Figure - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
The aim of this study was to develop a deep learning-based algorithm for fully automated spleen segmentation using CT images and to evaluate the performance in conditions directly or indirectly affecting the spleen (e.g., splenomegaly, ascites). For this, a 3D U-Net was trained on an in-house dataset (n = 61) including diseases with and without spl...
Context in source publication
Context 1
... Gabriel et al. reported bad and failed segmentation results in patients with splenic distortions, even if their model reached a DSC of 0.962 [34]. Table 3 shows that not all methods were assessed considering splenic abnormalities. The aim of this study was to develop a robust deep learning algorithm for spleen segmentation across various conditions that can alter or obscure the normal splenic anatomy. ...Similar publications
Objective
We aim to develop and validate a three-dimensional convolutional neural network (3D-CNN) model for automatic liver segment segmentation on MRI images.
Methods
This retrospective study evaluated an automated method using a deep neural network that was trained, validated, and tested with 367, 157, and 158 portal venous phase MR images, res...
Citations
... Moreover, established techniques such as the Deauville 5-point scale, directly utilise the [ 18 F]FDG uptake in normal liver and mediastinum as reference regions, to evaluate treatment response in lymphoma [1]. Even for non-[ 18 Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. regarding personalised treatment response and toxicity, such as in the emerging field of theranostics [5]. ...
... These provide information to enable segmentation of the investigated organ and the trained model can be rapidly applied to a new image [14]. Since no registration is required between the training dataset and the new image, this method has been broadly applied in CT segmentation of abdominal organs with studies reporting high similarity with manual segmentations [12,[15][16][17][18][19][20][21][22]. ...
Utilisation of whole-organ volumes to extract anatomical and functional information from computed tomography (CT) and positron emission tomography (PET) images may provide key information for the treatment and follow-up of cancer patients. However, manual organ segmentation, is laborious and time-consuming.
In this study, a CT-based deep learning method and a multi-atlas method were evaluated for segmenting the liver and spleen on CT images to extract quantitative tracer information from Fluorine-18 fluorodeoxyglucose ([ ¹⁸ F]FDG) PET images of 50 patients with advanced Hodgkin lymphoma (HL). Manual segmentation was used as the reference method. The two automatic methods were also compared with a manually defined volume of interest (VOI) within the organ, a technique commonly performed in clinical settings.
Both automatic methods provided accurate CT segmentations, with the deep learning method outperforming the multi-atlas with a DICE coefficient of 0.93 ± 0.03 (mean ± standard deviation) in liver and 0.87 ± 0.17 in spleen compared to 0.87 ± 0.05 (liver) and 0.78 ± 0.11 (spleen) for the multi-atlas. Similarly, a mean relative error of -3.2% for the liver and -3.4% for the spleen across patients was found for the mean standardized uptake value (SUV mean ) using the deep learning regions while the corresponding errors for the multi-atlas method were -4.7% and -9.2%, respectively. For the maximum SUV (SUV max ), both methods resulted in higher than 20% overestimation due to the extension of organ boundaries to include neighbouring, high-uptake regions. The conservative VOI method which did not extend into neighbouring tissues, provided a more accurate SUV max estimate.
In conclusion, the automatic, and particularly the deep learning method could be used to rapidly extract information of the SUV mean within the liver and spleen. However, activity from neighbouring organs and lesions can lead to high biases in SUV max and a manually defined volume of interest should be considered instead.
... Deep learning in biomedical applications has been utilized successfully, in various studies, for abdominal organ segmentation [15][16][17][18][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39]. In the present study, our focus is to employ deep-learning techniques for the precise segmentation of the spleen, and the calculation of the spleen volume, using the MRI scans of 20 patients with GD as a testing cohort. ...
... A variety of deep-learning algorithms have been applied for spleen segmentation, including 2D and 3D U-Net based models, some of which combine a post-processing pipeline [15][16][17]20,22,23,25,26,28,29,[32][33][34][35][36][37][38]. More advanced methods include the use of transformers [24,31]. ...
... In addition, researchers even designed a deep-learning neural network especially for spleen segmentation [27,30]. Most of the studies were carried out using CT [15][16][17]20,[22][23][24][25]28,29,31,32,34,35,37], and the accuracy (DC) obtained in these studies was mostly > 95.0%. When compared to MRI studies [24,26,27,33,34,36,38], it was found that MRI in general is less accurate, and the top-range accuracy is mostly around 94%. ...
The enlargement of the liver and spleen (hepatosplenomegaly) is a common manifestation of Gaucher disease (GD). An accurate estimation of the liver and spleen volumes in patients with GD, using imaging tools such as magnetic resonance imaging (MRI), is crucial for the baseline assessment and monitoring of the response to treatment. A commonly used method in clinical practice to estimate the spleen volume is the employment of a formula that uses the measurements of the craniocaudal length, diameter, and thickness of the spleen in MRI. However, the inaccuracy of this formula is significant, which, in turn, emphasizes the need for a more precise and reliable alternative. To this end, we employed deep-learning techniques, to achieve a more accurate spleen segmentation and, subsequently, calculate the resulting spleen volume with higher accuracy on a testing set cohort of 20 patients with GD. Our results indicate that the mean error obtained using the deep-learning approach to spleen volume estimation is 3.6 ± 2.7%, which is significantly lower than the common formula approach, which resulted in a mean error of 13.9 ± 9.6%. These findings suggest that the integration of deep-learning methods into the clinical routine practice for spleen volume calculation could lead to improved diagnostic and monitoring outcomes.
