Object knowledge-aware multiple instance learning for small tumor segmentation

Due to the low image contrast and high tortuosity of the small bowel on CT scans, it is difficult for even experienced radiologists to localize small bowel. To facilitate this, we propose a novel mesenteric vasculature-guided method for small bowel region segmentation on high-resolution contrast-enhanced CT angiography (CTA) scans. It is essential and useful for further localizing small bowel tumors. The mesenteric vasculature is first segmented using multi-view multi-scale vesselness enhancement filters on maximum intensity projection images. Second, the small bowel region is automatically segmented based on the map of mesenteric vasculature and its relationship to the small bowel. The method is applied to high-resolution contrast-enhanced abdominal CTA scans of patients suspected of having carcinoid tumors. The result, 79.3% volume overlap accuracy compared with the reference standard, shows it is feasible to segment the small bowel using the mesenteric vasculature as a roadmap on CTA scans.

To address the challenges of small intestinal tumor CT image labeling with a small number of data, class imbalance, and low segmentation accuracy. This paper introduces Dulecr-Net, a semi-supervised segmentation model. The model first employs the PolyMix algorithm combined with a dual-level consistency regularization approach to address the challenge of class imbalance in CT images of small intestinal tumors. Second, we designed a novel segmentation network, MSA-RFSEU-Net, in order to achieve high-precision segmentation of the lesion region in CT images of small intestinal tumors, by performing experimental analysis on the dataset provided by the Affiliated Hospital of Qingdao University. With only 15% labeled data, The Dice and Jaccard coefficients of the method reached 85.12% and 82.16% respectively, showing improvements of 19.86% and 28.62% compared to the fully supervised U-Net. The experimental results show that Dulecr-Net has a great performance improvement in the effectiveness of medical image segmentation compared with fully supervised and semi-supervised methods, and can effectively address imbalances in lesion area classes.

Recent advances in MRI‐guided radiation therapy (MRgRT) and deep learning techniques encourage fully adaptive radiation therapy (ART), real‐time MRI monitoring, and the MRI‐only treatment planning workflow. Given the rapid growth and emergence of new state‐of‐the‐art methods in these fields, we systematically review 197 studies written on or before December 31, 2022, and categorize the studies into the areas of image segmentation, image synthesis, radiomics, and real time MRI. Building from the underlying deep learning methods, we discuss their clinical importance and current challenges in facilitating small tumor segmentation, accurate x‐ray attenuation information from MRI, tumor characterization and prognosis, and tumor motion tracking. In particular, we highlight the recent trends in deep learning such as the emergence of multi‐modal, visual transformer, and diffusion models.

Segmentation is a routine and crucial procedure for the treatment of brain tumors. Deep learning based brain tumor segmentation methods have achieved promising performance in recent years. However, to pursue high segmentation accuracy, most of them require too much memory and computation resources. Motivated by a recently proposed partially reversible U-Net architecture that pays more attention to memory footprint, we further present a novel Memory-Efficient Cascade 3D U-Net (MECU-Net) for brain tumor segmentation in this work, which can achieve comparable segmentation accuracy with less memory and computation consumption. More specifically, MECU-Net utilizes fewer down-sampling channels to reduce the utilization of memory and computation resources. To make up the accuracy loss, MECU-Net employs multi-scale feature fusion module to enhance the feature representation capability. Additionally, a light-weight cascade model, which resolves the problem of small target segmentation accuracy caused by model compression to some extent, is further introduced into the segmentation network. Finally, edge loss and weighted dice loss are combined to refine the brain tumor segmentation results. Experiment results on BraTS 2019 validation set illuminate that MECU-Net can achieve average Dice coefficients of 0.902, 0.824 and 0.777 on the whole tumor, tumor core and enhancing tumor, respectively.

