Familial hypercholesterolemia (FH) is a common inherited lipid disorder associated with markedly elevated cholesterol levels and a substantially increased risk of premature cardiovascular disease. Advances in population genomics, including biobank sequencing, newborn screening initiatives, and genotype-first approaches, have created new opportunities for early identification of individuals with pathogenic FH variants and for systematic cascade testing of at-risk relatives. This narrative review synthesizes current evidence on three critical domains shaping the implementation of population genomics for FH: return of genomic results, cascade testing processes, and health-system readiness. The literature indicates that returning genomic results can enhance early diagnosis and preventive treatment, yet practices vary widely regarding disclosure methods, patient engagement, and integration with cardiovascular risk assessment. Cascade testing remains the most effective and cost-efficient strategy for identifying affected relatives, but uptake is consistently low due to communication barriers, limited digital infrastructure, and insufficient clinical coordination. Ethical, legal, and social considerations, including informed consent, privacy, potential stigma, and equitable access to testing, further complicate implementation. Evidence on health-system readiness highlights important gaps in laboratory capacity, workforce training, interoperability of genomic and clinical data systems, reimbursement models, and long-term sustainability planning. Overall, population genomics offers substantial promise for improving FH detection and prevention of cardiovascular disease at scale. However, successful implementation will require standardized return-of-results frameworks, strengthened cascade-testing pathways, equity-focused policies, and coordinated health-system investments. Future research should prioritize longitudinal outcome studies, evaluation of digital and patient-centered communication tools, and cross-jurisdictional implementation frameworks to ensure that population genomic screening for FH translates into measurable public-health benefits. Keywords: Familial hypercholesterolemia, Population genomics, Cascade testing, Return of genomic results, and Health system readiness

Putting together the pieces of the care pathway: The French National Health Data system and its applications in nephrology

Accurate pretreatment prostate cancer risk assessment is essential to balance long-term treatment benefits and harm. Although some clinical and pathological parameters are established prognostic factors, the role of imaging parameters in prognostication is unclear. To assess the prognostic value of pretreatment magnetic resonance imaging (MRI) parameters for oncological outcomes in men undergoing radical prostatectomy. A systematic literature search of MEDLINE, Embase, and Scopus was performed from inception through March 2025. Studies were included if they evaluated pretreatment prostate MRI in men undergoing radical prostatectomy and reported multivariable, time-to-event analyses for the outcomes of biochemical recurrence, metastatic failure, and prostate cancer-specific mortality. Two reviewers independently extracted data and assessed study quality using the Quality in Prognostic Studies tool. Random-effects meta-analysis was performed to pool hazard ratios (HRs). The primary outcome was biochemical recurrence. Secondary outcomes included metastatic failure and prostate cancer-specific mortality. Forty studies were included (comprising 24 941 patients). Extraprostatic extension (mrT3a disease) detected with MRI was independently associated with biochemical recurrence (pooled HR, 2.16 [95% CI, 1.84-2.54]), metastatic failure (HR, 3.18 [95% CI, 2.04-4.97]), and prostate cancer-specific mortality (HR, 10.93 [95% CI, 5.05-23.65]). Seminal vesicle invasion (mrT3b disease) detected with MRI was also independently associated with biochemical recurrence (HR, 2.74 [95% CI, 2.06-3.65]) and metastatic failure (HR, 5.58 [95% CI, 1.15-27.13]). The following quantitative MRI features were prognostic for biochemical recurrence: Prostate Imaging Reporting and Data System score of 4 or 5 (HR, 2.15 [95% CI, 1.82-2.55]), large tumor size (tumor diameter ≥20 mm; HR, 2.35 [95% CI, 1.71-3.24]), and apparent diffusion coefficient values less than 0.9 × 10-3 mm2/s (HR, 2.39 [95% CI, 1.82-3.14]). Heterogeneity was moderate (I2 < 65% for mrT3a and mrT3b disease) and no significant publication bias was detected. This systematic review and meta-analysis found that pretreatment MRI provides independent prognostic value for biochemical recurrence, metastatic failure, and prostate cancer-specific mortality in men undergoing radical prostatectomy, even when adjusted for established clinicopathologic factors.

