Lesion Boundary Research Articles

DBNet-SI: Dual branch network of shift window attention and inception structure for skin lesion segmentation

The U-shaped convolutional neural network (CNN) has attained remarkable achievements in the segmentation of skin lesion. However, given the inherent locality of convolution, this architecture cannot capture long-range pixel dependencies and multiscale global contextual information effectively. Moreover, repeated convolutions and downsampling operations can readily result in the omission of intricate local fine-grained details. In this paper, we proposed a U-shaped network (DBNet-SI) equipped with a dual-branch module that combines shift window attention and inception structures. First, we proposed a dual-branch module that combines shift window attention and inception structures (MSI) to better capture multiscale global contextual information and long-range pixel dependencies. Specifically, we have devised a cross-branch bidirectional interaction module within the MSI module to enable information complementarity between the two branches in the channel and spatial dimensions. Therefore, MSI is capable of extracting distinguishing and comprehensive features to accurately identify the skin lesion boundaries. Second, we have devised a progressive feature enhancement and information compensation module (PFEIC), which progressively compensates for fine-grained features through reconstructed skip connections and integrated global context attention modules. The results of the experiment show the superior segmentation performance of DBNet-SI compared with other deep learning models for skin lesion segmentation in the ISIC2017 and ISIC2018 datasets. Ablation studies demonstrate that our model can effectively extract rich multiscale global contextual information and compensate for the loss of local details.

An automatic histopathological image segmentation network based on global context-aware module and deep feature aggregation

Histopathological analysis is the gold standard for cancer diagnosis, but traditional manual identification methods are time-consuming, labor-intensive, and highly relied on the experience of pathologists. In recent years, deep learning methods can assist pathologists to achieve accurate and automated detection of lesion areas. However, lesion detection is a difficult task due to its various size, shape, and surrounding complex scenes. Existing methods ignore the global interaction between lesion area and normal area, so it is insufficient to extract important global features. In addition, most methods fail to effectively fuse deep and shallow features, resulting in the extracted features lacking discrimination, and these deficiencies lead to incomplete extraction and irregular boundaries of lesion. To address the above-mentioned issues, an end-to-end histopathological image segmentation network based on global context-aware module and deep feature aggregation (GCMDFA) is proposed for segmenting lesions in histopathological images (HIs). In the GCMDFA, a global context-aware module is applied to better explore the interactive relationship between any two pixels to directly capture global contextual information related to lesion tissue of HIs for achieving global features. To better decode the features of encoder, a deep feature aggregation structure is designed to effectively fuse deep and shallow layer features of HIs. Experimental results on three HI datasets (including public and external data) show that the proposed GCMDFA method is more competitive compared with some state-of-the-art (SOTA) methods. This research can assist pathologists to analyze HIs in clinical computer-aided diagnosis.

Potential Added Value of 18F-FDG PET Metabolic Parameters in Predicting Disease Relapse in Type 1 Autoimmune Pancreatitis

BackgroundThe predictive value of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) metabolic parameters for predicting AIP relapse is currently unknown. This study firstly explored the value of 18F-FDG PET/CT parameters as predictors of type 1 AIP relapse.MethodsThis multicenter retrospective cohort study analyzed 51 patients who received 18F-FDG PET/CT prior to treatment and did not receive maintenance therapy after remission. The study collected baseline characteristics and clinical data and conducted qualitative and semi-quantitative analysis of pancreatic lesions and extrapancreatic organs. The study used three thresholds to select the boundaries of pancreatic lesions to evaluate metabolic parameters, including the maximum standard uptake value (SUVmax), mean standard uptake value (SUVmean), total lesion glycolysis (TLG), metabolic tumor volume (MTV), and tumor-to-normal liver standard uptake value ratio (SUVR). Univariate and multivariate analyses were performed to identify independent predictors and build a recurrence prediction model. The model was internally validated using the bootstrap method and a nomogram was created for clinical application.ResultsIn the univariable analysis, the relapsed group showed higher levels of SUVmax (6.0 ± 1.6 vs. 5.2 ± 1.1; P = 0.047), SUVR (2.3 [2.0–3.0] vs. 2.0 [1.6–2.4]; P = 0.026), and TLG2.5 (234.5 ± 149.1 vs. 139.6 ± 102.5; P = 0.020) among the 18F-FDG PET metabolic parameters compared to the non-relapsed group. In the multivariable analysis, serum IgG4 (OR, 1.001; 95% CI, 1.000–1.002; P = 0.014) and TLG2.5 (OR, 1.007; 95% CI, 1.002–1.013; P = 0.012) were independent predictors associated with relapse of type 1 AIP. A receiver-operating characteristic curve of the predictive model with these two predictors demonstrated an area under the curve of 0.806.Conclusion18F-FDG PET/CT metabolic parameters, particularly TLG2.5, are potential predictors for relapse in patients with type 1 AIP. A multiparameter model that includes IgG4 and TLG2.5 can enhance the ability to predict AIP relapse.

