Med Phys. 2024 Dec 19. doi: 10.1002/mp.17580. Online ahead of print.

ABSTRACT

BACKGROUND: Manual contour corrections during fractionated magnetic resonance (MR)-guided radiotherapy (MRgRT) are time-consuming. Conventional population models for deep learning auto-segmentation might be suboptimal for MRgRT at MR-Linacs since they do not incorporate manual segmentation from treatment planning and previous fractions.

PURPOSE: In this work, we investigate patient-specific (PS) auto-segmentation methods leveraging expert-segmented planning and prior fraction MR images (MRIs) to improve auto-segmentation on consecutive treatment days.

MATERIALS AND METHODS: Data from 151 abdominal cancer patients treated at a 0.35 T MR-Linac (151 planning and 215 fraction MRIs) were included. Population baseline models (BMs) were trained on 107 planning MRIs for one-class segmentation of the aorta, bowel, duodenum, kidneys, liver, spinal canal, and stomach. PS models were obtained by fine-tuning the BMs using the planning MRI ( PS BM $\text{PS}_{\mathrm{BM}}$ ). Maximal improvement by continuously updating the PS models was investigated by adding the first four out of five fraction MRIs ( PS BM F4 $\text{PS}_{\mathrm{BM}}^{\operatorname{F4}}$ ). Similarly, PS models without BM were trained ( PS no BM $\text{PS}_{\mathrm{no BM}}$ and PS no BM F4 $\text{PS}_{\mathrm{no BM}}^{\operatorname{F4}}$ ). All hyperparameters were optimized using 23 patients, and the methods were tested on the remaining 21 patients. Evaluation involved Dice similarity coefficient (DSC), average ( HD avg $\text{HD}_{\rm avg}$ ) and the 95th percentile (HD95) Hausdorff distance. A qualitative contour assessment by a radiation oncologist was performed for BM, PS BM $\text{PS}_{\mathrm{BM}}$ , and PS no BM $\text{PS}_{\mathrm{no BM}}$ .

RESULTS: PS BM F4 $\text{PS}_{\mathrm{BM}}^{\operatorname{F4}}$ and PS BM $\text{PS}_{\mathrm{BM}}$ networks had the best geometric performance. PS no BM $\text{PS}_{\mathrm{no BM}}$ and BMs showed similar DSC and HDs values, however PS no BM F4 $\text{PS}_{\mathrm{no BM}}^{\operatorname{F4}}$ models outperformed BMs. PS BM $\text{PS}_{\mathrm{BM}}$ predictions scored the best in the qualitative evaluation, followed by the BMs and PS no BM $\text{PS}_{\mathrm{no BM}}$ models.

CONCLUSION: Personalized auto-segmentation models outperformed the population BMs. In most cases, PS BM $\text{PS}_{\mathrm{BM}}$ delineations were judged to be directly usable for treatment adaptation without further corrections, suggesting a potential time saving during fractionated treatment.

PMID:39699250 | DOI:10.1002/mp.17580

J Chem Theory Comput. 2024 Dec 19. doi: 10.1021/acs.jctc.4c01139. Online ahead of print.

ABSTRACT

We present flow matching for reaction coordinates (FMRC), a novel deep learning algorithm designed to identify optimal reaction coordinates (RC) in biomolecular reversible dynamics. FMRC is based on the mathematical principles of lumpability and decomposability, which we reformulate into a conditional probability framework for efficient data-driven optimization using deep generative models. While FMRC does not explicitly learn the well-established transfer operator or its eigenfunctions, it can effectively encode the dynamics of leading eigenfunctions of the system transfer operator into its low-dimensional RC space. We further quantitatively compare its performance with several state-of-the-art algorithms by evaluating the quality of Markov state models (MSM) constructed in their respective RC spaces, demonstrating the superiority of FMRC in three increasingly complex biomolecular systems. In addition, we successfully demonstrated the efficacy of FMRC for bias deposition in the enhanced sampling of a simple model system. Finally, we discuss its potential applications in downstream applications such as enhanced sampling methods and MSM construction.

PMID:39699247 | DOI:10.1021/acs.jctc.4c01139

Med Phys. 2024 Dec 19. doi: 10.1002/mp.17588. Online ahead of print.

ABSTRACT

BACKGROUND: Automation in radiotherapy presents a promising solution to the increasing cancer burden and workforce shortages. However, existing automated methods for breast radiotherapy lack a comprehensive, end-to-end solution that meets varying standards of care.

PURPOSE: This study aims to develop a complete portfolio of automated radiotherapy treatment planning for intact breasts, tailored to individual patient factors, clinical approaches, and available resources.

