Lesion Shape Research Articles

An MRI assessment of mechanisms underlying lesion growth and shrinkage in multiple sclerosis.

To assess the pathological mechanisms contributing to white matter (WM) lesion expansion or contraction and remyelination in multiple sclerosis (MS). We assessed 1,613 lesions in 49 people with relapsing-remitting MS in the CCMR-One bexarotene trial (EudraCT 2014-003145-99). We measured lesion orientation relative to WM tracts, surface-in gradients and veins. Jacobian deformation was used to assess lesion expansion over 6 months, while magnetization transfer ratio (MTR) imaging was used to assess remyelination. At baseline, 33% of lesions were aligned with veins, 2% along WM tracts, 0% with surface-in gradients, and 4% orthogonal to veins. No significant differences were observed in lesion shape, while lesions aligned with surface-in gradients and with veins had lower volume compared to all remaining orientations. At follow-up, 13% of lesions expanded and 7% contracted. The directions for both expansion and contraction were 18% and 8%, respectively, along WM tracts, 20% and 15% parallel to veins, 22% and 23% orthogonal to veins and 0% and 1% along surface-in gradients. Bexarotene had no effect on lesion expansion or contraction, but MTR significantly increased in lesions aligned with surface-in gradients and veins. Lesion expansion and shrinkage are affected by venous and WM tract factors, but these do not influence bexarotene's capacity to promote remyelination. This, instead, appears to be affected by surface-in factors. To limit lesion expansion and maximize tissue repair, multiple processes may need to be targeted.

Computed tomographic findings in 25 cats with ear canal neoplasia.

Computed tomography is commonly used to evaluate feline otic disease; however, published studies characterizing the CT appearance of ear canal neoplasia are limited. The purpose of this multicenter, retrospective, secondary analysis, cross-sectional study was to describe the CT features of histopathologically confirmed feline ear canal neoplasia. The CT studies of 25 cats with ear canal neoplasia were prospectively scored by consensus of two veterinary radiologists. Recorded parameters were the presence of focal or multifocal tissue enlargement (mass/masses), lesion shape, location of the center of mass, attenuation characteristics, features of contrast enhancement, involvement of otic structures, calvarial and brain changes, changes of nearby structures, and lymphadenopathy. There was a significant overlap of CT findings between cats with malignant ceruminous gland neoplasia, ceruminous gland adenoma, and squamous cell carcinoma (SCC). Ceruminous gland adenoma was typically homogeneous in attenuation with homogeneous contrast enhancement and no intralesional fluid accumulations (IFAs) or involvement of adjacent structures. In contrast, SCC consistently had heterogeneous attenuation, heterogeneous contrast enhancement, IFAs, and involvement/invasion of adjacent structures. Malignant ceruminous gland neoplasia had variable attenuation and pattern of contrast enhancement with occasional IFAs and occasional involvement/invasion of adjacent structures. Knowledge of these imaging features will inform the creation of prioritized differential diagnosis lists. However, a biopsy is required to confirm the diagnosis.

An adaptive and lightweight YOLOv5 detection model for lung tumor in PET/CT images

Multi-modal medical images are important in tumor lesion detection. However, the existing detection models only use single-modal to detect lesions, a multi-modal semantic correlation is not enough to consider and lacks ability to express the shape, size, and contrast degree features of lesions. A Cross Modal YOLOv5 model (CMYOLOv5) is proposed. Firstly, there are two networks, auxiliary network is consisted by dual-branch structure to extract semantic information from PET and CT, backbone network is consisted by YOLOv5 to extract semantic information from PET/CT. Secondly, Cross-modal Features Fusion (CFF) is designed in auxiliary network to fuse PET functional information and CT anatomical information. Self-Adaptive Attention Fusion (AAF) is designed in backbone network to fuse and enhance three-modal complementary information. Thirdly, Self-Adaptive Transformer (SAT) is designed in feature enhance neck. Using Transformer with deformable attention mechanism to focus on lung tumor region. Using MLP with channel attention mechanism to enhance features representation ability of lung tumor region. Finally, Reparameter Residual Block (RRB) and Reparameter Convolution operation (RC) are designed to fully learn richer PET, CT and PET/CT feature. Comparative experiments are conducted on clinical lung tumor PET/CT multi-modality dataset, the effectiveness of CMYOLOv5 is verified by Precision, Recall, mAP, F1, FPS, and training time, experimental results are 97.16%, 96.41%, 97.18%, 96.78%, 96.37 and 3912 s. CMYOLOv5 has high precision in the detection of irregular lung tumors, which is superior to the existing advanced methods.

