Pub Date : 2024-09-03DOI: 10.1088/2057-1976/ad72f8
Amit Kukker, Rajneesh Sharma, Gaurav Pandey, Mohammad Faseehuddin
This work proposes a novel technique called Enhanced JAYA (EJAYA) assisted Q-Learning for the classification of pulmonary diseases, such as pneumonia and tuberculosis (TB) sub-classes using chest x-ray images. The work introduces Fuzzy lattices formation to handle real time (non-linear and non-stationary) data based feature extraction using Schrödinger equation. Features based adaptive classification is made possible through the Q-learning algorithm wherein optimal Q-values selection is done via EJAYA optimization algorithm. Fuzzy lattice is formed using x-ray image pixels and lattice Kinetic Energy (K.E.) is calculated using the Schrödinger equation. Feature vector lattices having highest K.E. have been used as an input features for the classifier. The classifier has been employed for pneumonia classification (normal, mild and severe) and Tuberculosis detection (presence or absence). A total of 3000 images have been used for pneumonia classification yielding an accuracy, sensitivity, specificity, precision and F-scores of 97.90%, 98.43%, 97.25%, 97.78% and 98.10%, respectively. For Tuberculosis 600 samples have been used. The achived accuracy, sensitivity, specificity, precision and F-score are 95.50%, 96.39%, 94.40% 95.52% and 95.95%, respectively. Computational time are 40.96 and 39.98 s for pneumonia and TB classification. Classifier learning rate (training accuracy) for pneumonia classes (normal, mild and severe) are 97.907%, 95.375% and 96.391%, respectively and for tuberculosis (present and absent) are 96.928% and 95.905%, respectively. The results have been compared with contemporary classification techniques which shows superiority of the proposed approach in terms of accuracy and speed of classification. The technique could serve as a fast and accurate tool for automated pneumonia and tuberculosis classification.
本研究提出了一种名为增强型 JAYA(EJAYA)的新技术,可辅助 Q-Learning 利用胸部 X 光图像对肺炎和肺结核(TB)等肺部疾病进行分类。这项工作引入了模糊网格形成,利用薛定谔方程处理基于特征提取的实时(非线性和非稳态)数据。通过 Q-learning 算法实现了基于特征的自适应分类,其中最佳 Q 值的选择是通过 EJAYA 优化算法完成的。利用 X 射线图像像素形成模糊晶格,并利用薛定谔方程计算晶格动能(K.E.)。具有最高 K.E. 的特征向量晶格被用作分类器的输入特征。该分类器已用于肺炎分类(正常、轻度和重度)和肺结核检测(存在或不存在)。肺炎分类共使用了 3000 幅图像,准确率、灵敏度、特异性、精确度和 F 值分别为 97.90%、98.43%、97.25%、97.78% 和 98.10%。肺结核使用了 600 个样本。准确率、灵敏度、特异性、精确度和 F 分数分别为 95.50%、96.39%、94.40%、95.52% 和 95.95%。肺炎和肺结核分类的计算时间分别为 40.96 秒和 39.98 秒。肺炎类别(正常、轻度和重度)的分类器学习率(训练准确率)分别为 97.907%、95.375% 和 96.391%,肺结核类别(存在和不存在)的分类器学习率(训练准确率)分别为 96.928% 和 95.905%。将结果与当代分类技术进行比较后发现,所提出的方法在准确性和分类速度方面都更胜一筹。该技术可作为肺炎和肺结核自动分类的快速而准确的工具。
{"title":"Fuzzy lattices assisted EJAYA Q-learning for automated pulmonary diseases classification.","authors":"Amit Kukker, Rajneesh Sharma, Gaurav Pandey, Mohammad Faseehuddin","doi":"10.1088/2057-1976/ad72f8","DOIUrl":"10.1088/2057-1976/ad72f8","url":null,"abstract":"<p><p>This work proposes a novel technique called Enhanced JAYA (EJAYA) assisted Q-Learning for the classification of pulmonary diseases, such as pneumonia and tuberculosis (TB) sub-classes using chest x-ray images. The work introduces Fuzzy lattices formation to handle real time (non-linear and non-stationary) data based feature extraction using Schrödinger equation. Features based adaptive classification is made possible through the Q-learning algorithm wherein optimal Q-values selection is done via EJAYA optimization algorithm. Fuzzy lattice is formed using x-ray image pixels and lattice Kinetic Energy (K.E.) is calculated using the Schrödinger equation. Feature vector lattices having highest K.E. have been used as an input features for the classifier. The classifier has been employed for pneumonia classification (normal, mild and severe) and Tuberculosis detection (presence or absence). A total of 3000 images have been used for pneumonia classification yielding an accuracy, sensitivity, specificity, precision and F-scores of 97.90%, 98.43%, 97.25%, 97.78% and 98.10%, respectively. For Tuberculosis 600 samples have been used. The achived accuracy, sensitivity, specificity, precision and F-score are 95.50%, 96.39%, 94.40% 95.52% and 95.95%, respectively. Computational time are 40.96 and 39.98 s for pneumonia and TB classification. Classifier learning rate (training accuracy) for pneumonia classes (normal, mild and severe) are 97.907%, 95.375% and 96.391%, respectively and for tuberculosis (present and absent) are 96.928% and 95.905%, respectively. The results have been compared with contemporary classification techniques which shows superiority of the proposed approach in terms of accuracy and speed of classification. The technique could serve as a fast and accurate tool for automated pneumonia and tuberculosis classification.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142046228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-03DOI: 10.1088/2057-1976/ad7266
