Computer methods and programs in biomedicine最新文献

{"title":"Oral configuration-dependent variability of the metrics of exhaled respiratory droplets during a consecutive coughing event","authors":"Nguyen Dang Khoa ,&nbsp;Kazuma Nita ,&nbsp;Kazuki Kuga ,&nbsp;Kazuhide Ito","doi":"10.1016/j.cmpb.2025.108601","DOIUrl":"10.1016/j.cmpb.2025.108601","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Coughing events are eruptive sources of virus-laden droplets/droplet nuclei. These increase the risk of infection in susceptible individuals during airborne transmission. The oral cavity functions as an exit route for exhaled droplets. Thus, its configuration contributes significantly to the metrics of exhaled droplets.</div></div><div><h3>Methods</h3><div>In this study, two realistic numerical models were developed: the respiratory system from the throat to the second bifurcation and the oral cavity with different anatomical structures. A coupling of Eulerian Wall Film (EWF) – Discrete Phase Model (DPM) was employed to numerically describe the generation, absorption and exhalation properties. In addition, two sequential coughing episodes were considered with variegated profiles in the second cough.</div></div><div><h3>Results</h3><div>As a result, the enlargement of the oral cavity caused considerable alterations in the original spatial distribution and total number concentration of exhaled droplets: these were reduced by approximately 30 %. Considering the consecutive second cough, maintaining or decreasing the flow rates resulted in a decrease in the total quantity of exhalation droplets by 25–90 %. The variations in the oral structure or coughing flow profile also reallocated the local spatial and proportional distribution of exhaled droplet. The expelled droplets/droplet nuclei's size remained approximately 0.25–20 μm range with varied development trends even though the peak concentration reserved unchanged at approximately 5 μm.</div></div><div><h3>Conclusions</h3><div>This study is a substantial work emphasizing the dependent variability of oral geometry and coughing physiology related to the properties of exhaled droplets. It emphasizes the uncertainties in the input parameters required for indoor transmission risk studies related to intersubject variability.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108601"},"PeriodicalIF":4.9,"publicationDate":"2025-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143001505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Numerical simulation of fluid-structure interaction analysis for the performance of leaflet reimplantation with different types of artificial graft","authors":"Qian Wang ,&nbsp;Xinjin Luo ,&nbsp;Juntao Qiu ,&nbsp;Shengyi Hu ,&nbsp;Xuechao Ding ,&nbsp;Tianming Du ,&nbsp;Yanping Zhang ,&nbsp;Qianwen Hou ,&nbsp;Jianye Zhou ,&nbsp;Yiming Jin ,&nbsp;Aike Qiao","doi":"10.1016/j.cmpb.2025.108598","DOIUrl":"10.1016/j.cmpb.2025.108598","url":null,"abstract":"<div><div>Background and Objective: In clinical practice, valve-sparing aortic root replacement surgery primarily addresses left ventricular dysfunction in patients due to severe aortic regurgitation, but there is controversy regarding the choice of surgical technique. In order to investigate which type of valve-sparing aortic root replacement surgeries can achieve better blood flow conditions, this study examines the impact of changes in the geometric morphology of the aortic root on the hemodynamic environment through numerical simulation.</div><div>Methods: An idealized model of the aortic root was established based on data obtained from clinical measurements, including using the model of the aortic root without significant lesions as the control group (Model C), while using surgical models of leaflet reimplantation with tubular graft (Model T), leaflet reimplantation with Valsalva graft (Model V), and the Florida sleeve procedure (Model F) as the experimental groups. Fluid-structure interaction numerical simulations were conducted to assess the differences in blood flow between the three surgical techniques.</div><div>Results: Compared to the control group, all the three experimental groups showed no abnormal blood flow patterns in the aortic root. Additionally, the distribution of high-velocity blood flow was similar to that of the control group. Due to the changes in geometric shape after surgery, the impact locations of blood on the vessel wall varied, leading to different degrees of wall shear stress concentration at the sinus-conduit junction and the aortic valve ring in the three surgical models. During the peak systolic phase, the maximum opening area of the leaflets in the three surgical models (T, V, and F) differs from that of the control model, with the disparity in aortic valve leaflet opening area being 6.42 %, 9.17 %, and 8.63 %, respectively. When comparing the leaflet closure states, it was found that the closure velocity in Model V was close to that of Model C.</div><div>Conclusions: The changes in the geometry of the aortic sinus affect the hemodynamics within the aorta, and leaflet reimplantation with Valsalva graft and Florida sleeve procedures are more stable during blood flow impacts.