Manual Segmentation Research Articles

Comparison between Three Radiomics Models and Clinical Nomograms for Prediction of Lymph Node Involvement in PCa Patients Combining Clinical and Radiomic Features.

We aim to compare the performance of three different radiomics models (logistic regression (LR), random forest (RF), and support vector machine (SVM)) and clinical nomograms (Briganti, MSKCC, Yale, and Roach) for predicting lymph node involvement (LNI) in prostate cancer (PCa) patients. The retrospective study includes 95 patients who underwent mp-MRI and radical prostatectomy for PCa with pelvic lymphadenectomy. Imaging data (intensity in T2, DWI, ADC, and PIRADS), clinical data (age and pre-MRI PSA), histological data (Gleason score, TNM staging, histological type, capsule invasion, seminal vesicle invasion, and neurovascular bundle involvement), and clinical nomograms (Yale, Roach, MSKCC, and Briganti) were collected for each patient. Manual segmentation of the index lesions was performed for each patient using an open-source program (3D SLICER). Radiomic features were extracted for each segmentation using the Pyradiomics library for each sequence (T2, DWI, and ADC). The features were then selected and used to train and test three different radiomics models (LR, RF, and SVM) independently using ChatGPT software (v 4o). The coefficient value of each feature was calculated (significant value for coefficient ≥ ±0.5). The predictive performance of the radiomics models and clinical nomograms was assessed using accuracy and area under the curve (AUC) (significant value for p ≤ 0.05). Thus, the diagnostic accuracy between the radiomics and clinical models were compared. This study identified 343 features per patient (330 radiomics features and 13 clinical features). The most significant features were T2_nodulofirstordervariance and T2_nodulofirstorderkurtosis. The highest predictive performance was achieved by the RF model with DWI (accuracy 86%, AUC 0.89) and ADC (accuracy 89%, AUC 0.67). Clinical nomograms demonstrated satisfactory but lower predictive performance compared to the RF model in the DWI sequences. Among the prediction models developed using integrated data (radiomics and semantics), RF shows slightly higher diagnostic accuracy in terms of AUC compared to clinical nomograms in PCa lymph node involvement prediction.

Age-Related Retinal Layer Thickness Changes Measured by OCT in APPNL-F/NL-F Mice: Implications for Alzheimer's Disease.

In Alzheimer's disease (AD), transgenic mouse models have established links between abnormalities in the retina and those in the brain. APPNL-F/NL-F is a murine, humanized AD model that replicates several pathological features observed in patients with AD. Research has focused on obtaining quantitative parameters from optical coherence tomography (OCT) in AD. The aim of this study was to analyze, in a transversal case-control study using manual retinal segmentation via SD-OCT, the changes occurring in the retinal layers of the APPNL/F-NF/L AD model in comparison to C57BL/6J mice (WT) at 6, 9, 12, 15, 17, and 20 months of age. The analysis focused on retinal thickness in RNFL-GCL, IPL, INL, OPL, and ONL based on the Early Treatment Diabetic Retinopathy Study (ETDRS) sectors. Both APPNL-F/NL-F-model and WT animals exhibited thickness changes at the time points studied. While WT showed significant changes in INL, OPL, and ONL, the AD model showed changes in all retinal layers analyzed. The APPNL-F/NL-F displayed significant thickness variations in the analyzed layers except for the IPL compared to related WT. These thickness changes closely resembled those found in humans during preclinical stages, as well as during mild and moderate AD stages, making this AD model behave more similarly to the disease in humans.

Deep learning for intracranial aneurysm segmentation using CT angiography

Objective. This study aimed to employ a two-stage deep learning method to accurately detect small aneurysms (4–10 mm in size) in computed tomography angiography images. Approach. This study included 956 patients from 6 hospitals and a public dataset obtained with 6 CT scanners from different manufacturers. The proposed method consists of two components: a lightweight and fast head region selection (HRS) algorithm and an adaptive 3D nnU-Net network, which is used as the main architecture for segmenting aneurysms. Segments generated by the deep neural network were compared with expert-generated manual segmentation results and assessed using Dice scores. Main Results. The area under the curve (AUC) exceeded 79% across all datasets. In particular, the precision and AUC reached 85.2% and 87.6%, respectively, on certain datasets. The experimental results demonstrated the promising performance of this approach, which reduced the inference time by more than 50% compared to direct inference without HRS. Significance. Compared with a model without HRS, the deep learning approach we developed can accurately segment aneurysms by automatically localizing brain regions and can accelerate aneurysm inference by more than 50%.

