Manual segmentation of muscle and adipose compartments from computed tomography (CT) axial images is a potential bottleneck in early rapid detection and quantification of sarcopenia. A prototype deep learning neural network was trained on a multi-center collection of 3413 abdominal cancer surgery subjects to automatically segment truncal muscle, subcutaneous adipose tissue and visceral adipose tissue at the L3 lumbar vertebral level. Segmentations were externally tested on 233 polytrauma subjects. Although after severe trauma abdominal CT scans are quickly and robustly delivered, with often motion or scatter artefacts, incomplete vertebral bodies or arms that influence image quality, the concordance was generally very good for the body composition indices of Skeletal Muscle Radiation Attenuation (SMRA) (Concordance Correlation Coefficient (CCC) = 0.92), Visceral Adipose Tissue index (VATI) (CCC = 0.99) and Subcutaneous Adipose Tissue Index (SATI) (CCC = 0.99). In conclusion, this article showed an automated and accurate segmentation system to segment the cross-sectional muscle and adipose area L3 lumbar spine level on abdominal CT. Future perspectives will include fine-tuning the algorithm and minimizing the outliers.

Sarcopenia is a muscle disease that involves loss of muscle strength and physical function and is associated with adverse health effects. Even though sarcopenia has attracted increasing attention in the literature, many research findings have not yet been translated into clinical practice. In this article, we aim to validate a deep learning neural network for automated segmentation of L3 CT slices and aim to explore the potential for clinical utilization of such a tool for clinical practice. A deep learning neural network was trained on a multi-centre collection of 3413 abdominal cancer surgery subjects to automatically segment muscle, subcutaneous and visceral adipose tissue at the L3 lumbar vertebral level. 536 Polytrauma subjects were used as an independent test set to show generalizability. The Dice Similarity Coefficient was calculated to validate the geometric similarity. Quantitative agreement was quantified using Bland-Altman's Limits of Agreement interval and Lin's Concordance Correlation Coefficient. To determine the potential clinical usability, randomly selected segmentation images were presented to a panel of experienced clinicians to rate on a Likert scale. Deep learning results gave excellent agreement versus a human expert operator for all of the body composition indices, with Concordance Correlation Coefficient for skeletal muscle index of 0.92, Skeletal muscle radiation attenuation 0.94, Visceral Adipose Tissue index 0.99 and Subcutaneous Adipose Tissue Index 0.99. Triple-blinded visual assessment of segmentation by clinicians correlated only to the Dice coefficient, but had no association to quantitative body composition metrics which were accurate irrespective of clinicians' visual rating. A deep learning method for automatic segmentation of truncal muscle, visceral and subcutaneous adipose tissue on individual L3 CT slices has been independently validated against expert human-generated results for an enlarged polytrauma registry dataset. Time efficiency, consistency and high accuracy relative to human experts suggest that quantitative body composition analysis with deep learning should is a promising tool for clinical application in a hospital setting.

Background: This study aimed to evaluate whether hypertrophy after portal vein embolization (PVE) and maximum liver function capacity (LiMAx) are predictable by an artificial neural network (ANN) model based on computed tomography (CT) texture features. Methods: We report a retrospective analysis on 118 patients undergoing preoperative assessment by CT before and after PVE for subsequent extended liver resection due to a malignant tumor at RWTH Aachen University Hospital. The LiMAx test was carried out in a subgroup of 55 patients prior to PVE. Associations between CT texture features and hypertrophy as well as liver function were assessed by a multilayer perceptron ANN model. Results: Liver volumetry showed a median hypertrophy degree of 33.9% (16.5–60.4%) after PVE. Non-response, defined as a hypertrophy grade lower than 25%, was found in 36.5% (43/118) of the cases. The ANN prediction of the hypertrophy response showed a sensitivity of 95.8%, specificity of 44.4% and overall prediction accuracy of 74.6% (p < 0.001). The observed median LiMAx was 327 (248–433) μg/kg/h and was strongly correlated with the predicted LiMAx (R² = 0.89). Conclusion: Our study shows that an ANN model based on CT texture features is able to predict the maximum liver function capacity and may be useful to assess potential hypertrophy after performing PVE.