A multi-parameter intrahepatic cholangiocarcinoma scoring system based on modified contrast-enhanced ultrasound LI-RADS M criteria for differentiating intrahepatic cholangiocarcinoma from hepatocellular carcinoma.

To develop a multi-parameter intrahepatic cholangiocarcinoma (ICC) scoring system and compare its diagnostic performance with contrast-enhanced ultrasound (CEUS) liver imaging reporting and data system M (LR-M) criteria for differentiating ICC from hepatocellular carcinoma (HCC). This retrospective study enrolled 62 high-risk patients with ICCs and 62 high-risk patients with matched HCCs between January 2022 and December 2022 from two institutions. The CEUS LR-M criteria was modified by adjusting the early wash-out onset (within 45 s) and the marked wash-out (within 3 min). Then, a multi-parameter ICC scoring system was established based on clinical features, B-mode ultrasound features, and modified LR-M criteria. We found that elevated CA 19-9 (OR=12.647), lesion boundary (OR=11.601), peripheral rim-like arterial phase hyperenhancement (OR=23.654), early wash-out onset (OR=7.211), and marked wash-out (OR=19.605) were positive predictors of ICC, whereas elevated alpha-fetoprotein (OR=0.078) was a negative predictor. Based on these findings, an ICC scoring system was established. Compared with the modified LR-M and LR-M criteria, the ICC scoring system showed the highest area under the curve (0.911 vs. 0.831 and 0.750, both p<0.05) and specificity (0.935 vs. 0.774 and 0.565, both p<0.05). Moreover, the numbers of HCCs categorized as LR-M decreased from 27 (43.5%) to 14 (22.6%) and 4 (6.5%) using the modified LR-M criteria and ICC scoring system, respectively. The modified LR-M criteria-based multi-parameter ICC scoring system had the highest specificity for diagnosing ICC and reduced the number of HCC cases diagnosed as LR-M category.

SG-MIAN: Self-guided multiple information aggregation network for image-level weakly supervised skin lesion segmentation

Nowadays, skin disease is becoming one of the most malignant diseases that threaten people’s health. Computer aided diagnosis based on deep learning has become a widely used technology to assist medical professionals in diagnosis, and segmentation of lesion areas is one of the most important steps in it. However, traditional medical image segmentation methods rely on numerous pixel-level labels for fully supervised training, and such labeling process is time-consuming and requires professional competence. In order to reduce the costs of pixel-level labeling, we proposed a method only using image-level label to segment skin lesion areas. Due to the lack of lesion’s spatial and intensity information in image-level labels, and the wide distribution range of irregular shape and different texture on skin lesions, the algorithm must pay great attention to the automatic lesion localization and perception of lesion boundary. In this paper, we proposed a Self-Guided Multiple Information Aggregation Network (SG-MIAN). Our backbone network MIAN utilizes the Multiple Spatial Perceptron (MSP) solely using classification information as guidance to discriminate the key classification features of lesion areas, and thereby performing more accurate localization and activation of lesion areas. Additionally, adjunct to MSP, we also proposed an Auxiliary Activation Structure (AAS) and two auxiliary loss functions to further self-guided boundary correction, achieving the goal of accurate boundary activation. To verify the effectiveness of the proposed method, we conducted extensive experiments using the HAM10000 dataset and the PH2dataset, which demonstrated superior performance compared to most existing weakly supervised segmentation methods.

Overexpression of ameloblastin in secretory ameloblasts results in demarcated, hypomineralized opacities in enamel.

