👤 Alyson E Santoro

Early diagnosis of Alzheimer's disease (AD) and related dementias remains challenging because no single biomarker sufficiently captures the complex and multifactorial nature of the underlying pathology. In recent years, multimodal artificial intelligence (AI) models capable of integrating heterogeneous data sources-such as neuroimaging, fluid biomarkers, genetics, and cognitive assessments-have emerged as a promising strategy to improve early detection and risk stratification. We performed a PRISMA-guided systematic review (PROSPERO: CRD420251049848) of studies published from 2010 to 2025. We included 27 peer-reviewed studies applying AI/ML to ≥2 biomarker modalities for diagnostic classification or prognostic prediction (e.g., MCI-to-AD conversion), with an explicit emphasis on multimodal designs that incorporated at least one minimally invasive and/or widely deployable modality (e.g., cognitive tests, blood-based biomarkers, APOE/genetics, retinal imaging, or routine clinical features). Risk of bias was assessed using QUADAS-2. Across the 27 included studies, multimodal AI models generally outperformed the best unimodal baselines, particularly when combining complementary biological information (e.g., imaging with molecular or clinical features). Diagnostic tasks more often achieved high discrimination (frequently AUCs in the ~0.85-0.95 range under internal validation), whereas prognostic prediction-especially MCI-to-AD conversion-remained more challenging (typically ~0.75-0.85 AUC in the best-performing models). However, evidence for generalizability was limited, as external validation was uncommon and QUADAS-2 frequently highlighted concerns in the Index Test domain related to overfitting risk and incomplete validation. Overall, multimodal AI provides a more comprehensive representation of AD/MCI-related pathology than unimodal approaches and can improve early diagnostic classification and, to a lesser extent, prognostic prediction. However, translation to clinical practice is still constrained by limited external validation and heterogeneous reporting, which hamper generalizability and clinical trust. Future work should prioritize prospective multi-center studies, robust external validation, and transparent reporting (including interpretability analyses) to support real-world deployment. Show less

We evaluated the accuracy of standard machine learning (ML) algorithms in predicting 1-year cognitive decline in Alzheimer's disease patients with mild cognitive impairment (ADMCI) using resting-state electroencephalographic (rsEEG) biomarkers enriched with APOE genotype, sex, age, and educational attainment data. The study analyzed datasets from 63 ADMCI patients obtained from an international archive. The ML algorithms included Simple Logistic Regression, Model Trees, Logistic Regression, K-nearest neighbor, and Support Vector Machine. Input features comprised lobar rsEEG source activities across delta (<4 Hz) to alpha (≈10-12 Hz) bands, cerebrospinal fluid (CSF Aβ1-42/p-tau), and structural magnetic resonance imaging (sMRI) biomarkers. Cognitive decline was assessed over a 1-year follow-up ("stable" vs. "decliner") based on Mini-Mental State Examination (MMSE) scores. The four independent ML algorithms accurately predicted changes in the MMSE score over a 1-year follow-up, with accuracies of 77-78% in ADMCI participants aged ≥ 70 years and 74-77% in those aged < 70 years. These findings suggest that rsEEG biomarkers in ADMCI patients may not only reveal underlying pathophysiological mechanisms affecting cortical arousal and vigilance but also hold predictive value for cognitive outcomes. Show less

Hypertrophic cardiomyopathy (HCM), the most common inherited cardiomyopathy, represents a paradigmatic condition for precision cardiovascular medicine. Once regarded as a monogenic autosomal dominant disorder driven by rare sarcomeric variants, HCM is now recognized as a genetically complex disease characterized by incomplete penetrance, variable expressivity, and heterogeneous clinical trajectories. This review summarizes current evidence on the evolving genetic architecture of HCM, emphasizing the predominant role of definitively validated sarcomeric genes, particularly Show less