Development of an initial training and evaluation programme for manual lower limb muscle MRI segmentation

BackgroundMagnetic resonance imaging (MRI) quantification of intramuscular fat accumulation is a responsive biomarker in neuromuscular diseases. Despite emergence of automated methods, manual muscle segmentation remains an essential foundation. We aimed to develop a training programme for new observers to demonstrate competence in lower limb muscle segmentation and establish reliability benchmarks for future human observers and machine learning segmentation packages.MethodsThe learning phase of the training programme comprised a training manual, direct instruction, and eight lower limb MRI scans with reference standard large and small regions of interest (ROIs). The assessment phase used test–retest scans from two patients and two healthy controls. Interscan and interobserver reliability metrics were calculated to identify underperforming outliers and to determine competency benchmarks.ResultsThree experienced observers undertook the assessment phase, whilst eight new observers completed the full training programme. Two of the new observers were identified as underperforming outliers, relating to variation in size or consistency of segmentations; six had interscan and interobserver reliability equivalent to those of experienced observers. The calculated benchmark for the Sørensen-Dice similarity coefficient between observers was greater than 0.87 and 0.92 for individual thigh and calf muscles, respectively. Interscan and interobserver reliability were significantly higher for large than small ROIs (all p < 0.001).ConclusionsWe developed, implemented, and analysed the first formal training programme for manual lower limb muscle segmentation. Large ROI showed superior reliability to small ROI for fat fraction assessment.Relevance statementObservers competent in lower limb muscle segmentation are critical to application of quantitative muscle MRI biomarkers in neuromuscular diseases. This study has established competency benchmarks for future human observers or automated segmentation methods.Key points• Observers competent in muscle segmentation are critical for quantitative muscle MRI biomarkers.• A training programme for muscle segmentation was undertaken by eight new observers.• We established competency benchmarks for future human observers or automated segmentation methods.Graphical

Temporal and Spatial Dynamics of Ischemic Stroke Lesions after Acute Therapy: A Comprehensive Edema Assessment Using Combined 1H- and 23Na-MRI.

Ischemic cerebral stroke initiates a complex cascade of pathophysiological events, involving various forms of molecular shifts and edema. Early intervention is pivotal in minimizing tissue loss and improving clinical outcomes. This study explores the temporal and spatial evolution of tissue sodium concentration (TSC) in acute ischemic lesions after acute therapy using 23Na-MRI in addition to conventional 1H-MRI. Prospectively, from examined 58 patients with acute ischemic stroke with a combined 1H/23Na-MRI within 72 h of symptom onset after receiving acute therapy, 31 patients were included in the evaluation of this study. After co-registration of the 23Na-MRI images to the morphological 1H-MRI images, manual segmentation of the ischemic lesions was performed, and the ADC and TSC measurements were quantified and correlated with the time of onset and lesion volume. The mean TSC in ischemic lesions correlated positively with lesion volume (r = 0.52, p = 0.002) and showed a significant association with the time of stroke onset (r = 0.8, p &lt; 0.001). Patients who were treated only with intravenous rtPA showed homogenous sodium signal intensity in the ischemic lesions, whereas the patients who received mechanical recanalization exhibited distinctive sodium signal intensity patterns with focal significant TSC differences. The integration of 1H- and 23Na-MRI provides a nuanced understanding of temporal and spatial changes due to different types of edema in ischemic stroke lesions following acute treatment. Further exploration of these findings may enhance our understanding of stroke pathophysiology and guide personalized therapeutic interventions.