Rare diseases affect more than 300 million people globally, and only about 5% have approved therapies. Lysosomal storage disorders (LSDs) exemplify the diagnostic and long-term care complexity typical of rare diseases, and digital health technologies (DHTs), especially artificial intelligence (AI) and connected care (CC), are emerging tools to support LSD management. We aimed to map and synthesize peer-reviewed and gray literature from the past decade on DHTs relevant for LSD care, with a primary analytic focus on AI-enabled and CC solutions and a contextual mapping of other enabling DHTs. Evidence distribution was charted by population, care-journey phase, and outcome domains to identify gaps, methodological limitations, and timely priorities relevant for research, clinical practice implementation, and policies. We conducted a scoping review guided by a population, concept, context framework and operationalized through a Population, Intervention, Comparison, and Outcome (PICO)-informed data-charting structure to map study characteristics and reported outcomes, without causal or effectiveness assumptions and without risk-of-bias assessment. We searched PubMed, Google Scholar, and ClinicalTrials.gov for studies published between October 2015 and September 2024, complemented by AI-assisted discovery tools for citation extension. Reproducibility logs (search strings, run dates, filters, and stepwise counts) were maintained. Of 1751 records retrieved, 245 were included. Evidence was charted by LSD population, intervention class (AI, CC, and other enabling DHTs), outcome domains (patient, health care, and societal), and phase of the care journey. Among 245 included records, 92.2% (226/245) were peer-reviewed, and 7.8% (19/245) were gray literature; no completed and published randomized controlled trials or LSD-specific systematic reviews were identified, with evidence dominated by small, single-center observational studies. Overall, 40 peer-reviewed records reported AI-driven DHTs, 89 reported CC DHTs, and 144 reported other enabling DHTs (some multilabeled). Evidence was concentrated mostly in Gaucher and Fabry diseases. Nearly half of the mapped literature focused on screening and diagnosis, with fewer records addressing treatment intensification, rehabilitation, and end-of-life care. Outcomes were predominantly health care delivery performance measures, with fewer patient and societal outcomes. AI applications mainly supported diagnostic decision support, phenotyping, monitoring, tracking, and risk stratification; CC commonly involved telemedicine, remote monitoring, and patient-engagement platforms; enabling DHTs included interoperable data systems, registries, and digital infrastructures. The evidence base is appreciable for a niche field and reflects growing interest in AI and CC for LSD care, but heterogeneity and methodological limitations preclude inferences on effectiveness or routine implementation. This evidence map highlights relatively stronger areas and gaps, providing a structured foundation to inform timely expert consensus-building and research prioritization. Key priorities include interoperable data infrastructures and data availability, prospective multicenter evaluations, transparent reporting of algorithms and workflows, and implementation-relevant outcomes to support safe, equitable, and scalable adoption aligned with evolving European Union and global rare-disease priorities.

Kava (Piper methysticum), a central nervous system depressant derived from a plant in the pepper family native to the Pacific Islands, is traditionally consumed in religious, cultural, political, and social ceremonies. In the United States, kava emerged in the late 1990s and has experienced renewed growth and product diversification since the 2010s, with increasing availability of concentrated extracts and ready-to-drink beverages. These commercial products are commonly marketed as healthy alternatives to alcohol, sold near college campuses, and increasingly being combined with kratom, a psychoactive botanical with opioid-like effects, raising safety concerns. Data on kava-related use during January 2000-December 2025 that resulted in a report to the National Poison Data System (i.e., kava exposure report) were analyzed to assess trends by users' demographic characteristics, exposure type, and outcomes. Kava-related exposure reports declined sharply after a 2002 Food and Drug Administration advisory on kava-associated severe liver injury but have risen steadily since 2011, reaching 203 reported exposures in 2025. Reports primarily involved adults aged ≥20 years, but demographic characteristics have changed over time. During 2000-2001, reports primarily involved females and included more children aged ≤12 years, whereas exposure reports since 2013 have predominantly involved men; reports involving children have been rare. Since 2017, reports involving combined use of kava and kratom have increased, reaching 30% (61) of all kava reports in 2025. These increases have coincided with higher rates of serious reported clinical outcomes in recent years (32% in 2025 compared with 12% in 2000). These data indicate a resurgence of overall kava exposure reports to poison centers, as well as an increase in kratom-related kava reports, which has coincided with higher rates of serious clinical outcomes. The findings in this report suggest the need for enhanced surveillance for, clinical awareness of, and public education regarding commercial products containing kava.

Is It Justified to Order an Upfront Cytomegalovirus Immunohistochemical Stain in Patients with Severe Inflammatory Bowel Disease and Ulceration?

