Citations

Abstract Purpose: Mutant IDH1 (mIDH1) defines a therapeutically targetable subtype of intrahepatic cholangiocarcinoma (ICC), with the mIDH1 inhibitor ivosidenib approved for advanced disease. A subset of patients experiences prolonged disease stabilization; however, the molecular basis for eventual progression remains poorly defined. Experimental Design: We performed molecular profiling of matched baseline and postprogression circulating tumor DNA (ctDNA) samples from patients with mIDH1 ICC enrolled in the ClarIDHy phase III trial. Functional studies were conducted to characterize candidate resistance mechanisms. Results: Longitudinal ctDNA analysis of 18 patients treated with ivosidenib for >6 months revealed emergent genomic alterations in multiple cases. Acquired mutations in mitogen-activated protein kinase (MAPK) pathway genes (KRAS, NRAS, MAP2K1, NF1) were identified in five cases, with instances of concurrent alterations and/or high variant allele fractions (VAF). Additional candidate resistance events included a secondary IDH1 mutation and a hotspot IDH2 mutation, detected at low VAF in the same patient. Functional studies showed that these IDH mutations conferred sustained 2-hydroxyglutarate production and ivosidenib resistance, whereas MAPK activation blunted gene expression induced by ivosidenib plus IFNγ, a key therapeutic output of mIDH1 inhibition. In parallel, baseline ctDNA profiling of 81 patients revealed that ARID1A mutations and elevated mIDH1 VAF were associated with reduced clinical benefit. Conclusions: MAPK pathway alterations represent a recurrent mechanism of resistance to mIDH1 inhibition in ICC, whereas emergent IDH1/IDH2 mutations appear infrequent. Functional data suggest that MAPK-mediated resistance may involve impaired IFN signaling. These results support MAPK-directed combination strategies and highlight the utility of ctDNA profiling to identify predictive and resistance biomarkers in mIDH1-driven ICC.

BACKGROUND AND OBJECTIVES: Anterior lumbar interbody fusion (ALIF) is a powerful procedure for sagittal correction and indirect decompression in the lumbar spine, though aggressive correction can cause stretch neuropraxia and implant subsidence. The aim of this study was to identify radiographic and clinical predictors of segmental correction, as well as to evaluate the degree to which segmental correction on supine intraoperative radiographs translates to postoperative upright alignment changes. METHODS: We identified patients treated with 1- or 2-level ALIF at a tertiary care center and extracted data on demographics, spinopelvic parameters, and surgical details. Patients treated with isolated ALIF and ALIF with supplementary posterior instrumentation were compared on pre- and postoperative intradiscal lordosis (IDL) and change in IDL. Multivariable linear regression was performed to identify independent predictors of postoperative IDL. RESULTS: We included 204 patients (median age 57 years, 55.4% male) treated at 258 levels, of whom 114 (55.9%) underwent supplementary posterior fixation and 150 (73.5%) underwent single-level implantation. Across all patients, postoperative IDL was significantly greater than preoperative segmental lordosis (25.4° vs 15.3°; P < .001). Standalone ALIF achieved less ΔIDL than circumferential fusion for L5/S1 procedures (7.4° vs 11.6°; P < .001), though ΔIDL for L3/4 and L4/5 was similar between the standalone and circumferential groups. Multivariable analysis showed postoperative IDL to be predicted by greater preoperative IDL (β = 0.26° per degree; [0.15, 0.36]; P < .001), L5/S1 [vs L4/5] implantation (β = 3.97°; [2.72, 5.21]; P < .001), use of supplementary posterior fixation (β = 1.80°; [0.70, 2.89]; P < .001), and greater implant lordosis (β = 0.62° per degree; [0.44, 0.79]; P < .001). CONCLUSION: Postoperative IDL after 1- or 2-level ALIF is best predicted by placement of larger interbody sizes, L5/S1 implantation, greater baseline IDL, and use of adjunct posterior fixation. Patients with higher preoperative lordosis experienced smaller postoperative increases in IDL, though this was likely influenced by a treatment bias.