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108598"},"PeriodicalIF":4.9,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142977973","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BMA-Net: A 3D bidirectional multi-scale feature aggregation network for prostate region segmentation","authors":"Bangkang Fu ,&nbsp;Feng Liu ,&nbsp;Junjie He ,&nbsp;Zi Xu ,&nbsp;Yunsong Peng ,&nbsp;XiaoLi Zhang ,&nbsp;Rongpin Wang","doi":"10.1016/j.cmpb.2025.108596","DOIUrl":"10.1016/j.cmpb.2025.108596","url":null,"abstract":"<div><h3>Background and objective</h3><div>Accurate segmentation of the prostate region in magnetic resonance imaging (MRI) is crucial for prostate-related diagnoses. Recent studies have incorporated Transformers into prostate region segmentation to better capture long-range global feature representations. However, due to the computational complexity of Transformers, these studies have been limited to processing single slices. Incorporating multiple slices can facilitate more precise segmentation, but existing methods fail to effectively utilize both intra-slice and inter-slice multi-scale information.</div></div><div><h3>Methods</h3><div>To address these challenges, we propose a 3D bidirectional multi-scale feature aggregation network, called BMA-Net. This network employs a forward frequency-based global feature filtering branch to learn and filter highly correlated information both intra-slice and inter-slice. It also includes a reverse spatial attention branch, guided by Gaussian distance, to model spatial information within slices. Additionally, a convolutional neural network (CNN) branch is incorporated to supplement local feature information. To mitigate feature discrepancies among different branches, the network uses a multi-scale feature fusion module for feature interaction.</div></div><div><h3>Results</h3><div>Experiments on both public and in-house datasets were conducted. The results on the public dataset showed a Dice coefficient of 88.35 % in the central gland and 76.86 % in the peripheral zone. On the in-house dataset, the Dice coefficients were 85.85 % for the central gland and 74.50 % for the peripheral zone.</div></div><div><h3>Conclusions</h3><div>BMA-Net leverages multi-scale information both intra-slice and inter-slice to achieve more accurate segmentation of prostate regions. The experimental results demonstrate that our approach achieves superior segmentation performance compared to the current state-of-the-art methods.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108596"},"PeriodicalIF":4.9,"publicationDate":"2025-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143001189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Chained occurrences of early afterdepolarizations may create a directional triggered activity to initiate reentrant ventricular tachyarrhythmias","authors":"Kunichika Tsumoto ,&nbsp;Takao Shimamoto ,&nbsp;Yuma Aoji ,&nbsp;Yukiko Himeno ,&nbsp;Yuhichi Kuda ,&nbsp;Mamoru Tanida ,&nbsp;Akira Amano ,&nbsp;Yasutaka Kurata","doi":"10.1016/j.cmpb.2025.108587","DOIUrl":"10.1016/j.cmpb.2025.108587","url":null,"abstract":"<div><h3>Background and objective</h3><div>It has been believed that polymorphic ventricular tachycardia (VT) such as torsades de pointes (TdP) seen in patients with long QT syndromes is triggered by creating early afterdepolarization (EAD)-mediated triggered activity (TA). Although the mechanisms creating the TA have been studied intensively, characteristics of the arrhythmogenic (torsadogenic) substrates that link EAD developments to TA formation are still not well understood.</div></div><div><h3>Methods</h3><div>Computer simulations of excitation propagation in a homogenous two-dimensional ventricular tissue with an anisotropic conduction property were performed to characterize torsadogenic substrates that potentially form TA. We examined how the configuration of islands (clusters) of myocytes with synchronously chained occurrence of EADs within the tissue, each EAD cluster size and stimulation from different directions impact the TA creation.</div></div><div><h3>Results</h3><div>The presence of EAD clusters within the tissue created local regions of cardiomyocytes maintained at a depolarized membrane potential above 0 mV due to the chained occurrence of EADs. When the local area contained a concave surface border, the TA was created depending on its curvature. We found that the distance of EAD clusters was a critical factor for the development of EAD-mediated TA and polymorphic VT in long QT syndromes, that there existed a region of the distance favorable for the development of TA and VT, and that the TA was always created along the myocardial fiber orientation regardless of stimulating directions.</div></div><div><h3>Conclusion</h3><div>The chained occurrences of EADs may create a directional TA. Our findings provide deeper understandings of the cardiac arrhythmogenic substrates for preventing and treating arrhythmias.