Artificial Intelligence-Aided Diagnosis System for the Detection and Classification of Private-Part Skin Diseases: Decision Analytical Modeling Study.

Private-part skin diseases (PPSDs) can cause a patient's stigma, which may hinder the early diagnosis of these diseases. Artificial intelligence (AI) is an effective tool to improve the early diagnosis of PPSDs, especially in preventing the deterioration of skin tumors in private parts such as Paget disease. However, to our knowledge, there is currently no research on using AI to identify PPSDs due to the complex backgrounds of the lesion areas and the challenges in data collection. This study aimed to develop and evaluate an AI-aided diagnosis system for the detection and classification of PPSDs: aiding patients in self-screening and supporting dermatologists' diagnostic enhancement. In this decision analytical modeling study, a 2-stage AI-aided diagnosis system was developed to classify PPSDs. In the first stage, a multitask detection network was trained to automatically detect and classify skin lesions (type, color, and shape). In the second stage, we proposed a knowledge graph based on dermatology expertise and constructed a decision network to classify seven PPSDs (condyloma acuminatum, Paget disease, eczema, pearly penile papules, genital herpes, syphilis, and Bowen disease). A reader study with 13 dermatologists of different experience levels was conducted. Dermatologists were asked to classify the testing cohort under reading room conditions, first without and then with system support. This AI-aided diagnostic study used the data of 635 patients from two institutes between July 2019 and April 2022. The data of Institute 1 contained 2701 skin lesion samples from 520 patients, which were used for the training of the multitask detection network in the first stage. In addition, the data of Institute 2 consisted of 115 clinical images and the corresponding medical records, which were used for the test of the whole 2-stage AI-aided diagnosis system. On the test data of Institute 2, the proposed system achieved the average precision, recall, and F1-score of 0.81, 0.86, and 0.83, respectively, better than existing advanced algorithms. For the reader performance test, our system improved the average F1-score of the junior, intermediate, and senior dermatologists by 16%, 7%, and 4%, respectively. In this study, we constructed the first skin-lesion-based dataset and developed the first AI-aided diagnosis system for PPSDs. This system provides the final diagnosis result by simulating the diagnostic process of dermatologists. Compared with existing advanced algorithms, this system is more accurate in identifying PPSDs. Overall, our system can not only help patients achieve self-screening and alleviate their stigma but also assist dermatologists in diagnosing PPSDs.

The efficacy of a reporting and data system for computed tomography imaging in assessing the likelihood of malignancy in submandibular salivary gland tumors: a preliminary study.

To establish a submandibular gland imaging reporting and data system (SI-RADS) and explore its value in diagnosing submandibular gland (SMG) tumors. Patients with primary SMG tumors who underwent enhanced computed tomography (CT) scanning and surgical intervention at Peking University Hospital of Stomatology (2013-2022) were retrospectively included. The imaging data were assessed for acquisition of the relevant imaging features, including lesion boundary, shape, enhancement, calcification, lymph node, and adjacent tissue involvement, and vascular invasion. Thereafter, these lesions were categorized into 6 grades: RADS 1, normal; RADS 2, benign; RADS 3, probably benign; RADS 4a, low probability of malignancy; RADS 4b, moderate probability of malignancy; RADS 5, high probability of malignancy. In total, 336 SMG lesions (272 benign and 64 malignant) were included. The proportions of malignant tumors were 0%, 5.1%, 33.3%, 63.0%, and 100% for SI-RADS 2, 3, 4a, 4b, and 5 lesions, respectively. With an increase in SI-RADS grading, the proportion of malignancy increased significantly. Significant differences in malignancy proportions were observed between adjacent grades. The SI-RADS based on enhanced CT images exhibits significant importance for preoperative differentiation between benign and malignant SMG tumors. By precisely delineating the imaging characteristics, it provides clinicians with a tool for more accurate preoperative diagnosis and treatment planning.