Azmul H Siddique, Gary Ge, Jie Zhang
Purpose. Virtual Grid (VG) is an image processing technique designed to address scattered radiation from radiographic systems without a physical grid. It aims to eliminate artifacts caused by grid misalignment and enhance radiographic workflow efficiency. We intend to evaluate image quality between Virtual Grid and grid-based radiographic systems across various patient thicknesses.Methods. A Fujifilm Virtual Grid and GE AMX-4 portable radiographic system was used. Image quality was assessed using MTF, NPS, LCR, and CNR. MTF calculations employed an edge-device method with a 0.1 mmCu sheet. For NPS evaluation, uniform images were acquired with multiple 30 × 30 cm solid water blocks (2 cm thick), overlaid in 2 cm increments to simulate patient sizes from 2cm to 40 cm. LCR and CNR were evaluated using a CIRS test plate with 9-hole depths for a hole diameter of 0.375'. The test object was placed on top of the detector then water blocks, while maintaining the same SID, beam quality, and exposure between the units. Visual assessments were conducted by four readers, quantifying perceived hole numbers. The weighted Cohen's Kappa and Welch's T-test were utilized for statistical analysis.Results. At 80% MTF, VG exhibited high contrast resolution of 1.1 l p/mm compared to 1.2 l p/mm for the grid system. VG demonstrated lower noise levels across all frequencies for equivalent patient thicknesses. Welch's T-test indicated no significant differences in LCR (P = 0.31) and CNR (P = 0.34) between the systems. However, qualitative observation demonstrated VG's better low contrast response for patient sizes ≥10 cm. The average weighted Cohen's Kappa value was 0.78.Conclusion. This work indicates the Virtual Grid technology can effectively mitigate scattered radiation to improve granularity and low-contrast resolution in an image compared to a grid system. Furthermore, it can potentially reduce patient dose.
{"title":"Comparative evaluation of image quality between virtual grid and grid portable radiographic systems.","authors":"Azmul H Siddique, Gary Ge, Jie Zhang","doi":"10.1088/2057-1976/ad7266","DOIUrl":"10.1088/2057-1976/ad7266","url":null,"abstract":"<p><p><i>Purpose</i>. Virtual Grid (VG) is an image processing technique designed to address scattered radiation from radiographic systems without a physical grid. It aims to eliminate artifacts caused by grid misalignment and enhance radiographic workflow efficiency. We intend to evaluate image quality between Virtual Grid and grid-based radiographic systems across various patient thicknesses.<i>Methods</i>. A Fujifilm Virtual Grid and GE AMX-4 portable radiographic system was used. Image quality was assessed using MTF, NPS, LCR, and CNR. MTF calculations employed an edge-device method with a 0.1 mmCu sheet. For NPS evaluation, uniform images were acquired with multiple 30 × 30 cm solid water blocks (2 cm thick), overlaid in 2 cm increments to simulate patient sizes from 2cm to 40 cm. LCR and CNR were evaluated using a CIRS test plate with 9-hole depths for a hole diameter of 0.375'. The test object was placed on top of the detector then water blocks, while maintaining the same SID, beam quality, and exposure between the units. Visual assessments were conducted by four readers, quantifying perceived hole numbers. The weighted Cohen's Kappa and Welch's T-test were utilized for statistical analysis.<i>Results</i>. At 80% MTF, VG exhibited high contrast resolution of 1.1 l p/mm compared to 1.2 l p/mm for the grid system. VG demonstrated lower noise levels across all frequencies for equivalent patient thicknesses. Welch's T-test indicated no significant differences in LCR (P = 0.31) and CNR (P = 0.34) between the systems. However, qualitative observation demonstrated VG's better low contrast response for patient sizes ≥10 cm. The average weighted Cohen's Kappa value was 0.78.<i>Conclusion</i>. This work indicates the Virtual Grid technology can effectively mitigate scattered radiation to improve granularity and low-contrast resolution in an image compared to a grid system. Furthermore, it can potentially reduce patient dose.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142035137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-03DOI: 10.1088/2057-1976/ad7032
