Citations

Machine-learning (ML) models have the potential to transform health care by enabling more personalized and data-driven clinical decision making. However, their successful implementation in clinical practice requires careful consideration of factors beyond predictive accuracy. We provide an overview of essential considerations for developing clinically applicable ML models, including methods for assessing and improving calibration, selecting appropriate decision thresholds, enhancing model explainability, identifying and mitigating bias, as well as methods for robust validation. We also discuss strategies for improving accessibility to ML models and performing real-world testing. Highlights This tutorial provides clinicians with a comprehensive guide to implementing machine-learning classification models in clinical practice. Key areas covered include model calibration, threshold selection, explainability, bias mitigation, validation, and real-world testing, all of which are essential for the clinical deployment of machine-learning models. Following these guidance can help clinicians bridge the gap between machine-learning model development and real-world application and enhance patient care outcomes.

Background: Periacetabular osteotomy (PAO) is a complex procedure historically requiring inpatient stay. However, there is increased emphasis for other complex orthopaedic procedures (eg, arthroplasty) to be performed on an outpatient basis by insurance companies. Increasing resistance to insurance approval of inpatient listing for PAO patients at the participating institution suggests that this changing tide is affecting PAO surgery. Purpose: To (1) investigate the proportion of PAOs various surgical listing classifications (outpatient, outpatient-overnight, inpatient, etc), (2) determine the incidence of subsequent denial/request for additional documentation for approval of inpatient stay after PAO, and (3) characterize the accuracy of preoperative patient listing classification as well as mean length of stay (LOS) and incidence of successful same-day discharge after PAO at a single institution. Study Design: Case series; Level of evidence, 3. Methods: A retrospective chart review was performed to identify all PAOs, performed by 4 participating surgeons, at a single academic institution over a 2-year period. The initial listing status as an outpatient-overnight or inpatient procedure was identified. Whether a preoperative peer review was required for approval of inpatient listings as well as if additional documentation was necessary to convert outpatient-overnight listings to inpatient stays were also recorded. Results: A total of 140 PAOs among 117 patients were performed with 25 (17.9%) initially listed as an inpatient stay and 115 (82.1%) listed as outpatient-overnight. Of the 25 PAOs listed as inpatient, 2 (8.0%) required a preoperative peer review process to justify or clarify listing status. The mean LOS was 1.9 ± 1.4 days with 55.7% (78/140) of PAOs staying ≥2 nights in the hospital. Only 6 PAOs (4.3%) went home the same day as surgery. Patients who underwent PAO with concomitant hip arthroscopy stayed on average 2.4 ± 1.2 days in comparison with 1.6 ± 1.4 days for those who underwent PAO alone ( P < .001). Of the 115 PAOs listed as an outpatient-overnight, 53 (46.1%) converted to an inpatient stay, all of which required service documentation to support conversion to inpatient status. Conclusion: Over half of PAOs performed resulted in an inpatient stay with patients spending ≥2 nights in the hospital. Additionally, all conversions from outpatient-overnight to inpatient required additional service documentation to support status conversion. Understanding trends in postoperative hospitalizations and LOS may allow for better informed partnerships between surgeons and insurance companies creating more efficient preauthorizations, billing practices, and expected patterns of patient care and discharge.

Abstract Objective To examine the effectiveness of treatment strategies for medication‐overuse headache (MOH) in children and adolescents. Background There are no guidelines for the treatment of adult or pediatric patients with MOH. However, research pertaining to MOH in adults is far outpacing research in children and adolescents. Methods A systematic review of evidence on treatment for MOH in children and adolescents was performed via PubMed, Embase, Web of Science, PsycINFO, and Cochrane from database inception to October 2024 for articles with the subject headings and keywords for MOH and pediatric, children, and/or adolescent. Inclusion criteria were: (i) randomized controlled trials and retrospective and prospective observational studies, (ii) children and adolescents aged <18 years with chronic headaches and with medication overuse (MO) or MOH, and (iii) interventions addressing MO or MOH, including pharmacologic and behavioral interventions. At least two independent reviewers screened citations and full‐text articles independently. Results A total of 16 studies (one randomized control trial, six prospective observational studies and nine retrospective observational studies) were identified as meeting final inclusion criteria, resulting in 7978 total participants with a subset of those having MO, MOH or probable MOH. Participants ranged in mean age from 10 to 14.7 years, with 7.1%–46.2% male participants. All the studies recruited patients from an outpatient setting. Interventions used in the studies included withdrawal of the overused acute medication, initiation of preventive medication, patient education, physiotherapy and psychological treatment, and motivational interviewing. Multimodal therapy was studied more than single therapeutic intervention. There was vast heterogeneity across the studies, including study design, diagnostic criteria, intervention type, and outcomes. Conclusion From the data available, there does not seem to be a singular superior treatment approach. Accordingly, the most appropriate treatment modality will need to be tailored to the patient and their caregivers. Further research is required regarding management of MOH in children and adolescents.

