Surgery and postoperative complications disrupt circadian rhythms. While clinicians recognize the implicit value of circadian rhythms to evaluate postoperative recovery, the absence of a practical means to measure them has limited their exploration and integration into clinical practice in children. Consumer wearable device (Fitbit) was given to children 3-18 years old who underwent laparoscopic appendectomy for complicated appendicitis and data were collected during postoperative days (POD) 1 to 21. Three novel rest-activity rhythmicity (RAR) metrics were computed from minute-by-minute continuous Fitbit data: (1) Periodicity and (2) Amplitude of the 24-hour Activity Rhythm; and (3) Circadian quotient of the 24-hour heart rate (HR) Rhythm. RAR trajectories were aggregated for patients without postoperative complications and compared against trajectories for patients with complications. Ninety-four patients were included in the analysis (n=13 [14%] with complications). For patients without complications, three RAR metrics gradually increased in intensity until reaching plateaus between POD 11-20 (Periodicity of the 24-hour Activity Rhythm, 95% confidence interval [CI]: POD 9-14; Amplitude of the 24-hour Activity Rhythm, 95% CI: POD 16-24; Circadian quotient of the 24-hour HR Rhythm, 95% CI: POD 9-15). Although the RAR metrics for children with complications increased in intensity, none plateaued during the 21-day study period. Wearable-derived circadian rhythms demonstrated distinct patterns in RAR metrics for children with and without complications during the 21-day study period following appendectomy. This suggests that the RAR may be relevant digital biomarkers to track postoperative recovery and identify complications in pediatric populations.
Alzheimer's disease (AD) is a neurodegenerative disease that typically begins with mild cognitive impairment (MCI) and gradually worsens to mild, moderate, and severe dementia. There is increasing evidence that inflammatory responses in both peripheral and central compartments contribute to the pathophysiology of AD. However, meta-analyses that compare the changes in inflammatory biomarkers according to AD stages remain limited. This study aims to systematically evaluate the differences in inflammatory marker levels according to AD progression stages and to assess their potential utility as stage-specific biological indicators. This meta-analysis will be conducted according to the PRISMA-P guidelines. PubMed, Embase, Cochrane Library, and MEDLINE will be searched for studies involving elderly people aged 60 years or older with MCI or AD diagnosed through June 30, 2025. According to the PECO framework, the primary comparison will be each AD-related clinical stage, including MCI, mild AD, moderate AD, and severe AD, versus cognitively normal older adults. Secondary analyses will examine between-stage comparisons when sufficient data are available. The primary outcomes will be inflammatory biomarker levels measured in blood-derived samples, including serum and plasma. Cerebrospinal fluid inflammatory biomarkers will be included as predefined secondary outcomes and analyzed separately. Biomarkers of interest include IL-6, CRP, TNF-α, IL-1β, YKL-40, complement C3, C5a, IL-2, IFN-γ, GFAP, sTREM2, ferritin, and related inflammatory markers. Literature selection and data extraction will be performed independently by two reviewers. Statistical analysis will be performed using JASP to calculate the standardized mean differences (SMD) and 95% CI. A heterogeneity test, meta-regression, subgroup analysis, sensitivity analysis, and publication bias test will also be performed. This study will systematically evaluate stage-specific changes in inflammatory biomarkers and identify potential response patterns and biological relevance of each marker in relation to AD progression. This review will examine whether some biomarkers follow nonlinear trajectories rather than linear changes, which may help explore their association with pathological mechanisms such as microglial activation and peripheral-central immune axis changes. This meta-analysis may provide foundational data for the development of inflammatory fluid biomarkers for early diagnosis and stage-specific prognosis in AD. As this protocol is not eligible for registration in PROSPERO, a preliminary search has been performed to confirm that there are no duplicate systematic reviews.
Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and functional impairment, posing significant challenges for early detection and management. In recent years, digital health technologies have emerged as promising tools to enhance the diagnosis, monitoring, and treatment of AD. This review paper explores the multifaceted role of digital health technologies in the early detection and management of Alzheimer's disease. We examine the use of wearable devices that monitor cognitive function and daily activities, as well as mobile health applications designed for cognitive training and symptom tracking. Additionally, we analyze the impact of telehealth services in providing remote care, particularly for underserved populations. The integration of artificial intelligence and machine learning for analyzing behavioral and cognitive data to support early diagnosis and risk assessment is also discussed. Furthermore, we explore the concept of digital biomarkers and their potential to complement traditional diagnostic methods. Ethical considerations surrounding privacy, data security, and informed consent are addressed to ensure responsible implementation of these technologies. Finally, we highlight gaps in current research and propose future directions for integrating digital health technologies into Alzheimer's care, emphasizing the potential for personalized interventions tailored to individual patient needs. This review underscores the transformative potential of digital health to reshape Alzheimer's disease management and improve patient outcomes.
Digital pathology and artificial intelligence (AI) are transforming cancer diagnostics worldwide, yet their implementation in sub-Saharan Africa remains largely undocumented. The region faces a critical shortage of pathologists while bearing an increasing cancer burden, making AI-assisted diagnostics particularly relevant. However, practical deployment in resource-constrained environments raises unique technical, logistical, and educational challenges that differ substantially from those encountered in high-income settings. To systematically document the technical, logistical, and practical challenges encountered during the implementation of QuPath, an open-source digital pathology platform, for breast cancer immunohistochemical (IHC) biomarker assessment at a reference pathology lab in Cameroon, and to propose actionable solutions for similar resource-limited settings. We conducted a prospective implementation study at the Centre Pasteur du Cameroun, Yaoundé, involving 39 cases of invasive breast carcinoma with IHC for estrogen receptor, progesterone receptor, Ki67, and HER2. We documented all phases of the digital pathology workflow: pre-analytical slide preparation, slide digitization (performed remotely at Erasme University Hospital, Brussels), image transfer and storage, QuPath algorithm training, and automated analysis on a consumer-grade laptop (4 GB RAM, 500 GB storage). Challenges were categorized into four domains: hardware and infrastructure constraints, pre-analytical and scanning issues, software training and optimization, and human factors including the learning curve. Of 130 IHC slides, 18 (13.8%) required re-scanning due to detection failures or blurred images despite pre-scanning quality control. The absence of a local scanner necessitated international slide shipment, adding 8-12 weeks of delay and logistical complexity. Processing on a 4 GB RAM consumer laptop averaged 20 min per case (range: 5-60 min), with frequent application freezes on large tissue sections. Image files averaged 1.5 GB each at ×40 magnification, rapidly exhausting the 500 GB local storage capacity. The QuPath random tree classifier required manual annotation of representative tumor, stromal, and lymphoid regions on all 39 hematoxylin and eosin-stained cases before deployment on IHC slides, representing approximately 15-20 h of pathologist time. Exploratory concordance analysis suggested clinically meaningful agreement with expert pathologist scoring for three of the four biomarkers assessed, with detailed analytical validation reported separately. Implementing open-source digital pathology in sub-Saharan Africa is feasible but requires strategic planning around infrastructure, logistics, and training. We propose a practical framework addressing minimum hardware requirements, quality control protocols adapted to tropical environments, and a structured training program for pathologists. Our experience demonstrates that despite significant constraints, AI-assisted biomarker assessment can be successfully deployed in resource-limited settings, offering a pathway to improved diagnostic standardization where it is most needed.