Seyed Mohammad Mahdi Abtahi, Fatemeh Habibi
This study aims to evaluate the optical response dependence of the PAKAG polymer gel dosimeter on photon energy and dose rate. The produced gel dosimeters were irradiated using a Varian CL 21EX medical linear accelerator with delivered doses of 0, 2, 4, 6, 8, and 10 Gy. To examine the response dependence on the delivered dose rate, dose rates of 50, 100, 200, and 350 cGy min-1were investigated. Additionally, two incident beam qualities of 6 and 18 MV were examined to study the response dependence on the incident beam energy. The irradiated polymer gel dosimeters were readout using a UV-vis spectrophotometer in the 300 to 800 nm scan range. The results reveal that a wide variation in dose rate (50-350 cGy.min-1) influences the absorbance-dose response and the sensitivity of PAKAG gel. However, smaller variations did not show a significant effect on the response. Furthermore, the response changed insignificantly with beam quality for investigated energies. It was concluded that the optical reading response of the PAKAG polymer gel dosimeter is satisfactorily independent of external parameters, including dose rate and incident beam quality.
{"title":"Investigation of the beam quality and dose rate dependence of PAKAG polymer gel dosimeter in optical readout technique.","authors":"Seyed Mohammad Mahdi Abtahi, Fatemeh Habibi","doi":"10.1088/2057-1976/ad7032","DOIUrl":"10.1088/2057-1976/ad7032","url":null,"abstract":"<p><p>This study aims to evaluate the optical response dependence of the PAKAG polymer gel dosimeter on photon energy and dose rate. The produced gel dosimeters were irradiated using a Varian CL 21EX medical linear accelerator with delivered doses of 0, 2, 4, 6, 8, and 10 Gy. To examine the response dependence on the delivered dose rate, dose rates of 50, 100, 200, and 350 cGy min<sup>-1</sup>were investigated. Additionally, two incident beam qualities of 6 and 18 MV were examined to study the response dependence on the incident beam energy. The irradiated polymer gel dosimeters were readout using a UV-vis spectrophotometer in the 300 to 800 nm scan range. The results reveal that a wide variation in dose rate (50-350 cGy.min<sup>-1</sup>) influences the absorbance-dose response and the sensitivity of PAKAG gel. However, smaller variations did not show a significant effect on the response. Furthermore, the response changed insignificantly with beam quality for investigated energies. It was concluded that the optical reading response of the PAKAG polymer gel dosimeter is satisfactorily independent of external parameters, including dose rate and incident beam quality.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141995224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-02DOI: 10.1088/2057-1976/ad7609
Sibel Cendere, Ceren Yuksel, Ercument Ovali, Beste Kinikoglu, Ozgul Gok
In the innate immune system, natural killer (NK) cells are effector lymphocytes which control several tumor types and microbial infections by limiting disease spread and tissue damage. With tumor cell killing abilities, with no priming or prior activation, NKs are potential anti-cancer therapies. In clinical practice, NKs are used in intravenous injections as they typically grow as suspension, similar to other blood cells. In this study, we designed a novel and effective biomaterial-based platform for NK cell delivery, which included in-situ NK cell encapsulation into three-dimensional (3D) biocompatible polymeric scaffolds for potential anti-cancer treatments. Depending on physical cross-linking between an alginate (ALG) polymer and a divalent cation, two natural polymers (gelatin (GEL) and hyaluronic acid (HA)) penetrated into pores and generated an inter-penetrating hydrogel system with improved mechanical properties and stability. After extensive characterization of hydrogels, NK cells were encapsulated inside using our in-situ gelation procedure to provide a biomimetic microenvironment.
.
