Towards linking histological changes to liver viscoelasticity: a hybrid analytical-computational micromechanics approach.","authors":"Haritya Shah, Murthy N Guddati","doi":"10.1088/1361-6560/adaad3","DOIUrl":"10.1088/1361-6560/adaad3","url":null,"abstract":"<p><p>Motivated by elastography that utilizes tissue mechanical properties as biomarkers for liver disease, with the eventual objective of quantitatively linking histopathology and bulk mechanical properties, we develop a micromechanical modeling approach to capture the effects of fat and collagen deposition in the liver. Specifically, we utilize computational homogenization to convert the microstructural changes in hepatic lobule to the effective viscoelastic modulus of the liver tissue, i.e. predict the bulk material properties by analyzing the deformation of repeating unit cell. The lipid and collagen deposition is simulated with the help of ad hoc algorithms informed by histological observations. Collagen deposition is directly included in the computational model, while composite material theory is used to convert fat content to the microscopic mechanical properties, which in turn is included in the computational model. The results illustrate the model's ability to capture the effect of both fat and collagen deposition on the viscoelastic moduli and represents a step towards linking histopathological changes in the liver to its bulk mechanical properties, which can eventually provide insights for accurate diagnosis with elastography.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11829796/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143010063","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":"On the partial volume effect in magnetic particle imaging.","authors":"Hayden J Good, Toby Sanders, Andrii Melnyk, A Rahman Mohtasebzadeh, Eric Daniel Imhoff, Patrick Goodwill, Carlos M Rinaldi-Ramos","doi":"10.1088/1361-6560/ada417","DOIUrl":"https://doi.org/10.1088/1361-6560/ada417","url":null,"abstract":"<p><p><i>Objective.</i>Magnetic particle imaging (MPI) is an emerging tomographic 'hot spot' imaging modality with potential to visualize superparamagnetic iron oxide nanoparticle tracer distributions with high sensitivity and quantitative accuracy. MPI shares many similarities with positron emission tomography (PET), where the partial volume effect (PVE) can result in signal under- and over-quantification due to spill-over of signal arising from limited resolution. While the PVE has been alluded to in the MPI literature it has not been previously studied nor characterized. The objective of this study was to systematically characterize this PVE in MPI.<i>Approach.</i>This contribution characterizes the PVE using models of varying size and shape filled with a uniform concentration of tracer. The effect of object size on signal distribution was analyzed after application of a new image post-processing filter.<i>Main results.</i>As object size increased, signal distribution increased to a maximum signal value independent of object geometry and proportional to tracer concentration. Furthermore, for small objects with characteristic dimensions below the resolution of the tracer at the scanning conditions used, signal suppression was observed. These results are consistent with foundational observations of PVE in PET, suggesting that approaches to overcome the PVE in PET may be applicable to MPI.<i>Significance.</i>This finding has significant impact on the MPI field by demonstrating the presence of the PVE phenomenon that can directly influence imaging results.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":"70 4","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143189945","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":"Proton arc therapy plan optimization with energy layer pre-selection driven by organ at risk sparing and delivery time.","authors":"Sophie Wuyckens, Guillaume Janssens, Macarena Chocan Vera, Johan Sundström, Dario Di Perri, Edmond Sterpin, Kevin Souris, John A Lee","doi":"10.1088/1361-6560/adad2d","DOIUrl":"10.1088/1361-6560/adad2d","url":null,"abstract":"<p><p><i>Objective.</i>As proton arc therapy (PAT) approaches clinical implementation, optimizing treatment plans for this innovative delivery modality remains challenging, especially in addressing arc delivery time. Existing algorithms for minimizing delivery time are either optimal but computationally demanding or fast but at the expense of sacrificing many degrees of freedom. In this study, we introduce a flexible method for pre-selecting energy layers (EL) in PAT treatment planning before the actual robust spot weight optimization.<i>Approach.</i>Our EL pre-selection method employs metaheuristics to minimize a bi-objective function, considering a dynamic delivery time proxy and tumor geometrical coverage penalized as a function of selected organs-at-risk crossing. It is capable of parallelizing multiple instances of the problem. We evaluate the method using three different treatment sites, providing a comprehensive dosimetric analysis benchmarked against dynamic proton arc plans generated with early energy layer selection and spot assignment (ELSA) and IMPT plans in RayStation TPS.<i>Result.