Existing literature primarily focuses on artificial intelligence (AI) ability to detect malignant breast tumors, often neglecting or limiting analysis to Breast Imaging-Reporting and Data System (BI-RADS) categories 4 and 5. The diagnostic performance of AI for other BI-RADS categories remains understudied. The objective of this study is to compare the diagnostic accuracy of three mammographic AI services in predicting individual BI-RADS categories and definition of opportunity integration of AI into routine clinical practice. Anonymized mammograms were obtained from the Unified Radiological Information Service of Moscow. Inclusion criteria: screening mammogram, radiology report from an AI and a human radiologist, age patients 40-75 years. Exclusion criteria: mammograms without BI-RADS categories, BI-RADS categories 0 and 6. The AI performance was assessed by calculating their diagnostic performance using the radiologists' opinion as the ground truth together with the calibration tests. The study sample consisted of 81,895 mammograms. Median accuracy was 76.9%, with a positive predictive value (PPV) of 11.8%. The highest negative predictive value (NPV) was observed for BI-RADS 2 (78.5-83.4%). The second highest NPVs were observed for BI-RADS 1, 3, 4, and 5 (over 84.7%). Binary classification yielded median accuracy and PPV values of 80.5% and 98.6% respectively, compared to the calibration testing (76.0% and 84.7%). Most AI service metrics were suboptimal for individual BI-RADS prediction, potentially due to reliance on variable radiologist conclusions and lack of histological calibration. Binary classification demonstrated higher performance metrics, and no significant differences in NPV were observed across AI applications, which means they can be recommended to confirm the absence of pathology. Successful integration of AI into routine clinical practice requires consideration of various diagnostic accuracy assessment methods, tailored to specific use cases.

To evaluate the effects of nurse-led shared decision-making (SDM) on lung cancer screening outcomes, including low-dose CT (LDCT) uptake, benign findings, early cancer detection and willingness to participate among high-risk populations. Systematic review and meta-analysis. PubMed, Medline via OvidSP, Cochrane Central Register of Controlled Trials, EMBASE via OvidSP, Web of Science, Scopus, grey literature databases and clinical trial registries were searched from inception to March 2025. Studies evaluating nurse-led SDM interventions in high-risk lung cancer populations, reporting outcomes including LDCT uptake rates, screening results (Lung-RADS (Lung Imaging Reporting and Data System) classifications), early-stage cancer detection or willingness to participate. Randomised controlled trials, quasi-experimental studies and observational studies were included. Two reviewers independently extracted data and assessed risk of bias using the Risk of Bias in Non-randomised Studies of Interventions (for non-randomised studies) and Cochrane Risk of Bias 2.0 (for randomised controlled trials). Meta-analyses were conducted using random-effects models. Meta-regression explored sources of heterogeneity. 13 studies (n=13 608 participants) were included, comprising 10 single-arm studies and three comparative studies. In single-arm studies without control groups, nurse-led SDM programmes achieved a pooled LDCT uptake rate of 98% (95% CI 28% to 100%; I²=99%), and willingness to participate was 68% (95% CI 24% to 93%; I²=98%). In comparative studies, nurse-led SDM showed no significant difference in LDCT uptake compared with usual care (RR 1.00, 95% CI 0.99 to 1.02; I²=0%), suggesting non-inferiority rather than superiority. Among individuals who completed screening, 81% (95% CI 77% to 85%) had benign or low-risk findings (Lung-RADS [Lung Imaging Reporting and Data System] I/II), and 2% (95% CI 1% to 3%) were diagnosed with early-stage lung cancer, rates consistent with benchmark screening trials. Meta-regression identified female sex as positively associated with uptake (β=0.54, p<0.001), while current tobacco use was negatively associated (β=-0.37, p=0.033). The risk of bias was moderate to serious across studies. Comparative evidence suggests that nurse-led SDM achieves equivalent LDCT uptake to standard care approaches, indicating feasibility as an alternative service delivery model. However, the predominance of single-arm studies, high heterogeneity and moderate-to-serious risk of bias limit causal inference. High uptake rates in single-arm studies likely reflect selection bias rather than intervention effectiveness. Current evidence supports the feasibility but not the superiority of nurse-led SDM. Well-designed randomised controlled trials are needed to establish comparative effectiveness and cost-effectiveness before recommending widespread integration of nurse-led SDM into lung cancer screening programmes. PROSPERO CRD420251033595. https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=1033595.