ABSTRACTBackground/ObjectivesImmune‐mediated skin disorders, such as immunobullous dermatoses and leukocytoclastic vasculitis, rarely affect children. While direct immunofluorescence (DIF) biopsy is a standard diagnostic tool, limited data exist on pediatric DIF patterns, rates of positivity, pretest diagnostic concordance, and the relevance of biopsy site. This study sought to address these gaps.MethodsDIF data from all skin and mucosal specimens interpreted at Mayo Clinic's reference immunodermatology laboratory for patients aged 0–18 years (August 22, 2017 to November 30, 2023) were reviewed. DIF results were classified as positive if a characteristic pattern was seen and negative if the findings were negative or nondiagnostic.ResultsOf 986 pediatric DIF studies, 153 (15.5%) were positive and comparable to adult positivity rates (20.9%) during the same period. The most frequent DIF patterns were IgA‐predominant vasculitis (N = 85/153; 55.5%) and lichenoid tissue reaction (N = 21/153;13.7%). Concordance between pretest diagnosis and positive DIF results was highest for linear IgA bullous dermatosis (N = 7/7; 100%) and dermatitis herpetiformis (N = 6/6; 100%). Excluding these entities, DIF changed the pretest diagnosis in 16.7% (N = 19/114) of cases. While lower extremity biopsies were initially more likely to yield positive DIF results, this association disappeared when IgA vasculitis cases were excluded.ConclusionThe most frequent DIF pattern in children was that of IgA‐predominant vasculitis. Pediatric DIF positivity rates closely mirrored those of adults, supporting similar biopsy thresholds. DIF results differed from the pretest impression in a substantial percentage, supporting the value of DIF in select situations in the pediatric population. After controlling for IgA vasculitis, biopsy site was not associated with DIF positivity.

Abstract Introduction Limited data exist on the natural history and treatment persistence in ulcerative colitis (UC) patients on 5-aminosalicylic acid (5-ASA) in the era of biologic agents and small molecules. Methods This was a retrospective cohort study utilizing the US Collaborative Network in patients with UC who were initiated on 5-ASA between 2015 and 2019. The 1-, 3-, and 5-year incidence rates of advanced therapy use and colectomy were calculated. Predictors of advanced therapy use and colectomy within 5 years were analyzed and expressed as adjusted odds ratios (aORs) with 95% confidence intervals (CIs) after 1:1 propensity score matching (PSM). Results Among 15 177 patients in the UC 5-ASA cohort (mean age 47.2 ± 19.1 years old, 49.2% female and 70.3% White), the 1-, 3- and 5-year incidence of advanced therapy use was 5.37%, 11.56% and 15.8%, respectively. After PSM, age &lt; 40 years old (aOR 1.93, 95% CI, 1.73-2.15), male sex (aOR 1.13, 1.02-1.25), obesity (aOR 1.18, 95% CI, 1.05-1.31), pancolitis (aOR 1.18, 95% CI, 1.03-1.36), steroid use (aOR 1.79, 95% CI, 1.53-2.09), and CRP ≥ 12 mg/L (aOR 1.39, 95% CI, 1.05-1.83) prior to 5-ASA initiation were associated with advanced therapy use within 5 years. The 1-, 3- and 5-year incidence of colectomy was 0.74%, 1.28%, and 1.64% respectively. After PSM, male sex (aOR 1.36, 95% 1.06-1.74), obesity (aOR 1.32, 95% CI, 1.02-1.69), primary sclerosing cholangitis (aOR 2.37, 95% CI, 1.11-5.06), pancolitis (aOR 1.90, 95% CI, 1.29-2.78), and steroid use (aOR 1.96, 95% CI, 1.35-2.84) were associated with colectomy within 5 years. Conclusion Our study provides updated estimates of advanced therapy utilization and colectomy as well as predictors of disease progression in patients with UC on 5-ASA in the biologic era.

Parkinson’s disease (PD) is the second-most common neurodegenerative disease. With more than 20,000 new diagnosed cases each year, PD affects millions of individuals worldwide and is most prevalent in the elderly population. The current clinical methods for the diagnosis and severity assessment of PD rely on the visual and physical examination of subjects and identifying key disease motor signs and symptoms such as bradykinesia, rigidity, tremor, and postural instability. In the present study, we developed a method for the diagnosis and severity assessment of PD using Gramian Angular Fields (GAFs) in combination with deep Convolutional Neural Networks (CNNs). Our model was applied to PD gait signals captured using pressure sensors embedded into insoles. Our results indicated an accuracy of 98.6%, a true positive rate (TPR) of 99.2%, and a true negative rate (TNR) of 98.5%, showcasing superior classification performance for PD diagnosis compared to the methods used in recent studies in the literature. The estimation of disease severity scores using gait signals showed a high accuracy for the Hoehn and Yahr score as well as the Timed Up and Go (TUG) test score (R2 > 0.8), while we achieved a lower prediction performance for the Unified Parkinson’s Disease Rating Scale (UPDRS) and its motor component (UPDRSM) scores (R2 < 0.2). These results were achieved using gait signals recorded in time windows as small as 10 s, which may pave the way for shorter, more accessible assessment tools for diagnosis and severity assessment of PD.