Introduction: Developmental defects of the enamel manifest before tooth eruption and include amelogenesis imperfecta, a rare disease of underlying gene mutations, and molar-incisor hypomineralization (MIH), a prevalent disease in children originating from environmental and epigenetic factors. MIH enamel presents as the abnormal enamel marked by loss of translucency, demarcation between the healthy and affected enamel, and reduced mineral content. The pathophysiology of opaque, demarcated enamel lesions is not understood; however, the retention of enamel proteins in the matrix has been suggested. Ameloblastin (Ambn) is an enamel protein of the secreted calcium-binding phosphoproteins (SCPPs) critical for enamel formation. When the Ambn gene is mutated or deleted, teeth are affected by hypoplastic amelogenesis imperfecta. Methods: In this study, enamel formation in mice was analyzed when transgenic Ambn was overexpressed from the amelogenin promoter encoding full-length Ambn. Ambn was under- and overexpressed at six increasing concentrations in separate mouse lines. Results: Mice overexpressing Ambn displayed opaque enamel at low concentrations and demarcated lesions at high concentrations. The severity of enamel lesions increased starting from the inner enamel close to the dentino-enamel junction (DEJ) to span the entire width of the enamel layer in demarcated areas. Associated with the opaque enamel were 17-kDa Ambn cleavage products, a prolonged secretory stage, and a thin basement membrane in the maturation stage. Ambn accumulations found in the innermost enamel close to the DEJ and the mineralization front correlated with reduced mineral content. Demarcated enamel lesions were associated with Ambn species of 17kDa and higher, prolonged secretory and transition stages, a thin basement membrane, and shortened maturation stages. Hypomineralized opacities were delineated against the surrounding mineralized enamel and adjacent to ameloblasts detached from the enamel surface. Inefficient Ambn cleavage, loss of contact between ameloblasts, and the altered basement membrane curtailed the endocytic activity; thus, enamel proteins remained unresorbed in the matrix. Ameloblasts have the ability to distinguish between Ambn concentration and Ambn cleavage products through finely tuned feedback mechanisms. The under- or overexpression of Ambn in murine secretory ameloblasts results in either hypoplastic amelogenesis imperfecta or hypomineralization with opaque or sharply demarcated boundaries of lesions, similar to MIH.

The efficacy and safety analysis of endoscopic resection for infratemporal fossa benign mass

Objective: To investigate the efficacy and safety of endoscopic resection of infratemporal fossa mass and to determine the indications for surgery. Methods: A retrospective case series study was conducted, including a total of 29 patients who underwent endoscopic surgery to treat infratemporal fossa mass in the Department of Rhinology of Beijing Tongren Hospital, Capital Medical University, from April 2008 to December 2021. Ten males and 19 females were included in the study, with age of (46.5±13.7) years. Pre-and post-operative sinus CT, sinus or nasopharyngeal enhanced MRI were evaluated, respectively. The main outcome measurements were the total resection of mass and the incidence of surgery-related complications. Results: Among the 29 cases of infratemporal fossa mass, 22 were schwannomas, 3 were cysts, 2 were neurofibromas, 1 was pleomorphic adenoma and 1 was basal cell adenoma. Preoperative imaging showed well-defined lesion boundaries, and postoperative pathology confirmed the benign nature of all cases. The endoscopic transnasal approach was used in 28 patients, while the combination of the transnasal approach and the transoral approach was used in 1 patient. Complete tumor removal was achieved in all cases with a 100% resection rate. The average follow-up time was 38 months (7-168 months), and no tumor recurrence was observed. Conclusions: The Endoscopic transnasal approach is a safe and effective surgical approach for the treatment of benign tumors or masses in the infratemporal fossa.

A conditional generative adversarial networks and Yolov5 Darknet -based skin lesion localization and classification using independent component analysis model

Automated diagnosis systems significantly improve the early detection of skin cancer, thereby contributing to its successful treatment. Skin lesion segmentation represents a crucial yet challenging issue within the realm of computer-assisted dermoscopy image analysis. Numerous segmentation techniques reliant on convolutional neural networks falter in achieving precise lesion boundaries due to the diminishing spatial scale of feature maps as they traverse through network layers. Nonetheless, neural networks demand a substantial amount of data, which presents a notable obstacle in the medical field. More data, and possibly a pre-trained model tailored to medical images, are imperative. Researchers have resorted to extensive augmentation and preprocessing methods, as well as fine-tuning the network using pre-trained models on unrelated images, in order to address the challenge of lesion segmentation as the initial step in skin cancer analysis. This arises from the fact that datasets are limited in size and encompass a diverse array of images, varying in terms of lighting, color, scale, and markings. The proposed method addresses the issue of precisely locating and classifying skin lesions in dermatological images. The architecture of Conditional Generative Adversarial Networks (cGAN) enables the generation of high-resolution, contextually pertinent images, thereby enhancing the discriminative properties of skin lesions. YOLOv5 Darknet has an effective object detection mechanism that aids in precise lesion localization. Independent component analysis (ICA), a signal processing technique that identifies statistically separate components from diverse data sources, is employed to improve the model's robustness and interpretability further. This facilitates extracting latent features from the combined dermatological images and data collection, thereby enhancing the model's understanding of complex lesion characteristics. According to experimental results on an extensive dermatological data set, the proposed cGAN-YOLOv5 technique is effective. In pinpointing and classifying lesion locations, the hybrid model outperforms the competition. The synergistic integration of cGANs, YOLOv5 Darknet, and ICA demonstrates the effectiveness of multifaceted integration in enhancing medical image analysis. Using high-resolution feature blocks and a noise-suppressing attention method, cGAN-YOLOv5 is proposed. Therefore, the proposed strategy segments skin lesions effectively.