Noninvasive tracking of cervical dilation could reduce discomfort and infection risk from repeated digital examinations during labor. We present an electrohysterography (EHG)-based model framed as a digital biomarker of labor progression that leverages objective physiological signals with minimal clinical context. We analyzed 72 ten-minute single-channel EHG recordings from low-risk labor cases, yielding 648 segments of 120 s. Signals were filtered into three sub-bands. Twenty-one linear and nonlinear EHG descriptors were combined with two clinical variables, maternal age and gestational age, and two EHG-derived contraction-count features, namely counts of low (LC) and high (HC) uterine contractions, to form 25 predictors. Segments were labeled as low (1-4 cm), moderate (5-6 cm), or advanced (7-10 cm) dilation. Data were split 70/30 into training (n = 454) and independent test (n = 194) sets. Feature importance was estimated using χ2, ANOVA, and Kruskal-Wallis ranking. Thirty-three classifiers were evaluated using five-fold cross-validation within the training set, with the 10 top-ranked features. Among all models, a bagged tree ensemble achieved the highest macro-averaged F1 score and was therefore selected as the baseline classifier for this study. We then used a "Genetic Algorithm Ensemble Bagged Tree (GA-EBT)" approach, in which a binary-encoded genetic algorithm optimizes the bagged tree classifier's feature combination using stratified five-fold cross-validation with 50 repetitions on the training set. The best cross-validated model was a bagged tree ensemble. Performance plateaued at 17 predictors (median macro-F1 = 0.898) under progressive inclusion. The GA-EBT identified a four-feature subset - maternal age, gestational age, LC count, and HC count - that achieved F1, recall, precision, specificity, and accuracy of 1.000 on the independent test set for classifying cervical dilation stage (low, moderate, advanced). An EHG-derived digital biomarker combining a minimal set of clinical variables and EHG-derived contraction-count features enables accurate classification of cervical dilation stages from single-channel recordings. This pilot-stage classification approach showed maximal internal and independent test performance and may support real-time, noninvasive intrapartum monitoring while potentially reducing repeated digital examinations.
Endocrine resistance occurs in nearly all patients with hormone receptor-positive/human epidermal growth factor receptor 2-negative (HR+/HER2-) early breast cancer (EBC), which can develop local and distant recurrence. The situation highlights the need to explore biomarkers for the efficacy of endocrine therapy (ET). We performed digital spatial profiling on postoperative tumor samples from 20 HR+/HER2- EBC patients receiving adjuvant tamoxifen therapy, based on a designated panel comprising 235 ET-related genes. Paired patients were stratified into resistant and sensitive groups based on ET response. A total of 111 regions in three tissue compartments defined by morphology markers [tumor (PANCK+), leucocytes (CD45+), and nonimmune stroma (CD45-/PANCK-)] were investigated for immune and transcriptomic biomarkers. A total of 27, 13, and 5 differentially expressed genes (DEGs) were identified in PANCK+, CD45+, and CD45-/PANCK- regions. In the PANCK+ regions, mRNA expression of fourteen DEGs was significantly associated with disease-free survival (DFS), among which seven DEGs were further selected to construct a model for DFS. In the model, patients with low-risk scores had a median DFS of 55.77 months, significantly longer than 21.67 months among those with high-risk scores [P=2.1e-4, hazard ratio (HR)=6.73, 95% confidence interval (95% CI) =2.20-20.60)]. The area under curve for 1-year, 3-year, and 5-year DFS was 0.98, 0.95, and 0.91, indicating its superior efficacy. ET-sensitive patients had significantly higher non-classical monocyte infiltration in the CD45+ regions (P=0.03), whereas ET-resistant patients had significantly higher plasma cell infiltration in the CD45-/PANCK- regions (P=0.01). Our study has first demonstrated the spatial transcriptomic landscapes of patients with different responses to ET, which may help stratify patients who are responsive to ET and motivate the exploration of the molecular mechanisms of endocrine resistance.
Patients with critical illness often exhibit profound biological heterogeneity, complicating the identification of effective interventions. Resolving distinct molecular profiles to enable timely treatment decisions remains challenging, as integrating biomarker-guided care into routine monitoring is often hindered by fragmented, batch-based workflows. These manual operations decouple molecular data from the acute clinical timeline and are a barrier to reliable, real-time, near to patient, multi-center implementation. To address this gap, we developed an integrated microfluidic digital immunoassay system that achieves high analytical fidelity through a fully automated, simple workflow. The system utilizes a monolithic disposable cartridge to automate bead-based analyte capture, oil-phase partitioning, and signal amplification, eliminating the manual handling and emulsion steps that typically compromise digital assay robustness. The platform enables protein measurement within 45 minutes, achieving sub-picogram limit of detection ( < 0.13 pg/mL), a dynamic range spanning three orders of magnitude, and strong analytical reproducibility. We demonstrate the clinical utility of the system by profiling a validated panel of inflammatory biomarkers in plasma from critically ill pediatric patients, using low sample volumes. Results show strong agreement with gold-standard multiplex assays (R2 = 0.925-0.979). By providing a scalable framework for high-fidelity molecular profiling, this system supports the broader goal of accessible, multi-center biomarker validation and the practical implementation of precision medicine in critical care.