Estimating Uncertainty of Geographic Atrophy Segmentations with Bayesian Deep Learning

To apply methods for quantifying uncertainty of deep learning segmentation of geographic atrophy (GA). Retrospective analysis of OCT images and model comparison. One hundred twenty-six eyes from 87 participants with GA in the SWAGGER cohort of the Nonexudative Age-Related Macular Degeneration Imaged with Swept-Source OCT (SS-OCT) study. The manual segmentations of GA lesions were conducted on structural subretinal pigment epithelium en face images from the SS-OCT images. Models were developed for 2 approximate Bayesian deep learning techniques, Monte Carlo dropout and ensemble, to assess the uncertainty of GA semantic segmentation and compared to a traditional deep learning model. Model performance (Dice score) was compared. Uncertainty was calculated using the formula for Shannon Entropy. The output of both Bayesian technique models showed a greater number of pixels with high entropy than the standard model. Dice scores for the Monte Carlo dropout method (0.90, 95% confidence interval 0.87-0.93) and the ensemble method (0.88, 95% confidence interval 0.85-0.91) were significantly higher (P<0.001) than for the traditional model (0.82, 95% confidence interval 0.78-0.86). Quantifying the uncertainty in a prediction of GA may improve trustworthiness of the models and aid clinicians in decision-making. The Bayesian deep learning techniques generated pixel-wise estimates of model uncertainty for segmentation, while also improving model performance compared with traditionally trained deep learning models. Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Precision imaging of cardiac function and scar size in acute and chronic porcine myocardial infarction using ultrahigh-field MRI

Background7 T cardiac magnetic resonance imaging (MRI) studies may enable higher precision in clinical metrics like cardiac function, ventricular mass, and more. Higher precision may allow early detection of functional impairment and early evaluation of treatment responses in clinical practice and pre-clinical studies. Methods: Seven female German Landrace pigs were scanned prior to and at three time points (3–4 days, 7–10 days, and ~60 days) post myocardial infarction using a whole body 7 T system and three radiofrequency (RF) coils developed and built in-house to accompany animal growth. Results: The combination of dedicated RF hardware and 7 T MRI enables a longitudinal study in a pig model of acute and chronic infarction, providing consistent blood tissue contrast and high signal-to-noise ratio (SNR) in measurements of cardiac function, as well as low coefficients of variation (CoV) for ejection fraction (CoVintra-observer: 2%, CoVinter-observer: 3.8%) and infarct size (CoVintra-observer: 8.4%, CoVinter-observer: 3.8%), despite drastic animal growth. Conclusions: Best results are achieved via manual segmentation. We define state-of-the-art procedures for large animal studies at 7 T.

Finite element models with automatic computed tomography bone segmentation for failure load computation

Bone segmentation is an important step to perform biomechanical failure load simulations on in-vivo CT data of patients with bone metastasis, as it is a mandatory operation to obtain meshes needed for numerical simulations. Segmentation can be a tedious and time consuming task when done manually, and expert segmentations are subject to intra- and inter-operator variability. Deep learning methods are increasingly employed to automatically carry out image segmentation tasks. These networks usually need to be trained on a large image dataset along with the manual segmentations to maximize generalization to new images, but it is not always possible to have access to a multitude of CT-scans with the associated ground truth. It then becomes necessary to use training techniques to make the best use of the limited available data. In this paper, we propose a dedicated pipeline of preprocessing, deep learning based segmentation method and post-processing for in-vivo human femurs and vertebrae segmentation from CT-scans volumes. We experimented with three U-Net architectures and showed that out-of-the-box models enable automatic and high-quality volume segmentation if carefully trained. We compared the failure load simulation results obtained on femurs and vertebrae using either automatic or manual segmentations and studied the sensitivity of the simulations on small variations of the automatic segmentation. The failure loads obtained using automatic segmentations were comparable to those obtained using manual expert segmentations for all the femurs and vertebrae tested, demonstrating the effectiveness of the automated segmentation approach for failure load simulations.

Deep Learning Models for Abdominal CT Organ Segmentation in Children: Development and Validation in Internal and Heterogeneous Public Datasets.

