Diagnostics最新文献

Background/Objectives: Obstructive sleep apnea (OSA) is a prevalent, yet underdiagnosed, disorder associated with cardiovascular and cognitive risks. While overnight polysomnography (PSG) remains the diagnostic gold standard, it is resource-intensive and impractical for large-scale rapid screening. Methods: This study extends prior work on feature extraction and binary classification using tracheal breathing sounds (TBS) and anthropometric data by introducing a meta-modeling framework that utilizes machine learning (ML) and aggregates six one-vs.-one classifiers for multi-class OSA severity prediction. We employed out-of-bag (OOB) estimation and three-fold cross-validation to assess model generalization performance. To enhance reliability, the framework incorporates conformal prediction to provide calibrated confidence sets. Results: In the three-class setting (non, mild, moderate/severe), the model achieved 76.7% test accuracy, 77.7% sensitivity, and 87.1% specificity, with strong OOB performance of 91.1% accuracy, 91.6% sensitivity, and 95.3% specificity. Three-fold confirmed stable performance across folds (mean accuracy: 77.8%; mean sensitivity: 78.6%; mean specificity: 76.4%) and conformal prediction achieved full coverage with an average set size of 2. In the four-class setting (non, mild, moderate, severe), the model achieved 76.7% test accuracy, 75% sensitivity, and 92% specificity, with OOB performance of 88.2% accuracy, 91.6% sensitivity, and 88.2% specificity. Conclusions: These findings support the potential of this non-invasive system as an efficient and rapid OSA severity assessment whilst awake, offering a scalable alternative to PSG for large-scale screening and clinical triaging.

Pediatric odontogenic tumors are rare but are frequently overlooked because they often mimic simple cysts on routine radiographic examinations. The radiographic appearance on panoramic imaging and cone-beam computed tomography (CBCT) frequently does not correlate with the true biological nature of these lesions. On CBCT, classic odontogenic tumors often demonstrate mixed radiolucent-radiopaque patterns with ill-defined borders, internal calcifications, septations, or other structural features. The diagnostic challenge arises when an odontogenic tumor mimics a unilateral, well-defined radiolucent area or a cystic lesion with clear borders and no associated tooth displacement, erosion, root resorption, or cortical bone dehiscence. Panoramic radiography has inherent diagnostic limitations but remains widely used for routine dental screening. CBCT provides enhanced three-dimensional assessment and improves diagnostic accuracy in the evaluation of jaw lesions. A marked increase in dental follicle diameter necessitates differentiation between cystic transformation, inflammatory processes, and other odontogenic pathologies. Cortical swelling and bone asymmetry warrant careful evaluation. In this context, an atypical cyst-like lesion detected on routine panoramic radiography prompted a needle aspiration biopsy, which revealed a dry tap and suggested a solid lesion. This prompted CBCT evaluation. Two juvenile cases are presented in which clinical findings, panoramic radiography, and CBCT provided discordant diagnostic impressions of cystic-appearing lesions with well-defined borders and bone expansion. These cases illustrate a diagnostic pathway in which imaging demonstrates a cyst-like appearance with benign radiological features, fine-needle aspiration biopsy reveals the absence of cystic fluid, and histopathology confirms that radiology alone cannot reliably distinguish true cysts from solid odontogenic tumors in pediatric patients.

Background: Osteoporosis (OP) is characterized by reduced bone mineral density and increased fracture risk. Many spinal surgery patients have undiagnosed OP due to the lack of preoperative screening, leading to postoperative complications. Magnetic resonance imaging (MRI), a routine, non-invasive tool for spinal assessment, offers potential for opportunistic OP detection. This study aimed to develop deep learning models to identify OP using lumbar MRI. Methods: We retrospectively enrolled 218 patients (≥50 years) who underwent both lumbar MRI and dual-energy X-ray absorptiometry (DXA). After segmentation of vertebral bodies from T1- and T2-weighted MRI images, 738 images per sequence were extracted. Separate convolutional neural network (CNN) models were trained for each sequence. Model performance was evaluated using receiver operating characteristic curves and area under the curve (AUC). Results: Among tested classifiers, EfficientNet b4 showed the best performance. For the T1-weighted model, it achieved an AUC of 82%, with a sensitivity of 85% and specificity of 79%. For the T2-weighted model, the AUC was 83%, with a sensitivity of 86% and specificity of 80%. These results were superior to those of InceptionResNet v2 and ResNet-50 for both sequences. Conclusions: The AI models provided reliable OP classification without additional imaging or radiation. AI-based analysis of standard lumbar MRI sequences can accurately identify OP. These models may assist in early detection of undiagnosed OP in surgical candidates, enabling timely treatment and perioperative strategies to improve outcomes and reduce healthcare burden.

