Citations

Abstract Background Delirium is often underdiagnosed in clinical practice and is not routinely coded for billing. While manual chart review can identify delirium, it is labor-intensive and impractical for large-scale studies. Natural language processing (NLP) can analyze unstructured text in electronic health records (EHRs) to extract meaningful clinical information. Methods To support national integration of NLP for EHR-based delirium identification across different institutions, we launched the Delirium Interest Group within the national Evolve to Next-Gen Accrual to Clinical Trials (ENACT) NLP Working Group. This paper outlines our initial efforts to standardize, evaluate, and translate an NLP-based delirium detection model into the i2b2/ENACT platform. Results Multisite contextual inquiry identified several key challenges, including variations in local screening practices (eg, tools used, documentation frequency, and quality control), the need for harmonized definitions in the context of EHRs, and the complexity of modeling temporal logic. Multisite NLP evaluation revealed variable performance degradation driven by differences in delirium screening practices, clinical documentation patterns and semantics, and note syntactic structures. Conclusion Our work represents an important first step toward enabling scalable and standardized NLP-based delirium detection across institutions. By engaging diverse institutions through the ENACT NLP Working Group, we identified shared challenges and site-specific variations that impact model implementation and performance. Our collaborative approach enabled the development of a more robust framework for delirium identification across heterogeneous EHR systems. Future efforts will build on this foundation to enhance the validity, usability, and translational impact of delirium detection.

ABSTRACT Objectives To characterize practice patterns and outcomes in the management of low‐ and intermediate‐grade salivary gland carcinoma prior to the existence of treatment guidelines. Methods Analysis of a registry of patients who underwent parotid and submandibular gland resections for low‐and intermediate‐grade carcinomas between 2010 and 2019. Results Of all 786 patients included in the study, 726 (92%) had preoperative imaging and 653 (83%) had preoperative biopsy. Of the 729 patients with parotid gland cancer, the majority underwent superficial ( n = 384, 53%) or total ( n = 254, 35%) parotidectomy. In patients with facial nerve preservation, total parotidectomy was associated with a significant increase in transient facial weakness (72/177 (41%) vs. 82/311 (26%), RR 0.65, 95% CI 0.50–0.84, p < 0.05) and permanent facial nerve weakness (23/176 (13%) vs. 16/301 (5%), RR 0.41, 95% CI 0.22–0.75, p < 0.05) compared to superficial parotidectomy. Adjuvant radiation therapy ( RT ) was delivered to 285 (36%) patients. The proportion of patients receiving adjuvant RT declined significantly over the time period from 2015 to 2019 compared to 2010 to 2014 at 162/504 (32%) and 123/282 (44%), respectively ( RR 0.74, 95% CI 0.61–0.89, p < 0.05). When comparing the time periods from 2015 to 2019 and 2010 to 2014, there was no significant difference in local control rates ( RR 0.52, 95% CI 0.26–1.04, p = 0.06) or regional control rates ( RR 0.75, 95% CI 0.26–2.13, p = 0.58). Conclusions Management of low‐ and intermediate‐grade salivary cancer from 2010 to 2019 was variable, which is expected given the rarity and heterogeneity of the disease and the lack of treatment guidelines prior to 2021. Most patients with parotid malignancies underwent superficial or total parotidectomy. The extent of parotidectomy had an impact on facial nerve function outcomes. Delivery of adjuvant radiation trended down with time. The data presented here will support dissemination of the guidelines and provide data that could inform future trials. Level of Evidence 2b.

Background: The COVID-19 pandemic brought profound changes to health care systems worldwide. This study examines the adaptive responses of a sleep medicine practice that successfully pivoted to telemedicine, remote diagnostics, and agile operations, resulting in sustained growth and improved patient accessibility. Methods: We analyzed organizational and cultural elements enabling responsiveness, including telemedicine infrastructure, adaptive leadership, quality-focused processes, and effective communication. Objective outcomes were assessed by analyzing patient demographics, telemedicine visits, and sleep testing patterns during three phases: pre-pandemic (June 1, 2018 to March 8, 2020), acute COVID-19 (March 9, 2020 to April 19, 2020), and post-acute COVID-19 (April 20, 2020 to December 31, 2021). Results: Across 105,199 encounters, monthly sleep medicine visits and testing declined ∼50% during acute COVID-19, but rebounded to surpass baseline in the post-acute phase. Virtual visits increased significantly, replacing many face-to-face encounters. Home sleep apnea test use rose sharply, driven by disposable WatchPAT tests, while polysomnography use decreased. Testing volumes increased overall post-acute compared with pre-pandemic. Patients served during acute and post-acute phases were younger, more often female, non-White, and postgraduate-educated (all p < 0.001). Nonlocal patients increased during the post-acute phase, reflecting telemedicine expansion. Conclusions: Proactive telemedicine investments, adaptive leadership, quality-focused processes, and effective communication enabled this practice to adapt successfully during the pandemic. These resilience strategies provide a model for navigating future health care challenges, underscoring telemedicine and e-health’s critical roles in maintaining care delivery during crises.