在先天性免疫系统中,自然杀伤(NK)细胞是一种效应淋巴细胞,可通过限制疾病扩散和组织损伤来控制多种肿瘤类型和微生物感染。NK 细胞具有杀伤肿瘤细胞的能力,无需启动或事先激活,是一种潜在的抗癌疗法。在临床实践中,NK 通常以悬浮液的形式生长,与其他血细胞相似,因此被用于静脉注射。在这项研究中,我们设计了一种新颖有效的基于生物材料的 NK 细胞递送平台,其中包括将 NK 细胞原位封装到三维(3D)生物相容性聚合物支架中,用于潜在的抗癌治疗。根据藻酸盐(ALG)聚合物和二价阳离子之间的物理交联,两种天然聚合物(明胶(GEL)和透明质酸(HA))渗透到孔隙中,生成了一种具有更好机械性能和稳定性的相互渗透的水凝胶系统。在对水凝胶进行广泛表征后,利用我们的原位凝胶化程序将 NK 细胞封装在水凝胶中,以提供仿生微环境。
{"title":"Encapsulation of human natural killer cells into novel gelatin-based polymeric hydrogel networks.","authors":"Sibel Cendere, Ceren Yuksel, Ercument Ovali, Beste Kinikoglu, Ozgul Gok","doi":"10.1088/2057-1976/ad7609","DOIUrl":"https://doi.org/10.1088/2057-1976/ad7609","url":null,"abstract":"<p><p>In the innate immune system, natural killer (NK) cells are effector lymphocytes which control several tumor types and microbial infections by limiting disease spread and tissue damage. With tumor cell killing abilities, with no priming or prior activation, NKs are potential anti-cancer therapies. In clinical practice, NKs are used in intravenous injections as they typically grow as suspension, similar to other blood cells. In this study, we designed a novel and effective biomaterial-based platform for NK cell delivery, which included in-situ NK cell encapsulation into three-dimensional (3D) biocompatible polymeric scaffolds for potential anti-cancer treatments. Depending on physical cross-linking between an alginate (ALG) polymer and a divalent cation, two natural polymers (gelatin (GEL) and hyaluronic acid (HA)) penetrated into pores and generated an inter-penetrating hydrogel system with improved mechanical properties and stability. After extensive characterization of hydrogels, NK cells were encapsulated inside using our in-situ gelation procedure to provide a biomimetic microenvironment.
.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142118913","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this study, an individualized and stable passive-control lower-limb exoskeleton robot was developed. Users' joint angles and the center of pressure (CoP) of one of their soles were input into a convolutional neural network (CNN)-long short-term memory (LSTM) model to evaluate and adjust the exoskeleton control scheme. The CNN-LSTM model predicted the fitness of the control scheme and output the results to the exoskeleton robot, which modified its control parameters accordingly to enhance walking stability. The sole's CoP had similar trends during normal walking and passive walking with the developed exoskeleton; they-coordinates of the CoPs with and without the exoskeleton had a correlation of 91%. Moreover, electromyography signals from the rectus femoris muscle revealed that it exerted 40% less force when walking with a stable stride length in the developed system than when walking with an unstable stride length. Therefore, the developed lower-limb exoskeleton can be used to assist users in achieving balanced and stable walking with reduced force application. In the future, this exoskeleton can be used by patients with stroke and lower-limb weakness to achieve stable walking.