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108587"},"PeriodicalIF":4.9,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143001316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SD-LayerNet: Robust and label-efficient retinal layer segmentation via anatomical priors","authors":"Botond Fazekas ,&nbsp;Guilherme Aresta ,&nbsp;Dmitrii Lachinov ,&nbsp;Sophie Riedl ,&nbsp;Julia Mai ,&nbsp;Ursula Schmidt-Erfurth ,&nbsp;Hrvoje Bogunović","doi":"10.1016/j.cmpb.2025.108586","DOIUrl":"10.1016/j.cmpb.2025.108586","url":null,"abstract":"<div><h3>Background and objectives:</h3><div>Automated, anatomically coherent retinal layer segmentation in optical coherence tomography (OCT) is one of the most important components of retinal disease management. However, current methods rely on large amounts of labeled data, which can be difficult and expensive to obtain. In addition, these systems tend often propose anatomically impossible results, which undermines their clinical reliability.</div></div><div><h3>Methods:</h3><div>This study introduces a semi-supervised approach to retinal layer segmentation that leverages large amounts of unlabeled data and anatomical prior knowledge related to the structure of the retina. During training, we use a novel topological engine that converts inferred retinal layer boundaries into pixel-wise structured segmentations. These compose a set of anatomically valid disentangled representations which, together with predicted style factors, are used to reconstruct the input image. At training time, the retinal layer boundaries and pixel-wise predictions are both guided by reference annotations, where available, but more importantly by innovatively exploiting anatomical priors that improve the performance, robustness and coherence of the method even if only a small amount of labeled data is available.</div></div><div><h3>Results:</h3><div>Exhaustive experiments with respect to label efficiency, contribution of unsupervised data and robustness to different acquisition settings were conducted. The proposed method showed state of-the-art performance on all the studied public and internal datasets, specially in low annotated data regimes. Additionally, the model was able to make use of unlabeled data from a different domain with only a small performance drop in comparison to a fully-supervised setting.</div></div><div><h3>Conclusion:</h3><div>A novel, robust, label-efficient retinal layer segmentation method was proposed. The approach has shown state-of-the-art layer segmentation performance with a fraction of the training data available, while at the same time, its robustness against domain shift was also shown.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108586"},"PeriodicalIF":4.9,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142982642","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DECT sparse reconstruction based on hybrid spectrum data generative diffusion model","authors":"Jin Liu ,&nbsp;Fan Wu ,&nbsp;Guorui Zhan ,&nbsp;Kun Wang ,&nbsp;Yikun Zhang ,&nbsp;Dianlin Hu ,&nbsp;Yang Chen","doi":"10.1016/j.cmpb.2025.108597","DOIUrl":"10.1016/j.cmpb.2025.108597","url":null,"abstract":"<div><h3>Purpose</h3><div>Dual-energy computed tomography (DECT) enables the differentiation of different materials. Additionally, DECT images consist of multiple scans of the same sample, revealing information similarity within the energy domain. To leverage this information similarity and address safety concerns related to excessive radiation exposure in DECT imaging, sparse view DECT imaging is proposed as a solution. However, this imaging method can impact image quality. Therefore, this paper presents a hybrid spectrum data generative diffusion reconstruction model (HSGDM) to improve imaging quality.</div></div><div><h3>Method</h3><div>To exploit the spectral similarity of DECT, we use interleaved angles for sparse scanning to obtain low- and high-energy CT images with complementary incomplete views. Furthermore, we organize low- and high-energy CT image views into multichannel forms for training and inference and promote information exchange between low-energy features and high-energy features, thus improving the reconstruction quality while reducing the radiation dose. In the HSGDM, we build two types of diffusion model constraint terms trained by the image space and wavelet space. The wavelet space diffusion model exploits mainly the orientation and scale features of artifacts. By integrating the image space diffusion model, we establish a hybrid constraint for the iterative reconstruction framework. Ultimately, we transform the iterative approach into a cohesive sampling process guided by the measurement data, which collaboratively produces high-quality and consistent reconstructions of sparse view DECT.</div></div><div><h3>Results</h3><div>Compared with the comparison methods, this approach is competitive in terms of the precision of the CT values, the preservation of details, and the elimination of artifacts. In the reconstruction of 30 sparse views, with increases of 3.51 dB for the peak signal-to-noise ratio (PSNR), 0.03 for the structural similarity index measure (SSIM), and a reduction of 74.47 for the Fréchet inception distance (FID) score on the test dataset. In the ablation study, we determined the effectiveness of our proposed hybrid prior, consisting of the wavelet prior module and the image prior module, by comparing the visual effects and quantitative results of the methods using an image space model, a wavelet space model, and our hybrid model approach. Both qualitative and quantitative analyses of the results indicate that the proposed method performs well in sparse DECT reconstruction tasks.