... Lymph nodes, which are glands dispersed throughout your body, are via which the lymph fluid travels. Because of this, lymphoma is sometimes referred to as an immune system cancer [4] Only a few lymphomas are spleen cancerous and these are-1) Haemangioma: The most frequent primary benign tumour of the spleen, a hemangioma is made up of blood-filled, endothelium-lined arterial channels. A rare benign vascular disorder called diffuse hemangiomatosis of the spleen can present as systemic angiomatosis or, less frequently, solely affect the spleen. ...
Spleen is the largest secondary lymphoid organ and play a crucial role in the regulation of innate and adaptive immune. Among all cancers, spleen cancer is one of the serious diseases which effects minor population across the globe but is potentially fatal, especially if diagnosed in a later stage of development with a 20% survival rate at 6 months. Given the intricacy of the problem, several computer-aided diagnostic methods have been proposed and developed to increase the survival rate Spleen Cancer is a malignancy of white blood cells involving tumour deposits in the spleen. Most splenic cancer do not start in the spleen, and those that do, are almost always lymphomas. Lymphoma is a type of blood cancer that develops in the lymphatic system. Although there are many systems available with the medical industry, but this research is proposed to improve the performance of existing system by employing deep learning feature of Artificial Intelligence (AI) using Convolution Neural Networks (CNN), where Convolution Neural Networks (CNN) has been implemented using DenseNet-201. This research accomplished the desired parameters with values to achieve 99.60% accuracy and 0.1240 % loss while training and testing the model. The research has been supported with the datasets from Kaggle, IEEE transactions on information technology in biomedicine and IEEE International Symposium on Biomedical Imaging. The data set contains three clinical types spleen cancer, such as CLL (chronic lymphocytic leukemia); FL (follicular lymphoma); MCL (mantle cell lymphoma).
... In recent years, deep learning achieved high performance in segmentation and classification tasks in medical imaging, with some algorithms being successfully implemented in clinical routine [13][14][15]. In oncologic imaging, the accurate and automated segmentation of abdominal organs is a critical first step for the detection and delineation of tumors and metastases, and for surgical preplanning. ...
... The deep learning-based model was developed on the open source MONAI Framework (Medical Open Network for AI, version 0.8.0) [24]. The automated segmentation of the spleen in CT images was performed using a 3D U-Net architecture and already presented in our previous study [13]. Briefly, the model consisted of an enhanced version of 3D U-Net with residual units, which was trained on an open dataset and an inhouse dataset (a total of 122 patients). ...
... In the first step of our study, splenomegaly was automatically segmented with a 3D U-Net model, reaching a dice score of 0.94, which is comparable with the results of Humpire-Mamani et al. [13,19]. More recently, modified U-Net architectures, such as Attentionbased U-Net, TransUNet, and Swin-UNet achieved higher scores in medical imaging segmentation, as demonstrated by Gulzar et al. [38], but these models are more complex than the U-Net and necessitate more computational power. ...
Splenomegaly is a common cross-sectional imaging finding with a variety of differential diagnoses. This study aimed to evaluate whether a deep learning model could automatically segment the spleen and identify the cause of splenomegaly in patients with cirrhotic portal hypertension versus patients with lymphoma disease. This retrospective study included 149 patients with splenomegaly on computed tomography (CT) images (77 patients with cirrhotic portal hypertension, 72 patients with lymphoma) who underwent a CT scan between October 2020 and July 2021. The dataset was divided into a training (n = 99), a validation (n = 25) and a test cohort (n = 25). In the first stage, the spleen was automatically segmented using a modified U-Net architecture. In the second stage, the CT images were classified into two groups using a 3D DenseNet to discriminate between the causes of splenomegaly, first using the whole abdominal CT, and second using only the spleen segmentation mask. The classification performances were evaluated using the area under the receiver operating characteristic curve (AUC), accuracy (ACC), sensitivity (SEN), and specificity (SPE). Occlusion sensitivity maps were applied to the whole abdominal CT images, to illustrate which regions were important for the prediction. When trained on the whole abdominal CT volume, the DenseNet was able to differentiate between the lymphoma and liver cirrhosis in the test cohort with an AUC of 0.88 and an ACC of 0.88. When the model was trained on the spleen segmentation mask, the performance decreased (AUC = 0.81, ACC = 0.76). Our model was able to accurately segment splenomegaly and recognize the underlying cause. Training on whole abdomen scans outperformed training using the segmentation mask. Nonetheless, considering the performance, a broader and more general application to differentiate other causes for splenomegaly is also conceivable.
Objective
To establish a deep learning model for the detection of hypoxic–ischemic encephalopathy (HIE) features on CT scans and to compare various networks to determine the best input data format.
Methods
168 head CT scans of patients after cardiac arrest were retrospectively identified and classified into two categories: 88 (52.4%) with radiological evidence of severe HIE and 80 (47.6%) without signs of HIE. These images were randomly divided into a training and a test set, and five deep learning models based on based on Densely Connected Convolutional Networks (DenseNet121) were trained and validated using different image input formats (2D and 3D images).
Results
All optimized stacked 2D and 3D networks could detect signs of HIE. The networks based on the data as 2D image data stacks provided the best results (S100: AUC: 94%, ACC: 79%, S50: AUC: 93%, ACC: 79%). We provide visual explainability data for the decision making of our AI model using Gradient-weighted Class Activation Mapping.
Conclusion
Our proof-of-concept deep learning model can accurately identify signs of HIE on CT images. Comparing different 2D- and 3D-based approaches, most promising results were achieved by 2D image stack models. After further clinical validation, a deep learning model of HIE detection based on CT images could be implemented in clinical routine and thus aid clinicians in characterizing imaging data and predicting outcome.