BACKGROUND. Deep learning abdominal organ segmentation algorithms have shown excellent results in adults; validation in children is sparse. OBJECTIVE. The purpose of this article is to develop and validate deep learning models for liver, spleen, and pancreas segmentation on pediatric CT examinations. METHODS. This retrospective study developed and validated deep learning models for liver, spleen, and pancreas segmentation using 1731 CT examinations (1504 training, 221 testing), derived from three internal institutional pediatric (age ≤ 18 years) datasets (n = 483) and three public datasets comprising pediatric and adult examinations with various pathologies (n = 1248). Three deep learning model architectures (SegResNet, DynUNet, and SwinUNETR) from the Medical Open Network for Artificial Intelligence (MONAI) framework underwent training using native training (NT), relying solely on institutional datasets, and transfer learning (TL), incorporating pretraining on public datasets. For comparison, TotalSegmentator, a publicly available segmentation model, was applied to test data without further training. Segmentation performance was evaluated using mean Dice similarity coefficient (DSC), with manual segmentations as reference. RESULTS. For internal pediatric data, the DSC for TotalSegmentator, NT models, and TL models for normal liver was 0.953, 0.964-0.965, and 0.965-0.966, respectively; for normal spleen, 0.914, 0.942-0.945, and 0.937-0.945; for normal pancreas, 0.733, 0.774-0.785, and 0.775-0.786; and for pancreas with pancreatitis, 0.703, 0.590-0.640, and 0.667-0.711. For public pediatric data, the DSC for TotalSegmentator, NT models, and TL models for liver was 0.952, 0.871-0.908, and 0.941-0.946, respectively; for spleen, 0.905, 0.771-0.827, and 0.897-0.926; and for pancreas, 0.700, 0.577-0.648, and 0.693-0.736. For public primarily adult data, the DSC for TotalSegmentator, NT models, and TL models for liver was 0.991, 0.633-0.750, and 0.926-0.952, respectively; for spleen, 0.983, 0.569-0.604, and 0.923-0.947; and for pancreas, 0.909, 0.148-0.241, and 0.699-0.775. The DynUNet TL model was selected as the best-performing NT or TL model considering DSC values across organs and test datasets and was made available as an open-source MONAI bundle (https://github.com/cchmc-dll/pediatric_abdominal_segmentation_bundle.git). CONCLUSION. TL models trained on heterogeneous public datasets and fine-tuned using institutional pediatric data outperformed internal NT models and Total-Segmentator across internal and external pediatric test data. Segmentation performance was better in liver and spleen than in pancreas. CLINICAL IMPACT. The selected model may be used for various volumetry applications in pediatric imaging.

Unsupervised shape-and-texture-based generative adversarial tuning of pre-trained networks for carotid segmentation from 3D ultrasound images.

Vessel-wall volume and localized three-dimensional ultrasound (3DUS) metrics are sensitive to the change of carotid atherosclerosis in response to medical/dietary interventions. Manual segmentation of the media-adventitia boundary (MAB) and lumen-intima boundary (LIB) required to obtain these metrics is time-consuming and prone to observer variability. Although supervised deep-learning segmentation models have been proposed, training of these models requires a sizeable manually segmented training set, making larger clinical studiesprohibitive. We aim to develop a method to optimize pre-trained segmentation models without requiring manual segmentation to supervise the fine-tuningprocess. We developed an adversarial framework called the unsupervised shape-and-texture generative adversarial network (USTGAN) to fine-tune a convolutional neural network (CNN) pre-trained on a source dataset for accurate segmentation of a target dataset. The network integrates a novel texture-based discriminator with a shape-based discriminator, which together provide feedback for the CNN to segment the target images in a similar way as the source images. The texture-based discriminator increases the accuracy of the CNN in locating the artery, thereby lowering the number of failed segmentations. Failed segmentation was further reduced by a self-checking mechanism to flag longitudinal discontinuity of the artery and by self-correction strategies involving surface interpolation followed by a case-specific tuning of the CNN. The U-Net was pre-trained by the source dataset involving 224 3DUS volumes with 136, 44, and 44 volumes in the training, validation and testing sets. The training of USTGAN involved the same training group of 136 volumes in the source dataset and 533 volumes in the target dataset. No segmented boundaries for the target cohort were available for training USTGAN. The validation and testing of USTGAN involved 118 and 104 volumes from the target cohort, respectively. The segmentation accuracy was quantified by Dice Similarity Coefficient (DSC), and incorrect localization rate (ILR). Tukey's Honestly Significant Difference multiple comparison test was employed to quantify the difference of DSCs between models and settings, where was considered statisticallysignificant. USTGAN attained a DSC of % in LIB and % in MAB, improving from the baseline performance of % in LIB (p ) and % in MAB (p ). Our approach outperformed six state-of-the-art domain-adaptation models (MAB: , LIB: ). The proposed USTGAN also had the lowest ILR among the methods compared (LIB: 2.5%, MAB: 1.7%). Our framework improves segmentation generalizability, thereby facilitating efficient carotid disease monitoring in multicenter trials and in clinics with less expertise in 3DUSimaging.