Background/Objectives: Multimodal large language models (MLLMs) offer potential for automated fracture detection, yet their diagnostic stability under repeated inference remains underexplored. This study evaluates the diagnostic accuracy, stability, and intra-model consistency of four MLLMs in detecting hand fractures on plain radiographs. Methods: In total, images of hand radiographs of 65 adult patients with confirmed hand fractures (30 phalangeal, 30 metacarpal, 5 scaphoid) were evaluated by four models: GPT-5 Pro, Gemini 2.5 Pro, Claude Sonnet 4.5, and Mistral Medium 3.1. Each image was independently analyzed five times per model using identical zero-shot prompts (1300 total inferences). Diagnostic accuracy, inter-run reliability (Fleiss' κ), case-level agreement profiles, subgroup performance, and exploratory demographic inference (age, sex) were assessed. Results: GPT-5 Pro achieved the highest accuracy (64.3%) and consistency (κ = 0.71), followed by Gemini 2.5 Pro (56.9%, κ = 0.57). Mistral Medium 3.1 exhibited high agreement (κ = 0.88) despite low accuracy (38.5%), indicating systematic error ("confident hallucination"). Claude Sonnet 4.5 showed low accuracy (33.8%) and consistency (κ = 0.33), reflecting instability. While phalangeal fractures were reliably detected by top models, scaphoid fractures remained challenging. Demographic analysis revealed poor capabilities, with age estimation errors exceeding 12 years and sex prediction accuracy near random chance. Conclusions: Diagnostic accuracy and consistency are distinct performance dimensions; high intra-model agreement does not imply correctness. While GPT-5 Pro demonstrated the most favorable balance of accuracy and stability, other models exhibited critical failure modes ranging from systematic bias to random instability. At present, MLLMs should be regarded as experimental diagnostic reasoning systems rather than reliable standalone tools for clinical fracture detection.

Background/Objectives: Optoacoustic imaging technologies are emerging as promising tools for clinical practice. Several systems have the potential to fill specific niches in the medical imaging landscape thanks to a unique performance based on the combination of rich optical absorption contrast and high ultrasonic penetration-to-resolution ratios. However, current optoacoustic methods rely on tomographic reconstructions, which impose significant complexity on the systems in terms of number and distribution of transducers, acquisition electronics, and general operation. As a result, optoacoustic tomography apparatus are generally expensive and bulky and require intensive training for their operation. Here, we report on an optoacoustic imaging method that uses a single ultrasound transducer and non-tomographic image formation to overcome the drawbacks of classical tomographic methods. The method is designed for retrieving layered slab-like biological structures like tattoo ink, subcutaneous fat, muscles, or cerebrospinal fluid below the fontanelle. Moreover, it can be adapted to other geometries. Methods: We have implemented the method in a user-friendly, compact, simple, and low-cost system and tested its performance using simulations, synthetic phantoms, and biological phantoms containing tattoo ink. Results: Our results indicate that the system can discriminate slab-like structures from other shapes and recover them with the axial resolution of tomographic optoacoustic methods. The findings also suggest that the system has the potential to improve tattoo removal procedures. We further discuss its implications for pediatrics, traumatology, or endocrinology. Conclusions: This work paves the way for a new generation of simple, easy-to-use and low-cost imaging systems with the potential to impact several medical fields.