Background: Early detection of patients at high risk for recurrence or local tumor progression (LTP) following thermal ablation of hepatocellular carcinoma (HCC) is essential for treatment selection and individualized follow-up. This systematic review aims to assess and compare the performance of prognostic models predicting recurrence or LTP in patients with HCC treated with thermal ablation. Methods: PubMed, Web of Science, Cochrane, and Embase were searched for studies developing models to predict recurrence after thermal ablation in treatment-naïve HCC patients, using imaging and clinical data with reported test set performance. Risk of bias and applicability were assessed by the Prediction model Risk of Bias Assessment Tool. Data on model performance, feature extraction and modeling technique was collected. Results: In total, 16 studies comprising 39 prognostic models were included, all developed using retrospective data from China or Korea. Outcomes included recurrence-free survival, (intrahepatic) early recurrence, LTP, late recurrence and aggressive intrasegmental recurrence. Predictive parameters varied across models addressing identical outcomes. Outcome definitions also differed. Nine models were externally validated. Most studies had a high risk of bias due to methodological limitations. Conclusions: Variability in model development methodology and type of predictors was found. Models that integrated multiple types of predictors consistently outperformed those relying on one type. To advance predictive tools toward clinical implementation, future research should prioritize standardized outcome definitions, external testing, and transparent reporting. Until these challenges are addressed, current evaluated models should be regarded as promising but preliminary tools.

Certainty of classifications is crucial when it comes to the practical application of machine learning models. Model performance measures such as accuracy are focused on the average performance of a model. However, when a model is used in a practical setting, such as a medical clinic, it is more important to know how certain the model is of a given prediction or classification than its average performance. Unfortunately, often models only provide a final classification label, usually of the class with the highest probability. This output, however, is not sufficiently informative of the certainty of this particular classification, especially in the presence of multiple classes: the highest probability might be only barely higher than the second highest. Even when a probability distribution is provided, there is no established metric to determine if a particular classification is more certain than a different one. In this article we propose a novel metric we have termed Multinomial Classification Certainty, to represent the certainty of model predictions. We discuss why existing methods cannot represent this type of certainty and we show the mathematical meaning behind important thresholds for this new measure.

Deep learning (DL) techniques may support localizing the epileptogenic zone (EZ) and improve surgical outcomes in drug-resistant epilepsy. This systematic review synthesizes current evidence on DL-assisted EZ localization from neuroimaging acquisitions, aiming to outline methodological trends, limitations, and future directions that bridge the gap between clinical translation and technological advances. We systematically searched PubMed, Scopus, and Embase (via Ovid) on April 15, 2025, for studies applying DL to localize the EZ using neuroimaging data. The bias and applicability of studies was assessed using the PROBAST+AI tool. We extracted methodological details, as well as key performance metrics. Thirty-six studies met the eligibility criteria, most focusing on segmenting epileptogenic lesions using structural MRI. Focal cortical dysplasia was the most commonly targeted pathology, with fully convolutional networks being the predominant DL architecture. Approximately two-thirds of the studies showed high risk of bias and clinical applicability concerns, limited by non-representative cohorts and suboptimal evaluation methods. Five studies reported promising EZ detection rate in MRI-negative cases using large multi-center cohorts, yet progress in fine-grained localization tasks, such as lesion segmentation, remained moderate. This review highlights methodological limitations hindering the clinical translation of current DL approaches for EZ localization and provides a comprehensive set of recommendations to address them. Future work should prioritize developing standardized, clinically informative evaluation frameworks and explore research avenues aligned with modern DL practices, spanning from uncertainty quantification to large-scale vision foundation models and synthetic data generation.