METHODS: We developed five automated conventional treatment approaches and utilized an established RapidPlan model for volumetric arc therapy. These approaches include conventional tangents for whole breast treatment, two variants for supraclavicular nodes (SCLV) treatment with/without axillary nodes, and two options for comprehensive regional lymph nodes treatment. The latter consists of wide tangents photon fields with a SCLV field, and a photon tangents field with a matched electron field to treat the internal mammary nodes (IMNs), and a SCLV field. Each approach offers the choice of a single or two isocenter setup (with couch rotation) to accommodate a wide range of patient sizes. All algorithms start by automatically generating contours for breast clinical target volume, regional lymph nodes, and organs at risk using an in-house nnU-net deep learning models. Gantry angles and field shapes are then automatically generated and optimized to ensure target coverage while limiting the dose to nearby organs. The dose is optimized using field weighting for the lymph nodes fields and an automated field-in-field approach for the tangents. These algorithms were integrated into the RayStation treatment planning system and tested for clinical acceptability on 15 internal whole breast patients (150 plans) and 40 external patients from four different institutions in Switzerland, Argentina, Iran, and the USA (360 plans). Evaluation criteria included ensuring adequate coverage of targets and adherence to dose constraints for normal structures. A breast radiation oncologist reviewed the single institution dataset for clinical acceptability (5-point scale) and a physicist evaluated the multi-institutional dataset (use as is or edit).

RESULTS: The dosimetric evaluation across all datasets (510 plans) showed that 100% of the automated plans met the dose coverage requirements for the breast, 99% for the SCLV, 98% for the axillary nodes, and 91% for the IMN. As expected, hot spots were more prevalent when multiple fields were combined. For the heart, ipsilateral lung, and contralateral breast, automated plans met constraints for 95%, 92%, and 95% of the plans, respectively. Physician evaluation of the 15 internal patients indicated that all automated plans were clinically acceptable with minor edits. Notably, the use of automated contours with the RapidPlan model resulted in plans that were immediately ready for use in 73% of cases (95% confidence interval, 95% CI [51- 96]) of patients, with the remaining cases requiring minor stylistic edits. Similarly, the physicist's review of the 40 multi-institution patients showed that the auto-plans were ready for use 79% (95% CI [73,85]) of the time (95% CI [73,85]), with edits needed for the remaining cases.

CONCLUSION: This study demonstrates the feasibility of a comprehensive automated treatment planning model for whole breast radiotherapy, effectively accommodating diverse treatment paradigms.

PMID:39699058 | DOI:10.1002/mp.17588

Adv Sci (Weinh). 2024 Dec 19:e2407888. doi: 10.1002/advs.202407888. Online ahead of print.

ABSTRACT

To implement digital-twin smart home applications, the mat sensing system based on triboelectric sensors is commonly used for gait information collection from daily activities. Yet traditional mat sensing systems often miss upper body motions and fail to adequately project these into the virtual realm, limiting their specific application scenarios. Herein, triboelectric mat multimodal sensing system is designed, enhanced with a commercial infrared imaging sensor, to capture diverse sensory information for sleep and emotion-relevant activity monitoring without compromising privacy. This system generates pixel-based area ratio mappings across the entire mat array, solely based on the integral operation of triboelectric outputs. Additionally, it utilizes multimodal sensory intelligence and deep-learning analytics to detect different sleeping postures and monitor comprehensive sleep behaviors and emotional states associated with daily activities. These behaviors are projected into the metaverse, enhancing virtual interactions. This multimodal sensing system, cost-effective and non-intrusive, serves as a functional interface for diverse digital-twin smart home applications such as healthcare, sports monitoring, and security.

PMID:39698892 | DOI:10.1002/advs.202407888

J Chem Inf Model. 2024 Dec 19. doi: 10.1021/acs.jcim.4c01065. Online ahead of print.

ABSTRACT

The application of deep learning technology in the field of materials science provides a new method for predicting the adsorption energy of high-performance alloy catalysts in hydrogen evolution reactions and material discovery. The activity and selectivity of catalytic materials are mainly influenced by the properties and positions of active sites and adsorption sites. However, current deep learning models have not sufficiently focused on the importance of active atoms and adsorbates, instead placing more emphasis on the overall structure of the catalytic materials. In this paper, the overall molecular graph and a masked graph, which ignores fixed atoms, are separately input into the Masked Graph Transformer (MGT) network to enhance the model's ability to recognize key sites in catalytic reactions. Second, we introduce a nonlinear message-passing mechanism to improve the dot-product attention in the Transformer and capture the directional information on the relative positions of nodes by integrating molecular geometric information through deep tensor products. Subsequently, we constructed the NLMP-TransNet framework, which combines MPNN and Transformer and optimizes the model's learning and prediction capabilities through weight sharing and residual connections. The MGT achieves an error rate of 0.5447 eV on the small data set OC20-Ni, surpassing existing technologies. Ablation studies confirm the necessity of focusing on site features for accurate adsorption energy prediction. Code is available at https://github.com/KristinSun/OCP-MGT.git.