192. External validation of the Ovarian-Adnexal Reporting and Data System (O-RADS) Lexicon, IOTA Simple Rules and IOTA Two-Step Strategy to classify ovarian tumours in a paediatric population

A Low-Risk Nodule With a High-Risk Secret: Tall Cell Variant Papillary Thyroid Cancer Unveiled in a 2 cm Thyroid Imaging, Reporting and Data System 3 Lesion

Automated breast ultrasound systems (ABUS) have been increasingly used for breast lesion detection; however, standardized and interpretable prediction models based on ABUS lexicon features for differentiating benign from malignant lesions remain limited. Therefore, this study aimed to establish a prediction model that differentiates malignant from benign breast lesions using an ABUS. This prospective study enrolled patients with single breast lesions identified by handheld ultrasound (HHUS) in Peking University Cancer Hospital between June 2010 and December 2012. The ABUS was performed and the ultrasonic features were described based on the fifth edition of the Breast Imaging Reporting and Data System (BI-RADS) ultrasound lexicon and related literature of lesions and a pathology examination was performed to confirm pathological status. The model performance was evaluated using the area under the curve (AUC). There were 2,090 patients with single breast lesions included and 630 cases (30.1%) were selected for pre-test. Patients were then randomly split 1:1 into a training set (n=315) and a validation set (n=315). Univariate analysis revealed that lesion form, shape, orientation, margin, boundary, and posterior acoustic feature were significantly different between the benign lesion and malignant lesion group (all P<0.05). The lesion form, boundary, and margin were further selected for multivariate model development. The equation was Y=1.604 × boundary + 1.045 × lesion form - 5.436 × margin (A) - 2.166 × margin (B), the AUC reached 0.882 [95% confidence interval (CI): 0.844-0.919] in training set and 0.866 (95% CI: 0.824-0.928) in validation set. Lesion form, boundary, and margin information might be associated with benign and malignant lesions on ABUS. The malignant lesion prediction model based on lesion form, boundary, and margin showed great diagnostic performance in both training and validation cohorts.

Global mammographic asymmetry (GA) is generally considered benign, and its association with subsequent breast cancer risk is unclear. We examined whether GA on screening mammography predicts short-term and long-term breast cancer and whether this varies by Breast Imaging Reporting and Data System (BI-RADS) breast density. In this retrospective cohort study using the Korean National Health Insurance Service screening programme, we included women aged ≥40 years who underwent screening mammography in 2009-2010 and had no prior breast cancer. GA and BI-RADS density were recorded on baseline mammograms; incident invasive breast cancer through December 31, 2019 was ascertained from insurance claims. Cox proportional hazards models estimated adjusted hazard ratios (aHRs) for breast cancer associated with GA overall and by BI-RADS density and follow-up interval (<1, 1-2 and ≥2 years), adjusting for demographic, reproductive and lifestyle factors. Among 5,475,113 women, GA was present in 4.0%. Overall, GA was associated with a modestly increased breast cancer risk (aHR 1.15; 95% CI 1.11-1.19), strongest within 1 year of screening (aHR 1.90; 95% CI 1.70-2.12). In women with BI-RADS 1 breasts, GA doubled overall risk (aHR 2.03) and quadrupled short-term risk (<1 year: aHR 4.14), whereas in BI-RADS 4 breasts GA did not increase overall risk (aHR 0.94). GA is uncommon but identifies women at substantially elevated short-term breast cancer risk, particularly those with non-dense breasts, and has limited long-term prognostic value in extremely dense breasts. These findings support consideration of short-interval follow-up or supplemental imaging when GA is reported in non-dense breasts.

For management of hepatocellular carcinoma (HCC), locoregional therapies are commonly used as curative interventions, bridging strategies before a liver transplant, or methods to reduce the tumor burden. After treatment, imaging surveillance is essential to guiding subsequent patient treatment decisions by enabling timely detection of recurrence or residual tumor viability. However, imaging interpretation can be challenging due to variability in expected findings, based on the treatment modality, response achieved, and time since therapy. The CT and MRI Liver Imaging Reporting and Data System (LI-RADS) treatment response assessment algorithm was developed to standardize imaging interpretation and reporting after locoregional therapy for primary liver malignancies. Integrating emerging evidence and feedback from clinical users, the updated 2024 treatment response assessment algorithm establishes distinct criteria for nonradiation- and radiation-based therapies and includes ancillary MRI features to improve detection of tumor viability. The authors review the updated LI-RADS treatment response assessment algorithm, with particular attention to clinical applicability and implications for patient care. © The Author(s) 2026. Published by the Radiological Society of North America under a CC BY 4.0 license. Supplemental material is available for this article.

