Clarisa I. Sánchez

Background: Artificial intelligence (AI) allows automated evaluation of intracoronary optical coherence tomography (OCT) images. However, algorithms are mostly developed and validated on well‐curated data sets, which may not represent real‐world data. We sought to externally validate a previously developed algorithm performing full‐vessel segmentation for OCT in an unselected consecutive real‐world data set. Methods: This was a retrospective, single‐center, external validation study comprising 100 consecutive patients undergoing clinically indicated OCT. A previously developed AI algorithm (OCT‐AID) was used for automated pixelwise labeling of OCT images, distinguishing among lumen, guidewire artifact, intima, media, lipid plaque, calcium plaque, side branch, plaque rupture, thrombus, microvessel, and background. The AI‐based predictions were compared on a frame level to the reference standard obtained through manual OCT image analysis by expert readers. Results: Among 2560 analyzable frames, the agreement between the automated OCT image analysis and the reference standard was excellent for calcified plaque identification (κ=0.88 [95% CI, 0.84–0.92]) and quantification (intraclass correlation coefficient values ranged between 0.79 and 0.93), with a performance close to interobserver variability. For lipid plaque identification and quantification, the model performance was reasonable (κ=0.68 [95% CI, 0.64–0.72]; intraclass correlation coefficient for lipid arc, 0.79 [95% CI, 0.76–0.81]; intraclass correlation coefficient for minimum fibrous cap thickness, 0.59 [95% CI, 0.55–0.63]) and largely superior to interobserver variability. The algorithm performance for low‐prevalence features (e.g., plaque rupture) was limited. Conclusions: AI‐based fully automated evaluation of OCT images is feasible with performances consistent with interobserver variability in a real‐world data set of consecutive patients, supporting generalizability of the proposed methodology.

Background/Objectives: Computer-aided vessel occlusion detection for acute ischemic stroke is typically developed on early-phase CT angiography (CTA). In late-phase CTA, enhanced venous contrast can mimic arterial occlusions and lead to incorrect occlusion detections. To counteract this, we propose a constraint-based post-processing framework that leverages anatomical and temporal vascular information from CT perfusion (CTP) scans. Methods: Vascular anatomy and relative contrast-arrival times were encoded as time–vessel maps. Three complementary constraints worked on these maps to remove incorrect detections based on distance-to-vessel truncation points, contrast-arrival time, and spatial priors. Constraints were fixed on a 32-case late-phase development subset and applied to an external held-out cohort of late-phase CTA (N = 354). To enable application without CTP, an auxiliary CTA-to-time–vessel map generator was designed to estimate maps directly from CTA. Results: In the external cohort, the framework reduced incorrect occlusions per scan from 3.9 (95% CI 3.6–4.2) to 2.1 (95% CI 1.9–2.3), preserving occlusion-level sensitivity at 80% (95% CI 69–91). The patient-level AUROC was 83 for the baseline versus 88 with post-processing. Improvements were also observed in detectors exposed to late-phase CTA, where post-processing reduced incorrect occlusions per scan from 2.8 to 1.9. As only 51 occlusion-positive cases were present in the external cohort, confidence intervals were wide. Consequently, these estimates should be interpreted with caution. Conclusions: Anatomically and temporally informed post-processing reduced incorrect occlusion detections in late-phase CTA without measurable loss of sensitivity, potentially improving robustness against CTA phase changes in external institutions.

Background: Recent advances in deep learning have led to the development of ECG foundation models (ECG-FMs) trained with self-supervised learning, which can extract generalizable representations from large-scale data. In this study, we evaluated the performance of these ECG-FMs in detecting atrial fibrillation and sinus rhythm during continuous monitoring of intensive care unit (ICU) patients, using data from Amsterdam University Medical Center (Amsterdam UMC). Our three-stage study: (i) used within-dataset classification performance on public PhysioNet datasets as an indirect proxy for label consistency, (ii) tested off-the-shelf fine-tuned ECG-FMs on ICU data, and (iii) fine-tuned ECG-FM on curated public datasets, with and without weakly labelled in-house ICU data. Results: Off-the-shelf model showed limited transferability (F1 = 0.68 for sinus rhythm (SR), 0.52 for atrial fibrillation (AF)). Fine-tuning on consistent public datasets achieved robust performance (F1 = 0.91 for SR, 0.82 for AF), while adding weakly labelled ICU data further improved recall at some cost to precision. Performance remained stable across varying ECG lead configurations, supporting adaptability to heterogeneous inputs. Conclusions: These findings indicate that off-the-shelf pre-trained ECG-FMs do not generalise well to continuous monitoring in ICU populations; however, after targeted fine-tuning, they can achieve strong performance even in unseen cohorts, underscoring the importance of high-quality fine-tuning data and rigorous local validation.