PMID:39698829 | DOI:10.1021/acs.jcim.4c01065

Biom J. 2025 Feb;67(1):e70023. doi: 10.1002/bimj.70023.

ABSTRACT

In a longitudinal clinical registry, different measurement instruments might have been used for assessing individuals at different time points. To combine them, we investigate deep learning techniques for obtaining a joint latent representation, to which the items of different measurement instruments are mapped. This corresponds to domain adaptation, an established concept in computer science for image data. Using the proposed approach as an example, we evaluate the potential of domain adaptation in a longitudinal cohort setting with a rather small number of time points, motivated by an application with different motor function measurement instruments in a registry of spinal muscular atrophy (SMA) patients. There, we model trajectories in the latent representation by ordinary differential equations (ODEs), where person-specific ODE parameters are inferred from baseline characteristics. The goodness of fit and complexity of the ODE solutions then allow to judge the measurement instrument mappings. We subsequently explore how alignment can be improved by incorporating corresponding penalty terms into model fitting. To systematically investigate the effect of differences between measurement instruments, we consider several scenarios based on modified SMA data, including scenarios where a mapping should be feasible in principle and scenarios where no perfect mapping is available. While misalignment increases in more complex scenarios, some structure is still recovered, even if the availability of measurement instruments depends on patient state. A reasonable mapping is feasible also in the more complex real SMA data set. These results indicate that domain adaptation might be more generally useful in statistical modeling for longitudinal registry data.

PMID:39698740 | DOI:10.1002/bimj.70023

Quant Imaging Med Surg. 2024 Dec 5;14(12):8414-8428. doi: 10.21037/qims-24-650. Epub 2024 Nov 11.

ABSTRACT

BACKGROUND: Accurate assessment of aortic root is crucial for the preprocedural planning of transcatheter aortic valve replacement (TAVR). A variety software is emerging for the semiautomated or automated measurements during TAVR planning. This study evaluated a new deep-learning (DL) tool based on cardiac computed tomography angiography (CCTA) for fully automatic assessment of aortic root.

METHODS: The study included 126 patients with CCTA, 63 of whom underwent TAVR. In the overall population, the DL method was compared to manual measurements of the annulus dimensions. Within the TAVR group, the DL method was also compared to 3mensio software-derived aortic root measure, including the annulus, left ventricular outflow tract (LVOT), sinotubular junction (STJ), ascending aorta (AAo), and the heights of both the coronary ostia.

RESULTS: Data were successfully analyzed using the DL method in 122 (96.8%) of patients. The correlation of annular diameters between the DL and manual methods was good to excellent for the overall cohort (n=118; r=0.83), the TAVR group (n=59, r=0.86), and its subgroups [bicuspid aortic valve (BAV): n=12, r=0.74; tricuspid aortic valve (TAV): n=47, r=0.93]. In the comparison of the DL method with 3mensio, the highest correlation was found for AAo (r=0.99). Among the four diameter indices [minimum, maximum, perimeter-derived diameter (pDD), and area-derived diameter (aDD)], excellent correlation was observed for aDD (LVOT: r=0.92; annulus: r=0.89).

CONCLUSIONS: The DL method offers an effective and efficient tool for the quantification of aortic roots for TAVR planning.

PMID:39698729 | PMC:PMC11651967 | DOI:10.21037/qims-24-650

Quant Imaging Med Surg. 2024 Dec 5;14(12):9374-9384. doi: 10.21037/qims-24-1764. Epub 2024 Nov 13.

ABSTRACT

BACKGROUND: A change in the output of deep neural networks (DNNs) via the perturbation of a few pixels of an image is referred to as an adversarial attack, and these perturbed images are known as adversarial samples. This study examined strategies for compromising the integrity of DNNs under stringent conditions, specifically by inducing the misclassification of medical images of disease with minimal pixel modifications.

METHODS: This study used the following three publicly available datasets: the chest radiograph of emphysema (cxr) dataset, the melanocytic lesion (derm) dataset, and the Kaggle diabetic retinopathy (dr) dataset. To attack the medical images, we proposed a method termed decrease group differential evolution (DGDE) for generating adversarial images. Under this method, a noise matrix of the same size as the input image is first used to attack the image sample several times until the initial adversarial perturbation s0 is obtained. Next, a subset s1 is randomly picked from the initial adversarial perturbation s0 that is still able to cause the samples to be misclassified by the classifier. A new subset s2 is subsequently randomly selected from the adversarial perturbation subset s1, which can still cause the adversarial samples to be misclassified by the classifier. Finally, the adversarial perturbation subset sn with the minimum number of elements is obtained by continuous reduction of the number of perturbed pixels.