Multiphoton endomicroscopy (MPEM) has recently become a key development in optical biomedical diagnostics, providing histologically relevant in vivo images that are eliminating both the need for tissue damage during biopsy sampling and the need for dye injections. Due to its ability to visualize structures at the epithelial, extracellular matrix, and subcellular levels, MPEM offers a promising diagnostic method for precancerous conditions and early forms of gastrointestinal (GI) cancer. The high specificity of multiphoton signals-the two-photon fluorescence response of endogenous fluorophores (NADH, FAD), the second-harmonic generation signal from collagen, and others-makes this method a promising alternative to both traditional histology and confocal endoscopy, enabling real-time assessment of metabolic status, intestinal epithelial cell status, and stromal remodeling. Despite the promising prospects of multiphoton microscopy, its practical implementation is progressing extremely slowly. The main factors here include the difficulty of delivering ultrashort pulses with high peak power, which is necessary for multiphoton excitation (MPE), and obtaining these pulses at the required wavelengths to activate the autofluorescence mechanism. One of the most promising solutions is the use of specialized multimode optical fibers that can both induce beam self-cleaning (BSC), which allows for the formation of a stable beam profile close to the fundamental mode, and significantly broaden the optical spectrum, which can ultimately cover the entire region of interest. This review presents the biophysical foundations of multiphoton microscopy of GI tissue, existing endoscopic architectures for MPE, and an analysis of the potential for using novel nonlinear effects in multimode optical fibers, such as the BSC effect and supercontinuum generation. It is concluded that the use of optical fibers in which the listed effects are realized in the tracts of multiphoton endomicroscopes can become a key step in the creation of a new generation of high-resolution instruments for the early detection of malignant neoplasms of the GI tract.

Background: Unmeasured ions (UIs) contribute significantly to acid-base disturbances in critically ill patients, yet the optimal parameter for their estimation remains uncertain. The most widely used indicators are the albumin-corrected anion gap (AGc), the strong ion gap (SIG), and the base excess gap (BEGap). Methods: In this retrospective cohort study, a total of 2274 ICU patients (2018-2022) were included in the development cohort, and an independent external validation cohort of 1202 patients (2023-2025) was used to assess temporal generalizability. Three approaches to blood gas analysis-traditional (PaCO₂, HCO₃^-, AGc), Stewart (PaCO₂, SIDa, ATOT, SIG), and partitioned base excess (PaCO₂, BECl, BEAlb, BELac, BEGap)-were evaluated. Multivariable linear regression (MLR) and machine learning (ML, random forest [RF], extreme gradient boosting [XGBoost], and support vector regression [SVR]) were applied to evaluate the explanatory performance of analytical approaches with respect to arterial pH. Model performance was assessed using adjusted R², RMSE, and MAE. Variable importance was quantified with tree-based methods, SHAP values, and permutation importance. All modeling and reporting steps followed the PROBAST-AI guideline. Results: In multiple linear regression (MLR), the partitioned base excess (BE) approach achieved the highest explanatory performance (adjusted R² = 0.949), followed by the traditional (0.929) and Stewart approaches (0.926). In ML analyses, model fit was high across all approaches. For the traditional approach, R² values were 0.979 with RF, 0.974 with XGBoost, and 0.934 with SVR. The Stewart's approach showed lower overall explanatory performance, with R² values of 0.876 (RF), 0.967 (XGBoost), and 0.996 (SVR). The partitioned BE approach again demonstrated the strongest explanatory performance, achieving R² values of 0.975 with XGBoost and 0.989 with SVR. Across all analytical models, BEGap consistently emerged as a strong and independent determinant of arterial pH, outperforming SIG and AGc. SIG showed an intermediate contribution, whereas AGc provided minimal independent explanatory value. Among ML models, XGBoost showed the most stable and accurate explanatory performance across approaches. Conclusions: This study demonstrates that BEGap is a practical, physiologically informative, and bedside-applicable parameter for assessing unmeasured ions, outperforming both AGc and SIG across linear and non-linear analytical models.

