Citations

Background: Varicose veins (VVs) of the lower limbs, characterized by palpable, dilated, and tortuous veins, affect 2–73% of the global population. Artificial intelligence (AI) offers significant potential to enhance healthcare efficiency and decision-making, particularly in managing VVs through improved risk factor identification, diagnosis, and treatment planning. Objective: This abstract explores the role of AI in VV management, focusing on its applications in risk detection, image analysis, treatment planning, and surgical interventions, while addressing challenges to its widespread adoption. Methods: AI leverages advanced techniques such as computer vision and deep learning to analyze patient data, including medical history, symptoms, physical examinations, and imaging (e.g., ultrasounds, venography). It identifies patterns in large datasets to support personalized treatment plans, early risk detection, and disease severity assessment. Results: AI demonstrates promise in automating VV detection and classification, assessing disease severity, and aiding treatment planning. It enhances surgical interventions through preoperative planning, intraoperative navigation, and recurrence risk prediction. However, its adoption is limited by a lack of large-scale studies, concerns over accuracy, and the need for regulatory and ethical oversight. Conclusion: AI has the potential to revolutionize VV management by improving diagnosis, treatment precision, and patient outcomes. Further research, validation, and integration are critical to overcoming current limitations and fully realizing AI’s capabilities in clinical practice.

Abstract Introduction: Hemophagocytic lymphohistiocytosis (HLH) is a syndrome of aggressive inflammation and immune system activation. Its diagnosis is often hindered by its clinical overlap with sepsis, multiorgan failure, and other severe diseases common in the intensive care unit. We present a case of HLH diagnosed in the setting of multiple confounding conditions. Case Presentation: 24-year-old Caucasian male with common variable immunodeficiency presented to local hospital after a week of diffuse swelling and shortness of breath, initially found to be hypoxemic, hypoglycemic, hypovolemic, and febrile to 105°F. He was treated for septic shock, requiring norepinephrine, vasopressin, angiotensin, and stress dose hydrocortisone to maintain perfusion pressures. For work of breathing, he failed noninvasive ventilation and was intubated. Initial labs showed shock liver, acute kidney injury with severe metabolic acidosis, and MSSA bacteremia without endocarditis on transesophageal echocardiogram. He met HLH diagnostic criteria with high fever and ferritin 44,838 ng/mL; other inflammatory markers were not tested. Over two weeks the patient developed significant left lower extremity soft tissue infection concerning for Staphylococcus toxic shock syndrome, cholecystitis with cholecystostomy tube placement, hemorrhagic bullae on all limbs, left thigh hematoma, and new liver lesion that underwent biopsy complicated by significant bleeding requiring interventional radiology embolization. After transfer to our institution, he remained on fluctuating doses of norepinephrine with recurring fevers despite broad spectrum antibiotic and antifungal coverage. Repeat blood cultures were negative, MRI negative for osteomyelitis, pan-CT negative for other occult sources of infection, broad laboratory workup negative for atypical infections or antibodies related to rheumatologic diseases. At this point, he fulfilled HLH criteria given high fever, anemia and thrombocytopenia, splenomegaly, triglyceride 417 mg/dL, ferritin 1,953 ng/mL, and soluble IL-2 receptor >4,000 pg/mL. A multidisciplinary decision was made to start methylprednisone. Patient's fever and hemodynamic instability quickly resolved. He was then transitioned to anakinra with downtrend in his inflammatory markers. Discussion: HLH syndrome is often a delayed diagnosis as clinical manifestations mimic infections, can be triggered by such infections, or mimic other diseases in which a robust inflammatory response occurs. HLH laboratory findings are nonspecific or can have long turnaround time. Even its namesake of “hemophagocytosis” is not pathognomic for the disease. It is only in the case of excessive pathologic immune system activation, i.e. HLH syndrome, that immunosuppression is beneficial. Diagnosing HLH requires a curious clinical mind while treating concomitant urgent conditions, to avoid the “tunnel vision” of the often-cited occam's razor.