RESULTS: In this study, the DGDE method was used to attack the images of the cxr dataset, the derm dataset, and the dr dataset; and the minimum number of pixels required to be considered a successful attack was 11, 7, and 7, respectively, while the maximum number of pixel changes was 55, 35, and 21, respectively. The average number of pixel changes was 30, 18, and 11, respectively, in the cxr dataset, the derm dataset, and the dr dataset, respectively, while the percentages of the average number of pixel changes among the total number of pixels of the image were 0.0598%, 0.0359%, and 0.0219, respectively.

CONCLUSIONS: Unlike the traditional differential evolution (DE) method, the proposed DGDE method modifies fewer pixels to generate adversarial samples by introducing a variable population number and a novel crossover and selection strategy. However, the success rate of the initial attack on different image datasets varied greatly. In future studies, we intend to identify the reasons for this phenomenon and improve the success rate of the initial attack.

PMID:39698721 | PMC:PMC11651935 | DOI:10.21037/qims-24-1764

Quant Imaging Med Surg. 2024 Dec 5;14(12):8927-8941. doi: 10.21037/qims-24-907. Epub 2024 Nov 29.

ABSTRACT

BACKGROUND: Deep learning has developed rapidly, and deep learning reconstruction (DLR) methods in magnetic resonance imaging (MRI) are gaining attention for their potential to improve efficacy in clinical work. The preoperative MRI assessment of rectal cancer is crucial for patient management, but the imaging quality is currently limited by a number of factors. DLR could be applied to the preoperative MRI assessment of primary rectal cancer, but research about its specific reliability is limited. Thus, this study aimed to evaluate the reliability of DLR in the preoperative MRI examination of primary rectal cancer.

METHODS: This cross-sectional study was conducted at Ruijin Hospital, Shanghai Jiaotong University School of Medicine from March 2022 to October 2022. Patients with primary rectal cancer underwent routine MRI scans on a 3.0T magnetic resonance scanner (SIGNA Architect, GE Healthcare, USA) with 32-channels flexible coil with conventional reconstruction (ConR) and DLR. The DLR method had three noise reduction levels: DLR-H: 75% noise reduction reconstruction; DLR-M: 50% noise reduction reconstruction; and DLR-L: 25% noise reduction reconstruction. Three components were evaluated: objective image quality; subjective image quality; and diagnostic performance. The objective image quality assessment included the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). The subjective image quality assessment involved evaluating five subjective image quality parameters based on a 4-point Likert scale. The diagnostic performance assessment included tumour (T) staging, node (N) staging, as well as the circumferential resection margin and extramural vascular invasion evaluation. The images were evaluated in a blinded manner by two radiologists with different levels of experience. The paired sample Wilcoxon signed-rank test, Kappa test, interclass correlation coefficient, Chi-square test, Friedman test, and weighted kappa coefficients were used for the statistical analysis.

RESULTS: In total, 61 patients (mean age: 65±12 years; 38 men) were enrolled in the study. The DLR method improved the SNR and CNR values of the images relative to the ConR method, while the DLR-H produced the greatest improvement (P<0.040). The subjective image quality of the DLR-H images was superior to that of the ConR images (P<0.001), but there was no significant difference between the DLR-H and DLR-M images (P≥0.075). The evaluators showed good agreement in subjective scoring, and in the DLR image scoring, the evaluators have the best consistency in the DLR-H images scoring (kappa =0.921, P<0.001). The diagnostic efficacy of the DLR images was comparable to that of the ConR images in terms of T staging [Reader 1 (R1): P=0.603; Reader 2 (R2): P=0.206] and N staging (R1: P=0.990; R2: P=0.884).

CONCLUSIONS: The DLR method improved the quality of the images, and had comparable diagnostic efficacy without additional scanning time to that of the ConR method, and thus could be a feasible option for replacing the ConR method in the preoperative MRI examination of primary rectal cancer.

PMID:39698686 | PMC:PMC11651964 | DOI:10.21037/qims-24-907

Quant Imaging Med Surg. 2024 Dec 5;14(12):8999-9020. doi: 10.21037/qims-24-596. Epub 2024 Nov 29.

ABSTRACT

BACKGROUND: Robust registration of thoracic computed tomography (CT) images is strongly impacted by motion during acquisition, high-density objects, and noise, particularly in lower-dose acquisitions. Despite the enhanced registration speed achieved by popular deep learning (DL) methods, their robustness is often neglected. This study aimed to develop a robust thoracic CT image registration algorithm to address the aforementioned issues.

METHODS: A novel, anatomical structure-aware hierarchical registration. By this method, employing a divide-and-conquer approach, dissimilarity metrics, and regularization terms are selected for different regions based on their distinct image features and motion patterns. These terms are then innovatively reconstructed using the Welsch's function, which allows control over the penalty distribution on the loss values. Subsequently, a novel Welsch parameter update strategy is designed for the task of thoracic CT image registration, enabling dynamic sparsity in registration from coarse to fine levels to accommodate various levels of noise and sliding motion. Moreover, the majorization-minimization (MM) algorithm is used to handle the Welsch terms by constructing surrogate functions based on the current variable values for variable update, thereby reducing the complexity of optimization.