Cardiovascular disease remains the leading cause of global morbidity and mortality. The original My Heart Counts smartphone application demonstrated the feasibility of large-scale, fully digital recruitment and trial conduct, but was limited by platform exclusivity and the need for human experts to create text-based behavioral interventions. The next-generation My Heart Counts smartphone application is a prospective, observational cohort study with an embedded randomized crossover trial, evaluating personalized text-based coaching prompts, available in both English and Spanish. All study and trial operations will be conducted via the My Heart Counts smartphone application, re-designed using the open-source Stanford Spezi framework to support iOS, with a planned Android release in 2027. The target enrollment is N = 15,000 adults across the United States and United Kingdom. The study establishes a comprehensive digital biobank by synthesizing passive mobile health data (steps, flights climbed, heart rate, sleep, workouts), raw sensor data (e.g., accelerometry), longitudinal clinical surveys, active tasks (6-minute walk test and 12-minute Cooper run test), electrocardiograms (ECG), and electronic health record (EHR) data integrated via HL7 FHIR protocols. The embedded trial evaluates the effect of text-based coaching prompts generated by a large language model (LLM) grounded in the Transtheoretical Model of Change on daily physical activity, as compared to generic prompts. The primary endpoint of the randomized crossover trial is change in daily step count between LLM-driven and generic text-based intervention arms, analyzed using mixed-effects models. Secondary endpoints include change in mean active minutes and calorie burn over each intervention week. Other exploratory analyses include the changes in submaximal (6-minute walk test) and maximal (Cooper 12-minute run test) cardiorespiratory fitness, changes to sensor-derived biomarkers (e.g., sleep quality, resting heart rate, and heart rate variability), and association of sensor-derived biomarkers with EHR-confirmed clinical outcomes. By utilizing autonomous, LLM-driven coaching, modular software design, and cross-platform accessibility, our smartphone application-based study will provide a scalable model for inclusive and decentralized preventive care of patients with cardiovascular disease. Recruitment commenced in March 2026 and is ongoing.
Accurate and comparable quantification of somatic mutations is essential for precision oncology, as clinical decision-making increasingly relies on the quantification of molecular biomarkers. Despite major technological advances, inter-laboratory variability and the lack of metrological traceability remain significant barriers to harmonization and confidence in mutation testing results. Reference Measurement Procedures (RMPs) represent a critical framework to address these challenges by anchoring molecular measurements to common quantitative standards. Here, we describe the development and validation of a candidate RMP for the detection and quantification of the clinically relevant NRAS p.Q61R mutation using digital PCR (dPCR). The assay was systematically optimized to maximize specificity and minimize cross-reactivity between wild-type and mutant alleles. Analytical characterization demonstrated excellent linearity across a broad range of variant allele frequencies (vAF), with a limit of detection of 0.1 %. Precision studies performed on commercially available circulating tumour DNA reference materials (RM) showed good repeatability and intermediate precision, while a full measurement uncertainty budget confirmed the robustness of the approach. Comparison with a commercial dPCR assay provided independent support for assay comparability and consistent vAF estimates across the investigated range. Preliminary inter-laboratory assessment supported transferability of the candidate RMP and comparability of the resulting measurements. Overall, this work establishes a metrologically characterized and transferable dPCR-based RMP for NRAS p.Q61R quantification. Its implementation can support the harmonization of molecular measurements, the value assignment of RM, and the alignment of routine and secondary methods, thereby strengthening the reliability of quantitative biomarker assessment in precision oncology.