</div></div><div><h3>Conclusion</h3><div>We have developed a unified optimized mathematical model that integrates the image space and wavelet space prior knowledge into an iterative model. This model is more practical and interpretable than existing approaches are. The experimental results demonstrate the competitive performance of the proposed model.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108597"},"PeriodicalIF":4.9,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142982135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Medical image segmentation model with auto-dynamic convolution and location attention mechanism","authors":"Yuenan Wang ,&nbsp;Hua Wang ,&nbsp;Fan Zhang","doi":"10.1016/j.cmpb.2025.108593","DOIUrl":"10.1016/j.cmpb.2025.108593","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Medical image segmentation is a technique used to identify and locate anatomical structures or diseased areas from medical images with high accuracy. Accurate image segmentation is crucial in medical applications such as clinical diagnosis, surgical planning, and treatment monitoring. It provides reliable quantitative information, which helps in making decisions. The current models used for medical image segmentation are good at capturing far-reaching and global context information. However, these models have limited resolution due to their high computing requirements. Additionally, many of the models lack important information between slices, which reduces overall performance.</div></div><div><h3>Methods:</h3><div>To address these challenges, we introduced a new architecture: auto-dynamic convolution with location attention former(AD-LA Former). We have developed a novel method that utilizes auto-dynamic convolution and location attention mechanism to dynamically adapt to different data patterns. Our model incorporates an internal scaling layer to enhance the dynamics of training, integrates auto-dynamic convolution to comprehensively learn the different attention of convolution kernel, and introduces location attention to obtain more precise spatial location information and dependency.</div></div><div><h3>Results:</h3><div>We evaluated our model against leading methods on popular medical segmentation datasets such as synapse, ISIC2017 and ISIC2018 datasets, and it has demonstrated better performance compared to other methods. On the synapse dataset, DSC can reach 83.48; on the ISIC2017 dataset, ACC, SE and SP can respectively reach 0.9703, 0.9865 and 0.9295. On the ISIC2018 dataset, ACC, SE and SP can respectively reach 0.9646, 0.9904 and 0.9124.</div></div><div><h3>Conclusions:</h3><div>AD-LA Former can solve the problem of redundant information between channels and realize the ability to capture cross-channel information, so that more accurate and efficient segmentation results can be obtained.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108593"},"PeriodicalIF":4.9,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143001508","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Neurophysiological data augmentation for EEG-fNIRS multimodal features based on a denoising diffusion probabilistic model","authors":"Li Chen ,&nbsp;Zhong Yin ,&nbsp;Xuelin Gu ,&nbsp;Xiaowen Zhang ,&nbsp;Xueshan Cao ,&nbsp;Chaojing Zhang ,&nbsp;Xiaoou Li","doi":"10.1016/j.cmpb.2025.108594","DOIUrl":"10.1016/j.cmpb.2025.108594","url":null,"abstract":"<div><h3>Background and objective</h3><div>The hybrid brain computer interfaces (BCI) combining electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) have attracted extensive attention for overcoming the decoding limitations of the single-modality BCI. With the deepening application of deep learning approaches in BCI systems, its significant performance improvement has become apparent. However, the scarcity of brain signal data limits the performance of deep learning models.</div></div><div><h3>Methods</h3><div>In this study, we proposed an EEG-fNIRS data augmentation framework based on the combination of denoising diffusion probabilistic model (DDPM) and adding Gaussian noise (EFDA-CDG), for enhancing the performance of hybrid BCI systems. Firstly, we unified the temporal and spatial dimensions of EEG and fNIRS by manually extracting features and spatial mapping interpolation to create EEG-fNIRS joint distribution samples. Then, the DDPM generative model was combined with the traditional method of adding Gaussian noise to provide richer training data for the classifier. Finally, we constructed a classification module that applies EEG feature attention and fNIRS terrain attention to improve classification accuracy.</div></div><div><h3>Results</h3><div>In order to evaluate the effectiveness of EFDA-CDG framework, experiments were conducted and fully validated on three publicly available databases and one self-collected database. In the context of a participant-dependent training approach, our method achieves accuracy rates of 82.02% for motor imagery, 91.93% for mental arithmetic, and 90.54% for n-back tasks on public databases. Additionally, our method boasts an accuracy rate of 97.82% for drug addiction discrimination task on the self-collected database.