Rare variants in melanocortin 4 receptor gene (MC4R) result in a severe form of early-onset obesity; however, it is unclear how these variants may affect abdominal fat distribution, intrahepatic fat accumulation, and related metabolic sequelae. Eight hundred seventy-seven youth (6-21 years) with overweight/obesity, recruited from the Yale Pediatric Obesity Clinic in New Haven, CT, underwent genetic analysis to screen for functionally damaging, rare variants (MAF < 0.01) in MC4R. Participants were assigned to a Pathogenic Variant or No Pathogenic Variant group and completed a 10-timepoint 180-min oral glucose tolerance test (OGTT) and abdominal MRI. Compared to the No Pathogenic Variant group, the Pathogenic Variant group demonstrated significantly greater glucose concentrations (AUC Pathogenic variants in MC4R are associated with increased VAT, HFF%, and insulin resistance, independent from the degree of obesity in youth. Show less

Molecular analysis of Each participating institution was requested to apply its own diagnostic testing strategy on 8 sections obtained from artificial reference specimens built to harbor Overall, cell resuspension yielded higher amounts of DNA and RNA (SNU16 61.5 ng/µl, 38100.0 pg/µl; RT112 118.0/µl, 2140.0 pg/µl, respectively) in comparison with SNU16+ RT112 mixing cell block (0.7 ng/µl DNA and 412.0 pg/µl RNA). Moreover, FFPE samples showed a higher fragmentation index (DIN 1.2 and RIN not calculated) compared with cell line resuspension (DIN 2.2 and 9.5 for SNU16 and RT112; RIN 3.9 and 6.8 for SNU16 and RT112). All participating institutions identified NGS represents the most suitable approach in molecular profiling of Show less

Sarcopenic obesity (SO) is characterized by the presence of both obesity and sarcopenia and is related to disability and loss of independence in older adults. The extent to which time spent in light physical activity (LPA), or moderate-to-vigorous physical activity (MVPA) is associated with SO risk in older adults remains unclear. The aim of this study was (a) to examine the association between the level of adherence to recommended amounts of MVPA and the risk of SO in older adults and (b) to determine whether time spent in LPA is associated with SO risk independently of time spent in MVPA. This cross-sectional study involved 862 community-dwelling older adults (58% women; aged 65-79 years) from four European countries. Accelerometer-determined time in MVPA was categorized as follows: inactive (< 75 min/week), moderately active (75-149 min/week), active (150-299 min/week) and highly active (≥ 300 min/week). Time in LPA was expressed in tertiles. The outcome measure SO risk was determined based on appendicular lean mass, waist circumference, handgrip strength and the 5-times sit-to-stand test. Odds ratios (OR) with a 95% confidence interval (95% CI) of high SO risk across levels of MVPA and LPA were determined by binary logistic regression adjusted for the level of systemic inflammation (high-sensitivity C-reactive protein) and dietary protein intake. Compared to the inactive group, ORs of having a high SO risk were about 50%-80% lower, depending on the MVPA level, with the largest risk reduction in the highly active group (OR: 0.23, 95% CI: 0.13-0.39; p < 0.05). The likelihood of having a high SO risk was significantly lower among the highly active group compared to the active group (OR: 0.50; 95% CI: 0.33-0.77; p < 0.05). More time in LPA was associated with a significantly lower likelihood of having high SO risk (highest vs. lowest tertile: OR: 0.52, 95% CI: 0.30-0.89; p < 0.05) only in participants with low amounts of MVPA. In contrast, LPA was not associated with SO risk among participants meeting the MVPA recommendation. MVPA is strongly associated with a lower likelihood of having a high SO risk in older adults, independently of the level of systemic inflammation and intakes of dietary proteins. LPA is related to SO risk in sedentary older adults, which supports the promotion of physical activity regardless of intensity for mitigating SO. Show less