LeSAM: Adapt Segment Anything Model for Medical Lesion Segmentation.

The Segment Anything Model (SAM) is a foundational model that has demonstrated impressive results in the field of natural image segmentation. However, its performance remains suboptimal for medical image segmentation, particularly when delineating lesions with irregular shapes and low contrast. This can be attributed to the significant domain gap between medical images and natural images on which SAM was originally trained. In this paper, we propose an adaptation of SAM specifically tailored for lesion segmentation termed LeSAM. LeSAM first learns medical-specific domain knowledge through an efficient adaptation module and integrates it with the general knowledge obtained from the pre-trained SAM. Subsequently, we leverage this merged knowledge to generate lesion masks using a modified mask decoder implemented as a lightweight U-shaped network design. This modification enables better delineation of lesion boundaries while facilitating ease of training. We conduct comprehensive experiments on various lesion segmentation tasks involving different image modalities such as CT scans, MRI scans, ultrasound images, dermoscopic images, and endoscopic images. Our proposed method achieves superior performance compared to previous state-of-the-art methods in 8 out of 12 lesion segmentation tasks while achieving competitive performance in the remaining 4 datasets. Additionally, ablation studies are conducted to validate the effectiveness of our proposed adaptation modules and modified decoder.

Mitigation of Motion Artifacts in Handheld Laser Speckle Contrast Imaging Illustrated on Psoriasis Lesions.

Handheld laser speckle contrast imaging (LSCI) is crucial in clinical settings, but motion artifacts (MA) can compromise perfusion image reliability. Current prevention and suppression methods are often impractical or complex. Machine vision techniques, promising in medical imaging, could improve signal quality, but their use in suppressing MA is still unexplored. We propose an innovative method based on linear regression for MA correction (MAC) in LSCI and validate it in vivo. We performed paired handheld and mounted LSCI measurements on 14 subjects with psoriasis using the previously validated handheld perfusion imager (HAPI). By marking lesion boundaries for clinical purposes, the HAPI used a monochromatic camera for both speckle imaging and motion detection, simplifying hardware requirements. Accurate estimation of relative displacements between the test object and LSCI probe allowed us to apply MAC to the perfusion images. Local perfusion values correlated with applied speed were used to calculate and correct MA. The difference between mean perfusion in handheld and mounted modes after MAC significantly decreased (median error 14.2 perfusion units (p.u.) on lesions before correction ( ) and 0.5 p.u. after correction ( p=0.2)). The findings provide evidence for robust handheld LSCI and validate the MA technique in psoriasis case. Of the two causes of MA-onsurface speeds and wavefront tilt-we address the former and correct mean perfusion, assuming constant temporal perfusion at each location. We describe a practical, non-contact, marker-free technique for reliable handheld perfusion imaging, supporting further clinical translation in plastic surgery and burns.

A Bi-Directionally Fused Boundary Aware Network for Skin Lesion Segmentation.