BACKGROUND: Patients with lung disease, sleep apnea, and chronic thromboemboli can develop pulmonary hypertension, currently classified as group 3 or 4. Many of these patients also have risk factors for heart failure with preserved ejection fraction (HFpEF), but the optimal approach to identify the disease overlap remains unclear. METHODS: Pretest probability for HFpEF was determined using the HFpEF-ABA (age, body mass index, atrial fibrillation) algorithm among adjudicated group 3 or 4 patients at risk for pulmonary hypertension in the PVDOMICS study (Redefining Pulmonary Hypertension Through Pulmonary Vascular Disease Phenomics). Patients were stratified by current resting right heart catheterization criteria, and in a separate analysis, stratified only by HFpEF-ABA probability into low (<25%), intermediate, and high (≥75%) HFpEF probability groups. RESULTS: Among 598 patients with group 3 disease, 27% (n=161) had elevated pulmonary capillary wedge pressure (PCWP) with postcapillary pulmonary hypertension even at rest, which was associated with the highest exercise PCWP. However, regardless of this resting PCWP-based classification, a larger subset had intermediate-to-high HFpEF-ABA probabilities (32% [n=197] high and 57% [n=358] intermediate HFpEF probability). High HFpEF probability in group 3 disease was associated with higher resting and dynamic PCWP response to NO, fluid, and exercise ( P <0.0001 for all). These changes were comparable with more traditionally defined HFpEF without pulmonary vascular disease (n=61) but less severe than those with combined precapillary and postcapillary pulmonary hypertension HFpEF (n=31; interaction P =0.006). Increasing HFpEF probability in group 3 disease was associated with worse left heart remodeling, quality of life, 6-minute walk distance, and peak VO 2 ( P <0.0001 for all). Findings were replicated in group 4 disease (n=102). CONCLUSIONS: Quantifying pretest probability for HFpEF in patients with sleep apnea, lung disease, or chronic thromboemboli identifies a progressive gradient for dynamic PCWP abnormalities with worse functional status and quality of life. These subclinical left heart abnormalities are not universally detectable by resting right heart catheterization alone and call for further study of whether strategies to prevent or treat HFpEF might improve functional status in these patients with high risk of occult HFpEF.

Objective: This study aimed to evaluate the characteristics, clinical outcomes, and resource use of patients with unresectable hepatocellular carcinoma (uHCC) treated with first-line (1L) atezolizumab plus bevacizumab (A+B) at five United States (US) institutions: the Mayo Clinic, Houston Methodist, Moffitt Cancer Center, Mays Cancer Center, and University of Arizona. Methods: Treating oncologists extracted data from medical charts of patients with uHCC who were treated with A+B after 1 January 2019. Real-world progression-free survival (rwPFS) and overall survival (OS) were assessed using the Kaplan–Meier method for the overall cohort and for a “trial-like” subgroup with characteristics similar to those in the IMbrave150 trial (Eastern Cooperative Oncology Group Performance Status [ECOG PS] 0–1, Child–Pugh [CP] class A, albumin–bilirubin grade 1–2). Results: Of the 300 patients in the overall cohort (median age of 68 years; 12% ECOG PS ≥ 2; 73% CP A; 26% CP B; median follow-up of 8.7 months), the median rwPFS was 6.8 (95% confidence interval [CI]: 5.8, 8.4) months, and the median OS was 14.4 (95% CI: 12.3, 18.2) months. In the trial-like subgroup (n = 194), the median rwPFS was 8.8 (95% CI: 7.6, 12.1) months and the median OS was 19.5 (95% CI: 14.6, 24.7) months. A significantly lower proportion of patients with CP A compared with CP B (39.7% vs. 73.4%) experienced hospitalization within one year of A+B initiation, whereas hospitalizations due to treatment-related adverse events were similar. Conclusions: This study provides insights into the real-world effectiveness of 1L A+B in a diverse US patient cohort, with results from trial-like patients supporting the reproducible efficacy of A+B in clinical practice.

Klippel–Trenaunay (KT) syndrome and Cardiac Sarcoidosis (CS) are two distinct medical conditions that rarely coexist, each presenting unique challenges in diagnosis and management. Here, we present a rare case of a 30-year-old male with a history of KT syndrome complicated by recurrent deep vein thrombosis, who presented with symptoms of acute heart failure including shortness of breath, fatigue, dizziness, palpitations, and chest pain and was subsequently diagnosed with isolated CS. We discuss the importance of thorough clinical evaluation and multimodal diagnostic approaches in this complex scenario with overlapping symptoms and diagnostic dilemmas.