The diagnosis of lymphoplasmacytic lymphoma (LPL) in the bone marrow (BM) is challenged by aberrant phenotypes and by overlapping histological features with marginal zone lymphoma (MZL). To address these issues, we (i) assessed LPL immunophenotype on a large series of BM samples, (ii) drew possible correlations between LPL phenotype and clinical/molecular data and (iii) investigated the role of new phenotypical markers in the differential diagnosis between LPL and MZL. The study retrospectively considered 81 clinically annotated LPL diagnosed at Padua University Hospital (Padua, Italy) during a 5-year period. BM findings were correlated with clinical laboratory findings and with MYD88 and CXCR4 mutational status. The obtained results were compared with a series of 77 MZL in the BM, including 46 splenic MZL (SMZL), 14 nodal MZL (NMZL) and 17 extra-nodal MZL (EMZL). The LPL cohort included 52 males and 29 females (median age at diagnosis = 71 years). Aberrant CD10 and CD5 positivity was documented in 3 of 81 (3.7%) and 13 of 81 (16.1%) cases, respectively. CD23 positivity occurred in 56 of 81 (69.1%) cases, being usually partial/focal. CD23 expression did not correlate with any specific clinical-pathological parameter. Comparison with SMZL, NMZL and EMZL highlighted less frequent splenomegaly, higher serum paraprotein, higher CD23 expression and fewer follicular dendritic cell networks in LPL. A combined clinical-pathological score supported the differential diagnosis between LPL and MZL of any type. The highest diagnostic yield was obtained for the differential diagnosis between LPL and SMZL. Partial positivity for CD23 is a common feature of LPL in the BM. Together with other clinical and histological parameters, CD23 expression supports the differential diagnosis between LPL and MZL. Show less

The antisense oligonucleotide against APOC3 mRNA volanesorsen was recently introduced to treat Familial Chylomicronemia Syndrome (FCS). Cases of decreased platelet count are reported among patients treated with volanesorsen. The aim of the study was to evaluate platelet function and thrombin generation (TG) assessment in FCS patients receiving volanesorsen. We performed a cross-sectional study on FCS patients treated with volanesorsen. Changes in platelet count PLC were assessed from baseline to Tw12 and Tw36. To assess TG, samples were processed by CAT (with PPP-reagent LOW). The results were expressed by the thrombogram graphic (thrombin variation over time); LagTime; endogenous thrombin potential (ETP); peak; time to reach peak (ttpeak), StartTail and Velocity Index. Platelet aggregation was assessed by testing different agonists using the turbidimetry method. Four FCS patients and four matched healthy controls were included in the present study. Changes in PLC were 30% at Tw12 and 34% at Tw36. Thrombin generation results showed values in the normal range (for patients and controls, respectively, LagTime:10.42 ± 4.40 and 9.25 ± 0.99; ttPeak:14.33 ± 4.01 and 13.10 ± 0.67; StartTail: 32.13 ± 3.54 and 29.46 ± 1.69; Velocity Index: 20.21 ± 3.63 and 33.05 ± 13.21; ETP: 599.80 ± 73.47 and 900.2 ± 210.99; peak value: 76.84 ± 1.07 and 123.30 ± 39.45) and no significant difference between cases and controls. Platelet aggregation test showed values in range, with no significant difference compared to healthy controls. Our study showed for the first time that no significant changes in general hemostasis assessed by TG and in platelet function were observed in FCS patients receiving volanesorsen. Show less

Genetic alterations activating the MAPK pathway are common in non-small cell lung cancer (NSCLC). Patients with NSCLC may benefit from treatment with the pan-RAF inhibitor naporafenib (LXH254) plus the ERK1/2 inhibitor rineterkib (LTT462) or MEK1/2 inhibitor trametinib. This first-in-human phase 1b dose-escalation/dose-expansion study investigated the combinations of naporafenib (50-350 mg once daily [QD] or 300-600 mg twice daily [BID]) with rineterkib (100-300 mg QD) in patients with KRAS-/BRAF-mutant NSCLC and naporafenib (200 mg BID or 400 mg BID) with trametinib (0.5 mg QD, 1 mg QD or 1 mg QD 2 weeks on/2 weeks off) in patients with KRAS-/BRAF-mutant NSCLC and NRAS-mutant melanoma. The primary objectives were to identify the recommended dose for expansion (RDE) and evaluate tolerability and safety. Secondary objectives included antitumor activity and pharmacodynamics. Overall, 216 patients were treated with naporafenib plus rineterkib (NSCLC: n = 101) or naporafenib plus trametinib (NSCLC: n = 79; melanoma: n = 36). In total, 10 of 62 (16%) patients experienced at least one dose-limiting toxicity. The RDEs were established as naporafenib 400 mg BID plus rineterkib 200 mg QD, naporafenib 200 mg BID plus trametinib 1 mg QD and naporafenib 400 mg BID plus trametinib 0.5 mg QD. The most frequent grade ≥ 3 treatment-related adverse event was increased lipase (8/101 [7.9%] patients) for naporafenib plus rineterkib and rash (22/115 [19.1%] patients) for naporafenib plus trametinib. Among patients with NSCLC, partial response was observed in three patients (one with KRAS-mutant, two with BRAF Both naporafenib combinations had acceptable safety profiles. Antitumor activity was limited in patients with NSCLC, despite the observed on-target pharmacodynamic effect. gov identifier: NCT02974725. Show less