Petal segmentation in CT images based on divide-and-conquer strategy.

Manual segmentation of the petals of flower computed tomography (CT) images is time-consuming and labor-intensive because the flower has many petals. In this study, we aim to obtain a three-dimensional (3D) structure of Camellia japonica flowers and propose a petal segmentation method using computer vision techniques. Petal segmentation on the slice images fails by simply applying the segmentation methods because the shape of the petals in CT images differs from that of the objects targeted by the latest instance segmentation methods. To overcome these challenges, we crop two-dimensional (2D) long rectangles from each slice image and apply the segmentation method to segment the petals on the images. Thanks to cropping, it is easier to segment the shape of the petals in the cropped images using the segmentation methods. We can also use the latest segmentation method for the task because the number of images used for training is augmented by cropping. Subsequently, the results are integrated into 3D to obtain 3D segmentation volume data. The experimental results show that the proposed method can segment petals on slice images with higher accuracy than the method without cropping. The 3D segmentation results were also obtained and visualized successfully.

Combined MR quantitative susceptibility mapping and multi-shell diffusion in Parkinson's disease.

Quantitative susceptibility mapping (QSM), neurite orientation dispersion and density imaging (NODDI), and the g-ratio have separately shown differences between Parkinson's disease (PD) and healthy controls. The g-ratio has, however, not been studied in PD in the substantia nigra (SN) and the putamen. A combination of these methods could also potentially be a complementary imaging biomarker for PD. This study aimed to assess the diagnostic performance of QSM, NODDI, the g-ratio, and a combined QSM-NODDI imaging marker in the SN and putamen of PD patients. In this prospective study, the diagnostic performance of median region of interest values was compared in a cohort of 15 participants with PD and 14 healthy controls after manual segmentation. The diagnostic performance was assessed using the area under curve (AUC) for the receiving operator characteristic. Median QSM in the contralateral SN identified PD with AUC 0.77, and median isotropic volume fraction identified PD in the ipsilateral SN with AUC 0.68. A combined NODDI-QSM marker improved diagnostic performance (AUC 0.80). No significant differences were found in the g-ratio. A combination of median QSM and median isotropic volume fraction improves the differentiation of PD from healthy controls and is a potential biomarker in the diagnostics of PD. This confirms previously reported results indicating that combining QSM and NODDI modestly improves differentiation of PD.

Segmentation-Free Outcome Prediction from Head and Neck Cancer PET/CT Images: Deep Learning-Based Feature Extraction from Multi-Angle Maximum Intensity Projections (MA-MIPs).

We introduce an innovative, simple, effective segmentation-free approach for survival analysis of head and neck cancer (HNC) patients from PET/CT images. By harnessing deep learning-based feature extraction techniques and multi-angle maximum intensity projections (MA-MIPs) applied to Fluorodeoxyglucose Positron Emission Tomography (FDG-PET) images, our proposed method eliminates the need for manual segmentations of regions-of-interest (ROIs) such as primary tumors and involved lymph nodes. Instead, a state-of-the-art object detection model is trained utilizing the CT images to perform automatic cropping of the head and neck anatomical area, instead of only the lesions or involved lymph nodes on the PET volumes. A pre-trained deep convolutional neural network backbone is then utilized to extract deep features from MA-MIPs obtained from 72 multi-angel axial rotations of the cropped PET volumes. These deep features extracted from multiple projection views of the PET volumes are then aggregated and fused, and employed to perform recurrence-free survival analysis on a cohort of 489 HNC patients. The proposed approach outperforms the best performing method on the target dataset for the task of recurrence-free survival analysis. By circumventing the manual delineation of the malignancies on the FDG PET-CT images, our approach eliminates the dependency on subjective interpretations and highly enhances the reproducibility of the proposed survival analysis method. The code for this work is publicly released.