DESIGN & DEVELOPMENT OF DIFFERENTIAL DIAGNOSIS SOFTWARE FOR SALIVARY GLANDS DISEASE AND COMPARISON WITH CLINICIAN'S RESULTS

Introduction: Due to increase of population, increasing number of less-experienced dentists more associated information and consequent updates of such information; the need for design of information processing systems to help in better diagnosis of patients with oral disorders becomes more evident. Among other factors, some factors such as age, clinical symptoms, location and shape of lesion could influence the decision to treat salivary gland disorders and determine of their possibility. Up to now, there is no software for differential diagnosis of salivary gland diseases. Objectives: The objective of present study is to design and develop a software so as to help to differential diagnosis of salivary gland diseases and determine the accuracy and effectiveness of its results. Methods: Software based on Bayes' theorem was designed with the help of editor application into which the information of 60 salivary gland diseases was inserted. Then, the information of 40reports of salivary gland diseases was added to the software. In the next stage, the information of the chosen cases given to two expert clinicians so as to develop a differential diagnosis. The result was analyzed using menemar and Wilcoxon test. Results: The results were compared with results of the reports. Almost 60 percent of generated answers by the software and two clinicians included absolute diagnosis of statistically insignificant difference (P-value=0.96). Conclusion: The results showed the acceptable performance of the software. Although the software cannot act instead of a clinician due to the fact that the software user should be quite familiar with salivary gland diseases and methods of examination, this software can be used to help the clinicians in differential diagnosis and training objectives.

Epidemiological features of cutaneous leishmaniosis in North Khorasan Province in Iran: A cross-sectional study

Objective: To investigate the trend and epidemiological status of cutaneous leishmaniasis in the North of Khorasan province, Iran from 2014 to 2022. Methods: This a cross-sectional study collected and analyzed the information of 849 patients with cutaneous leishmaniasis (CL) through the health center portal from 2014 to 2022. The relationship between sex, number of lesions, and lesion shape was investigated. Results: Totally 849 participants were included and their mean age was 23.0 ± 20.2 years. The most common parasite species found was Leishmania major (n=714, 84.1%). The number of CL in Esfarayen County increased from 2014 (n=83) to 2022 (n=147) and was more prevalent in men than in women. Furthermore, there was no significant difference in the number and shape of lesions between men and women. Conclusions: As leishmania major is the predominant type in Esfarayen County of Iran and the number of cases is increasing, it is important to control the reservoir and disease vector in this area. Therefore, a more comprehensive surveillance is suggested.

NCOG-25. STEREOTACTIC RADIOSURGERY VERSUS SURGICAL MICRODISSECTION FOR JUGULAR FORAMEN SCHWANNOMAS: A RETROSPECTIVE COMPARATIVE STUDY