Ultrasonography-Based Application of Node-RADS in Differentiating Malignant and Benign Superficial Lymph Nodes: A Retrospective Study.

Ovarian-Adnexal Reporting & Data System Ultrasound (O-RADS US) is a validated scoring system in adult women with adnexal lesions to help assess the risk of potential malignancy. Limited data exists for children in whom malignancy is rare. To evaluate inter-radiologist agreement and diagnostic performance when using the O-RADS US in pediatric patients with ovarian lesions. Retrospective IRB-approved study included pelvic ultrasounds (US) from 2015 to 2020 in pediatric patients (<18years). Pelvic US with ovarian lesions measuring >3cm in premenarchal patients and >5cm in menarchal patients were included. Three pediatric radiologists reviewed each US and recorded imaging characteristics and O-RADS classification. Diagnostic performance was assessed, and agreement among radiologists was calculated. In total, 160 pelvic US exams were included in 160 patients, with a mean patient age of 12.1years (SD=4.9). Most lesions were classified as O-RADS 2 (almost certainly benign), and fewer cases as O-RADS 4 (intermediate risk) or O-RADS 5 (high risk). Inter-radiologist agreement for O-RADS category was moderate (κ=0.42). Diagnostic performance of US O-RADS demonstrated high sensitivity and NPV (100% for all three reviewers). Specificities were 74-82%, and PPV was low at 5-7% for distinguishing malignant/borderline lesions from benign lesions. Application of the O-RADS US system in pediatric patients may be challenging due to the low overall malignancy rate. Nevertheless, an O-RADS 2 classification provides meaningful reassurance, reflecting minimal malignancy risk in children. Larger studies are needed to determine the clinical utility of O-RADS US and whether pediatric-specific modifications are required.

To develop a practical consensus-based framework for 10 steps to improve sepsis care in low-resource settings (LRSs), aligned with the sepsis chain of survival and informed by global expertise. We reviewed peer-reviewed literature on sepsis epidemiology, prevention, recognition, and management in LRS; international guidelines, including the Surviving Sepsis Campaign; and prior "10-step" consensus frameworks for resuscitation and emergency care. A Task Force representing adult and pediatric sepsis care, emergency care, critical care, infectious diseases, public health, and implementation science identified key domains from the above data sources. With guidance from methodologists and implementation science experts, we utilized an iterative, consensus-based process-literature review, Delphi survey, Utstein-style conference, stakeholder input, and public comment-to first define and then refine steps and implementation strategies. The process resulted in 10 nonsequential, actionable steps covering governance and commodities, provider and caregiver education, community and facility prevention, early recognition and rapid response, timely guideline-based interventions, structured post-sepsis care, data systems, quality improvement, a culture of excellence and respect, and holistic well-being of patients, caregivers, and providers. Each step includes a rationale and potential implementation strategies adaptable to local resources and needs. Collectively, the ten steps emphasize integration across the continuum of care, equitable access to essential interventions, and the role of emerging technologies to prevent, recognize, monitor, and follow-up sepsis. The 10 steps provide a consensus-driven roadmap for health leaders, clinicians, and policymakers to improve sepsis care, strengthen the sepsis chain of survival, reduce preventable morbidity and mortality, and address global inequities in sepsis outcomes.