RESULTS: Experimental results on publicly available deformable image registration lab four-dimensional CT (DIR-Lab 4DCT) and chronic obstructive pulmonary disease (COPD) datasets with and without noise, showed that our proposed method achieves comparable performance to state-of-the-art methods in noise-free scenarios [1.14 and 1.19 mm compared to 1.14 and 1.35 mm target registration errors (TREs)], while demonstrating superior robustness in the presence of noise (1.78 and 2.38 mm compared to 2.00 and 3.31 mm TREs). Ablation studies also validated the effectiveness of each component in the method.

CONCLUSIONS: A novel and robust algorithm for thoracic CT image registration has been proposed, which has significant potential for valuable clinical applications, including surgical quantitative imaging.

PMID:39698683 | PMC:PMC11652047 | DOI:10.21037/qims-24-596

Quant Imaging Med Surg. 2024 Dec 5;14(12):8331-8346. doi: 10.21037/qims-24-1022. Epub 2024 Nov 1.

ABSTRACT

BACKGROUND: The use of U-Net and its variations has led to significant advancements in medical image segmentation. However, the encoder-decoder structures of these models often lose spatial information during downsampling. Skip connections can help address this issue; however, they may also introduce excessive irrelevant background information. Additionally, medical images display significant scale variations and complex tissue structures, making it challenging for existing models to accurately separate tissues from the background. To address these issues, we developed the Res2Net-ConvFormer-Dilation-UNet (Res2-CD-UNet), a multi-scale feature extraction network for medical image segmentation.

METHODS: This study presents a novel U-shaped segmentation network that employs Res2Net as the backbone and incorporates a convolution-style transformer in the encoding stage to enhance global attention. Additionally, a novel channel feature fusion block (CFFB) has been introduced in the skip connection stage to minimize the effects of background noise.

RESULTS: The proposed model was evaluated using publicly available datasets, Synapse and Seg.A.2023. Using the Synapse dataset, the average dice similarity coefficient (DSC) reached 83.92%, which was 1.96% higher than the suboptimal model, and the average Hausdorff distance (HD) was 14.51 mm. Among the eight organs evaluated, optimal results were achieved for four organs. Similarly, using the Seg.A.2023 dataset, the proposed model also achieved the best results with an average DSC of 93.27%.

CONCLUSIONS: The results of this study indicate that the proposed model can more accurately segment regions of interest and better extract multi-scale features in medical images than existing deep-learning algorithms.

PMID:39698667 | PMC:PMC11652013 | DOI:10.21037/qims-24-1022

Quant Imaging Med Surg. 2024 Dec 5;14(12):9258-9275. doi: 10.21037/qims-24-595. Epub 2024 Nov 8.

ABSTRACT

BACKGROUND: Rapidly progressive interstitial lung disease (RP-ILD) significantly impacts the prognosis of patients with idiopathic inflammatory myopathies (IIM). High-resolution computed tomography (HRCT) is a crucial noninvasive technique for evaluating interstitial lung disease (ILD). Utilizing quantitative computed tomography (QCT) enables accurate quantification of disease severity and evaluation of prognosis, thereby serving as a crucial computer-aided diagnostic method. This study aimed to establish and validate a machine learning (ML) model to predict RP-ILD in patients with idiopathic inflammatory myopathy-related interstitial lung disease (IIM-ILD) based on QCT and clinical features.

METHODS: A total of 514 patients (367 females, median age 54 years) with IIM-ILD in the China-Japan Friendship Hospital were retrospectively included, out of which 249 cases (165 females, median age 55 years) were identified as having RP-ILD. To extract the quantitative features on HRCT, deep learning (DL) methods were employed, along with demographic factors, pulmonary function test results, and blood gas analysis results; these factors were integrated into a final prediction model.

RESULTS: Logistic regression was chosen as the final model due to its superior area under the curve (AUC) and explainability compared to the other seven ML models. The validation dataset yielded an AUC of 0.882 [95% confidence interval (CI): 0.797-0.967], indicating that the combined QCT and clinical features model outperformed both the QCT-only model and the clinically-only model. In calibration and clinical decision curve analysis, the final model demonstrated minimal prediction bias (concordance index: 0.887, 95% CI: 0.800-0.974, P<0.001) and provided greater net benefit across most thresholds. The nomogram encompassed the incorporation of the following variables: subtype, gender, forced expiratory volume in one second (FEV1%), diffusing capacity for carbon monoxide (DLCO%), oxygenation index (OI), and quantitative ground-glass opacities (GGOs), consolidation, pulmonary vascular, and branches on HRCT.