Background: Trustworthy AI serves as a foundational pillar for two major AI ethics conferences: AIES and FAccT. Current research often adopts techno-centric approaches, focusing primarily on technical attributes such as accuracy, reliability, robustness, and fairness, while overlooking the sociotechnical dimensions critical to understanding AI trustworthiness in real-world contexts. Objectives: This scoping review aims to examine how the AIES and FAccT communities conceptualize, measure, and validate AI trustworthiness, identifying major gaps and opportunities for advancing a holistic understanding of trustworthy AI systems. Methods: We conduct a scoping review of the AIES and FAccT conference proceedings to date, systematically analyzing how trustworthiness is defined, operationalized, and applied across different research domains. Our analysis focuses on conceptualization approaches, measurement methods, verification and validation techniques, application areas, and underlying values. Results: While significant progress has been made in defining technical attributes such as transparency, accountability, and robustness, our findings reveal critical gaps. Current research often predominantly emphasizes technical precision at the expense of social and ethical considerations. The sociotechnical nature of AI systems remains less explored and trustworthiness emerges as a contested concept shaped by those with the power to define it. Conclusions: An interdisciplinary approach combining technical rigor with social, cultural, and institutional considerations is essential for advancing trustworthy AI. We propose actionable measures for the AI ethics community to adopt holistic frameworks that genuinely address the complex interplay between AI systems and society, ultimately promoting responsible technological development that benefits all stakeholders.

Deep learning models are increasingly applied to electrocardiogram (ECG) analysis to optimise cardiovascular care. However, potential biases within these models may impact their reliability and clinical applicability. This study investigates potential sex bias in deep learning models for 12-lead ECG classification of Sinus Rhythm (SR), Atrial Fibrillation (AF), and Myocardial Infarction (MI). We evaluate three model architectures—Convolutional Neural Network, xResNet101, and a Residual Network with an attention mechanism—under varying sex ratios in the training data. Among these models, the attention-based Residual Network demonstrated the highest and most equitable performance, particularly in SR and AF classification. MI classification exhibited pronounced sex-based disparities, even with balanced training data. These findings underscore the importance of incorporating fairness considerations in the development of clinical deep learning systems to ensure reliable and unbiased performance across diverse patient populations. Moreover, optimising lead selection may further enhance both fairness and overall model performance.

In this paper, we present a new approach for uncertainty-aware retinal layer segmentation in Optical Coherence Tomography (OCT) scans using probabilistic signed distance functions (SDF). Traditional pixel-wise and regression-based methods primarily encounter difficulties in precise segmentation and lack of geometrical grounding respectively. To address these shortcomings, our methodology refines the segmentation by predicting a signed distance function (SDF) that effectively parameterizes the retinal layer shape via level set. We further enhance the framework by integrating probabilistic modeling, applying Gaussian distributions to encapsulate the uncertainty in the shape parameterization. This ensures a robust representation of the retinal layer morphology even in the presence of ambiguous input, imaging noise, and unreliable segmentations. Both quantitative and qualitative evaluations demonstrate superior performance when compared to other methods. Additionally, we conducted experiments on artificially distorted datasets with various noise types-shadowing, blinking, speckle, and motion-common in OCT scans to showcase the effectiveness of our uncertainty estimation. Our findings demonstrate the possibility to obtain reliable segmentation of retinal layers, as well as an initial step towards the characterization of layer integrity, a key biomarker for disease progression. Our code is available at \url{https://github.com/niazoys/RLS_PSDF}.

Purpose: The purpose of this study was to develop a deep learning algorithm for detecting and quantifying incomplete retinal pigment epithelium and outer retinal atrophy (iRORA) and complete retinal pigment epithelium and outer retinal atrophy (cRORA) in optical coherence tomography (OCT) that generalizes well to data from different devices and to validate in an intermediate age-related macular degeneration (iAMD) cohort. Methods: The algorithm comprised a domain adaptation (DA) model, promoting generalization across devices, and a segmentation model for detecting granular biomarkers defining iRORA/cRORA, which are combined into iRORA/cRORA segmentations. Manual annotations of iRORA/cRORA in OCTs from different devices in the MACUSTAR study (168 patients with iAMD) were compared to the algorithm's output. Eye level classification metrics included sensitivity, specificity, and quadratic weighted Cohen's κ score (κ_w). Segmentation performance was assessed quantitatively using Bland-Altman plots and qualitatively. Results: For ZEISS OCTs, sensitivity and specificity for iRORA/cRORA classification were 38.5% and 93.1%, respectively, and 60.0% and 96.4% for cRORA. For Spectralis OCTs, these were 84.0% and 93.7% for iRORA/cRORA, and 62.5% and 97.4% for cRORA. The κ_w scores for 3-way classification (none, iRORA, and cRORA) were 0.37 and 0.73 for ZEISS and Spectralis, respectively. Removing DA reduced κ_w from 0.73 to 0.63 for Spectralis. Conclusions: The DA-enabled iRORA/cRORA segmentation algorithm showed superior consistency compared to human annotations, and good generalization across OCT devices. Translational Relevance: The application of this algorithm may help toward precise and automated tracking of iAMD-related lesion changes, which is crucial in clinical settings and multicenter longitudinal studies on iAMD.