Abstract Background Immunohistochemically detected phosphorylated α‐synuclein (p‐α‐syn) in cutaneous nerves has been found in >95% of patients meeting consensus criteria for dementia with Lewy bodies (DLB); however, its performance in prodromal stages (mild cognitive impairment‐Lewy body [MCI‐LB]) and clinical practice is unknown. Objective We evaluated skin biopsy for p‐α‐syn in a convenience cohort of patients with suspected MCI‐LB or DLB in a tertiary care setting. Methods 53 patients underwent commercially available skin biopsy for skin nerve p‐α‐syn as part of a clinical evaluation for suspected MCI‐LB/DLB. Consensus MCI‐LB/DLB criteria blinded to skin biopsy result were applied through retrospective chart review; those not definitively meeting criteria were considered questionable synucleinopathy (QS). Results Median age at skin biopsy was 70 years (range, 43–87 years) with 81% men. 36 patients (68%) had positive skin biopsies. 32/43 (74%) patients meeting clinical consensus criteria for DLB/MCI‐LB had a positive skin biopsy, with higher frequency in MCI‐LB (87%) compared with DLB (60%). 4/10 (40%) QS had positive biopsies. One patient with a positive skin biopsy had no evidence of central nervous system (CNS) α‐syn on autopsy. Conclusions Positive skin biopsy for p‐α‐syn occurs across the spectrum of cognitive presentations of Lewy body disease including prodromal DLB, albeit at a lower frequency in this retrospective review of atypical cases compared with typical clinical presentations. A positive skin biopsy may not reflect CNS synucleinopathy as the etiology for a patient's symptoms in some cases. Longitudinal studies of skin biopsies for p‐α‐syn in suspected MCI‐LB/DLB with autopsy confirmation are needed. © 2025 International Parkinson and Movement Disorder Society.

Abstract Molecular profiling of human primary cell types is essential for understanding human biology. We present a transcriptome and proteome map of 28 primary human cell types. Three major clusters of epithelial, endothelial, and mesenchymal cell types were observed in both the transcriptome and proteome levels along with the discovery of cell type enriched molecules including GRAP and C1orf116. The epithelial cell specific protein C1orf116 was further validated using immunohistochemistry across various human tissues. An exhaustive protein database search considering 39 post-translational modifications (PTMs) revealed novel insights into the PTM landscape including identification of understudied PTMs such as serine O-acetylation and histidine methylation. This also enabled comprehensive characterization of proteins with diverse PTMs. Interestingly, an unexpectedly higher frequency of dioxidation on tryptophan compared to methionine led to the identification of oxidative mitochondria complex subunit proteins. Further, a search strategy accounting for alternative translational start sites, splice junctions and translational readthrough refined genome annotation using proteomic evidence. For example, peptides from translational readthrough including extended sequence of LDHB and MDH1 were detected representing the first peptide-level evidence of these protein readthrough isoforms. Our comprehensive transcriptome and proteome data revealed cell type-specific molecular cues and heterogeneity, offering new insights into disease mechanisms often overlooked by tissue proteomics.

The application of artificial intelligence (AI), large language models and generative AI (GenAI) is accelerating rapidly across many industries worldwide. Innovative models and tools are being piloted to understand their business and societal value. Models such as ChatGPT (OpenAI), Gemini (Google) and Copilot (Microsoft) have become increasingly sophisticated and widely adopted. Understanding the advantages and drawbacks of GenAI models is crucial for organisations seeking to leverage GenAI to maximise their efficiency and expand their capabilities. This paper describes the strategic approach, implementation and findings of using GenAI in a health care business context. A GenAI model was used to analyse and summarise strategic and operational information submitted by external organisations seeking to collaborate with Mayo Clinic. The goal was to reduce process cycle time and manual resource intensity. In parallel, a comparative human review was conducted by functional experts. Both the AI-generated and human-reviewed results were evaluated using sentiment analysis (SA) and then categorised as positive, neutral or negative. Attention was paid to possible inaccuracies and biases associated with the use of the GenAI model. GenAI reduced the manual workload by approximately 20−25 per cent and presented information in a well-structured format. However, the SA showed the GenAI results as positively biased and the human-generated results as neutral. These results were most likely influenced by a positive bias in the input documents, the lack of contextual comparators/ standards in the GenAI model and an inherent bias within the model’s algorithm(s). Additionally, limitations in input size and language translation hindered the GenAI model’s effectiveness. The possibility of AI hallucinations/inaccuracies emphasised the importance of validation by a human expert. Furthermore, carefully formatting the input data enhanced the value and relevance of the GenAI model’s outputs. GenAI has the potential to increase efficiency and alleviate manual burden, but it must be thoughtfully implemented to address inaccuracies, biases and privacy concerns. This article is also included in The Business & Management Collection which can be accessed at https:// hstalks.com/business/.

Abstract INTRODUCTION Plasma phosphorylated tau at threonine 217 (p‐tau217) has emerged as one of the most promising blood‐based biomarkers for Alzheimer's disease (AD). While rare, heterophilic antibodies (HAb) are a persistent potential confounding factor in immunoassays. METHODS Potential HAb interference in the Lumipulse G pTau217 assay in samples exhibiting concentrations > 10 pg/mL were investigated. Samples were subjected to HAb blocking reagent (HBT), and in some cases, polyethylene glycol (PEG) precipitation and serial dilutions. RESULTS In 14 of the 15 suspected HAb cases, HBT treatment greatly reduced measured p‐tau217 concentrations. and in one of three specimens selected for further investigation, reduced Aβ42 concentrations PEG precipitation in the three selected samples also reduced p‐tau217 concentrations, while serial dilution yielded mixed effects. Control specimens were unaffected by sample treatment or dilutions. DISCUSSION HAb may result in falsely elevated p‐tau217 and p‐tau217/Aβ42 ratio. HAb interference should be considered in cases of unusually high p‐tau217 concentrations prior to clinical interpretation. Highlights Potential heterophile antibody (HAb) interference in the Lumipulse p‐tau217 immunoassay was investigated. Investigation was performed using HAb blocking reagent, and in some cases also using PEG precipitation . Patients with p‐tau217 concentrations > 10 pg/mL exhibited positive HAb interference in 14 of 15 cases. HAb interference can lead to falsely positive p‐tau217/Aβ42 ratios.