Abstract Small cell carcinoma of the ovary, hypercalcemic type (SCCOHT) is a rare, deadly form of ovarian cancer that uniformly harbors mutations in SMARCA4, a member of the SWI/SNF chromatin remodeling complex. SWI/SNF impacts RNA splicing, and dysregulation of splicing can generate immunogenic tumor antigens. Here, we explored the relationship between SMARCA4 loss and RNA splicing dysregulation. SCCOHT primary tumors harbored tumor-associated outlier splicing events compared to normal tissues. Many of the tumor events were retained introns encoding novel peptides predicted to bind to MHC-I complexes. Immune cells were observed in primary SCCOHT tumors, suggesting a potentially immune reactive tumor microenvironment. Mutations in several SWI/SNF subunits were associated with higher rates of outlier retained introns across tumor types in TCGA data. Interestingly, RNA sequencing of isogenic SCCOHT cell lines demonstrated a role for SMARCA4 in intron retention. Distinct protein-protein interactions between splicing factors identified in SCCOHT cell lines supported a role for SMARCA4 in splicing regulation. Further, SWI/SNF localized to genes which were differentially spliced. Mass spectrometry analyses confirmed expression of some of these novel peptides and a subset of these are predicted to bind to MHC-I complexes. A pool of these novel peptides derived from retained introns in SCCOHT triggered proliferation and expression of TNFa and INFb in primary human T cells. Together, these data suggest that SMARCA4 loss in SCCOHT leads to intron retention. Furthermore, T cell activation by novel peptides encoded by these tumor-specific splicing events suggests intron retention could be a source of tumor-associated antigens in SCCOHT. Citation Format: Elizabeth Raupach, Apurva Hegde, Krystine Garcia-Masnfield, Marice Alcantara, David Rose, Rebecca Halperin, Krystal Orlando, Jessica Lang, Ritin Sharma, Victoria David-Dirgo, Salvatore Facista, Rayvon Moore, Rochelle Kofman, Zoe Jensen, Victoria Zismann, Anthony Karnezis, Yemin Wang, Lynda Bennett, Timothy Whitsett, Marcin Kortylewski, William Hendricks, David Huntsman, Lorna Rodriguez-Rodriguez, Bernard Weissman, Jeffrey Trent, Patrick Pirrotte. Loss of SMARCA4 leads to intron retention and generation of tumor-associated antigens in small cell carcinoma of the ovary, hypercalcemic type [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Ovarian Cancer Research; 2025 Sep 19-21; Denver, CO. Philadelphia (PA): AACR; Cancer Res 2025;85(18_Suppl):Abstract nr B065.