</div></div><div><h3>Conclusions</h3><div>EFDA-CDG framework successfully facilitates data augmentation, thereby enhancing the performance of EEG-fNIRS hybrid BCI systems.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"261 ","pages":"Article 108594"},"PeriodicalIF":4.9,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143001497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Rad4XCNN: A new agnostic method for post-hoc global explanation of CNN-derived features by means of Radiomics","authors":"Francesco Prinzi ,&nbsp;Carmelo Militello ,&nbsp;Calogero Zarcaro ,&nbsp;Tommaso Vincenzo Bartolotta ,&nbsp;Salvatore Gaglio ,&nbsp;Salvatore Vitabile","doi":"10.1016/j.cmpb.2024.108576","DOIUrl":"10.1016/j.cmpb.2024.108576","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>In recent years, machine learning-based clinical decision support systems (CDSS) have played a key role in the analysis of several medical conditions. Despite their promising capabilities, the lack of transparency in AI models poses significant challenges, particularly in medical contexts where reliability is a mandatory aspect. However, it appears that explainability is inversely proportional to accuracy. For this reason, achieving transparency without compromising predictive accuracy remains a key challenge.</div></div><div><h3>Methods:</h3><div>This paper presents a novel method, namely Rad4XCNN, to enhance the predictive power of CNN-derived features with the inherent interpretability of radiomic features. Rad4XCNN diverges from conventional methods based on saliency maps, by associating intelligible meaning to CNN-derived features by means of Radiomics, offering new perspectives on explanation methods beyond visualization maps.</div></div><div><h3>Results:</h3><div>Using a breast cancer classification task as a case study, we evaluated Rad4XCNN on ultrasound imaging datasets, including an online dataset and two in-house datasets for internal and external validation. Some key results are: (i) CNN-derived features guarantee more robust accuracy when compared against ViT-derived and radiomic features; (ii) conventional visualization map methods for explanation present several pitfalls; (iii) Rad4XCNN does not sacrifice model accuracy for their explainability; (iv) Rad4XCNN provides a global explanation enabling the physician to extract global insights and findings.</div></div><div><h3>Conclusions:</h3><div>Our method can mitigate some concerns related to the explainability-accuracy trade-off. This study highlighted the importance of proposing new methods for model explanation without affecting their accuracy.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"260 ","pages":"Article 108576"},"PeriodicalIF":4.9,"publicationDate":"2025-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142969949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning and statistical shape modelling for real-time prediction of stent deployment in realistic anatomies","authors":"Beatrice Bisighini ,&nbsp;Miquel Aguirre ,&nbsp;Baptiste Pierrat ,&nbsp;Stéphane Avril","doi":"10.1016/j.cmpb.2024.108583","DOIUrl":"10.1016/j.cmpb.2024.108583","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>The rise in minimally invasive procedures has created a demand for efficient and reliable planning software to predict intra- and post-operative outcomes. Surrogate modelling has shown promise, but challenges remain, particularly in cardiovascular applications, due to the complexity of parametrising anatomical structures and the need for large training datasets. This study aims to apply statistical shape modelling and machine learning for predicting stent deployment in real time using patient-specific models. ►</div></div><div><h3>Methods:</h3><div>We built a statistical shape model starting from an open-source clinical dataset, which we then used to generate new synthetic cases. Finite element simulations of stent deployment were performed on these cases using an in-house software. A surrogate model was then trained to map the statistical features of the synthetic models to the corresponding stent configurations, evaluating sensitivity to dataset size. ►</div></div><div><h3>Results:</h3><div>Even with the smallest dataset (400 samples), the average prediction error in position among the tested cases never exceeded 8.6%, with a median one within the testing dataset of 1.6%. As the number of training samples increased (4900), we achieved a median position error lower than 0.1 mm (0.97%) and a maximum position error of 0.5 mm (4.8%). Notably, the largest errors occur in the radial direction of the stent, while the deployed length is accurately predicted in all the cases. ►</div></div><div><h3>Conclusions:</h3><div>The consistent success in performance strongly suggests that surrogate modelling represents a clinically valuable tool for accurately computing stent deployment outcomes in real time, even within complex anatomical scenarios.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"260 ","pages":"Article 108583"},"PeriodicalIF":4.9,"publicationDate":"2025-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142969944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}