It is quite challenging to visually identify skin lesions with irregular shapes, blurred boundaries and large scale variances. Convolutional Neural Network (CNN) extracts more local features with abundant spatial information, while Transformer has the powerful ability to capture more global information but with insufficient spatial details. To overcome the difficulties in discriminating small or blurred skin lesions, we propose a Bi-directionally Fused Boundary Aware Network (BiFBA-Net). To utilize complementary features produced by CNNs and Transformers, we design a dual-encoding structure. Different from existing dual-encoders, our method designs a Bi-directional Attention Gate (Bi-AG) with two inputs and two outputs for crosswise feature fusion. Our Bi-AG accepts two kinds of features from CNN and Transformer encoders, and two attention gates are designed to generate two attention outputs that are sent back to the two encoders. Thus, we implement adequate exchanging of multi-scale information between CNN and Transformer encoders in a bi-directional and attention way. To perfectly restore feature maps, we propose a progressive decoding structure with boundary aware, containing three decoders with six supervised losses. The first decoder is a CNN network for producing more spatial details. The second one is a Partial Decoder (PD) for aggregating high-level features with more semantics. The last one is a Boundary Aware Decoder (BAD) proposed to progressively improve boundary accuracy. Our BAD uses residual structure and Reverse Attention (RA) at different scales to deeply mine structural and spatial details for refining lesion boundaries. Extensive experiments on public datasets show that our BiFBA-Net achieves higher segmentation accuracy, and has much better ability of boundary perceptions than compared methods. It also alleviates both over-segmentation of small lesions and under-segmentation of large ones.

Effects of Contact Force on Lesion Size During Pulsed Field Catheter Ablation: Histochemical Characterization of Ventricular Lesion Boundaries.

Effects of contact force (CF) on lesion formation during pulsed field ablation (PFA) have not been well validated. The purpose of this study was to determine the relationship between average CF and lesion size during PFA using a swine-beating heart model. A 7F catheter with a 3.5-mm ablation electrode and CF sensor (TactiCath SE, Abbott) was connected to a PFA system (CENTAURI, Galvanize Therapeutics). In 5 closed-chest swine, biphasic PFA current was delivered between the ablation electrode and a skin patch at 40 separate sites in right ventricle (28 Amp) and 55 separate sites in left ventricle (35 Amp) with 4 different levels of CF: (1) low (CF range of 4-13 g; median, 9.5 g); (2) moderate (15-30 g; median, 21.5 g); (3) high (34-55 g; median, 40 g); and (4) no electrode contact, 2 mm away from the endocardium. Swine were sacrificed at 2 hours after ablation, and lesion size was measured using triphenyl tetrazolium chloride staining. In 1 additional swine, COX (cytochrome c oxidase) staining was performed to examine mitochondrial activity to delineate reversible and irreversible lesion boundaries. Histological examination was performed with hematoxylin and eosin and Masson trichrome staining. Ablation lesions were well demarcated with triphenyl tetrazolium chloride staining, showing (1) a dark central zone (contraction band necrosis and hemorrhage); (2) a pale zone (no mitochondrial activity and nuclear pyknosis, indicating apoptosis zone); and a hyperstained zone by triphenyl tetrazolium chloride and COX staining (unaffected normal myocardium with preserved mitochondrial activity, consistent with reversible zone). At constant PFA current intensity, lesion depth increased significantly with increasing CF. There were no detectable lesions resulting from ablation without electrode contact. Acute PFA ventricular lesions show irreversible and reversible lesion boundaries by triphenyl tetrazolium chloride staining. Electrode-tissue contact is required for effective lesion formation during PFA. At the same PFA dose, lesion depth increases significantly with increasing CF.

Application of artificial intelligence-based dual source CT scanning in the differentiation of lung adenocarcinoma in situ and minimally invasive adenocarcinoma.

Lung adenocarcinoma is the most common type of lung cancer with highly incidence and mortality. Due to the overlap of morphological features, it is difficult to distinguish clinically between preinvasive lesions (in situ adenocarcinoma, AIS) and invasive lesions (minimally invasive adenocarcinoma, MIA), which appear as ground glass cloudy nodules. This study was performed to probe the application value of artificial intelligence (AI)-based dual source CT scanning in the differentiation of AIS as well as MIA. The clinical data of 136 patients in Shanghai Baoshan Hospital of Integrated Traditional Chinese and Western Medicine from January 2019 to January 2022 were retrospectively analyzed. The accuracy of AI in distinguishing lung AIS (n=76) and MIA (n=60) were analyzed. The effectiveness of AI in detecting nodules and its diagnostic efficacy for AIS and MIA were explored. The proportion of patients with clear and regular lesion boundaries in AIS was higher than that in MIA. The mean lesion diameter of AIS patients was shorter than MIA patients. There was no difference in the CT value between AIS and MIA in the ground glass nodule density area of pure ground glass nodule and mixed ground glass nodule, but the CT value of the solid nodule density area in AIS was lower. The occurrence of pulmonary vascular abnormality, air bronchogram sign, and pleural depression in AIS patients were lower than MIA patients. The detection rate of AI for lung adenocarcinoma with nodule diameter ≤ 5 mm, complete solid nodules and ground glass nodules was significantly higher than radiologists. The sensitivity, specificity, positive prediction rate, negative prediction rate and accuracy of AI detection were significantly higher than radiologists. AI-based dual source CT scanning can clearly show the morphological characteristics of lung adenocarcinoma, which is helpful for the differential diagnosis of lung AIS as well as MIA.