CONCLUSIONS: When utilizing ML techniques, the baseline QCT has the potential to predict rapid progression in patients with IIM-ILD. The prediction performance will be further improved by incorporating clinical data alongside HRCT features.

KEYWORDS: Idiopathic inflammatory myopathy (IIM); rapidly progressive interstitial lung disease (RP-ILD); high-resolution computed tomography (HRCT); machine learning (ML); quantitative computed tomography (QCT).

PMID:39698644 | PMC:PMC11652001 | DOI:10.21037/qims-24-595

Quant Imaging Med Surg. 2024 Dec 5;14(12):8140-8154. doi: 10.21037/qims-24-489. Epub 2024 Nov 21.

ABSTRACT

BACKGROUND: Deep-learning-based denoising improves image quality and quantification accuracy for low count (LC) positron emission tomography (PET). Conventional deep-learning-based denoising methods only require single LC PET image input. This study aims to propose a deep-learning-based LC PET denoising method incorporating computed tomography (CT) priors to further reduce the dose level.

METHODS: Fifty patients who underwent their routine whole-body 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) PET/CT scans in March 2022 were retrospectively and non-consecutively recruited. For full count (FC) PET, patients were injected with 3.7 MBq/kg FDG and scanned for 5 bed positions with 2 min/bed. LC PET of 1/10 (LC-10) and 1/20 (LC-20) count levels of FC PET were obtained by randomly down-sampling the FC list mode data, which were then paired with FC PET for training U-Net (U-Net-1) and cGAN (cGAN-1). Networks incorporated CT images as input (U-Net-2 and cGAN-2) were also implemented. Quantitative analysis of physical and clinical indices was performed and statistically assessed with Wilcoxon sign-rank test with Bonferroni correction.

RESULTS: Mean square error and structural similarity index were the best for cGAN-2, followed by U-Net-2, cGAN-1 and U-Net-1. The errors of mean standardized uptake value (SUV) and maximum SUV were lowest for cGAN-2, followed by cGAN-1, U-Net-2 and U-Net-1. For cGAN-2, image quality and lesion detectability scores were 3.71±0.94 and 4.25±0.83 for LC-10, 3.57±0.79 and 3.61±1.21 for LC-20, while they were 3.49±0.92 and 4.42±0.08 for FC. Notably, some small lesions were "masked out" on cGAN/U-Net-1 but can be retrieved on cGAN/U-Net-2 denoised PET for LC-20.

CONCLUSIONS: Deep-learning-based LC PET denoising incorporating CT priors is more effective than conventional deep-learning-based denoising with single LC PET input, especially at lower dose levels.

PMID:39698633 | PMC:PMC11652058 | DOI:10.21037/qims-24-489

Quant Imaging Med Surg. 2024 Dec 5;14(12):8771-8784. doi: 10.21037/qims-24-1316. Epub 2024 Nov 13.

ABSTRACT

BACKGROUND: Extensive research on the application of artificial intelligence (AI) and radiomics in lung cancer has been published in recent years; however, it is necessary to identify the current status, hotspots, and trends in the field. Thus, this study conducted a bibliometric analysis of relevant studies to investigate the application of AI and radiomics in lung cancer.

METHODS: Related publications were retrieved from the Web of Science Core Collection (WoSCC). CiteSpace generated the associated co-occurrence network maps in terms of institutions, authors, and keywords. Bibliometrix was used to perform a bibliometric analysis of relevant countries/regions and journals. In addition, the collected information was used to generate figures using R.

RESULTS: A total of 2,989 publications were included in this study, of which 2,804 (93.8%) were articles and 185 (6.2%) were reviews. In 2016, there was a rapid increase in the number of publications in this field. Most of the research originated from China (n=1,365, 45.7%). While Fudan University (n=109, 3.6%) attracted the greatest attention among all institutions. In terms of the authors, Gillies (28 publications, 0.9%) published the greatest number of articles. In terms of the journals, Frontiers in Oncology (n=177, 5.9%) and Radiology (5,152 citations) had the greatest number of publications and influence, respectively. The main keywords identified were "lung cancer", "deep learning", "classification", "computed tomography", and "features". Burst detection suggested that "texture", "image classification", and "false positive reduction" have recently appeared at the frontier of research.

CONCLUSIONS: This study used bibliometric methods to analyze the relevant literature to discuss the current research hotspots and future trends in the application of AI and radiomics in lung cancer. This information may help relevant researchers to shape the direction of future studies, such as innovations in AI techniques standardized feature extraction, and extend understandings of epidermal growth factor receptor mutations.

PMID:39698627 | PMC:PMC11652003 | DOI:10.21037/qims-24-1316

Quant Imaging Med Surg. 2024 Dec 5;14(12):8864-8877. doi: 10.21037/qims-24-1075. Epub 2024 Nov 27.