Advanced U-Net Model for Precise Brain Tumor Segmentation in MRI Images

The accuracy of medical image segmentation is crucial for diagnosis and treatment planning in the modern healthcare system. Deep learning methods, like CNNs, UNETs, and Transformers, have completely changed this industry by automating labor-intensive manual segmentation procedures that were previously done by hand. However, problems like complex architectures and blurry characteristics continue, which causes issues with accuracy. Researchers are working hard to overcome these obstacles to fully realize the potential of medical image segmentation in the revolution of healthcare. Our paper presents an enhanced U-Net model specifically designed for brain tumour MRI image segmentation to improve precision. There are three primary components to our strategy. First, we prioritize feature augmentation using methods like CLAHE in the picture preprocessing phase. Second, we modify the U-Net model's architecture with an emphasis on a customized layered design in order to improve segmentation outcomes. Finally, we use a CNN model for post-processing to further optimize segmentation results using further convolutional layers. A total of 3,064 brain MRI pictures were used to test (612 images), validate (612 images), and train (1,840 images) our model. We obtained exceptional recall (93.66%), accuracy (97.79%), F-score (93.15%), and precision (92.66%). The Dice coefficient's training and validation curves showed little variation, with training reaching roughly 93% and validation 84%, suggesting good generalization ability. High accuracy was validated by visual review of the segmentation findings, albeit occasionally little mistakes like false positives were noticed.

Convolutional Neural Network-Based Automated Segmentation of Skeletal Muscle and Subcutaneous Adipose Tissue on Thigh MRI in Muscular Dystrophy Patients.

We aim to develop a deep learning-based algorithm for automated segmentation of thigh muscles and subcutaneous adipose tissue (SAT) from T1-weighted muscle MRIs from patients affected by muscular dystrophies (MDs). From March 2019 to February 2022, adult and pediatric patients affected by MDs were enrolled from Azienda Ospedaliera Universitaria Pisana, Pisa, Italy (Institution 1) and the IRCCS Stella Maris Foundation, Calambrone-Pisa, Italy (Institution 2), respectively. All patients underwent a bilateral thighs MRI including an axial T1 weighted in- and out-of-phase (dual-echo). Both muscles and SAT were manually and separately segmented on out-of-phase image sets by a radiologist with 6 years of experience in musculoskeletal imaging. A U-Net1 and U-Net3 were built to automatically segment the SAT, all the thigh muscles together and the three muscular compartments separately. The dataset was randomly split into the on train, validation, and test set. The segmentation performance was assessed through the Dice similarity coefficient (DSC). The final cohort included 23 patients. The estimated DSC for U-Net1 was 96.8%, 95.3%, and 95.6% on train, validation, and test set, respectively, while the estimated accuracy for U-Net3 was 94.1%, 92.9%, and 93.9%. Both of the U-Nets achieved a median DSC of 0.95 for SAT segmentation. The U-Net1 and the U-Net3 achieved an optimal agreement with manual segmentation for the automatic segmentation. The so-developed neural networks have the potential to automatically segment thigh muscles and SAT in patients affected by MDs.

A Preoperative Prediction Model for Lymph Node Metastasis in Patients with Gastric Cancer Using a Machine Learning-based Ultrasomics Approach.

This study aimed to develop an ultrasomics model for predicting lymph node metastasis preoperative in patients with gastric cancer (GC). This study enrolled GC patients who underwent preoperative ultrasound examination. Manual segmentation of the region of interest (ROI) was performed by an experienced radiologist to extract radiomics features using the Pyradiomics software. The Z-score algorithm was used for feature normalization, followed by the Wilcoxon test to identify the most informative features. Linear prediction models were constructed using the least absolute shrinkage and selection operator (LASSO). The performance of the ultrasomics model was evaluated using the area under curve (AUC), sensitivity, specificity, and the corresponding 95% confidence intervals (CIs). A total of 464 GC patients (mean age: 60.4 years ±11.3 [SD]; 328 men [70.7%]) were analyzed, of whom 291 had lymph node metastasis. The patients were randomly assigned to either the training (n=324) or test (n=140) sets, using a 7:3 ratio. An ultrasomics model that consisted of 19 radiomics features was developed using Wilcoxon and LASSO algorithms in the training set. Our ultrasomics model showed moderate performance for lymph node metastasis prediction in both the training (AUC: 0.802, 95%CI: 0.752-0.851, P<0.001) and test sets (AUC: 0.802, 95%CI: 0.724-0.879, P<0.001). The calibration curve analysis indicated good agreement between the predicted probabilities of ultrasomics and actual lymph node metastasis status. Our study highlights the potential of a machine learning-based ultrasomics model in predicting lymph node metastasis in GC patients, offering implications for personalized therapy approaches.