{"title":"Development of an individualized stable and force-reducing lower-limb exoskeleton.","authors":"Guo-Shing Huang, Meng-Hua Yen, Chia-Chun Chang, Chung-Liang Lai, Chi-Chun Chen","doi":"10.1088/2057-1976/ad686f","DOIUrl":"https://doi.org/10.1088/2057-1976/ad686f","url":null,"abstract":"<p><p>In this study, an individualized and stable passive-control lower-limb exoskeleton robot was developed. Users' joint angles and the center of pressure (CoP) of one of their soles were input into a convolutional neural network (CNN)-long short-term memory (LSTM) model to evaluate and adjust the exoskeleton control scheme. The CNN-LSTM model predicted the fitness of the control scheme and output the results to the exoskeleton robot, which modified its control parameters accordingly to enhance walking stability. The sole's CoP had similar trends during normal walking and passive walking with the developed exoskeleton; the<i>y</i>-coordinates of the CoPs with and without the exoskeleton had a correlation of 91%. Moreover, electromyography signals from the rectus femoris muscle revealed that it exerted 40% less force when walking with a stable stride length in the developed system than when walking with an unstable stride length. Therefore, the developed lower-limb exoskeleton can be used to assist users in achieving balanced and stable walking with reduced force application. In the future, this exoskeleton can be used by patients with stroke and lower-limb weakness to achieve stable walking.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142103985","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-30DOI: 10.1088/2057-1976/ad6dcd
Bao Ngoc Huynh, Aurora Rosvoll Groendahl, Oliver Tomic, Kristian Hovde Liland, Ingerid Skjei Knudtsen, Frank Hoebers, Wouter van Elmpt, Einar Dale, Eirik Malinen, Cecilia Marie Futsaether
Objective.Target volumes for radiotherapy are usually contoured manually, which can be time-consuming and prone to inter- and intra-observer variability. Automatic contouring by convolutional neural networks (CNN) can be fast and consistent but may produce unrealistic contours or miss relevant structures. We evaluate approaches for increasing the quality and assessing the uncertainty of CNN-generated contours of head and neck cancers with PET/CT as input.Approach.Two patient cohorts with head and neck squamous cell carcinoma and baseline18F-fluorodeoxyglucose positron emission tomography and computed tomography images (FDG-PET/CT) were collected retrospectively from two centers. The union of manual contours of the gross primary tumor and involved nodes was used to train CNN models for generating automatic contours. The impact of image preprocessing, image augmentation, transfer learning and CNN complexity, architecture, and dimension (2D or 3D) on model performance and generalizability across centers was evaluated. A Monte Carlo dropout technique was used to quantify and visualize the uncertainty of the automatic contours.Main results. CNN models provided contours with good overlap with the manually contoured ground truth (median Dice Similarity Coefficient: 0.75-0.77), consistent with reported inter-observer variations and previous auto-contouring studies. Image augmentation and model dimension, rather than model complexity, architecture, or advanced image preprocessing, had the largest impact on model performance and cross-center generalizability. Transfer learning on a limited number of patients from a separate center increased model generalizability without decreasing model performance on the original training cohort. High model uncertainty was associated with false positive and false negative voxels as well as low Dice coefficients.Significance.High quality automatic contours can be obtained using deep learning architectures that are not overly complex. Uncertainty estimation of the predicted contours shows potential for highlighting regions of the contour requiring manual revision or flagging segmentations requiring manual inspection and intervention.
{"title":"Deep learning with uncertainty estimation for automatic tumor segmentation in PET/CT of head and neck cancers: impact of model complexity, image processing and augmentation.","authors":"Bao Ngoc Huynh, Aurora Rosvoll Groendahl, Oliver Tomic, Kristian Hovde Liland, Ingerid Skjei Knudtsen, Frank Hoebers, Wouter van Elmpt, Einar Dale, Eirik Malinen, Cecilia Marie Futsaether","doi":"10.1088/2057-1976/ad6dcd","DOIUrl":"10.1088/2057-1976/ad6dcd","url":null,"abstract":"<p><p><i>Objective.</i>Target volumes for radiotherapy are usually contoured manually, which can be time-consuming and prone to inter- and intra-observer variability. Automatic contouring by convolutional neural networks (CNN) can be fast and consistent but may produce unrealistic contours or miss relevant structures. We evaluate approaches for increasing the quality and assessing the uncertainty of CNN-generated contours of head and neck cancers with PET/CT as input.<i>Approach.</i>Two patient cohorts with head and neck squamous cell carcinoma and baseline<sup>18</sup>F-fluorodeoxyglucose positron emission tomography and computed tomography images (FDG-PET/CT) were collected retrospectively from two centers. The union of manual contours of the gross primary tumor and involved nodes was used to train CNN models for generating automatic contours. The impact of image preprocessing, image augmentation, transfer learning and CNN complexity, architecture, and dimension (2D or 3D) on model performance and generalizability across centers was evaluated. A Monte Carlo dropout technique was used to quantify and visualize the uncertainty of the automatic contours.