Available treatments for colorectal cancer are limited. However, in the last few years several advances and new treatment options became available and expanded the continuum of care in metastatic colorectal cancer (mCRC). Fruquintinib, a tyrosine kinase inhibitor, has been shown to be effective in heavily pretreated mCRC progressing to trifluridine-tipiracil (FTD/TPI) or regorafenib or both. Preclinical studies have shown that fruquintinib inhibits with high selectivity VEGFR 1-2-3, leading to a blockade in angiogenesis process, but also acts, with weak inhibition, on RET, FGFR-1, and c-kit kinases. Fruquintinib demonstrated good efficacy and tolerance in chemorefractory mCRC in two phase III trial: FRESCO and FRESCO 2. These results led to FDA approval of fruquintinib for pretreated mCRC patients who received prior fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy. Fruquintinib is a valid therapeutic option for heavily pretreated mCRC patients. However, an optimal sequence of treatments is yet to be defined. In this review, we propose an algorithm for later lines of treatment to integrate fruquintinib as a standard of care together with the new therapeutic combinations that recently showed clinical benefit for chemorefractory mCRC, in both molecularly selected (e.g. KRAS Show less

Atherosclerosis, the major cause of myocardial infarction and stroke, results from converging inflammatory, metabolic, and biomechanical factors. Arterial lesions form at sites of low and disturbed blood flow but are suppressed by high laminar shear stress (LSS) mainly via transcriptional induction of the anti-inflammatory transcription factor, Kruppel-like factor 2 (Klf2). We therefore performed a whole genome CRISPR-Cas9 screen to identify genes required for LSS induction of Klf2. Subsequent mechanistic investigation revealed that LSS induces Klf2 via activation of both a MEKK2/3-MEK5-ERK5 kinase module and mitochondrial metabolism. Mitochondrial calcium and ROS signaling regulate assembly of a mitophagy- and p62-dependent scaffolding complex that amplifies MEKK-MEK5-ERK5 signaling. Blocking the mitochondrial pathway in vivo reduces expression of KLF2-dependent genes such as eNOS and inhibits vascular remodeling. Failure to activate the mitochondrial pathway limits Klf2 expression in regions of disturbed flow. This work thus defines a connection between metabolism and vascular inflammation that provides a new framework for understanding and developing treatments for vascular disease. Show less

Carfilzomib is a last generation proteasome inhibitor (PI) with proven clinical efficacy in the treatment of relapsed/refractory multiple myeloma. This drug is considered to be extremely specific in inhibiting the chymotrypsin-like activity of the 20S proteasome, encoded by the β5 subunit, overcoming some bortezomib limitations, the first PI approved for multiple myeloma therapy which is however burdened by a significant toxicity profile, due also to its off-target effects. Here, molecular approaches coupled with molecular docking studies have been used to unveil that the Insulin-Degrading Enzyme, a ubiquitous and highly conserved Zn Show less