Use of automated assessment for determining associations of low muscle mass and muscle loss with overall survival in patients with colorectal cancer – A validation study

BackgroundLow muscle mass and skeletal muscle mass (SMM) loss are associated with adverse patient outcomes, but the time-consuming nature of manual SMM quantification prohibits implementation of this metric in clinical practice. Therefore, we assessed the feasibility of automated SMM quantification compared to manual quantification. We evaluated both diagnostic accuracy for low muscle mass and associations of SMM (change) with survival in colorectal cancer (CRC) patients. MethodsComputed tomography (CT) images from CRC patients enrolled in two clinical studies were analyzed. We compared i) manual vs. automated segmentation of preselected slices at the third lumbar [L3] vertebra ("semi-automated"), and ii) manual L3-slice-selection + manual segmentation vs. automated L3-slice-selection + automated segmentation ("fully-automated"). Automated L3-selection and automated segmentation was performed with Quantib Body Composition v0.2.1. Bland-Altman analyses, within-subject coefficients of variation (WSCVs) and Intraclass Correlation Coefficients (ICCs) were used to evaluate the agreement between manual and automatic segmentation. Diagnostic accuracy for low muscle mass (defined by an established sarcopenia cut-off) was calculated with manual assessment as the “gold standard”. Using either manual or automated assessment, Cox proportional hazard ratios (HRs) were used to study the association between changes in SMM (>5% decrease yes/no) during first-line metastatic CRC treatment and mortality adjusted for prognostic factors. SMM change was also assessed separately in weight-stable (<5%, i.e. occult SMM loss) patients. ResultsIn total, 1580 CT scans were analyzed, while a subset of 307 scans were analyzed in the fully-automated comparison. Included patients (n=553) had a mean age of 63±9 years and 39% were female. The semi-automated comparison revealed a bias of -2.41 cm2, 95% limits of agreement [-9.02 to 4.20], a WSCV of 2.25%, and an ICC of 0.99 (95% confidence intervals (CI) 0.97 to 1.00). The fully-automated comparison method revealed a bias of -0.08 cm2 [-10.91 to 10.75], a WSCV of 2.85% and an ICC of 0.98 (95% CI 0.98 to 0.99). Sensitivity and specificity for low muscle mass were 0.99 and 0.89 for the semi-automated comparison and 0.96 and 0.90 for the fully-automated comparison. SMM decrease was associated with shorter survival in both manual and automated assessment (n=78/280, HR 1.36 [95% CI 1.03 to 1.80] and n=89/280, HR 1.38 [95% CI 1.05 to 1.81]). Occult SMM loss was associated with shorter survival in manual assessment, but not significantly in automated assessment (n=44/263, HR 1.43 [95% CI 1.01 to 2.03] and n=51/2639, HR 1.23 [95% CI 0.87 to 1.74]). ConclusionDeep-learning based assessment of SMM at L3 shows reliable performance, enabling the use of CT measures to guide clinical decision making. Implementation in clinical practice helps to identify patients with low muscle mass or (occult) SMM loss who may benefit from lifestyle interventions.

Artificial intelligence-enhanced opportunistic screening of osteoporosis in CT scan: a scoping Review.