Abstract Background The past two decades have witnessed tremendous growth in the appreciation and treatment of sacroiliac joint (SIJ) complex pain, including anatomical dissections that shed light on innervation, an appreciation for the contribution of extra-articular components to SIJ complex pain, the advent of radiofrequency ablation (RFA) and a host of minimally invasive surgical techniques. Yet, there is no standardization on diagnosis and treatment paradigms. Methods In February 2023, the Boards of Directors for the American Academy of Pain Medicine (AAPM) and American Society of Regional Anesthesia & Pain Medicine (ASRA-PM) approved the development of multispecialty guidelines on SIJ complex pain. Thirty partner organizations with clinical and scientific interests in SIJ complex pain were identified, and formal letters of request-for-participation were sent to each, along with a request for nominees to serve on the committee. Twenty five organizations agreed to participate in addition to the Departments of Defense and Veterans Affairs. A steering committee developed 21 questions, which spanned criteria for diagnosis, non-interventional and interventional treatments including surgery, technical parameters on how to optimize results, and what constitutes positive outcomes. Questions were methodically assigned to specialized modules comprising 4-5 members with complementary expertise, who collaborated with the Subcommittee Lead and one of three Committee Chairs to develop preliminary drafts. Following thorough revisions, these drafts were subsequently submitted to the full committee for comprehensive review. A modified Delphi method was used in which the answers to questions were sent to the committee en bloc and comments were returned in a non-blinded fashion to the Chairs, who incorporated the comments and sent out revised versions until consensus was achieved. During a committee meeting before commencement, it was agreed that recommendations would be noted when there was &gt;50% agreement among committee members, but that a formal recommendation would require ≥75% consensus. Results Twenty-one organizations formally endorsed the guidelines. The American Society of Anesthesiologists, and American Academy of Physical Medicine & Rehabilitation, and the North American Spine Society affirmed the benefit of the guidelines but did not officially endorse them. The American Academy of Neurology declined to affirm the benefit of the guidelines citing “lack of relevance to their membership.” Per policies, while the Departments of Defense and Veterans Affairs did not formally review the guidelines for endorsement, their representatives approved them. In addition to being endorsed or the benefit affirmed by all voting organizations, complete consensus from committee members was obtained on all 21 questions. On 2 recommendations, there were dissensions from 3 societies who thought that selecting patients for sacral lateral branch RFA and minimally invasive fusion should be based on ≥75% relief from 2 blocks instead of at least 50% relief from a single block before RFA, and greater than 50% pain relief with documented functional improvement after a single block before fusion. One additional society (Latin American Society of Regional Anesthesia) abstained on the statement that the evidence is stronger for extra-articular than intra-articular injections. The committee found that a battery of physical exam tests has reasonable sensitivity, but lower specificity, for identifying intra-articular but not extra-articular pathology, with negative tests having greater predictive value than positive ones. Intra-articular injections have diagnostic validity for SIJ intra-articular, but not extra-articular pain. There is unclear or negative evidence for imaging. The prevalence rates of intra-articular and extra-articular pathology are comparable, with both intra- and extra-articular steroid injections providing at least 4 weeks of relief in well-selected patients. However, the evidence is slightly stronger for extra-articular corticosteroid injections to provide short-term relief. The evidence base for non-interventional therapies is indirect, extrapolated mostly from low back pain studies. There is weak evidence supporting dextrose-based prolotherapy and platelet-rich plasma to provide at least 3 months of pain relief. There is strong evidence for sacral lateral branch RFA to provide relief for at least 6 months in individuals with extra-articular pathology, with face validity and indirect evidence from randomized trials supporting sacral lateral branch blocks as a prognostic tool. There is stronger evidence for larger lesions or more aggressive lesioning strategies than for less stringent techniques. There is weak evidence to support non-steroidal anti-inflammatory drugs to prevent neuritis after RFA, and in most cases anticoagulation does not require cessation in the periprocedural perioid. With an aggressive lesioning strategy, sensory stimulation provides minimal therapeutic benefit, with weak, extrapolated evidence that motor stimulation can provide safety benefit. The cutoff to designate diagnostic or prognostic blocks as positive is most commonly set at 50%, with higher values not shown to improve outcomes for more definitive procedures; for therapeutic treatment outcomes, the evidence supports a lower threshold of ≥30% pain relief or meaningful benefit on non-pain outcomes (eg, opioid cessation) for designation of a positive response. For carefully selected patients with intra-articular SIJ complex pain based on controlled blocks who have failed conservative therapies, there is weak or very weak evidence that minimally invasive SIJ fusion can provide benefit for at least one year. Conclusions SIJ complex pain remains an underappreciated source of chronic low back pain, affecting between 15% and 30% of patients with axial pain predominantly below L5. Answers to many questions were limited by low-quality evidence, indicating the need for better research. SIJ complex pain is a multifarious condition (ie, pain can be from different portions of both the intra- and extra-articular components of the joint) for which an interdisciplinary, multimodal treatment plan can optimize treatment outcomes.

Abstract Radiological protocol selection is a critical but time-consuming step in clinical workflow, requiring radiologists to match patient indications with the appropriate MRI or CT protocol. Manual selection can be prone to delays or potential errors, and automated approaches must contend with substantial data imbalance, site-specific variation, and evolving nomenclature. We investigated whether a large language model (LLM) can support reliable protocol selection at scale and whether retrieval-augmented generation (RAG) offers operational advantages over direct fine-tuning. Using 498,228 patient reports collected across three Mayo Clinic sites (Arizona, Florida, and Rochester) spanning six radiological divisions, we trained site-specific Llama 3.2 3B models for use with and without retrieval augmentation. Division-scoped Facebook AI Similarity Search (FAISS) indexes constructed from procedure and diagnosis text were used to supply contextual evidence in the RAG framework. Both fine-tuned and RAG-augmented models achieved strong performance across sites, with F1 scores of 0.88–0.90. RAG matched or modestly trailed direct fine-tuning overall but delivered consistent gains in specific divisions (e.g., musculoskeletal imaging). Importantly, the RAG model introduced abstention behavior (Not Enough Information), which concentrated in linguistically diverse divisions and provided an interpretable signal of uncertainty. These findings suggest that RAG-based models are viable for division-scoped protocol selection and offer practical advantages. Retrieval indexes can be refreshed far more easily and with fewer resources than retraining LLMs, enabling continual adaptation to evolving clinical workflows. Future prospective deployment will evaluate real-time accuracy, agreement with practitioners, and the role of abstention as a safety mechanism in clinical decision support.