Digital biomarkers offer promising avenues for enhancing the identification, prognosis, and phenotyping of Alzheimer's Disease (AD). This study evaluates the feasibility and utility of the MIND GamePack©, a digital game platform designed to measure leisure cognitive activity over time. Gameplay data were collected from 60 participants across two cohorts over 3-6 months: Cohort I (no cognitive complaints) and Cohort II (subjective complaints, mild cognitive impairment, mild dementia). Analyses focused on compliance, correlation with validated neuropsychological instruments (Montreal Cognitive Assessment [MoCA], Repeatable Battery for the Assessment of Neuropsychological Status [RBANS], and Trail Making Test [TMT]), test-retest reliability, and known groups comparison. The MIND GamePack© displayed high compliance and excellent test-retest reliability within 1 week (ICC=0.81-0.95) for most selected features. Several game features displayed moderate to strong correlations with standardized neuropsychological test performance. Features related to memory tasks across select games (Normalized Accuracy and Normalized Redundant Move Variability of Memory Match and Word Repetition Rate of Word Scramble) exhibited significant associations with RBANS Sum of Index Score, TMT Part A, and MoCA, respectively. Participants without cognitive impairment exhibited improvement in features over 3 months compared to those with cognitive impairment. Baseline game features differentiated cognitively normal [CN] from cognitively impaired [CI] participants across nearly all domains after controlling for age, sex, and education (p<.01). The MIND GamePack© offers a sensitive, engaging approach to cognitive monitoring and screening, with potential use for dense tracking in longitudinal research and interventional clinical trials.
Artificial intelligence applied to the ECG is expanding the clinical role of this widely available diagnostic tool beyond conventional waveform interpretation by enabling identification of electrophysiological patterns associated with cardiovascular structure, function and risk. Within this evolving framework, the artificial intelligence enhanced ECG is emerging as a scalable digital biomarker platform supporting screening, diagnosis and prognostic stratification across the cardiovascular disease continuum.This narrative review synthesises current evidence on the clinical applications of artificial intelligence enhanced ECG, including detection of cardiac rhythm disorders, early identification of structural heart disease, decision support in acute coronary syndromes, prediction of clinical outcomes and population level cardiovascular screening using wearable technologies. Deep learning models applied to standard and simplified electrocardiographic recordings demonstrate strong diagnostic performance in identifying atrial fibrillation, left ventricular dysfunction, hypertrophic cardiomyopathy, cardiac amyloidosis and heart failure with preserved ejection fraction. These approaches also enable estimation of long-term cardiovascular risk from apparently normal tracings and support identification of disease patterns extending beyond overt cardiac conditions.Despite these advances, clinical implementation remains limited by the need for prospective multicentre validation, improved model interpretability, standardised evaluation strategies and integration into routine clinical workflows. This review highlights current validation gaps and outlines key opportunities for translation of artificial intelligence-enhanced ECG into digital cardiology practice.
Thrombosis is a leading cause of global morbidity and mortality. Despite its importance, significant gaps persist in diagnostic tools among primary healthcare professionals. We developed the rationale and proposed validation protocol for a dual clinical prediction and triage framework: ARTI (Arterial Routine Thrombosis Risk) and VRTI (Venous Routine Thrombosis Risk), supported by an integrated AI agent designed for non-specialists. The framework integrates risk factors, symptoms, and simple biomarkers into rapid scoring algorithms, enabling AI-assisted assessment within 60 seconds. Prospective multicenter validation is required to evaluate performance, clinical utility, and feasibility.
Disease-modifying therapies for Alzheimer's disease (AD) heighten demand for scalable tools enabling primary care providers (PCPs) to detect cognitive impairment and triage patients for appropriate evaluation. The brief tablet-based Linus Health Core Cognitive Evaluation (CCE) integrates the Digital Clock and Recall (DCR) and Life and Health Questionnaire (LHQ) to generate clinical decision support (CDS) and decision-tree pathways. We conducted a retrospective specialist content-validity study (June 15-27, 2023) using a modified RAND/UCLA Appropriateness Method. Five board-certified cognitive/behavioral neurologists independently rated CDS recommendations and nine predefined pathway parts for patients aged ≥55 across 21 de-identified reports. Items scored 1-9 were summarized as pooled medians with interquartile ranges; medians ≥7 indicated appropriateness. Agreement among experts was quantified using ICC [2,k] and ICC [2,1]. All cognitive-impairment recommendations met the threshold. All seven borderline/impaired-DCR pathways were appropriate (median 7-8). Two pathways fell below threshold: cognitively unimpaired individuals with Green DCR scores (median 6) and a preliminary anti-amyloid treatment referral pathway (median 5). Agreement was moderate per patient [median ICC(2,k) = 0.61] and lower for individual diagnostic-concern recommendations [median ICC(2,k) = 0.25], reflecting specialist heterogeneity on borderline non-cognitive items and ceiling effects on high-rated items. Cognitive neurologists judged CCE-derived CDS appropriate for PCP workup and referral decisions in older adults with suspected cognitive impairment. Findings support initial content validity of assessment-linked CDS, identify refinement priorities in low-risk and emerging-therapy pathways, and motivate planned PCP appropriateness and prospective implementation studies.