<i>Main results</i>. CNN models provided contours with good overlap with the manually contoured ground truth (median Dice Similarity Coefficient: 0.75-0.77), consistent with reported inter-observer variations and previous auto-contouring studies. Image augmentation and model dimension, rather than model complexity, architecture, or advanced image preprocessing, had the largest impact on model performance and cross-center generalizability. Transfer learning on a limited number of patients from a separate center increased model generalizability without decreasing model performance on the original training cohort. High model uncertainty was associated with false positive and false negative voxels as well as low Dice coefficients.<i>Significance.</i>High quality automatic contours can be obtained using deep learning architectures that are not overly complex. Uncertainty estimation of the predicted contours shows potential for highlighting regions of the contour requiring manual revision or flagging segmentations requiring manual inspection and intervention.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141911569","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-30DOI: 10.1088/2057-1976/ad7594
Souha Nemri, Luc Duong
Echocardiography is one the most commonly used imaging modalities for the diagnosis
of congenital heart disease. Echocardiographic image analysis is crucial to obtaining
accurate cardiac anatomy information. Semantic segmentation models can be used
to precisely delimit the borders of the left ventricle, and allow an accurate and
automatic identification of the region of interest, which can be extremely useful for
cardiologists. In the field of computer vision, convolutional neural network (CNN)
architectures remain dominant. Existing CNN approaches have proved highly efficient
for the segmentation of various medical images over the past decade. However, these
solutions usually struggle to capture long-range dependencies, especially when it comes
to images with objects of different scales and complex structures. In this study, we
present an efficient method for semantic segmentation of echocardiographic images
that overcomes these challenges by leveraging the self-attention mechanism of the
Transformer architecture. The proposed solution extracts long-range dependencies and
efficiently processes objects at different scales, improving performance in a variety of
tasks. We introduce Shifted Windows Transformer models (Swin Transformers), which
encode both the content of anatomical structures and the relationship between them.
Our solution combines the Swin Transformer and U-Net architectures, producing a
U-shaped variant. The validation of the proposed method is performed with the
EchoNet-Dynamic dataset used to train our model. The results show an accuracy
of 0.97, a Dice coefficient of 0.87, and an Intersection over union (IoU) of 0.78.
Swin Transformer models are promising for semantically segmenting echocardiographic
images and may help assist cardiologists in automatically analyzing and measuring
complex echocardiographic images.
{"title":"Automatic segmentation of echocardiographic images using a Shifted Windows Vision Transformer architecture.","authors":"Souha Nemri, Luc Duong","doi":"10.1088/2057-1976/ad7594","DOIUrl":"https://doi.org/10.1088/2057-1976/ad7594","url":null,"abstract":"<p><p>Echocardiography is one the most commonly used imaging modalities for the diagnosis
of congenital heart disease. Echocardiographic image analysis is crucial to obtaining
accurate cardiac anatomy information. Semantic segmentation models can be used
to precisely delimit the borders of the left ventricle, and allow an accurate and
automatic identification of the region of interest, which can be extremely useful for
cardiologists. In the field of computer vision, convolutional neural network (CNN)
architectures remain dominant. Existing CNN approaches have proved highly efficient
for the segmentation of various medical images over the past decade. However, these
solutions usually struggle to capture long-range dependencies, especially when it comes
to images with objects of different scales and complex structures. In this study, we
present an efficient method for semantic segmentation of echocardiographic images
that overcomes these challenges by leveraging the self-attention mechanism of the
Transformer architecture. The proposed solution extracts long-range dependencies and
efficiently processes objects at different scales, improving performance in a variety of
tasks. We introduce Shifted Windows Transformer models (Swin Transformers), which
encode both the content of anatomical structures and the relationship between them.
Our solution combines the Swin Transformer and U-Net architectures, producing a
U-shaped variant. The validation of the proposed method is performed with the
EchoNet-Dynamic dataset used to train our model. The results show an accuracy
of 0.97, a Dice coefficient of 0.87, and an Intersection over union (IoU) of 0.78.