This scoping review aimed to assess the current research on artificial intelligence (AI)--enhanced opportunistic screening approaches for stratifying osteoporosis and osteopenia risk by evaluating vertebral trabecular bone structure in CT scans. PubMed, Scopus, and Web of Science databases were systematically searched for studies published between 2018 and December 2023. Inclusion criteria encompassed articles focusing on AI techniques for classifying osteoporosis/osteopenia or determining bone mineral density using CT scans of vertebral bodies. Data extraction included study characteristics, methodologies, and key findings. Fourteen studies met the inclusion criteria. Three main approaches were identified: fully automated deep learning solutions, hybrid approaches combining deep learning and conventional machine learning, and non-automated solutions using manual segmentation followed by AI analysis. Studies demonstrated high accuracy in bone mineral density prediction (86-96%) and classification of normal versus osteoporotic subjects (AUC 0.927-0.984). However, significant heterogeneity was observed in methodologies, workflows, and ground truth selection. The review highlights AI's promising potential in enhancing opportunistic screening for osteoporosis using CT scans. While the field is still in its early stages, with most solutions at the proof-of-concept phase, the evidence supports increased efforts to incorporate AI into radiologic workflows. Addressing knowledge gaps, such as standardizing benchmarks and increasing external validation, will be crucial for advancing the clinical application of these AI-enhanced screening methods. Integration of such technologies could lead to improved early detection of osteoporotic conditions at a low economic cost.

Radiomics by Quantitative Diffusion-weighted MRI for Predicting Response in Patients with Extremity Soft-tissue Undifferentiated Pleomorphic Sarcoma

Purpose: This study aimed to determine the relevance of first- and high-order radiomic features derived from Diffusion-Weighted Imaging (DWI) and Apparent Diffusion Coefficient (ADC) maps for predicting treatment response in patients with Undifferentiated Pleomorphic Sarcoma (UPS). Methods: This retrospective study included 33 extremity UPS patients with pre-surgical DWI/ADC and surgical resection. Manual volumetric tumor segmentation was performed on DWI/ADC maps acquired at Baseline (BL), Post-Chemotherapy (PC), and Post-Radiation Therapy (PRT). The percentage of pathology-assessed treatment effect (PATE) in surgical specimens categorized patients into responders (R; PATE ≥ 90%; 16 patients), partial-responders (PR; 89% - 31% PATE; 10 patients), and non-responders (NR; PATE ≤ 30%; 7 patients). 107 radiomic features were extracted from BL, PC, and PRT ADC maps. Statistical analyses compared R vs. PR/NR. Results: Pseudo-progression at PC and universal stability at PRT were observed in R and PR/NR based on RECIST, WHO, and volumetric assessments. At PRT, responders displayed a 35% increase in ADC mean (p = 0.0034), a 136% decrease in skewness (p = 0.0001), and a 363% increase in the 90th percentile proportion (p = 0.0009). Comparing R vs. PR/NR at BL, statistically significant differences were observed in glrlm_highgraylevelrunemphasis (p = 0.0081), glrlm_shortrunhighgraylevelemphasis (p = 0.0138), gldm_highgraylevelemphasis (p = 0.0138), glcm_sumaverage (p = 0.0164), glcm_jointaverage (p = 0.0164), and glcm_autocorrelation (p = 0.0193). At PC, firstorder_meanabsolutedeviation (p = 0.0078), firstorder_interquartilerange (p = 0.0109), firstorder_variance (p = 0.0109), and firstorder_robustmeanabsolutedeviation (p = 0.0151) provided statistically significant differences. Conclusion: Observing a high post-therapeutic ADC mean, low skewness, and high 90th percentile proportion with respect to baseline is predictive of successfully treated UPS patients presenting &gt; 90% PATE. Highly significant higher-order radiomic results include glrlm-highgraylevelrunemphasis (BL) and first-order-mean absolute deviation (PC).

Pixel-wise segmentation of cells in digitized Pap smear images

A simple and cheap way to recognize cervical cancer is using light microscopic analysis of Pap smear images. Training artificial intelligence-based systems becomes possible in this domain, e.g., to follow the European recommendation to screen negative smears to reduce false negative cases. The first step for such a process is segmenting the cells. A large and manually segmented dataset is required for this task, which can be used to train deep learning-based solutions. We describe a corresponding dataset with accurate manual segmentations for the enclosed cells. Altogether, the APACS23 (Annotated PAp smear images for Cell Segmentation 2023) dataset contains about 37 000 manually segmented cells and is separated into dedicated training and test parts, which could be used for an official benchmark of scientific investigations or a grand challenge.