Dyslipidemia is highly prevalent in youth with type 2 diabetes (T2D), yet the pathogenic components of dyslipidemia in youth with T2D are poorly understood. To evaluate the genetic determinants of lipid traits in youth with T2D through a genome-wide association study. We genotyped 206 928 variants and imputed 17 642 824 variants in 1076 youth (mean age 15.0 ± 2.48 years) with T2D from the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) and SEARCH for Diabetes in Youth (SEARCH) studies as part of the Progress in Diabetes Genetics in Youth (ProDiGY) consortium. We performed association testing for triglyceride and low-density lipoprotein cholesterol and high-density lipoprotein cholesterol (HDL-c) concentrations adjusted for the genetic relationship matrix within each substudy followed by meta-analyses for each trait. We identified a novel association between a deletion on chromosome 3 (3:67817380_AT/A_Deletion:RP11-81N13.1) and triglyceride levels at genome-wide level of significance ( Our genetic analyses of lipid traits in youth with T2D have identified 1 novel and 1 previously known locus. Additional studies are needed to further characterize the genetic architecture of dyslipidemia in youth with T2D. Show less

Photosynthetic picoeukaryotes contribute a significant fraction of primary production in the upper ocean. Micromonas pusilla is an ecologically relevant photosynthetic picoeukaryote, abundantly and widely distributed in marine waters. Grazing by protists may control the abundance of picoeukaryotes such as M. pusilla, but the diversity of the responsible grazers is poorly understood. To identify protists consuming photosynthetic picoeukaryotes in a productive North Pacific Ocean region, we amended seawater with living Show less

Basolateral polymerization of cellular fibronectin (FN) into a meshwork drives endothelial cell (EC) polarity and vascular remodelling. However, mechanisms coordinating α5β1 integrin-mediated extracellular FN endocytosis and exocytosis of newly synthesized FN remain elusive. Here we show that, on Rab21-elicited internalization, FN-bound/active α5β1 is recycled to the EC surface. We identify a pathway, comprising the regulators of post-Golgi carrier formation PI4KB and AP-1A, the small GTPase Rab11B, the surface tyrosine phosphatase receptor PTPRF and its adaptor PPFIA1, which we propose acts as a funnel combining FN secretion and recycling of active α5β1 integrin from the trans-Golgi network (TGN) to the EC surface, thus allowing FN fibrillogenesis. In this framework, PPFIA1 interacts with active α5β1 integrin and localizes close to EC adhesions where post-Golgi carriers are targeted. We show that PPFIA1 is required for FN polymerization-dependent vascular morphogenesis, both in vitro and in the developing zebrafish embryo. Show less

Recently, the single nucleotide polymorphism (SNP) identified as rs1260326, in the glucokinase regulatory protein (GCKR), was associated with hypertriglyceridemia in adults. Because accumulation of triglycerides in hepatocytes represents the hallmark of steatosis, we aimed to investigate whether this variant might be associated with fatty liver (hepatic fat content, HFF%). Moreover, because recently rs738409 in the PNPLA3 and rs2854116 in the APOC3 were associated with fatty liver, we explored how the GCKR SNP and these two variants jointly influence hepatosteatosis. We studied 455 obese children and adolescents (181 Caucasians, 139 African Americans, and 135 Hispanics). All underwent an oral glucose tolerance test and fasting lipoprotein subclasses measurement by proton nuclear magnetic resonance. A subset of 142 children underwent a fast gradient magnetic resonance imaging to measure the HFF%. The rs1260326 was associated with elevated triglycerides (Caucasians P = 0.00014; African Americans P = 0.00417), large very low-density lipoprotein (VLDL) (Caucasians P = 0.001; African Americans, P = 0.03), and with fatty liver (Caucasians P = 0.034; African Americans P = 0.00002; and Hispanics P = 0.016). The PNPLA3, but not the APOC3 rs2854116 SNP, was associated with fatty liver but not with triglyceride levels. There was a joint effect between the PNPLA3 and GCKR SNPs, explaining 32% of HFF% variance in Caucasians (P = 0.00161), 39.0% in African Americans (P = 0.00000496), and 15% in Hispanics (P = 0.00342). The rs1260326 in GCKR is associated with hepatic fat accumulation along with large VLDL and triglyceride levels. GCKR and PNPLA3 act together to convey susceptibility to fatty liver in obese youths. Show less