</i>The algorithm efficiently generates Pareto-optimal EL pre-selections in approximately 5 min. Subsequent PAT treatment plans derived from these selections and optimized within the TPS, demonstrate high-quality target coverage, achieving a high conformity index, and effective sparing of organs at risk. These plans meet clinical goals while achieving a 20%-40% reduction in delivery time compared to ELSA plans.<i>Significance.</i>The proposed algorithm offers speed and efficiency, producing high-quality PAT plans by placing proton arc sectors to efficiently reduce delivery time while maintaining good target coverage and healthy tissues sparing.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143024009","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":"Optimizing energy threshold selection for low-concentration contrast agent quantification in small animal photon-counting CT.","authors":"Xiaoyu Hu, Yuncheng Zhong, Xun Jia, Kai Yang","doi":"10.1088/1361-6560/adac9e","DOIUrl":"10.1088/1361-6560/adac9e","url":null,"abstract":"<p><p><i>Objective.</i>Gold nanoparticles (GNPs) are widely used for biological research and applications. The in-vivo concentration of GNPs is usually low due to biological safety concerns, thus posing a challenge for imaging. This work investigates on optimal energy threshold selection in photon-counting detector(PCD)-based CT (PCCT) for the quantification of low-concentration GNPs.<i>Approach.</i>We derived the mathematical expression of the upper bound of the material decomposition error in the gold image. Comprehensive simulations were implemented for cylindrical phantom with inserts of different GNP concentrations. CT scans of this phantom were simulated with a 140 kVp x-ray beam under a realistic pre-clinical CT dose range. The PCD energy thresholds from 30 to 110 keV were enumerated for 2,3-channel PCCT and the optimal energy thresholds were determined by searching for the lowest decomposition error.<i>Main results.</i>The optimal energy threshold(s) to minimize the decomposition error in gold image was 44 keV for the 2-channel PCCT and{34,40}keV for the 3-channel case. Numerical results also validated the derived upper bounds of the decomposition error.<i>Significance.</i>This work addressed the need for selecting appropriate energy thresholds for accurate quantification of contrast agent distributions in pre-clinical PCCT. Both the analytical expression of the upper bound of material decomposition error and simulation results showed that the balanced consideration on photon counting noise levels and the numerical properties of the decomposition matrix is required in selecting the appropriate energy thresholds to achieve the most accurate material decomposition.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11892678/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143010013","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":"Self-supervised parametric map estimation for multiplexed PET with a deep image prior.","authors":"Bolin Pan, Paul K Marsden, Andrew J Reader","doi":"10.1088/1361-6560/ada717","DOIUrl":"10.1088/1361-6560/ada717","url":null,"abstract":"<p><p>Multiplexed positron emission tomography (mPET) imaging allows simultaneous observation of physiological and pathological information from multiple tracers in a single PET scan. Although supervised deep learning has demonstrated superior performance in mPET image separation compared to purely model-based methods, acquiring large amounts of paired single-tracer data and multi-tracer data for training poses a practical challenge and needs extended scan durations for patients. In addition, the generalisation ability of the supervised learning framework is a concern, as the patient being scanned and their tracer kinetics may potentially fall outside the training distribution. In this work, we propose a self-supervised learning framework based on the deep image prior (DIP) for mPET image separation using just one dataset. In particular, we integrate the multi-tracer compartmental model into the DIP framework to estimate the parametric maps of each tracer from the measured dynamic dual-tracer activity images. Consequently, the separated dynamic single-tracer activity images can be recovered from the estimated tracer-specific parametric maps. In the proposed method, dynamic dual-tracer activity images are used as the training label, and the static dual-tracer image (reconstructed from the same patient data from the start to the end of acquisition) is used as the network input. The performance of the proposed method was evaluated on a simulated brain phantom for dynamic dual-tracer [¹⁸F]FDG+[¹¹C]MET activity image separation and parametric map estimation. The results demonstrate that the proposed method outperforms the conventional voxel-wise multi-tracer compartmental modeling method (vMTCM) and the two-step method DIP-Dn+vMTCM (where dynamic dual-tracer activity images are first denoised using a U-net within the DIP framework, followed by vMTCM separation) in terms of lower bias and standard deviation in the separated single-tracer images and also for the estimated parametric maps for each tracer, at both voxel and ROI levels.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142952995","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":"CoReSi: a GPU-based software for Compton camera reconstruction and simulation in collimator-free SPECT.","authors":"Vincent Lequertier, Étienne Testa, Voichiţa Maxim","doi":"10.1088/1361-6560/adaacc","DOIUrl":"10.1088/1361-6560/adaacc","url":null,"abstract":"<p><p><i>Objective.