Deep Learning for Detecting BRCA Mutations in High-Grade Ovarian Cancer Based on an Innovative Tumor Segmentation Method From Whole Slide Images

Simple SummaryRecent advancements in AI have revolutionized cancer research, especially in the analysis of histopathological imaging data with minimal human involvement. Early detection of lymph node metastasis in breast cancer is vital for treatment outcomes. This paper introduces a novel approach that combines representation learning and deep learning (DL) to detect small tumors (STs) without neglecting larger ones. The proposed method uses representation learning to identify STs in histopathology images, followed by DL algorithms for breast cancer detection. Extensive evaluation shows remarkable accuracy in detecting STs without compromising larger-lesion detection. This approach enables early detection, timely intervention, and potentially improved treatment outcomes. The integration of representation learning and DL offers a promising solution for ST detection in breast cancer. By reducing human involvement and leveraging AI capabilities, the proposed method achieves impressive accuracy in identifying STs. Further research and validation could enhance diagnostic capabilities and personalized treatment strategies, ultimately benefiting breast cancer patients.The early diagnosis of lymph node metastasis in breast cancer is essential for enhancing treatment outcomes and overall prognosis. Unfortunately, pathologists often fail to identify small or subtle metastatic deposits, leading them to rely on cytokeratin stains for improved detection, although this approach is not without its flaws. To address the need for early detection, multiple-instance learning (MIL) has emerged as the preferred deep learning method for automatic tumor detection on whole slide images (WSIs). However, existing methods often fail to identify some small lesions due to insufficient attention to small regions. Attention-based multiple-instance learning (ABMIL)-based methods can be particularly problematic because they may focus too much on normal regions, leaving insufficient attention for small-tumor lesions. In this paper, we propose a new ABMIL-based model called normal representative keyset ABMIL (NRK-ABMIL), which addresseses this issue by adjusting the attention mechanism to give more attention to lesions. To accomplish this, the NRK-ABMIL creates an optimal keyset of normal patch embeddings called the normal representative keyset (NRK). The NRK roughly represents the underlying distribution of all normal patch embeddings and is used to modify the attention mechanism of the ABMIL. We evaluated NRK-ABMIL on the publicly available Camelyon16 and Camelyon17 datasets and found that it outperformed existing state-of-the-art methods in accurately identifying small tumor lesions that may spread over a few patches. Additionally, the NRK-ABMIL also performed exceptionally well in identifying medium/large tumor lesions.

Accurate kidney and tumor segmentation of computed tomography (CT) scans is vital for diagnosis and treatment, but manual methods are time-consuming and inconsistent, highlighting the value of AI automation. This study develops a fully automated AI model using vision transformers (ViTs) and convolutional neural networks (CNNs) to detect and segment kidneys and kidney tumors in Contrast-Enhanced (CECT) scans, with a focus on improving sensitivity for small, indistinct tumors. The segmentation framework employs a ViT-based model for the kidney organ, followed by a 3D UNet model with enhanced connections and attention mechanisms for tumor detection and segmentation. Two CECT datasets were used: a public dataset (KiTS23: 489 scans) and a private institutional dataset (Private: 592 scans). The AI model was trained on 389 public scans, with validation performed on the remaining 100 scans and external validation performed on all 592 private scans. Tumors were categorized by TNM staging as small (≤4 cm) (KiTS23: 54%, Private: 41%), medium (>4 cm to ≤7 cm) (KiTS23: 24%, Private: 35%), and large (>7 cm) (KiTS23: 22%, Private: 24%) for detailed evaluation. Kidney and kidney tumor segmentations were evaluated against manual annotations as the reference standard. The model achieved a Dice score of 0.97 ± 0.02 for kidney organ segmentation. For tumor detection and segmentation on the KiTS23 dataset, the sensitivities and average false-positive rates per patient were as follows: 0.90 and 0.23 for small tumors, 1.0 and 0.08 for medium tumors, and 0.96 and 0.04 for large tumors. The corresponding Dice scores were 0.84 ± 0.11, 0.89 ± 0.07, and 0.91 ± 0.06, respectively. External validation on the private data confirmed the model's effectiveness, achieving the following sensitivities and average false-positive rates per patient: 0.89 and 0.15 for small tumors, 0.99 and 0.03 for medium tumors, and 1.0 and 0.01 for large tumors. The corresponding Dice scores were 0.84 ± 0.08, 0.89 ± 0.08, and 0.92 ± 0.06. The proposed model demonstrates consistent and robust performance in segmenting kidneys and kidney tumors of various sizes, with effective generalization to unseen data. This underscores the model's significant potential for clinical integration, offering enhanced diagnostic precision and reliability in radiological assessments.