This study evaluated the efficacy of the multiple sclerosis disease-modifying therapies, intramuscular interferon beta-1a (Avonex) and subcutaneous peginterferon beta-1a (Plegridy), using data from the United States Network of Pediatric Multiple Sclerosis Centers. In this retrospective analysis, 154 patients with multiple sclerosis were included who were treated with Avonex (n = 130), Plegridy (n = 23), or both treatments (n = 1) before the age of 18 years. After 3 months’ sustained use acclimation (“wash-in”), the probability of being relapse-free during the first year was 68.3% for Avonex-treated patients and 69.9% for Plegridy-treated patients; annualized relapse rates were 0.50 and 0.59, respectively. Both disease-modifying therapies demonstrated efficacy similar to that reported in adult populations. Despite the lack of formal approval for pediatric multiple sclerosis, these outcomes indicate that patients may benefit from treatment with Avonex or Plegridy. Understanding efficacy of specific disease-modifying therapies in pediatric multiple sclerosis is essential to making informed treatment decisions.

OBJECTIVE The Centers for Medicare & Medicaid Services (CMS) requires a low C-peptide level for insulin pump coverage unless the individual is β-cell autoantibody positive, which precludes coverage of automated insulin delivery (AID) systems for many people with type 2 diabetes. RESEARCH DESIGN AND METHODS In the Randomized Trial Evaluating the Efficacy and Safety of Control-IQ+ Technology in Adults With Type 2 Diabetes Using Basal-Bolus Insulin Therapy study evaluating the t:slim X2 insulin pump with Control-IQ+ technology, adults with insulin-treated type 2 diabetes were categorized into high C-peptide (n = 195) and low C-peptide (n = 59) groups based on CMS criteria. RESULTS In the AID group, mean HbA1c decreased from baseline by 0.8%, which was significantly greater than in the control group with both high (P &lt; 0.001) and low (P = 0.02) C-peptide levels. Results were similar in participants ≥65 years old. CONCLUSIONS The benefit of AID is present with high and low C-peptide levels. Thus, requiring a low C-peptide level as a prerequisite for AID therapy is not warranted.

Importance The Implantable Neurostimulator for the Treatment of Parkinson’s Disease (INTREPID) trial was a randomized, double-blind, sham-controlled study of subthalamic nucleus (STN) deep brain stimulation (DBS) for the treatment of Parkinson disease (PD). Objective To evaluate the long-term (5-year) outcomes and safety of STN-DBS for PD. Design, Setting, and Participants This was a prospective, randomized (3:1), 12-week double-blind sham-controlled study at 23 movement disorder centers across the US with an open-label 5-year follow-up. Patients were implanted and followed up with the Vercise DBS system from May 2013 to December 2022. Eligibility required diagnosis of bilateral idiopathic PD with more than 5 years of motor symptoms, more than 6 hours per day of poor motor function, modified Hoehn and Yahr Scale scores higher than 2, Unified Parkinson’s Disease Rating Scale (UPDRS-III) score of 30 or higher (medication-off state), and 33% or higher improvement in UPDRS-III medication-on score. Intervention Bilateral STN-DBS for moderate to advanced PD. Main Outcomes and Measures Primary outcomes included changes in UPDRS and dyskinesia scores, quality-of-life measures, and safety assessments. Exploratory analyses included medication reduction and DBS association with motor signs. Results A total of 313 patients were enrolled with 191 receiving the DBS system, and 137 participants (72%) completed the study. The study population had a mean (SD) age of 60 (7.9) years, with 139 (73%) male participants. Motor function without medication as measured by UPDRS-III improved from a mean (SD) of 42.8 (9.4) to 21.1 (10.6) at year 1 (51%; 95% CI, 49%-53%; P &amp;lt; .001) and 27.6 (11.6) at year 5 (36%; 95% CI, 33%-38%; P &amp;lt; .001). Activities of daily living without medication as measured by UPDRS-II improved from a mean (SD) of 20.6 (6.0) to 12.4 (6.1) at year 1 (41%; 95% CI, 38%-42%; P &amp;lt; .001) and 16.4 (6.5) at year 5 (22%; 95% CI, 18%-23%; P &amp;lt; .001). Dyskinesia scores decreased from 4.0 (5.1) to 1.0 (2.1) at year 1 (75%; 95% CI, 73%-75%; P &amp;lt; .001) and to 1.2 (2.1) at year 5 (70%; 95% CI, 63%-75%; P &amp;lt; .001). The levodopa equivalent dose was reduced by 28% at year 1, remaining stable at year 5 (28%; 95% CI, 26%-31%; P &amp;lt; .001). The most common serious adverse event was infection (9 participants). Ten deaths were reported, none related to the study. Conclusions and Relevance Although STN-DBS outcomes declined slightly, possibly due to the progressive nature of the disease, patients with PD sustained significant improvement in motor and activities of daily living scores, along with a stable reduction in anti-parkinsonian medication over the 5-year follow-up period.