ABSTRACT

BACKGROUND: The classification of Breast Imaging Reporting and Data System (BI-RADS) category 4A lesions in mammography is complicated by subjective interpretations and unclear criteria, which can lead to potential misclassifications and unnecessary biopsies. Thus, more accurate assessment methods need to be developed. This study aimed to improve the classification prediction of BI-RADS 4A positive lesions in mammography by combining deep learning (DL) technology with relevant clinical factors.

METHODS: A retrospective analysis of 590 patients diagnosed with BI-RADS 4A at Shenzhen People's Hospital and Shenzhen Luohu People's Hospital was conducted, and a multi-faceted approach was employed to construct a robust predictive model. The patients were divided into training, validation, and external validation sets. The classification results from a DL system applied to mammography were recorded, and data on relevant clinical factors were collected. Univariate and multivariate logistic regression analyses were performed to identify the independent predictive factors. A predictive model and nomogram integrating these factors were developed. Assessment metrics, such as the areas under the curve (AUCs), calibration curves, and a decision curve analysis (DCA), were employed to evaluate the diagnostic performance, calibration, and clinical net benefit of the model. External validation was conducted to assess the generalization ability of the model.

RESULTS: Four independent predictive factors (i.e., age, nipple discharge, ultrasound BI-RADS assessment, and DL system classification results) were identified and included in the predictive model. The model showed commendable diagnostic performance with AUC values of 0.85, 0.82, and 0.84 for the training, validation, and external validation sets, respectively. There were no statistically significant differences in the AUCs of the predictive model between the training set, and the internal and external validation sets (P=0.543 and 0.842, respectively). The calibration curves showed excellent calibration in the training, validation, and external validation sets, indicating a minimal deviation between the predicted and actual positive risk probabilities (P=0.906, 0.890, and 0.769, respectively). The DCA results illustrated the clinical net benefit of the model for risk thresholds greater than 0.15 and less than 0.70 in both the internal validation and external validation sets.

CONCLUSIONS: Our predictive model, which incorporated age, nipple discharge, ultrasound BI-RADS assessment, and DL system classification results, emerged as a powerful tool for accurately predicting BI-RADS 4A positive lesions. Its application holds significant promise in helping radiologists enhance diagnostic precision and reduce unnecessary biopsies in BI-RADS 4A positive lesion cases.

PMID:39698623 | PMC:PMC11652064 | DOI:10.21037/qims-24-1075

Nat Comput Sci. 2024 Dec 18. doi: 10.1038/s43588-024-00752-y. Online ahead of print.

ABSTRACT

Crystallization of amorphous precursors into metastable crystals plays a fundamental role in the formation of new matter, from geological to biological processes in nature to the synthesis and development of new materials in the laboratory. Reliably predicting the outcome of such a process would enable new research directions in these areas, but has remained beyond the reach of molecular modeling or ab initio methods. Here we show that candidates for the crystallization products of amorphous precursors can be predicted in many inorganic systems by sampling the local structural motifs at the atomistic level using universal deep learning interatomic potentials. We show that this approach identifies, with high accuracy, the most likely crystal structures of the polymorphs that initially nucleate from amorphous precursors, across a diverse set of material systems, including polymorphic oxides, nitrides, carbides, fluorides, chlorides, chalcogenides and metal alloys.

PMID:39695321 | DOI:10.1038/s43588-024-00752-y

Sci Rep. 2024 Dec 18;14(1):30554. doi: 10.1038/s41598-024-81132-4.

ABSTRACT

Diabetic Retinopathy (DR) stands as a significant global cause of vision impairment, underscoring the critical importance of early detection in mitigating its impact. Addressing this challenge head-on, this study introduces an innovative deep learning framework tailored for DR diagnosis. The proposed framework utilizes the EfficientNetB0 architecture to classify diabetic retinopathy severity levels from retinal images. By harnessing advanced techniques in computer vision and machine learning, the proposed model aims to deliver precise and dependable DR diagnoses. Continuous testing and experimentation shows to the efficiency of the architecture, showcasing promising outcomes that could help in the transformation of both diagnosing and treatment of DR. This framework takes help from the EfficientNet Machine Learning algorithms and employing advanced CNN layering techniques. The dataset utilized in this study is titled 'Diagnosis of Diabetic Retinopathy' and is sourced from Kaggle. It consists of 35,108 retinal images, classified into five categories: No Diabetic Retinopathy (DR), Mild DR, Moderate DR, Severe DR, and Proliferative DR. Through rigorous testing, the framework yields impressive results, boasting an average accuracy of 86.53% and a loss rate of 0.5663. A comparison with alternative approaches underscores the effectiveness of EfficientNet in handling classification tasks for diabetic retinopathy, particularly highlighting its high accuracy and generalizability across DR severity levels. These findings highlight the framework's potential to significantly advance the field of DR diagnosis, given more advanced datasets and more training resources which leads it to be offering clinicians a powerful tool for early intervention and improved patient outcomes.