Abstract INTRODUCTION Jugular foramen schwannomas (JFS) are rare, benign tumors arising from lower cranial nerves. Given their delicate location, surgical microdissection carries a substantial risk of neurological morbidity. Stereotactic radiosurgery (SRS) presents an alternative, particularly for smaller or residual tumors. OBJECTIVE This study aims to compare the outcomes and symptomatic improvements between these two treatment modalities. METHODS We conducted a retrospective analysis of 32 patients with JFS who underwent SRS (13 cases [40.6%]) or surgical dissection (19 cases [59.4%]) as their primary management at Stanford Medical Center between 1998 and 2024. Outcomes included progression-free survival (PFS), post-management adverse events based on common terminology criteria for adverse events (CTCAE), symptom improvement, overall survival (OS), and the necessity for secondary interventions. RESULTS Significant differences were observed in baseline characteristics between the SRS and surgery groups, including mean age (62.52 vs. 46.87 years, p<0.001), mean largest diameter (34.38 vs. 53.66 mm, p=0.02), and median total tumor volume (6.50 vs. 19.70 mm3, p=0.009). There were no significant differences in sex or lesion shape (dumbbell vs. non-dumbbell shaped). After adjusting for baseline characteristics, no significant differences were noted in PFS (85.71 vs. 77.77%), OS (92.3 vs 100%), symptom improvement (61.53 vs. 52.63%) or mean CTCAE grade (1.31 vs. 1.58) between the SRS and surgery groups. SRS patients had significantly lower odds of requiring secondary treatment procedures after their primary intervention as compared to those in the surgery group (OR=0.01, 95% CI: 0.0001-0.59, p=0.02). CONCLUSION Both SRS and surgery provide exceptional outcomes for JFS with no significant differences in PFS, OS, CTCAE grades, or symptom improvement between the two groups. However, SRS patients had lower odds of requiring secondary interventions for JFS. This is among the first studies comparing surgical microdissection and SRS for JFS. Future controlled prospective studies are necessary to derive more definitive conclusions.

Accuracy of imaging of BI-RADS 4 subcategorizations in breast lesion diagnosis: Radiologic-pathologic correlation.

To correlate breast imaging-reporting and data system (BI-RADS) category 4 lesions with histopathology results to assess the accuracy of subcategorization. A retrospective study was carried out from September 2021 to June 2022. A total of 247 breast lesions were reviewed categorized as BI-RADS 4 using ultrasound (US) and digital mammography. Feature analysis of the lesions were obtained using BI-RADS terminology and assigned to subcategories (4A, 4B, and 4C). Pathological analysis was carried out on tissue obtained through US-guided core biopsy. A p-value of <0.05 was considered significant. Of the 247 lesions, 135 were categorized as subcategory 4A, 68 as 4B, and 44 as 4C. Overall, 41 (16.6%) had malignant lesions, while 206 (83.4%) had benign lesions. The mean age of the patients with benign versus malignant lesions was (43.18±14.02 vs. 51.24±14.15 years; p<0.001). Mean size of benign versus malignant lesions was (1.93±1.65 vs. 3.82±3.89cm; p<0.001). Findings were compared with histopathology, and the positive predictive value fell within the reference range for subcategories 4C (>70%). High reliability was observed between the 2 readers, with a weighted Cohen's Kappa value of 0.79 (0.73-0.85). Significant disagreements in the assignment of features on radiological lesion characterization were observed between the 2 readers regarding lesion density, shape, echo pattern, vascularity, and borders. The results of this study contribute to the existing body of knowledge, emphasizing the need for standardized guidelines for the characterization of BI-RADS 4 subcategories and improved diagnostic accuracy in the management of breast lesions.

A Multi-label Artificial Intelligence Approach for Improving Breast Cancer Detection With Mammographic Image Analysis.

Breast cancer remains a major global health concern. This study aimed to develop a deep-learning-based artificial intelligence (AI) model that predicts the malignancy of mammographic lesions and reduces unnecessary biopsies in patients with breast cancer. In this retrospective study, we used deep-learning-based AI to predict whether lesions in mammographic images are malignant. The AI model learned the malignancy as well as margins and shapes of mass lesions through multi-label training, similar to the diagnostic process of a radiologist. We used the Curated Breast Imaging Subset of Digital Database for Screening Mammography. This dataset includes annotations for mass lesions, and we developed an algorithm to determine the exact location of the lesions for accurate classification. A multi-label classification approach enabled the model to recognize malignancy and lesion attributes. Our multi-label classification model, trained on both lesion shape and margin, demonstrated superior performance compared with models trained solely on malignancy. Gradient-weighted class activation mapping analysis revealed that by considering the margin and shape, the model assigned higher importance to border areas and analyzed pixels more uniformly when classifying malignant lesions. This approach improved diagnostic accuracy, particularly in challenging cases, such as American College of Radiology Breast Imaging-Reporting and Data System categories 3 and 4, where the breast density exceeded 50%. This study highlights the potential of AI in improving the diagnosis of breast cancer. By integrating advanced techniques and modern neural network designs, we developed an AI model with enhanced accuracy for mammographic image analysis.