Abstract RATIONALE The best described endotype of COPD is alpha-1 antitrypsin (AAT) deficiency, which is due to a genetic abnormality in the SERPINA1 gene. Homozygotes for the Z allele (PI[asterisk]ZZ individuals) carry the genotype most commonly associated with severe AAT deficiency, but a highly prevalent endotype is the heterozygous state (PI[asterisk]MZ individuals). Here we test the hypothesis that individuals with PI[asterisk]MZ genotype are at higher risk for COPD exacerbation and healthcare utilization compared to those with no pathogenic variants in the SERPINA1 gene (PI[asterisk]MM individuals). METHODS Real-world data on 4,456 individuals with available AAT genotyping were extracted from the Cleveland Clinic Health System (CCHS) electronic health records (EHR) and analyzed to compare the risk of moderate COPD exacerbation (defined as a short-term steroid prescription), all-cause emergent care (defined as an express care visit, urgent care visit, or emergency department visit), and all-cause hospitalization between individuals with PI[asterisk]MZ vs. PI[asterisk]MM genotypes. Cox proportional hazards models were used to adjust for age, sex, race, BMI, smoking status, comorbidity count, liver disease diagnosis, and median income by zip code.To assess the effect of AAT levels on outcomes, we categorized participants into three groups based on AAT level and used 90 mg/dL as a threshold given it is the lower limit of the normal reference range and genotype: MM individuals with an AAT level >90 mg/dL, MZ >90 mg/dL and MZ <90 mg/dL. To assess the effect of PI[asterisk]MZ genotype among individuals with pre-existing lung disease, we performed a subgroup analysis of individuals with at least a six-month inhaler prescription. RESULTS In this cohort, 308 individuals had a PI[asterisk]MZ genotype. PI[asterisk]MZ was associated with increased risk for moderate COPD exacerbations (adjusted HR [95% CI]: 1.66 [1.27, 2.17]) and hospitalizations (adjusted HR [95% CI]: 1.44 [1.19, 1.75]) compared to PI[asterisk]MM. The risk of hospitalization was higher among PI[asterisk]MZ individuals with AAT levels <90 mg/dL (adjusted HR [95% CI]: 1.59 [1.14, 2.23]) but not in those with AAT levels >90 mg/dL, as compared to PI[asterisk]MM. Sensitivity analysis in those treated with inhaler therapy showed overall similar results. CONCLUSIONS Increased hospitalizations and moderate COPD exacerbations among PI[asterisk]MZ individuals compared to PI[asterisk]MM individuals merit future prospective studies to better characterize the longitudinal course and healthcare utilization among those with an PI[asterisk]MZ genotype. Given the high prevalence of these carriers globally, PI[asterisk]MZ represents a COPD phenotype that invites additional investigation to identify biomarkers of worse disease and treatable traits.

AbstractINTRODUCTIONIdentifying individuals with dementia is crucial for prevalence estimation and service planning, but reliable, scalable methods are lacking. We developed novel set algorithms using both structured and unstructured electronic health record (EHR) data, applying Diagnostic and Statistical Manual of Mental Disorders criteria for dementia case identification.METHODSOur cohort (n = 1082) included individuals aged ≥ 60 with dementia identified through specialist clinics and a comparison group without dementia. Clinicians from Australia and the United States informed predictor selection. We developed algorithms through a biostatistics stream for structured data and a natural language processing (NLP) stream for text, synthesizing results via logistic regression.RESULTSThe final structured model retained 16 variables (area under the receiver operating characteristic curve [AUC] 0.853, specificity 72.2%, sensitivity 80.6%). NLP classifiers (logistic regression, support vector machine, and random forest models) performed comparably. The final, combined model outperformed all others (AUC = 0.951, P < 0.001 for comparison to structured model).DISCUSSIONEmbedding text‐derived insights within algorithms trained on structured medical data significantly enhances dementia identification capacity.Highlights Algorithmic tools for detection of individuals with dementia are available; however, previous work has used heterogeneous case definitions which are not clinically meaningful, and has relied on proxies such as diagnostic codes or medications for case ascertainment. We used a novel, dual‐stream algorithmic development approach, simultaneously and separately modeling a clinically meaningful outcome (diagnosis of dementia according to specialized clinical impression) using structured and unstructured electronic health record datasets. Our clinically grounded case definition supported the inclusion of key structured variables (such as dementia International Classification of Disease codes and medications) as modeling predictors rather than outcomes. Our algorithms, published in detail to support validation and replication, represent a major step forward in the use of routinely collected data for detection of diagnosed dementia.