Biological aging is increasingly understood as a heterogeneous, multi-system process marked by organ-level vulnerability, cross-organ coordination, and variation in resilience. Advances in plasma proteomics, metabolomics, imaging, DNA methylation, digital biomarkers, and genetic epidemiology have enabled organ-level, system-level, and cross-organ age models, but these measures are often interpreted more strongly than the evidence permits. In this Review, we synthesize evidence on biological age models derived from molecular, imaging, digital, clinical, and multi-omic data and introduce ORGAN-AGE as an interpretive framework for judging what these signals can and cannot establish. We distinguish organ-derived, organ-enriched, organ-informative, system-informative, and systemic biomarkers, because circulating molecular signals rarely prove tissue origin. We critically evaluate how age gaps are constructed, bias-corrected, validated, interpreted, and linked to mortality, frailty, dementia, cardiovascular disease, diabetes complications, multimorbidity, and functional decline. A central argument is that organ age gaps can localize apparent aging burden, whereas cross-organ coupling should be treated as a graded inference rather than evidence of direct biological propagation unless longitudinal, molecular, genetic, functional, or experimental support is available. Resilience should likewise be operationalized through recovery, adaptation, and trajectory change rather than inferred from static biomarkers alone. Finally, we outline a staged translational roadmap in which organ aging models may support risk interpretation, trial enrichment, target prioritization, and future digital-twin research only after calibration, transportability, incremental utility, and clinical end-use have been demonstrated.
Anal squamous cell carcinoma (ASCC) is a rare gastrointestinal cancer linked to high-risk human papillomavirus (HPV) infection in approximately 90% of cases. Current circulating tumor DNA (ctDNA) approaches in ASCC primarily rely on pathogenic HPV-based biomarkers; however, these methods do not detect all HPV strains and are not applicable to HPV-negative tumors. To address this limitation, we developed a ddPCR assay targeting hypermethylated genomic CpG biomarkers, allowing ctDNA detection independent of tumor HPV status. An anal cancer methylation-specific multiplex droplet digital PCR (AnMM-ddPCR) assay was developed to target five previously described CpG biomarkers hypermethylated in ASCC: ASCL1, LHX8, WDR17, ZIC1, and ZNF582, and the ALB reference gene. Patient samples and samples from non-cancer controls were analyzed using the BioRad QX600 ddPCR multiplexing platform. CpG-methylated biomarker levels were significantly higher in ASCC tissue compared with normal tissue and whole blood. The AnMM-ddPCR assay successfully detected all five ctDNA markers in plasma from ASCC patients, achieving an AUC of 0.72 (95% CI: 0.55-0.89; P = 0.018) in baseline plasma samples from ASCC patients with T1 or T2 tumors ≤4 cm, versus 0.90 (95% CI: 0.76-1.00; P = 0.0005) in plasma from patients with T2 tumors >4 cm or T3 tumors. At a specificity of 91.30%, sensitivity increased with stage from 52.94% (95% CI: 30.96-73.83) to 88.89% (95% CI: 56.50-99.43). The AnMM-ddPCR assay enables HPV-independent ctDNA detection in plasma samples from anal cancer patients and supports its evaluation in future liquid biopsy applications.