Acidic/abrasive challenges on simulated non-carious cervical lesions development and morphology

This in vitro investigation assessed how frequency of erosive challenges and duration of toothbrushing abrasion influenced non-carious cervical lesions (NCCLs) development and morphology. Experimental units were prepared using extracted human premolars assigned to four erosive-abrasive frequency protocols (n=16): F0. No acid exposure (negative control), F2.5 K. Acid exposure (1 % citric acid at natural pH) every 2500, F5K. 5000 and F15K. 15000 brushing-strokes. All groups were brushed for 55000 total brushing-strokes. Three-dimension images of the teeth were captured at baseline, after 15000, 35000 and 55000 brushing-strokes, using an intraoral scanner (TRIOS4, 3Shape). WearCompare software (Leeds Digital Dentistry) was used to analyze volumetric tooth loss (mm3) by superimposition followed by subtraction analysis. Lesion angle was measured (ImageJ, NIH) and morphology visually classified. Data were analyzed using ANOVA and Fisher's Exact tests adopting two-sided 5 % significance level. Tooth loss increased with brushing-strokes overall (p<0.001) and for each erosive-abrasive protocol (p<0.001). Acid exposure significantly increased tooth loss (p<0.001), regardless of brushing interval (p<0.001), however by 35000 strokes no tooth loss difference was observed among acid-exposed groups (p>0.05). Control had significantly sharper mean lesion angle (59°) than all acid-exposed groups (∼145°) (p<0.001), and significantly different lesion shape with 94 % wedge-shaped lesions versus 0 %, respectively (p<0.001). In contrast to the control, acid exposure was associated to more striated lesions. Simulated NCCLs developed and progressed differently and more rapidly in the presence of acidic challenges, regardless of their frequency. Exposure to acid impacted the morphology of lesions.

Multi-modal classification of breast cancer lesions in Digital Mammography and contrast enhanced spectral mammography images

Breast cancer ranks as the second most prevalent cancer in women, recognized as one of the most dangerous types of cancer, and is on the rise globally. Regular screenings are essential for early-stage treatment. Digital mammography (DM) is the most recognized and widely used technique for breast cancer screening. Contrast-Enhanced Spectral Mammography (CESM or CM) is used in conjunction with DM to detect and identify hidden abnormalities, particularly in dense breast tissue where DM alone might not be as effective. In this work, we explore the effectiveness of each modality (CM, DM, or both) in detecting breast cancer lesions using deep learning methods. We introduce an architecture for detecting and classifying breast cancer lesions in DM and CM images in Craniocaudal (CC) and Mediolateral Oblique (MLO) views. The proposed architecture (JointNet) consists of a convolution module for extracting local features, a transformer module for extracting long-range features, and a feature fusion layer to fuse the local features, global features, and global features weighted based on the local ones. This significantly enhances the accuracy of classifying DM and CM images into normal or abnormal categories and lesion classification into benign or malignant. Using our architecture as a backbone, three lesion classification pipelines are introduced that utilize attention mechanisms focused on lesion shape, texture, and overall breast texture, examining the critical features for effective lesion classification. The results demonstrate that our proposed methods outperform their components in classifying images as normal or abnormal and mitigate the limitations of independently using the transformer module or the convolution module. An ensemble model is also introduced to explore the effect of each modality and each view to increase our baseline architecture's accuracy. The results demonstrate superior performance compared with other similar works. The best performance on DM images was achieved with the semi-automatic AOL Lesion Classification Pipeline, yielding an accuracy of 98.85 %, AUROC of 0.9965, F1-score of 98.85 %, precision of 98.85 %, and specificity of 98.85 %. For CM images, the highest results were obtained using the automatic AOL Lesion Classification Pipeline, with an accuracy of 97.47 %, AUROC of 0.9771, F1-score of 97.34 %, precision of 94.45 %, and specificity of 97.23 %. The semi-automatic ensemble AOL Classification Pipeline provided the best overall performance when using both DM and CM images, with an accuracy of 94.74 %, F1-score of 97.67 %, specificity of 93.75 %, and sensitivity of 95.45 %. Furthermore, we explore the comparative effectiveness of CM and DM images in deep learning models, indicating that while CM images offer clearer insights to the human eye, our model trained on DM images yields better results using Attention on Lesion (AOL) techniques. The research also suggests a multimodal approach using both DM and CM images and ensemble learning could provide more robust classification outcomes.