Characterization of white matter lesions in multiple sclerosis using proton density and T1-relaxation measures

PurposeAlthough lesion dissemination in time is a defining characteristic of multiple sclerosis (MS), there is a limited understanding of lesion heterogeneity. Currently, conventional sequences such as fluid attenuated inversion recovery (FLAIR) and T1-weighted (T1W) data are used to assess MS lesions qualitatively. Estimating water content could provide a measure of local tissue rarefaction, or reduced tissue density, resulting from chronic inflammation. Our goal was to utilize the proton spin density (PD), derived from a rapid, multi-contrast STAGE (strategically acquired gradient echo) protocol to characterize white matter (WM) lesions seen on T2W, FLAIR and T1W data. Materials and methodsTwenty (20) subjects with relapsing-remitting MS were scanned at 3 T using T1W, T2-weighted, FLAIR and strategically acquired gradient echo (STAGE) sequences. PD and T1 maps were derived from the STAGE data. Disease severity scores, including Extended Disability Status Scale (EDSS) and Multiple Sclerosis Functional Composite (MSFC), were correlated with total, high PD and high T1 lesion volumes. A probability map of high PD regions and all lesions across all subjects was generated. Five perilesional normal appearing WM (NAWM) bands surrounding the lesions were generated to compare the median PD and T1 values in each band with the lesional values and the global WM. ResultsT1W intensity was negatively correlated with PD as expected (R = −0.87, p < 0.01, R2 = 0.756) and the FLAIR signal was suppressed for high PD volumes within the lesions, roughly for PD ≥ 0.85. The threshold for high PD and T1 regions was set to 0.909 and 1953.6 ms, respectively. High PD regions showed a high probability of occurrence near the boundary of the lateral ventricles. EDSS score and nine-hole peg test (dominant and non-dominant hand) were significantly correlated with the total lesion volume and the volumes of high PD and T1 regions (p < 0.05). There was a significant difference in PD/T1 values between the high PD/T1 regions within the lesions and the remaining lesional tissue (p < 0.001). In addition, the PD values of the first NAWM perilesional band directly adjacent to the lesional boundary displayed a significant difference (p < 0.05) compared to the global WM. ConclusionLesions with high PD and T1s had the highest probability of occurrence at the boundary of the lateral ventricles and likely represent chronic lesions with significant local tissue rarefaction. Moreover, the perilesional NAWM exhibited subtly increasing PD and T1 values from the NAWM up to the lesion boundary. Unlike on the T1 maps, the perilesional band adjacent to the lesion boundary possessed a significantly higher PD value than the global WM PD values. This shows that PD maps were sensitive to the subtle changes in NAWM surrounding the lesions.

A Microenvironment-responsive small-molecule probe and application in quick acute myocardial infarction imaging

Acute myocardial infarction (AMI) patients are at an elevated risk for life-threatening myocardial ischemia/reperfusion injury. Early-stage nonradioactive and noninvasive diagnosis of AMI is imperative for the subsequent disease treatment, yet it presents substantial challenges. After AMI, the myocardium typically exhibits elevated levels of peroxynitrite (ONOO−), constituting a distinct microenvironmental feature. In this context, the near-infrared imaging probe (BBEB) is employed to precisely delineate the boundaries of AMI lesions with a high level of sensitivity and specificity by monitoring endogenous ONOO−. This probe allows for the early detection of myocardial damage at cellular and animal levels, providing exceptional temporal and spatial resolution. Notably, BBEB enables visualization of ONOO− level alterations during AMI treatment incorporating antioxidant drugs. Overall, BBEB can rapidly and accurately visualize myocardial injury, particularly in the early stages, and can further facilitate antioxidant drug screening.