Post-transplant malignancies are a major cause of long-term morbidity and mortality in kidney transplant recipients. Reliable non-invasive biomarkers may improve post-transplant oncological surveillance. Circulating microRNAs have emerged as promising candidate biomarkers in several malignancies, but their role in kidney transplant recipients remains poorly characterized. To compare circulating serum microRNA concentrations between kidney transplant recipients with and without post-transplant malignancies and to identify candidate microRNAs associated with the presence of malignancy. This exploratory pilot study included 40 kidney transplant recipients (27 with post-transplant malignancies and 13 cancer-free controls). Serum concentrations of five preselected microRNAs (miR-21, miR-146a, miR-196a, miR-203a, and miR-221) were quantified using droplet digital PCR. Group comparisons were performed using the Mann-Whitney U test, and receiver operating characteristic (ROC) curve analysis was used to evaluate discriminatory performance. Serum concentrations of miR-21 and miR-196a were significantly higher in recipients with post-transplant malignancies than in cancer-free controls (both p = 0.003). Serum miR-221 concentrations were also significantly elevated (p = 0.020), whereas no significant differences were observed for miR-146a or miR-203a. ROC analysis demonstrated acceptable discriminatory performance for miR-21 (AUC 0.742, 95%CI 0.601-0.883) and miR-196a (AUC 0.771, 95%CI 0.619-0.922). Circulating serum miR-21 and miR-196a were significantly associated with the presence of post-transplant malignancy in kidney transplant recipients, while miR-221 showed an additional exploratory association. These findings support further investigation of circulating microRNAs as candidate non-invasive biomarkers in transplant oncology. Larger prospective studies are required to validate these observations.
Oral implantology has experienced substantial digital transformation over the past two decades. The adoption of cone-beam computed tomography (CBCT), computer-assisted surgery, and artificial intelligence (AI) has improved diagnostic accuracy, treatment planning, and procedural precision. However, clinical decision-making remains largely dependent on conventional radiographic interpretation and clinician experience. We propose the concept of Implanomics (Implant-Omics) as an integrative framework combining radiomics, AI, and digital workflows. Radiomics, which converts medical images into quantitative data, offers new opportunities to extract imaging biomarkers from routine CBCT examinations. By characterizing bone morphology, trabecular architecture, and image texture beyond visual assessment, radiomics may support more objective evaluation of implant sites and improve prediction of treatment outcomes. When combined with machine-learning algorithms, these imaging-derived features can be incorporated into predictive models for risk assessment, implant planning, and longitudinal monitoring. In this Perspective, we discuss the evolution of implant dentistry from experience-based planning toward imaging-driven decision support and examine the emerging role of radiomics and AI in precision implant care. We highlight current applications, key challenges, and future directions for integrating quantitative imaging analysis into clinical workflows. Although substantial validation and implementation challenges remain, the convergence of radiomics, artificial intelligence, and digital implant technologies may contribute to more individualized and evidence-informed treatment strategies in implant dentistry.
Nutritional deterioration (ND) during radiotherapy (RT) for esophageal cancer is common, yet the optimal timing for proactive supportive care remains unclear. We characterized time-resolved nutritional trajectories, examined baseline factors associated with deterioration, and translated the findings into a patient-facing digital prototype. In this prospective longitudinal cohort, 123 patients underwent nutritional assessment at five timepoints (admission, RT1, RT14, RT20, and discharge) using the Patient-Generated Subjective Global Assessment (PG-SGA), serum albumin, and total protein. ND was defined as an upward shift in PG-SGA risk category (0-3/4-8/≥9) from admission to discharge. Sleep was assessed by baseline interview, the Pittsburgh Sleep Quality Index (PSQI), and exploratory wearable measures. Logistic regression was used to identify baseline factors associated with ND. Complementary continuous-change analysis and linear mixed-effects modeling were additionally performed to characterize deterioration magnitude and longitudinal trajectories. Exploratory correlations assessed sleep-nutrition associations. Based on the findings, a digital prototype ("Esophageal Care Companion") was developed with risk-aware onboarding and phase-specific monitoring prompts. PG-SGA increased from 4.59 ± 3.66 at admission to 6.95 ± 3.48 at RT20 (p < 0.001), with concordant declines in albumin and total protein. ND occurred in 40.7% (50/123). In the primary categorical model, baseline sleep problems were associated with higher odds of ND [adjusted odds ratio (aOR) 2.48; p = 0.034], as was N3 nodal stage (aOR 3.56; p = 0.022). In complementary analyses, N3 remained associated with greater PG-SGA worsening and a distinct longitudinal nutritional trajectory, whereas the sleep signal was less consistent across continuous and longitudinal models. Sleep measures showed exploratory phase-specific associations with nutritional biomarkers, including RT14 rapid eye movement sleep duration with albumin (r = 0.258; p = 0.004) and discharge PSQI with PG-SGA (r = 0.276; p = 0.002). Nutritional risk worsened during esophageal cancer RT, particularly in the mid-to-late treatment phase, supporting RT14-RT20 as a clinically relevant window for intensified assessment and supportive-care escalation. Advanced nodal stage emerged as a robust marker of less favorable nutritional trajectory. Baseline sleep problems may serve as a low-burden observational risk signal but should be interpreted cautiously. The prototype demonstrates the feasibility of translating phase-aware risk monitoring into digital supportive care.