Обзор экспериментальных моделей исследования эффективности и безопасности титановых имплантатов

The use of three-dimensional titanium implants is becoming a common approach in reconstructive surgery for bone lesions. Computer modeling and selection of printing method parameters make it possible to create an individual implant design that takes into account the structural features, shape and size of the bone lesion. The safety requirements for implants made of titanium and its alloys as medical devi­ces are regulated by relevant documents (laws, government regulations, GOST), which require confirmation of the absence of harmful effects of the material on the body. The review provides generalized data from experimental studies of the safety properties and effectiveness of titanium implants, presented in electronic databases of scientific literature (PubMed, Embase, Google Scholar, Elibrary) for the period from 2015 to December 2023. The results of a study of the occurrence of general and local (in the implantation zone) reaction of the body, toxicity, activity of neutrophils and macrophages, the influence of biomaterial on surrounding soft tissues in the short and long term after implantation in experimental animals, study of blood parameters of animals confirming the absence of toxicity and good tolerability of implanted titanium structures. A conclusion is presented on the relevance of experimental models for studying physical, mechanical, biomechanical properties, compatibility, and safety. Each specific model has advantages that are advisable to use to solve certain problems in preclinical research. It was found that the largest number of studies are carried out on rabbits, followed by rats, dogs, sheep and goats. These studies have proven that the elastic modulus and strength of titanium alloy have values close to the values of these parameters of native bone. The object of the defect in experimental animals is most often the femur, less often the tibia, radius, etc. Differences in the ability and time of bone healing in animals and humans require extrapolation of the results to the reaction of the human body. Based on the results of the review, it can be said that there is a need to continue experimental studies on large animals (sheep) to be able to model large defects. In addition, it is necessary to standardize parameters that determine bone-implant contact and quantify bone ingrowth.

MLFEU-NET: A Multi-scale Low-level Feature Enhancement Unet for breast lesions segmentation in ultrasound images

Breast lesions constantly threaten the health of females. Segmentation methods of breast lesions are important for clinical diagnosis, and neural networks have been widely used and played a significant role in this field. However, false detections and miss detections still exist due to the variability in the shape and location of breast lesions, artifacts and noise in breast ultrasound (BUS) images, and structural defects in conventional segmentation networks. In this paper, a multi-scale low-level feature enhancement Unet (MLFEU-net) structure is presented, consisting of the U-net structure, low-level feature enhancement block (LFEB), and a parallel multiscale feature fusion (PMFF) module. Specifically, LFEB enhances the detail information during shallow downsampling and further feature selection. Meanwhile, in the neck of Unet, PMFF module uses different scales of dilation convolution to provide different sizes of sensory fields, and to efficiently merge shallow and deep features filtered, leads to more accurate segmentation results. To evaluate the MLFEU-net’s segmentation ability, five quantitative metrics were used on two breast ultrasound datasets, and it was compared with several advanced segmentation techniques. The results illustrate that our approach surpasses other methods in performance. Moreover, robustness experiments validate the effectiveness of our approach in achieving robust segmentation of breast lesions.