Objectives: To develop a simplified MRI-based shorthand assessment method, referred to as the Sagittal Tibial Epi-Physis (STEP) Shorthand, for skeletal age assessment in skeletally immature patients with anterior cruciate ligament (ACL) injuries. This study aimed to elaborate a single-plane MRI-based skeletal age estimation tool and to explore its feasibility and inter-rater reliability in comparison with existing MRI-based shorthands. Methods: This prospective study included 130 knee MRIs (79% males) from 97 skeletally immature patients (overall average age of 14.0 ± 2.1 years) with ACL injuries treated between February 2022 and January 2025. A new shorthand assessment method was developed based on sagittal T1-weighted MRI evaluation of the proximal tibial epiphysis. A validation cohort of 74 MRIs was independently evaluated by four raters with different levels of expertise using the STEP, Meza, and Politzer shorthand atlases. Inter-rater reliability (ICC), intra-rater agreement (Cohen's kappa), and association with chronological age (Spearman rho) were calculated. Results: The STEP Shorthand tool demonstrated a strong association with chronological age (rho = 0.890, p < 0.001) with consistent associations across sex subgroups. Inter-rater reliability was high and comparable to established MRI-based shorthands. The use of a focused sagittal T1-weighted evaluation allowed for a simplified and reproducible assessment across raters with varying experience levels. Conclusions: The STEP Shorthand represents a pragmatic and reliable tool for MRI-based skeletal age assessment in pediatric and adolescent patients with ACL injuries. The STEP Shorthand can support timely decision-making in surgical planning and enhance standardization across different levels of clinical expertise.

Background/Objectives: Considering the excretion pathways and administered gadolinium dose, our institution has developed a tailored gadolinium-based contrast agents (GBCAs) administration protocol for patients with renal impairment to facilitate more rapid elimination and minimal retention of gadolinium. This study aims to evaluate the 8-year clinical outcomes and safety of this institutional protocol. Methods: This single-center retrospective study included patients with renal impairment who underwent GBCA-enhanced MRI between January 2015 and December 2022. The protocol recommended specific GBCAs and adjusted doses based on chronic kidney disease (CKD) stage and serum bilirubin levels: gadoxetate disodium was used for normal serum bilirubin level due to its dual excretion pathway, while macrocyclic agents were used for those with elevated serum bilirubin levels. During the follow-up period, occurrence of nephrogenic systemic fibrosis (NSF) and evidence of gadolinium deposition in brain tissues were evaluated. Results: A total of 288 patients (age, 64.6 ± 11.7 years; male, 64.9%) underwent 716 GBCA-enhanced MRI examinations in accordance with the institutional protocol. The cohort included 62 patients with CKD stage 4 and 131 patients with CKD stage 5 or undergoing hemodialysis. In patients with CKD stage 4 and 5 and those undergoing hemodialysis, 597 examinations were performed using gadoxetate disodium, and 119 used macrocyclic agents. No cases of NSF or gadolinium deposition in brain tissues were identified over mean follow-up intervals of 27.5 and 27.8 months, respectively. Conclusions: The tailored GBCA administration protocol, considering the excretion pathways and administered gadolinium dose, appears to be safe with respect to NSF for patients with renal impairment, and no evidence of brain gadolinium deposition was observed in the evaluated subset of patients.

Background: Persistent metopic suture represents a normal anatomical variant that may persist into adulthood and can be misinterpreted as a frontal skull fracture, particularly in trauma-related forensic cases. Despite its clinical and medico-legal relevance, data derived from autopsy-based evaluations remain limited, with most prevalence studies relying on dry skull collections or radiological series. This study aimed to determine the prevalence and morphological characteristics of persistent metopic suture in adult autopsy cases and to evaluate its distribution according to age, sex, and cause of death. Methods: This cross-sectional study included 500 consecutive adult autopsy cases (≥18 years). The frontal bone was directly inspected during autopsy for the presence of metopic suture, which was classified as complete or incomplete. Descriptive statistics were applied, and associations between metopism and sex, age group, and cause of death were analyzed using chi-square or Fisher's exact test, as appropriate. Results: Complete metopism was identified in 7 of 500 cases, corresponding to a prevalence of 1.4% (95% confidence interval: approximately 0.6-2.9%). No incomplete metopic sutures were observed. Metopism was slightly more frequent in females than males; however, no statistically significant association was found between metopism and sex, age group, or cause of death (p > 0.05). Conclusions: Persistent metopic suture is an uncommon but clinically and forensically relevant anatomical variant in adults. Its recognition during forensic autopsy is essential to avoid misinterpretation as a cranial fracture, particularly in trauma-related deaths, thereby preventing diagnostic and medico-legal errors.