Neuropsychiatric disorders are characterized by substantial biological heterogeneity, with patients sharing the same diagnosis often exhibiting distinct symptom profiles, treatment responses and longitudinal course of the disease. This heterogeneity limits the clinical utility of traditional symptom-based classifications, currently available biomarkers, and traditional group-level neuroimaging analyses, creating a major challenge for precision medicine. In this review, we provide a conceptual overview of recent paradigms for analyzing disease heterogeneity and integrate them into a coherent, comprehensive conceptual framework for the ultimate goal of disease progression modeling. We identify three gradual shifts of research paradigms in neuroimaging-based studies: (1) moving from group-level case-control analyses to normative modeling of individual variability; (2) transitioning from traditional subtype-oriented clustering to continuous dimensional generative modeling; (3) shifting from disease course analyses to virtual transition modeling according to digital twin brain models. These paradigm shifts are not isolated but can be integrated into a layered and interrelated logical framework, aiming to quantify individual deviations, characterize heterogeneous pathological dimensions and simulate disease progression trajectories over time. We further discuss how these paradigms can facilitate the development of biologically informed biomarkers, the formulation of personalized treatment plans, and the implementation of trajectory-based intervention measures. Finally, we critically examine the key challenges that must be addressed before clinical translation, including mechanistic interpretability, longitudinal validation, multimodal and multisite data integration, model reproducibility and prospective cohort validation. Our goal is to promote the development of novel approaches for disease progression modeling, thereby accelerating the translation of experimental research into clinical practice.
The diagnosis and monitoring of Alzheimer disease (AD) currently rely on clinician-administered, in-person, and cross-sectional pen-and-paper cognitive assessments. While clinically validated, these measures are time-intensive, infrequently administered, and limited in their ability to detect early, subtle, or short-term cognitive changes. Thus, more frequent, ecologically valid assessments are critical to improving sensitivity to early cognitive impairment and disease progression. This study aims to develop and pilot a smartphone-based assessment battery that combines active cognitive assessments with passive smartphone sensor data (eg, steps, sleep) and survey data to identify and longitudinally characterize cognitive impairment associated with AD. We developed a suite of digitized versions of standard cognitive tests alongside novel, game-based cognitive tests within the mindLAMP platform. Uniquely, these tests integrate into the platform's mobile survey and digital phenotyping capabilities to produce a comprehensive assessment tool capable of simultaneously tracking self-reported, behavioral, and cognitive symptoms in real time. These tools were unified within the Smartphone Monitoring Assessment in Real Time-Alzheimer's framework. Across a 6-month pilot study involving individuals with mild cognitive impairment or mild AD, we will examine the feasibility, acceptability, and longitudinal adherence to these assessments. We will compare digital cognitive and passive data streams against standard clinical assessments to evaluate their usefulness in detecting cognitive impairment and change over time. Recruitment began in April of 2025. As of February 2026, 13 participants with mild cognitive impairment or AD (mean age 72.8, SD 6.5 y, 8 male) and 12 controls (mean age 71.6, SD 7.8 y, 6 male) have been enrolled; recruitment is ongoing. Preliminary analyses on participant compliance, passive data, and variations in game scores are in progress. Data analysis is expected to be completed by mid-2026, and we anticipate results to be published in 2027. This study is funded by the a2 Pilot Awards, a subaward of funding given to the Trustees of the University of Pennsylvania under the a2 Collective, beginning in April 2024. Smartphone-based cognitive assessments, when combined with digital phenotyping, offer a scalable and ecologically valid approach to detecting and monitoring AD in real-world settings. This framework has the potential to enhance early detection, enable continuous monitoring, and support future machine learning-based automated identification of cognitive impairment, ultimately facilitating earlier and more personalized care.