</i>Compton cameras (CCs) are imaging devices that may improve observation of sources of<i>γ</i>photons. The images are obtained by solving a difficult inverse problem. We present CoReSi, a Compton reconstruction and simulation software implemented in Python and powered by PyTorch to leverage multi-threading and to easily interface with image processing and deep learning algorithms. The code is mainly dedicated to medical imaging and near-field experiments where images are reconstructed in 3D.<i>Approach.</i>The code was developed over several years in C++, with the initial version being proprietary. We have since redesigned and translated it into Python, adding new features to improve its adaptability and performances. This paper reviews the literature on CC mathematical models, explains the implementation strategies we have adopted and presents the features of CoReSi.<i>Main results.</i>The code includes state-of-the-art mathematical models from the literature, from the simplest, which allow limited knowledge of the sources, to more sophisticated ones with a finer description of the physics involved. It offers flexibility in defining the geometry of the CC and the detector materials. Several identical cameras can be considered at arbitrary positions in space. The main functions of the code are dedicated to the computation of the system matrix, leading to the forward and backward projector operators. These are the cornerstones of any image reconstruction algorithm. A simplified Monte Carlo data simulation function is provided to facilitate code development and fast prototyping.<i>Significance.</i>As far as we know, there is no open source code for CC reconstruction, except for MEGAlib, which is mainly dedicated to astronomy applications. This code aims to facilitate research as more and more teams from different communities such as applied mathematics, electrical engineering, physics, medical physics get involved in CC studies. Implementation with PyTorch will also facilitate interfacing with deep learning algorithms.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143009981","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":"A robust auto-contouring and data augmentation pipeline for adaptive MRI-guided radiotherapy of pancreatic cancer with a limited dataset.","authors":"Mehdi Shojaei, Björn Eiben, Jamie R McClelland, Simeon Nill, Alex Dunlop, Arabella Hunt, Brian Ng-Cheng-Hin, Uwe Oelfke","doi":"10.1088/1361-6560/adabac","DOIUrl":"10.1088/1361-6560/adabac","url":null,"abstract":"<p><p><i>Objective.</i>This study aims to develop and evaluate a fast and robust deep learning-based auto-segmentation approach for organs at risk in MRI-guided radiotherapy of pancreatic cancer to overcome the problems of time-intensive manual contouring in online adaptive workflows. The research focuses on implementing novel data augmentation techniques to address the challenges posed by limited datasets.<i>Approach.</i>This study was conducted in two phases. In phase I, we selected and customized the best-performing segmentation model among ResU-Net, SegResNet, and nnU-Net, using 43 balanced 3DVane images from 10 patients with 5-fold cross-validation. Phase II focused on optimizing the chosen model through two advanced data augmentation approaches to improve performance and generalizability by increasing the effective input dataset: (1) a novel structure-guided deformation-based augmentation approach (sgDefAug) and (2) a generative adversarial network-based method using a cycleGAN (GANAug). These were compared with comprehensive conventional augmentations (ConvAug). The approaches were evaluated using geometric (Dice score, average surface distance (ASD)) and dosimetric (D2% and D50% from dose-volume histograms) criteria.<i>Main results.</i>The nnU-Net framework demonstrated superior performance (mean Dice: 0.78 ± 0.10, mean ASD: 3.92 ± 1.94 mm) compared to other models. The sgDefAug and GANAug approaches significantly improved model performance over ConvAug, with sgDefAug demonstrating slightly superior results (mean Dice: 0.84 ± 0.09, mean ASD: 3.14 ± 1.79 mm). The proposed methodology produced auto-contours in under 30 s, with 75% of organs showing less than 1% difference in D2% and D50% dose criteria compared to ground truth.<i>Significance.</i>The integration of the nnU-Net framework with our proposed novel augmentation technique effectively addresses the challenges of limited datasets and stringent time constraints in online adaptive radiotherapy for pancreatic cancer. Our approach offers a promising solution for streamlining online adaptive workflows and represents a substantial step forward in the practical application of auto-segmentation techniques in clinical radiotherapy settings.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11783596/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143009976","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":"Augmenting motion artifacts to enhance auto-contouring of complex structures in cone-beam computed tomography imaging.","authors":"Angelo Genghi, Mário João Fartaria, Anna Siroki-Galambos, Simon Flückiger, Fernando Franco, Adam Strzelecki, Pascal Paysan, Julius Turian, Zhen Wu, Luca Boldrini, Giuditta Chiloiro, Thomas Costantino, Justin English, Tomasz Morgas, Thomas Coradi","doi":"10.1088/1361-6560/ada0a0","DOIUrl":"https://doi.org/10.1088/1361-6560/ada0a0","url":null,"abstract":"<p><p><i>Objective</i>. To develop an augmentation method that simulates cone-beam computed tomography (CBCT) related motion artifacts, which can be used to generate training-data to increase the performance of artificial intelligence models dedicated to auto-contouring tasks.