Abstract Background and Aims Classification and risk stratification in aortic (AR), mitral (MR), and tricuspid regurgitation (TR) remains a significant clinical challenge. This study aimed to develop an artificial intelligence (AI) system to assess valvular regurgitation and stratify MR-progression risk. Methods Using transthoracic echocardiograms (TTEs) at two sites (internal development/test, external test), the DELINEATE-Regurgitation system was developed to classify AR, MR, and TR severity using colour Doppler videos. Methods of summating video-level classifications into study-level predictions were tested, comparing single-view with multiview approaches integrating predictions across multiple videos. Model agreement with cardiologists was assessed by weighted kappa. A separate AI system (DELINEATE-MR-Progression) analysing colour Doppler videos was developed to predict which patients with mild, mild–moderate, and moderate MR were most likely to progress to moderate–severe or severe MR with analysis by Kaplan–Meier and Cox proportional hazards models. Results A total of 71 660 TTEs with 1 203 980 colour Doppler videos were included. The weighted kappa in internal/external test sets for regurgitation classification was 0.81/0.76 for AR, 0.76/0.72 for MR, and 0.73/0.64 for TR using a multiview approach taking all colour Doppler videos in a study, demonstrating substantial agreement with cardiologist interpretation with superiority of multiview over single view approaches. In the progression analysis, the AI score stratified MR-progression risk even when controlled for clinical factors known to be associated with MR progression [internal test set hazard ratio 4.1 (95% confidence interval 2.5–6.6)]. Conclusions An AI system can accurately classify AR, MR, and TR and predict MR progression beyond currently known risk factors.

Abstract Background Projectile and target fragmentation are nuclear phenomena that can influence the computation of the linear energy transfer (LET) and the relative biological effectiveness (RBE) in external radiotherapy with accelerated ions. Correlations between these two quantities are routinely established during radiobiological experiments to interpret the results and to develop and calibrate RBE models. Purpose This study systematically evaluates the impact of secondary fragments and irradiation scenarios on the dose‐averaged LET, the RBE, and their correlation in the case of exposures to clinically relevant ion beams. Methods 57 600 microdosimetric lineal energy spectra and corresponding LET distributions were simulated with the Monte Carlo code PHITS across different scenarios, including track segment calculations, pristine, and spread‐out Bragg peaks of 1 H, 4 He, 12 C, 16 O, and 20 Ne ions within water phantoms. The LET distributions were analyzed to calculate the dose‐average LET, both including or excluding the contribution of secondary ions of an element different from the primary beam. Similarly, the lineal energy distributions were processed in conjunction with the Mayo Clinic Florida microdosimetric kinetic model to compute the RBE for two theoretical cell lines ( α / β = 2 and 10 Gy in the case of 6 MV x‐rays). Correlations between the RBE and dose‐averaged LET were established by analyzing the simulation results within water phantoms and then compared to the corresponding trends from the track segment calculations. Results Excluding secondary fragments had a pronounced impact on the dose‐averaged LET and the RBE, particularly in the entrance regions of proton beams and close to the distal edge of heavier ions. The correlations between the RBE and the dose‐averaged LET were not universal, but highly dependent on the irradiation scenario. For proton beams only, the dose‐averaged LET of hydrogen ions served as a practical first‐order descriptor of the RBE. Track segment simulations, commonly used for calibrating and benchmarking RBE models, provided a reasonable approximation for low‐energy beams but failed to fully capture the complexity of polyenergetic radiation fields. Conclusions Secondary fragments can substantially affect the dose‐averaged LET and the RBE, even in proton beams. The dose‐averaged LET, including or not the contributions from secondary fragments, was generally unable to adequately capture RBE variations across different scenarios. A more comprehensive approach, considering microdosimetric distributions, is necessary to accurately describe RBE variations in ion therapy.