BCUIS‐Net: A breast cancer ultrasound image segmentation network via boundary‐aware and shape feature fusion

AbstractBreast cancer is a highly lethal disease with the highest mortality rate among women worldwide. Breast tumor segmentation from ultrasound images plays a critical role in enabling early detection, leading to a reduction in mortality rates. However, the challenge of ultrasound breast cancer segmentation arises from factors such as indistinct lesion boundaries, noise artifacts, and inhomogeneous intensity distribution within the lesion region. To address the bottlenecks, a novel boundary‐aware shape feature fusion network (BCUIS‐Net) is proposed to segment breast lesion in ultrasound images. Firstly, a boundary‐aware module (BAM) is put forward to accurately localize the ambiguous tumor regions and boundaries by embedding the horizontal and vertical position information into the channel attention. Subsequently, a shape feature fusion (SFF) module is presented to fuse shape features and segmentation features, in order to adaptively extract their complementary features by aggregating contextual information in an attention module. Specifically, the different levels of features from the encoder are up‐sampled to the original image size and fed into the BAM to predict the boundary map. The boundary and decoder‐generated feature maps are thereafter fused by the SFF module to exploit the complementarity between them to correct errors in segmentation and shape features, effectively eliminating false detections and noise in the features to achieve accurate segmentation of pathological regions. Finally, the shape fusion loss is derived from a combination of the binary cross‐entropy loss and the distance map loss to intelligently penalize incorrect predictions and thus improve the attention to boundary locations. The performance of the network is evaluated in two public breast ultrasound datasets. Experimental results verify that the proposed method obtains superior segmentation results and outperforms the most recent state‐of‐the‐art, in which IOU is increased by 2.15% and 2.59% on UDIAT and BUSI, respectively.

90Y SPECT scatter estimation and voxel dosimetry in radioembolization using a unified deep learning framework

Purpose90Y SPECT-based dosimetry following radioembolization (RE) in liver malignancies is challenging due to the inherent scatter and the poor spatial resolution of bremsstrahlung SPECT. This study explores a deep-learning-based absorbed dose-rate estimation method for 90Y that mitigates the impact of poor SPECT image quality on dosimetry and the accuracy–efficiency trade-off of Monte Carlo (MC)-based scatter estimation and voxel dosimetry methods.MethodsOur unified framework consists of three stages: convolutional neural network (CNN)-based bremsstrahlung scatter estimation, SPECT reconstruction with scatter correction (SC) and absorbed dose-rate map generation with a residual learning network (DblurDoseNet). The input to the framework is the measured SPECT projections and CT, and the output is the absorbed dose-rate map. For training and testing under realistic conditions, we generated a series of virtual patient phantom activity/density maps from post-therapy images of patients treated with 90Y-RE at our clinic. To train the scatter estimation network, we use the scatter projections for phantoms generated from MC simulation as the ground truth (GT). To train the dosimetry network, we use MC dose-rate maps generated directly from the activity/density maps of phantoms as the GT (Phantom + MC Dose). We compared performance of our framework (SPECT w/CNN SC + DblurDoseNet) and MC dosimetry (SPECT w/CNN SC + MC Dose) using normalized root mean square error (NRMSE) and normalized mean absolute error (NMAE) relative to GT.ResultsWhen testing on virtual patient phantoms, our CNN predicted scatter projections had NRMSE of 4.0% ± 0.7% on average. For the SPECT reconstruction with CNN SC, we observed a significant improvement on NRMSE (9.2% ± 1.7%), compared to reconstructions with no SC (149.5% ± 31.2%). In terms of virtual patient dose-rate estimation, SPECT w/CNN SC + DblurDoseNet had a NMAE of 8.6% ± 5.7% and 5.4% ± 4.8% in lesions and healthy livers, respectively; compared to 24.0% ± 6.1% and 17.7% ± 2.1% for SPECT w/CNN SC + MC Dose. In patient dose-rate maps, though no GT was available, we observed sharper lesion boundaries and increased lesion-to-background ratios with our framework. For a typical patient data set, the trained networks took ~ 1 s to generate the scatter estimate and ~ 20 s to generate the dose-rate map (matrix size: 512 × 512 × 194) on a single GPU (NVIDIA V100).ConclusionOur deep learning framework, trained using true activity/density maps, has the potential to outperform non-learning voxel dosimetry methods such as MC that are dependent on SPECT image quality. Across comprehensive testing and evaluations on multiple targeted lesions and healthy livers in virtual patients, our proposed deep learning framework demonstrated higher (66% on average in terms of NMAE) estimation accuracy than the current “gold-standard” MC method. The enhanced computing speed with our framework without sacrificing accuracy is highly relevant for clinical dosimetry following 90Y-RE.