PMID:39695310 | DOI:10.1038/s41598-024-81132-4

Nature. 2024 Dec 18. doi: 10.1038/s41586-024-08334-8. Online ahead of print.

ABSTRACT

Old age is associated with a decline in cognitive function and an increase in neurodegenerative disease risk1. Brain ageing is complex and is accompanied by many cellular changes2. Furthermore, the influence that aged cells have on neighbouring cells and how this contributes to tissue decline is unknown. More generally, the tools to systematically address this question in ageing tissues have not yet been developed. Here we generate a spatially resolved single-cell transcriptomics brain atlas of 4.2 million cells from 20 distinct ages across the adult lifespan and across two rejuvenating interventions-exercise and partial reprogramming. We build spatial ageing clocks, machine learning models trained on this spatial transcriptomics atlas, to identify spatial and cell-type-specific transcriptomic fingerprints of ageing, rejuvenation and disease, including for rare cell types. Using spatial ageing clocks and deep learning, we find that T cells, which increasingly infiltrate the brain with age, have a marked pro-ageing proximity effect on neighbouring cells. Surprisingly, neural stem cells have a strong pro-rejuvenating proximity effect on neighbouring cells. We also identify potential mediators of the pro-ageing effect of T cells and the pro-rejuvenating effect of neural stem cells on their neighbours. These results suggest that rare cell types can have a potent influence on their neighbours and could be targeted to counter tissue ageing. Spatial ageing clocks represent a useful tool for studying cell-cell interactions in spatial contexts and should allow scalable assessment of the efficacy of interventions for ageing and disease.

PMID:39695234 | DOI:10.1038/s41586-024-08334-8

Sci Rep. 2024 Dec 18;14(1):30542. doi: 10.1038/s41598-024-81961-3.

ABSTRACT

Skin cancer is a significant global health concern, with timely and accurate diagnosis playing a critical role in improving patient outcomes. In recent years, computer-aided diagnosis systems have emerged as powerful tools for automated skin cancer classification, revolutionizing the field of dermatology. This survey analyzes 107 research papers published over the last 18 years, providing a thorough evaluation of advancements in classification techniques, with a focus on the growing integration of computer vision and artificial intelligence (AI) in enhancing diagnostic accuracy and reliability. The paper begins by presenting an overview of the fundamental concepts of skin cancer, addressing underlying challenges in accurate classification, and highlighting the limitations of traditional diagnostic methods. Extensive examination is devoted to a range of datasets, including the HAM10000 and the ISIC archive, among others, commonly employed by researchers. The exploration then delves into machine learning techniques coupled with handcrafted features, emphasizing their inherent limitations. Subsequent sections provide a comprehensive investigation into deep learning-based approaches, encompassing convolutional neural networks, transfer learning, attention mechanisms, ensemble techniques, generative adversarial networks, vision transformers, and segmentation-guided classification strategies, detailing various architectures, tailored for skin lesion analysis. The survey also sheds light on the various hybrid and multimodal techniques employed for classification. By critically analyzing each approach and highlighting its limitations, this survey provides researchers with valuable insights into the latest advancements, trends, and gaps in skin cancer classification. Moreover, it offers clinicians practical knowledge on the integration of AI tools to enhance diagnostic decision-making processes. This comprehensive analysis aims to bridge the gap between research and clinical practice, serving as a guide for the AI community to further advance the state-of-the-art in skin cancer classification systems.

PMID:39695157 | DOI:10.1038/s41598-024-81961-3

Sci Data. 2024 Dec 18;11(1):1376. doi: 10.1038/s41597-024-04206-y.

ABSTRACT

The growing popularity of machine learning (ML) and deep learning (DL) in scientific fields is hindered by the scarcity of high-quality datasets. While quantum mechanical (QM) predictions using DL techniques such as graph neural networks (GNNs) and generative models are gaining traction, insufficient training data remains a bottleneck. The QM40 dataset addresses this challenge by representing 88% of the FDA-approved drug chemical space. It includes molecules containing 10 to 40 atoms and composed of elements commonly found in drug molecular structures (C, O, N, S, F, Cl). QM40 offers valuable resources for researchers which include the core QM40 main dataset, containing 16 key quantum mechanical parameters for 162,954 molecules calculated using the B3LYP/6-31G(2df,p) level of theory in Gaussian16, ensuring consistency with established datasets like QM9 and Alchemy. This compatibility allows for future concatenation of QM40 with these datasets. In addition to other valuable information, the QM40 dataset offers the initial and optimized Cartesian coordinates, Mulliken charges, and detailed bond information, including local vibrational mode force constants, which serve as indicators of bond strength. QM40 can be used to benchmark both existing and new methods for predicting QM calculations using ML and DL techniques.

PMID:39695146 | DOI:10.1038/s41597-024-04206-y