The Liver Imaging Reporting and Data System (LI-RADS) assessment is subject to inter-reader variability. The present study aimed to evaluate the impact of an artificial intelligence (AI) system on the accuracy and inter-reader agreement of LI-RADS classification based on contrast-enhanced magnetic resonance imaging among radiologists with varying experience levels. This single-center, multi-reader, multi-case retrospective study included 120 patients with 200 focal liver lesions who underwent abdominal contrast-enhanced magnetic resonance imaging examinations between June 2023 and May 2024. Five radiologists with different experience levels independently assessed LI-RADS classification and imaging features with and without AI assistance. The reference standard was established by consensus between two expert radiologists. Accuracy was used to measure the performance of AI systems and radiologists. Kappa or intraclass correlation coefficient was utilized to estimate inter-reader agreement. The LI-RADS categories were as follows: 33.5% of LR-3 (67/200), 29.0% of LR-4 (58/200), 33.5% of LR-5 (67/200), and 4.0% of LR-M (8/200) cases. The AI system significantly improved the overall accuracy of LI-RADS classification from 69.9 to 80.1% (p < 0.001), with the most notable improvement among junior radiologists from 65.7 to 79.7% (p < 0.001). Inter-reader agreement for LI-RADS classification was significantly higher with AI assistance compared to that without (weighted Cohen's kappa, 0.655 vs. 0.812, p < 0.001). The AI system also enhanced the accuracy and inter-reader agreement for imaging features, including non-rim arterial phase hyperenhancement, non-peripheral washout, and restricted diffusion. Additionally, inter-reader agreement for lesion size measurements improved, with intraclass correlation coefficient changing from 0.857 to 0.951 (p < 0.001). The AI system significantly increases accuracy and inter-reader agreement of LI-RADS 3/4/5/M classification, particularly benefiting junior radiologists.

Although pediatric thyroid cancer is rare, it has characteristics distinct from those of adult thyroid cancer. Thyroid nodules in children present a higher risk of malignancy, more frequent lymph node and distant metastases, and distinct molecular profiles compared to adults. Despite a more aggressive initial presentation, the long-term prognosis for children is excellent, with paradoxically low mortality rates, even in patients with distant metastases. Therefore, it is questionable whether ultrasound-based risk-stratification systems primarily developed for adults can be directly applied to children. The 2021 Korean Thyroid Imaging Reporting and Data System (K-TIRADS) introduced pediatric-specific biopsy cut-offs and risk-adapted considerations, improving sensitivity, specificity, and overall accuracy. Nevertheless, challenges remain in achieving better diagnostic performance. Specific considerations must also be noted when evaluating pediatric thyroid nodules, such as the diffuse sclerosing subtype of papillary thyroid cancer and intrathyroidal ectopic thymus. Overdiagnosis and age-related heterogeneity further complicate risk assessment and management. Future guidelines could adopt stratified approaches based on patient age and sonographic mimickers, with additional integration of molecular profiling and artificial intelligence-assisted decision support. This review summarizes the current state of ultrasonographic evaluation of pediatric thyroid nodules, including the 2021 K-TIRADS, and discusses future refinements for pediatric-specific ultrasound risk-stratification systems.

Limitations of Large Language Models in Assisting PI-RADS Scoring on Prostate Biparametric MRI Text Reports.

Exosome-surface enhanced Raman spectroscopy-artificial intelligence platform (exosome-SERS-AI) is an innovative liquid biopsy method that acquires SERS signals from plasma exosomes and analyzes them using deep learning models to diagnose cancer. This study aimed to evaluate whether exosome-SERS-AI could increase the diagnostic accuracy of ultrasonography (US) for suspicious breast lesions. This prospective multicenter study enrolled 500 patients between November 2024 and December 2025. Eligible participants will be women aged ≥ 40 years who will undergo US performed by specialized breast radiologists and have suspicious breast lesions assigned to a Breast Imaging Reporting and Data System (BI-RADS) category 3-5 assessment. A 6 mL sample of whole blood was collected from each participant. After plasma separation from blood, SERS, which is highly sensitive to exosomes, was employed to measure Raman signals, and the acquired data were processed using artificial intelligence algorithms. Following blood sampling, all patients underwent US-guided core needle biopsy for breast lesions classified as BI-RADS category 4 and 5, and 12-months of follow-up US for lesions classified as BI-RADS category 3. Histopathological examination was used as the reference standard for BI-RADS 4 and 5 lesions, whereas stability on 12-month follow-up US was used as the reference standard for BI-RADS 3 lesions. The enrolled cohort is expected to have an equal distribution of benign and malignant cases. The following outcome measures were compared between US alone and the combination of exosome-SERS-AI with US: sensitivity, specificity, positive predictive value, negative predictive value, and the area under the receiver operating characteristic curve. Enrollment is expected to be completed by 2025, and the study results are expected to be presented in 2026. This prospective multicenter study will evaluate the performance of exosome-SERS-AI compared to US in women with BI-RADS categories 3-5. Participant enrollment is ongoing. ClinicalTrials.gov Identifier: NCT06672302. Registered on November 4, 2024.