BackgroundCholangiolocellular carcinoma (CoCC) is a relatively rare primary liver tumor. We present a literature review and case report of a patient who presented with a slow-growing CoCC that was completely resected after a 5-year follow-up period.Case presentationThe patient was a 66-year-old man with a history of inflammatory thoracic and intra-abdominal pseudo-tumors. He was regularly followed up at our hospital for partial dilation of the pancreatic duct branch located in the body of the pancreas. Five years earlier, computed tomography (CT) demonstrated a small tumor in liver segment 4. Radiological findings were suggestive of hemangioma. Tumor size gradually increased during the 5-year follow-up period. CT scans showed that the tumor had progressed in size from 10 to 20 mm. Positron emission tomography CT revealed an accumulation of fluorodeoxyglucose (standardized uptake value max 5.3) at the tumor site. The tumor exhibited high intensity on T2-weighted and diffusion-weighted images of ethoxybenzyl magnetic resonance imaging. The tumor showed high intensity during the early phase but low intensity during the hepatobiliary phase. Tumor markers were within their respective normal ranges. Suspecting intrahepatic cholangiocarcinoma, left hepatectomy was performed. The tumor was diagnosed as CoCC based on pathological findings. The patient’s post-operative course was uneventful. The patient survived for a year, without any recurrence.ConclusionsIn cases dealing with small tumor sizes, it is difficult to distinguish between CoCC and hemangioma due to their similar radiological findings. Thus, it is important to consider the diagnosis of CoCC in small benign hepatic tumors. As such, follow-up radiological examination is recommended.

Liver and tumor segmentation from CT images is a complex and crucial step in achieving full-course adaptive radiotherapy and also plays an essential role in computer-aided clinical diagnosis systems. Deep learning-based methods play an important role in achieving automatic segmentation. This research aims to improve liver tumor detection performance by proposing a dual path feature extracting strategy and employing Swin-Transformer. The hierarchical Swin-Transformer is embedded into the encoder and decoder and combined with CNN to form a dual coding path structure incorporating long-range dependencies and multi-scale contextual connections to capture coarse-tuned features at different semantic scales. The features of the two encoding paths and the upsampling path are fused, tested and validated with LITS and in-house datasets. The proposed method has a DG of 97.95% and a DC of 96.2% for liver segmentation; a DG of 80.6% and a DC of 68.1% for tumor segmentation; and a classification study of the tumor dataset shows a DG of 91.1% and a DC of 87.2% for large and continuous tumors and a DG of 71.7% and a DC of 66.4% for small and scattered tumors. Swin-Transformer can be used as a robust encoder for medical image segmentation networks and, combined with CNN networks, can better recover local spatial information and enhance feature representation. Accurate localization before segmentation can achieve better results for small and scattered tumors.

To establish the effects of size and segmentation methods on intra-reader reliability of primary tumor metabolic tumor volume (MTV) and total glycolytic activity (TGA) in human solid tumors. This is a retrospective study of 121 patients who had a baseline FDG PET/CT scan for oncologic staging. Volumetric parameter readings were performed in random order on two separate occasions, 12 weeks apart, by the same reader. The MTV and TGA were segmented using gradient and fixed maximum standardized uptake value (SUVmax) threshold methods. Intra-reader reliability was established by the intraclass correlation coefficient (ICC) and Bland-Altman analysis. The biases for MTV were 2.95, 14.76 and 11.13% for gradient segmentation, 38 and 50% SUVmax fixed threshold segmentations, respectively (p < 0.0001). For TGA, the corresponding biases were 0.76, 10.36 and 7.46% (p < 0.0001). There were no statistically significant differences in the biases between the first and second reads for MTV segmented for small and large volume tumors by the gradient method (p < 0.34) or 50% SUVmax threshold segmentation (p < 0.08). However, there were statistically significant differences in the corresponding biases for the 38% SUVmax threshold segmentation (p < 0.04). There were no statistically significant differences in the biases between the first and second reads for TGA segmented for small and large volume tumors (p < 0.98). Intra-reader reliability for primary tumor FDG MTV and TGA is affected by the tumor size and segmentation methods. The segmentation bias was smaller for gradient method than percentage fixed threshold method for MTV. The segmentation biases were smaller for TGA than MTV.