AYURVEDIC PROPRIETARY INTERVENTION IN LEUCODERMA (SHVITRA): A CASE STUDY

Ayurveda mentions all types of skin disorders under one common term i.e., Kushtha Roga. Shvitra is explained in the classics as a type of Kushtha. The disease is said to be result of impairment of Tridosha and Dhatu such as Rakta, Mamsa and Meda. Signs and symptoms of Shvitra and vitiligo resemble closely, hence the two diseases are co-related as well. The cause of vitiligo is yet unclear and is believed to be an auto-immune disease. Modern science has limited treatment options to address this concern. Whereas, in Ayurveda, although the disease is termed as Kricchasadhya (difficult to treat) or Asaadhya, yet many treatment options are advised. This case report elucidates the therapeutic effect of Ayurveda treatment modality in a patient having progressive leucoderma. A 21-year-old female patient with complaints of whitish-lesions over forehead, neck and back having moderate itching and pre-diagnosed with leucoderma visited our OPD. After the treatment for 11 months, significant changes were observed in the shape and size of lesions.

FTSNet: Fundus Tumor Segmentation Network on Multiple Scales Guided by Classification Results and Prompts.

The segmentation of fundus tumors is critical for ophthalmic diagnosis and treatment, yet it presents unique challenges due to the variability in lesion size and shape. Our study introduces Fundus Tumor Segmentation Network (FTSNet), a novel segmentation network designed to address these challenges by leveraging classification results and prompt learning. Our key innovation is the multiscale feature extractor and the dynamic prompt head. Multiscale feature extractors are proficient in eliciting a spectrum of feature information from the original image across disparate scales. This proficiency is fundamental for deciphering the subtle details and patterns embedded in the image at multiple levels of granularity. Meanwhile, a dynamic prompt head is engineered to engender bespoke segmentation heads for each image, customizing the segmentation process to align with the distinctive attributes of the image under consideration. We also present the Fundus Tumor Segmentation (FTS) dataset, comprising 254 pairs of fundus images with tumor lesions and reference segmentations. Experiments demonstrate FTSNet's superior performance over existing methods, achieving a mean Intersection over Union (mIoU) of 0.8254 and mean Dice (mDice) of 0.9042. The results highlight the potential of our approach in advancing the accuracy and efficiency of fundus tumor segmentation.

Applications of fractal analysis techniques in magnetic resonance imaging and computed tomography for stroke diagnosis and stroke-related brain damage: a narrative review

The fractal analysis technique has emerged as a novel and promising method in mathematical analysis, providing valuable insights across various fields of neuroimaging. The fractal analysis technique allows for the quantitative characterization of complex geometric structures that traditional Euclidean geometry-based morphometric methods fail to describe adequately. This review provides an overview of the principles, characteristics, and main applications of the fractal analysis technique, focusing on its applications and perspectives in stroke diagnosis based on neuroimaging data. In stroke research, the fractal analysis technique has been used to characterize brain tissue, pathological foci, and the vascular network, providing critical diagnostic and prognostic information. Researchers have applied the fractal analysis technique to brain lesions resulting from ischemic strokes to conduct geometric analyses of lesion shapes, indicating its diagnostic and prognostic values. Fractal properties have been used to study the texture of lesions, healthy tissue, and penumbra zones, which is essential for determining the presence and boundaries of damaged brain tissue. Additionally, fractal analysis of intracerebral hemorrhages has shown that hemorrhage geometry is correlated with prognosis and survival rates. This method has been used to assess cortex and white matter configurations in stroke patients, highlighting brain remodeling and compensatory changes. It has also been proven effective in detecting morphological alterations in brain structures during transient ischemic attacks. Moreover, fractal analysis of the brain vasculature revealed changes associated with ischemic stroke and hemorrhage. Overall, the fractal analysis technique in brain magnetic resonance imaging and computed tomography is an informative and sensitive imaging analysis method that, with further development, can significantly improve stroke diagnosis and prognosis on the basis of neuroimaging data.

Computed tomography-based radial endobronchial ultrasound image simulation of peripheral pulmonary lesions using deep learning.