Swin Transformer models are promising for semantically segmenting echocardiographic
images and may help assist cardiologists in automatically analyzing and measuring
complex echocardiographic images.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142103977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-30DOI: 10.1088/2057-1976/ad6b31
Sanjay S Yengul, Paul E Barbone, Bruno Madore
Background:Dispersion presents both a challenge and a diagnostic opportunity in shear wave elastography (SWE).Shear Wave Rheometry(SWR) is an inversion technique for processing SWE data acquired using an acoustic radiation force impulse (ARFI) excitation. The main advantage of SWR is that it can characterize the shear properties of homogeneous soft media over a wide frequency range. Assumptions associated with SWR include tissue homogeneity, tissue isotropy, and axisymmetry of the ARFI excitation).Objective:Evaluate the validity of the SWR assumptions in ex vivo bovine liver.Approach:SWR was used to measure the shear properties of bovine liver tissue as function of frequency over a large frequency range. Assumptions associated with SWR (tissue homogeneity, tissue isotropy, and axisymmetry of the ARFI excitation) were evaluated through measurements performed at multiple locations and probe orientations. Measurements focused on quantities that would reveal violations of the assumptions.Main results:Measurements of shear properties were obtained over the 25-250 Hz range, and showed a 4-fold increase in shear storage modulus (from 1 to 4 kPa) and over a 10-fold increase in the loss modulus (from 0.2 to 3 kPa) over that decade-wide frequency range. Measurements under different conditions were highly repeatable, and model error was low in all cases.Significance and Conclusion:SWR depends on modeling the ARFI-induced shear wave as a full vector viscoelastic shear wave resulting from an axisymmetric source; it is agnostic to any specific rheological model. Despite this generality, the model makes three main simplifying assumptions. These results show that the modeling assumptions used in SWR are valid in bovine liver over a wide frequency band.
{"title":"Characterizing dispersion in bovine liver using ARFI-based shear wave rheometry.","authors":"Sanjay S Yengul, Paul E Barbone, Bruno Madore","doi":"10.1088/2057-1976/ad6b31","DOIUrl":"10.1088/2057-1976/ad6b31","url":null,"abstract":"<p><p><i>Background:</i>Dispersion presents both a challenge and a diagnostic opportunity in shear wave elastography (SWE).<i>Shear Wave Rheometry</i>(SWR) is an inversion technique for processing SWE data acquired using an acoustic radiation force impulse (ARFI) excitation. The main advantage of SWR is that it can characterize the shear properties of homogeneous soft media over a wide frequency range. Assumptions associated with SWR include tissue homogeneity, tissue isotropy, and axisymmetry of the ARFI excitation).<i>Objective:</i>Evaluate the validity of the SWR assumptions in ex vivo bovine liver.<i>Approach:</i>SWR was used to measure the shear properties of bovine liver tissue as function of frequency over a large frequency range. Assumptions associated with SWR (tissue homogeneity, tissue isotropy, and axisymmetry of the ARFI excitation) were evaluated through measurements performed at multiple locations and probe orientations. Measurements focused on quantities that would reveal violations of the assumptions.<i>Main results:</i>Measurements of shear properties were obtained over the 25-250 Hz range, and showed a 4-fold increase in shear storage modulus (from 1 to 4 kPa) and over a 10-fold increase in the loss modulus (from 0.2 to 3 kPa) over that decade-wide frequency range. Measurements under different conditions were highly repeatable, and model error was low in all cases.<i>Significance and Conclusion:</i>SWR depends on modeling the ARFI-induced shear wave as a full vector viscoelastic shear wave resulting from an axisymmetric source; it is agnostic to any specific rheological model. Despite this generality, the model makes three main simplifying assumptions. These results show that the modeling assumptions used in SWR are valid in bovine liver over a wide frequency band.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141892792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Low-frequency sonophoresis has emerged as a promising minimally invasive transdermal delivery method. However, effectively inducing cavitation on the skin surface with a compact, low-frequency ultrasound transducer poses a significant challenge. This paper presents a modified design of a low-frequency ultrasound transducer capable of generating ultrasound cavitation on the skin surfaces. The transducer comprises a piezoelectric ceramic disk and a bowl-shaped acoustic resonator. A conical slit structure was incorporated into the modified transducer design to amplify vibration displacement and enhance the maximum sound pressure. The FEM-based simulation results confirmed that the maximum sound pressure at the resonance frequency of 78 kHz was increased by 1.9 times that of the previous design. Ultrasound cavitation could be experimentally observed on the gel surface. Moreover, 3 minutes of ultrasound treatment significantly improved the caffeine permeability across an artificial membrane. These results demonstrated that this transducer holds promise for enhancing drug permeation by generating ultrasound cavitation on the skin surface.