Breast tumor segmentation is a critical aspect of magnetic resonance imaging (MRI)-based breast disease diagnosis. Numerous networks and algorithms, including U-Net and its enhancements, have been proposed for breast tumor segmentation. However, existing methods have certain shortcomings and limitations, including insufficient extraction of multi-scale contextual information, which poses challenges in adapting to tumors of different sizes and distinguishing tumor boundaries from surrounding tissues. Additionally, the feature extraction process lacks specificity and is prone to interference from irrelevant information outside the tumor region. This study aimed to address these challenges, and achieve the accurate and automated segmentation of breast tumors in MRI scans. A new three-dimensional (3D) breast tumor segmentation network named the multi-scale hybrid attention U-shaped network (MHAU-Net) was designed. The network used four sets of atrous convolutions with different dilated ratios to extract multi-scale context information. Global pooling and single-channel convolution structures were employed to construct channel and spatial blocks. Subsequently, the network integrated four sets of atrous convolutions with spatial and channel attention blocks to extract hybrid attention features. Compared to existing MRI segmentation networks for breast tumors, the MHAU-Net demonstrated superior performance in extracting informative features and adapting to tumors of diverse sizes and shapes. To evaluate the proposed approach, we curated a large-scale breast MRI dataset comprising 906 3D images. A comparative analysis with seven commonly used segmentation networks revealed the superior performance of our method. Our network had a dice similarity coefficient (DSC) and intersection over union (IoU) of 84.1%±2.1% and 74.2%±3.4%, respectively, representing a 6.0% and 7.1% improvement over the baseline 3D U-Net. Additionally, our method had DSC values of 85.7%±1.6%, 84.3%±2.8%, 86.7%±1.7%, and 86.3%±1.5% for single, small, large, and mass tumors, respectively. Our results highlight the superior overall performance of the proposed method, and show its ability to adapt to various types of tumor images. This study establishes a solid foundation for further exploring the application value of deep learning in breast cancer diagnosis.

Gastrointestinal stromal tumor (GIST) is the most common mesenchymal tumor in the digestive tract. Recurrences may occur even after radical resection; however, recurrence later than 10 years after surgery is rare. We report a case of GIST with recurrence of liver metastasis 25 years after surgery. A 56-year-old man complained of sudden epigastric pain and was transferred to the emergency department. He had undergone partial resection of the small intestine for leiomyosarcoma 25 years previously. Abdominal computed tomography showed multiple liver tumors with massive hemorrhage. Ultrasound-guided percutaneous biopsy was performed for the 15-mm hepatic tumor in segment 2. Pathological findings revealed proliferation of spindle-shaped atypical cells, and immunostaining for c-kit and CD34 was both positive; the patient was therefore diagnosed with GIST. He then underwent chemotherapy for 7 years but died of multiple organ failure due to GIST. Autopsy revealed GIST occupying the entire liver with peritoneal dissemination, and minute lung metastases that could not be identified by CT were also detected. This case is interesting in considering the recurrence of GIST, and we will report it together with the literature review.

Computer-aided diagnosis is of great significance to improve the diagnostic accuracy of pancreatic cancer that has an insidious course and does not have obvious symptoms at first. However, segmentation of pancreatic cancer is challenging because the tumors vary in size with the smallest tumor having a size of about 0.5 in diameter, and most of them have irregular shapes and unclear boundaries. In this study, we developed a deep learning architecture Multi-Scale Channel Attention Unet (MSCA-Unet) for pancreatic tumor segmentation and collected CT images of 419 patients from The Affiliated Hospital of Qingdao University and a public dataset. We embedded the multi-scale network into the encoder to extract semantic information at different scales and the decoder to provide supplemental information to overcome the loss of information in the upsampling and the drift of the localized tumor due to the upsampling and skip connections. We adopted the channel attention unit after the multi-scale convolution to emphasize the informative channels, which was observed to have the effects of accelerating the positioning process, reducing false positives, and improving the accuracy of outlining very small, irregular pancreatic tumors. Our results show that our network outperformed the other current mainstream segmentation networks and achieved a Dice index of 68.03%, a Jaccard of 59.31%, and an FPR of 1.36% on the private dataset Task-01 without data pre-processing. Compared with the other pancreatic tumor segmentation networks on the public dataset Task-02, our network produced the best Dice index, 80.12%, with the assistance of the data pre-processing scheme. This study strategically utilizes the multi-scale convolution and channel attention mechanism of the architecture to provide a dedicated network for segmentation of small and irregular pancreatictumors.

Intercostal and transthoracic trocars enable easier laparoscopic resection of dome liver lesions

Segmentation of lymphoma tumor in PET images using cellular automata: A preliminary study