Radial endobronchial ultrasound (R-EBUS) plays an important role during transbronchial sampling of peripheral pulmonary lesions (PPLs). However, existing navigational bronchoscopy systems provide no guidance for R-EBUS. To guide intraoperative R-EBUS probe manipulation, we aimed to simulate R-EBUS images of PPLs from preoperative computed tomography (CT) data using deep learning. Preoperative CT and intraoperative ultrasound data of PPLs in 250 patients who underwent R-EBUS-guided transbronchial lung biopsy were retrospectively collected. Two-dimensional CT sections perpendicular to the biopsy path were transformed into ultrasonic reflection and transmission images using an ultrasound propagation model to obtain the initial simulated R-EBUS images. A cycle generative adversarial network was trained to improve the realism of initial simulated images. Objective and subjective indicators were used to evaluate the similarity between real and simulated images. Wasserstein distances showed that utilizing the cycle generative adversarial network significantly improved the similarity between real and simulated R-EBUS images. There was no statistically significant difference in the long axis, short axis, and area between real and simulated lesions (all P > 0.05). Based on the experts' evaluation, a median similarity score of ≥4 on a 5-point scale was obtained for lesion size, shape, margin, internal echoes, and overall similarity. Simulated R-EBUS images of PPLs generated by our method can closely mimic the corresponding real images, demonstrating the potential of our method to provide guidance for intraoperative R-EBUS probe manipulation.

HDB-Net: hierarchical dual-branch network for retinal layer segmentation in diseased OCT images.

Optical coherence tomography (OCT) retinal layer segmentation is a critical procedure of the modern ophthalmic process, which can be used for diagnosis and treatment of diseases such as diabetic macular edema (DME) and multiple sclerosis (MS). Due to the difficulties of low OCT image quality, highly similar retinal interlayer morphology, and the uncertain presence, shape and size of lesions, the existing algorithms do not perform well. In this work, we design an HDB-Net network for retinal layer segmentation in diseased OCT images, which solves this problem by combining global and detailed features. First, the proposed network uses a Swin transformer and Res50 as a parallel backbone network, combined with the pyramid structure in UperNet, to extract global context and aggregate multi-scale information from images. Secondly, a feature aggregation module (FAM) is designed to extract global context information from the Swin transformer and local feature information from ResNet by introducing mixed attention mechanism. Finally, the boundary awareness and feature enhancement module (BA-FEM) is used to extract the retinal layer boundary information and topological order from the low-resolution features of the shallow layer. Our approach has been validated on two public datasets, and Dice scores were 87.61% and 92.44, respectively, both outperforming other state-of-the-art technologies.

Intelligent skin lesion segmentation using deformable attention Transformer U-Net with bidirectional attention mechanism in skin cancer images.

In recent years, the increasing prevalence of skin cancers, particularly malignant melanoma, has become a major concern for public health. The development of accurate automated segmentation techniques for skin lesions holds immense potential in alleviating the burden on medical professionals. It is of substantial clinical importance for the early identification and intervention of skin cancer. Nevertheless, the irregular shape, uneven color, and noise interference of the skin lesions have presented significant challenges to the precise segmentation. Therefore, it is crucial to develop a high-precision and intelligent skin lesion segmentation framework for clinical treatment. A precision-driven segmentation model for skin cancer images is proposed based on the Transformer U-Net, called BiADATU-Net, which integrates the deformable attention Transformer and bidirectional attention blocks into the U-Net. The encoder part utilizes deformable attention Transformer with dual attention block, allowing adaptive learning of global and local features. The decoder part incorporates specifically tailored scSE attention modules within skip connection layers to capture image-specific context information for strong feature fusion. Additionally, deformable convolution is aggregated into two different attention blocks to learn irregular lesion features for high-precision prediction. A series of experiments are conducted on four skin cancer image datasets (i.e., ISIC2016, ISIC2017, ISIC2018, and PH2). The findings show that our model exhibits satisfactory segmentation performance, all achieving an accuracy rate of over 96%. Our experiment results validate the proposed BiADATU-Net achieves competitive performance supremacy compared to some state-of-the-art methods. It is potential and valuable in the field of skin lesion segmentation.