{"title":"A compact and low-frequency drive ultrasound transducer for facilitating cavitation-assisted drug permeation via skin.","authors":"Shinya Yamamoto, Naohiro Sugita, Keita Tomioka, Tadahiko Shinshi","doi":"10.1088/2057-1976/ad7596","DOIUrl":"https://doi.org/10.1088/2057-1976/ad7596","url":null,"abstract":"<p><p>Low-frequency sonophoresis has emerged as a promising minimally invasive transdermal delivery method. However, effectively inducing cavitation on the skin surface with a compact, low-frequency ultrasound transducer poses a significant challenge. This paper presents a modified design of a low-frequency ultrasound transducer capable of generating ultrasound cavitation on the skin surfaces. The transducer comprises a piezoelectric ceramic disk and a bowl-shaped acoustic resonator. A conical slit structure was incorporated into the modified transducer design to amplify vibration displacement and enhance the maximum sound pressure. The FEM-based simulation results confirmed that the maximum sound pressure at the resonance frequency of 78 kHz was increased by 1.9 times that of the previous design. Ultrasound cavitation could be experimentally observed on the gel surface. Moreover, 3 minutes of ultrasound treatment significantly improved the caffeine permeability across an artificial membrane. These results demonstrated that this transducer holds promise for enhancing drug permeation by generating ultrasound cavitation on the skin surface.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142103976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-30DOI: 10.1088/2057-1976/ad7592
Mohammad Amin Abazari, M Soltani, Faezeh Eydi, Arman Rahmim, Farshad Moradi Kashkooli
18F-Fluoromisonidazole (18F-FMISO) is a highly promising positron emission tomography radiopharmaceutical for identifying hypoxic regions in solid tumors. This research employs spatiotemporal multi-scale mathematical modeling to explore how different levels of angiogenesis influence the transport of radiopharmaceuticals within tumors. In this study, two tumor geometries with heterogeneous and uniform distributions of capillary networks were employed to incorporate varying degrees of microvascular density. The synthetic image of the heterogeneous and vascularized tumor was generated by simulating the angiogenesis process. The proposed multi-scale spatiotemporal model accounts for intricate physiological and biochemical factors within the tumor microenvironment, such as the transvascular transport of the radiopharmaceutical agent, its movement into the interstitial space by diffusion and convection mechanisms, and ultimately its uptake by tumor cells. Results showed that both quantitative and semi-quantitative metrics of 18F-FMISO uptake differ spatially and temporally at different stages during tumor growth. The presence of a high microvascular density in uniformly vascularized tumor increases cellular uptake, as it allows for more efficient release and rapid distribution of radiopharmaceutical molecules. This results in enhanced uptake compared to the heterogeneous vascularized tumor. In both heterogeneous and uniform distribution of microvessels in tumors, the diffusion transport mechanism has a more pronounced than convection. The findings of this study shed light on the transport phenomena behind 18F-FMISO radiopharmaceutical distribution and its delivery in the tumor microenvironment, aiding oncologists in their routine decision-making processes.
{"title":"Mathematical Modeling of 18F-Fluoromisonidazole (18F-FMISO) Radiopharmaceutical Transport in Vascularized Solid Tumors.","authors":"Mohammad Amin Abazari, M Soltani, Faezeh Eydi, Arman Rahmim, Farshad Moradi Kashkooli","doi":"10.1088/2057-1976/ad7592","DOIUrl":"https://doi.org/10.1088/2057-1976/ad7592","url":null,"abstract":"<p><p>18F-Fluoromisonidazole (18F-FMISO) is a highly promising positron emission tomography radiopharmaceutical for identifying hypoxic regions in solid tumors. This research employs spatiotemporal multi-scale mathematical modeling to explore how different levels of angiogenesis influence the transport of radiopharmaceuticals within tumors. In this study, two tumor geometries with heterogeneous and uniform distributions of capillary networks were employed to incorporate varying degrees of microvascular density. The synthetic image of the heterogeneous and vascularized tumor was generated by simulating the angiogenesis process. The proposed multi-scale spatiotemporal model accounts for intricate physiological and biochemical factors within the tumor microenvironment, such as the transvascular transport of the radiopharmaceutical agent, its movement into the interstitial space by diffusion and convection mechanisms, and ultimately its uptake by tumor cells. Results showed that both quantitative and semi-quantitative metrics of 18F-FMISO uptake differ spatially and temporally at different stages during tumor growth. The presence of a high microvascular density in uniformly vascularized tumor increases cellular uptake, as it allows for more efficient release and rapid distribution of radiopharmaceutical molecules. This results in enhanced uptake compared to the heterogeneous vascularized tumor. In both heterogeneous and uniform distribution of microvessels in tumors, the diffusion transport mechanism has a more pronounced than convection. The findings of this study shed light on the transport phenomena behind 18F-FMISO radiopharmaceutical distribution and its delivery in the tumor microenvironment, aiding oncologists in their routine decision-making processes.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142103981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}