<i>Approach.</i>The augmentation technique generates data that simulates artifacts typically present in CBCT imaging. The simulated pseudo-CBCT (pCBCT) is created using interleaved sequences of simulated breath-hold and free-breathing projections. Neural networks for auto-contouring of head and neck and bowel structures were trained with and without pCBCT data. Quantitative and qualitative assessment was done in two independent test sets containing CT and real CBCT data focus on four anatomical regions: head, neck, abdomen, and pelvis. Qualitative analyses were conducted by five clinical experts from three different healthcare institutions.<i>Main results.</i>The generated pCBCT images demonstrate realistic motion artifacts comparable to those observed in real CBCT data. Training a neural network with CT and pCBCT data improved Dice similarity coefficient (DSC) and average contour distance (ACD) results on CBCT test sets. The results were statistically significant (<i>p</i>-value ⩽.03) for bone-mandible (model without/with pCBCT: 0.91/0.92 DSC,<i>p</i>⩽ .01; 0.74/0.66 mm ACD,<i>p</i>⩽.01), brain (0.34/0.93 DSC,<i>p</i>⩽ 1 × 10^-5; 17.5/2.79 mm ACD,<i>p</i>= 1 × 10^-5), oral-cavity (0.81/0.83 DSC,<i>p</i>⩽.01; 5.11/4.61 mm ACD,<i>p</i>= .02), left-submandibular-gland (0.58/0.77 DSC,<i>p</i>⩽.001; 3.24/2.12 mm ACD,<i>p</i>⩽ .001), right-submandibular-gland (0.00/0.75 DSC,<i>p</i>⩽.1 × 10^-5; 17.5/2.26 mm ACD,<i>p</i>⩽ 1 × 10^-5), left-parotid (0.68/0.78 DSC,<i>p</i>⩽ .001; 3.34/2.58 mm ACD,<i>p</i>⩽.01), large-bowel (0.60/0.75 DSC,<i>p</i>⩽ .01; 6.14/4.56 mm ACD,<i>p</i>= .03) and small-bowel (3.08/2.65 mm ACD,<i>p</i>= .03). Visual evaluation showed fewer false positives, false negatives, and misclassifications in artifact-affected areas. Qualitative analyses demonstrated that, auto-generated contours are clinically acceptable in over 90% of cases for most structures, with only a few requiring adjustments.<i>Significance.</i>The inclusion of pCBCT improves the performance of trainable auto-contouring approaches, particularly in cases where the images are prone to severe artifacts.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":"70 3","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143067298","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":"Automated estimation of individualized organ-specific dose and noise from clinical CT scans.","authors":"Sen Wang, Maria Jose Medrano, Abdullah Al Zubaer Imran, Wonkyeong Lee, Jennie Jiayi Cao, Grant M Stevens, Justin Ruey Tse, Adam S Wang","doi":"10.1088/1361-6560/ada67f","DOIUrl":"https://doi.org/10.1088/1361-6560/ada67f","url":null,"abstract":"<p><p><i>Objective</i>. Radiation dose and diagnostic image quality are opposing constraints in x-ray computed tomography (CT). Conventional methods do not fully account for organ-level radiation dose and noise when considering radiation risk and clinical task. In this work, we develop a pipeline to generate individualized organ-specific dose and noise at desired dose levels from clinical CT scans.<i>Approach</i>. To estimate organ-specific dose and noise, we compute dose maps, noise maps at desired dose levels and organ segmentations. In our pipeline, dose maps are generated using Monte Carlo simulation. The noise map is obtained by scaling the inserted noise in synthetic low-dose emulation in order to avoid anatomical structures, where the scaling coefficients are empirically calibrated. Organ segmentations are generated by a deep learning-based method (TotalSegmentator). The proposed noise model is evaluated on a clinical dataset of 12 CT scans, a phantom dataset of 3 uniform phantom scans, and a cross-site dataset of 26 scans. The accuracy of deep learning-based segmentations for organ-level dose and noise estimates was tested using a dataset of 41 cases with expert segmentations of six organs: lungs, liver, kidneys, bladder, spleen, and pancreas.<i>Main results</i>. The empirical noise model performs well, with an average RMSE approximately 1.5 HU and an average relative RMSE approximately 5% across different dose levels. The segmentation from TotalSegmentator yielded a mean Dice score of 0.8597 across the six organs (max = 0.9315 in liver, min = 0.6855 in pancreas). The resulting error in organ-level dose and noise estimation was less than 2% for most organs.<i>Significance</i>. The proposed pipeline can output individualized organ-specific dose and noise estimates accurately for personalized protocol evaluation and optimization. It is fully automated and can be scalable to large clinical datasets. This pipeline can be used to optimize image quality for specific organs and thus clinical tasks, without adversely affecting overall radiation dose.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":"70 3","pages":""},"PeriodicalIF":3.3,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143056100","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}