Epidermis lesion detection via optimized distributed capsule neural network

Skin cancer, encompassing various forms such as melanoma, basal cell carcinoma, and others, remains a significant global health concern, often proving fatal if not diagnosed and treated in its early stages. The challenge of accurately diagnosing skin cancer, particularly melanoma, persists even for experienced dermatologists due to the intricate and unpredictable nature of its symptoms. To address the need for more accurate and efficient skin cancer detection, a novel Golden Hawk Optimization-based Distributed Capsule Neural Network (GHO-DCaNN) is proposed. This novel technique leverages advanced computational methods to improve the reliability and precision of skin cancer diagnosis. An optimized clustering-based segmentation approach is introduced, integrating the innovative Sewer Shad Fly Optimization (SSFO), which combines elements of both mayfly and moth flame optimization. This integration enhances the accuracy of lesion boundary delineation and feature extraction. The core of the innovation lies in the optimized distributed capsule neural network, which is trained using the Hybrid GHO. This optimizer, inspired by the behaviors of the golden eagle and fire hawk, ensures the effectiveness of epidermis lesion detection, pushing the boundaries of skin cancer diagnosis methods. The achievements based on the metrics, like specificity, sensitivity, and accuracy show 97.53%, 99.05%, and 98.83% for 90% of training and 97.83%, 99.50%, and 99.06% for k-fold of 10, respectively.

Evaluation of virtual monochromatic imaging with dual-energy computed tomography of small liver metastases from malignant abdominal tumours: Quantitative and qualitative analyses

Background The assessment of small metastatic liver tumours using dual-energy computed tomography (DECT) has not been fully established. Purpose To assess the effect of low-keV virtual monochromatic imaging (VMI) with non-contrast and contrast-enhanced DECT on the qualitative and quantitative image parameters of small liver metastases. Material and methods Two radiologists retrospectively evaluated 92 metastatic liver tumours (5–20 mm) in 32 patients. Non-contrast and contrast-enhanced VMI were reconstructed at seven energy levels (40–100 keV) with 10-keV intervals. Lesion boundary, lesion delineation, image noise, and overall image quality were evaluated using the visual analogue scale. A high subjective score indicates good overall image quality, clear nodal boundaries and delineation, and less noticeable image noise. Subjective scores were compared using the Kruskal–Wallis test. A quantitative analysis involving the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) was performed. Results The lesion boundary was highest at 40 keV and significantly improved during the non-contrast portal venous phase compared to that at higher keV ( p &lt; .005). The lesion delineation score was significantly higher at 40 keV and tended to decrease at higher keV. Image noise and overall image quality were rated low at low keV; however, those at 80, 90, and 100 keV were rated the highest ( p &lt; .005). The CNR and SNR were highest for non-contrast CT at 100 keV. During the portal venous phase, no significant differences were observed in CNR and SNR at each keV. Conclusion Low-keV imaging using non-contrast and contrast-enhanced DECT is useful for delineating small hepatic metastatic tumours.

Effects of bone mineral density at the lateral sclerotic boundary on the femoral head collapse onset in osteonecrosis of the femoral head: A preliminary study

BackgroundIn the natural course of osteonecrosis of the femoral head, sclerotic changes at the boundary of necrotic lesion gradually occur until femoral head collapse. This study aims to examine the effects of bone mineral density at the lateral boundary of necrotic lesion on a subsequent femoral head collapse. MethodsWe developed patient-specific finite element models of 9 hips with subsequent collapse and 10 hips without subsequent collapse. Cubic regions of interest were selected at both subchondral areas of the lateral boundary and the adjacent necrotic lesion. Bone mineral density values of the regions of interest were quantitatively measured, and a ratio of bone mineral density values (lateral boundary/necrotic lesion) was calculated. Stress values at the lateral boundary were also evaluated. FindingsThe ratio of bone mineral density values was significantly higher in hips with subsequent collapse than that without subsequent collapse (p = 0.0016). The median equivalent stress and shear stress were significantly higher in hips with subsequent collapse than that without subsequent collapse (p = 0.0071, and p = 0.0143, respectively). The ratio of bone mineral density values showed a promising value in predicting the occurrence of subsequent femoral head collapse (AUC = 0.97). InterpretationOur results indicated that bone mineral density value at the lateral boundary of necrotic lesion may be associated with the occurrence of subsequent femoral head collapse in pre-collapse stage osteonecrosis of the femoral head.