Parkinson’s Disease (PD) is a neurodegenerative disorder in which gait abnormalities serve as key indicators of motor impairment and disease progression. Although wearable sensor-based gait analysis has advanced, existing methods still face challenges in modeling multi-sensor spatial relationships, extracting adaptive multi-scale temporal features, and effectively integrating time–frequency information. To address these issues, this paper proposes a multi-sensor gait neural network that integrates biomechanical priors with time–frequency collaborative learning for the automatic assessment of PD gait severity. The framework consists of three core modules: (1) BGS-GAT (Biomechanics-Guided Graph Attention Network), which constructs a sensor graph based on plantar anatomy and explicitly models inter-regional force dependencies via graph attention; (2) AMS-Inception1D (Adaptive Multi-Scale Inception-1D), which employs dilated convolutions and channel attention to extract multi-scale temporal features adaptively; and (3) TF-Branch (Time–Frequency Branch), which applies Real-valued Fast Fourier Transform (RFFT) and frequency-domain convolution to capture rhythmic and high-frequency components, enabling complementary time–frequency representation. Experiments on the PhysioNet multi-channel foot pressure dataset demonstrate that the proposed model achieves 0.930 in accuracy and 0.925 in F1-score for four-class severity classification, outperforming state-of-the-art deep learning models.

Abstract Rapid and profound reduction of free light chains (FLCs) improves outcomes in AL amyloidosis; however, early FLC kinetics with daratumumab‐based frontline therapy remain undefined. We retrospectively analyzed 315 patients treated with daratumumab–bortezomib–cyclophosphamide–dexamethasone (Dara–VCd) or Dara–Vd at eight centers. Involved FLC (iFLC), the difference between iFLC and uninvolved FLC (dFLC), and hematologic response (≥very good partial response [VGPR]) were assessed at baseline and at 1, 3, and 6 months, and correlated with light chain isotype, disease burden (bone‐marrow plasma‐cell percentage [%BMPC], baseline dFLC), and cytogenetics. Hematological ≥VGPR increased from 49.8% at 1 month to 66.0% at 3 months. The kappa isotype showed slower responses, with higher iFLC/dFLC and lower ≥VGPR at each time point compared to lambda. Higher %BMPC or baseline dFLC predicted persistently elevated iFLC/dFLC and reduced ≥VGPR. The t(11;14) translocation was associated with lower baseline FLC but similar subsequent kinetics. BMPC < 10% and dFLC < 18 mg/dL independently predicted early ≥VGPR (in ≤1 month). Early ≥VGPR conferred superior hematologic event‐free survival (heme‐EFS) and overall survival (OS) versus later responders (P < 0.001). At a 3‐month landmark, achieving dFLC < 1 mg/dL further stratified prognosis, conferring longer heme‐EFS and OS (P < 0.01). Frontline Dara‐based therapy elicits rapid, deep responses in AL amyloidosis; early ≥VGPR (within 1 month) and dFLC < 1 mg/dL by 3 months predict durable EFS and OS, whereas higher baseline disease burden delays hematologic response.

Abstract Variants in the cystic fibrosis transmembrane conductance regulator ( CFTR ) gene, which causes cystic fibrosis, have become increasingly well‐characterized in recent decades. However, clinicians remain challenged by CFTR variants that result in variable expressivity and reduced penetrance, including the variant R117H (c.350G>A) and the intron 9 poly‐thymidine (poly‐T) and poly‐thymidine‐guanosine (TG) tracts, highlighting a knowledge gap that impacts the delivery of accurate information to patients and families. To address this gap, we developed an educational resource that provides descriptive likelihoods of specific phenotypes that may result from genotypes including R117H and the poly‐T and TG tracts. Development of the resource was grounded in an extensive literature review, clinical expertise, and publicly available CFTR variant databases, with feedback incorporated from CF clinicians, genetic counselors, and CF content experts. Genotypes were determined to have very low, low, moderate, high, or very high likelihood of resulting in CF, CFTR‐related disorder (CFTR‐RD), no symptoms, or the designation of CFTR‐related metabolic syndrome/CF‐screen positive, inconclusive diagnosis (CRMS/CFSPID) following CF newborn screening. The resource described in this report is intended to help clinicians navigate the uncertainty surrounding potential diagnostic scenarios to present accurate assessments to patients and families, a critical component of CF genetic testing and counseling.