AbstractBreast cancer affects millions globally, necessitating precise biomarker testing for effective treatment. HER2 testing is crucial for guiding therapy, particularly with novel antibody‐drug conjugates (ADCs) like trastuzumab deruxtecan, which shows promise for breast cancers with low HER2 expression. Current HER2 testing methods, including immunohistochemistry (IHC) and in situ hybridization (ISH), have limitations. IHC, a semi‐quantitative assay, is prone to interobserver variability. While ISH provides higher precision than IHC, it remains more resource‐intensive in terms of cost and workflow. However, turnaround time is typically faster than that of other advanced molecular methods such as next‐generation sequencing. We adapted the clustering‐constrained‐attention multiple‐instance deep learning model to improve IHC testing and reduce dependence on reflex fluorescence ISH (FISH) tests. Using 5,731 HER2 IHC images, including 592 cases with FISH testing, we trained two models: one for predicting HER2 scores from IHC images and another for predicting FISH scores from equivocal cases. The HER2 IHC score prediction model achieved 91% ± 0.01 overall accuracy and a receiver operating characteristic (ROC) area under the curve (AUC) of 0.98 ± 0.01. The FISH score prediction model had an ROC AUC of 0.84 ± 0.07, with sensitivity at 0.37 ± 0.13 and specificity at 0.96 ± 0.03. External validation on cases from 203 institutions showed similar performance. The HER2 IHC model maintained a 91% ± 0.01 accuracy and an ROC AUC of 0.98 ± 0.01, while the FISH model had an ROC AUC of 0.75 ± 0.03, with sensitivity at 0.28 ± 0.04 and specificity at 0.93 ± 0.01. Our model advances HER2 scoring by reducing subjectivity and variability in current scoring methods. Despite lower accuracy and sensitivity in the FISH prediction model, it may be a beneficial option for settings where reflex FISH testing is unavailable or prohibitive. With high specificity, our model can serve as an effective screening tool, enhancing breast cancer diagnosis and treatment selection.

Abstract The substantial carbon footprint imparted by medical services warrants increased attention to their environmental impact. National guideline organizations such as the US Preventive Services Task Force (USPSTF) recommend multiple modalities for average-risk colorectal cancer (CRC) screening with varying resource intensity. The aim of this study was to quantify the environmental burden for 2 of the most used CRC screening modalities, colonoscopy and the multi-target stool DNA (mt-sDNA) test. A validated CRC microsimulation model was used to estimate the number of screening and follow-up tests for a cohort of 1 million average-risk individuals who underwent screening between ages 45 and 75. Component resources used for mt-sDNA, including waste products, energy, and transportation for colonoscopy and mt-sDNA, were collected from January 1, 2023, to January 1, 2024, and converted to carbon-equivalent emissions. Resources used for colonoscopy were captured from the literature. Resources devoted to screening colonoscopy were substantially (59%) higher than those to mt-sDNA, even when including follow-up colonoscopy. Of note, follow-up colonoscopy accounted for the majority (64%) of total emissions for the mt-sDNA screening strategy. Compared with colonoscopy screening, mt-sDNA substantially reduces the carbon emissions attributable to population-level CRC screening. Environmental impact should be included as a factor when choosing among guideline-recommended CRC screening strategies.