Prospective Validation Research Articles

Feasibility of a Pediatric Acute Video Consultation Process Among Health Care Professionals in Primary Care in a Rural Setting: Protocol for a Prospective Validation Study.

For years, in Catalonia and in the rest of Spain, there has been a deficit and an unequal geographical distribution of health professionals specializing in pediatrics, especially in rural areas. Among the proposals to improve this situation is the promotion of the use of information and communication technologies (ICT) among users and professionals. Moreover, with the outbreak of COVID-19, the use of telehealth has become an essential tool, with an overall increase in non-face-to-face visits, including in primary care pediatrics. In this context, telemedicine, when used in primary care pediatrics, can be an effective means of improving families' access to medical care. Currently, in Catalonia, telemedicine involving patients and health professionals is used in pediatric primary care through telephone consultation and asynchronous teleconsultation (eConsulta). Video consultation is in practice not used, although it could have different applications. The aim of this study is to evaluate the feasibility of a video consultation process with physical examination in acute pediatric pathology in rural areas among primary care professionals. In addition, the level of satisfaction with these remote consultations will be assessed from the perspective of both the users and the health care professionals. We will conduct a prospective experimental study to analyze the possibility of using video consultation in pediatric acute care in primary care in central Catalonia (Spain). A minimum of 170 children aged between 0 and 14 years attending the primary care center (PCC) for acute illness for a period of 1 year will be included in the study. Initially, the telemetric visit, including a physical examination, will include a nurse at the patient and family's side and a pediatrician who will participate remotely. Subsequently, the pediatrician will visit the patient in person and the physical examination and diagnosis made during the remote visit will be compared with the physical examination and diagnosis of the face-to-face visit, which is considered the gold standard. Recruitment was planned to begin in the second half of 2023 and continue for at least 1 year. It is anticipated to be a good resource for a variety of acute pediatric conditions in primary care. The evaluation will focus on the feasibility of performing live remote visits and comparing their diagnostic accuracy with that of face-to-face visits. We believe that this study could provide evidence on the feasibility and diagnostic accuracy of video consultation in pediatric acute primary care in a rural setting, as well as on satisfaction with video consultations among both users and professionals. If proven useful in addressing the acute needs of children in a variety of situations, it could become a digital health tool that improves the overall pediatric primary care service in rural areas, for both families and professionals. PRR1-10.2196/52946.

Design of a targeted blood transcriptional panel for monitoring immunological changes accompanying pregnancy.

Immunomodulatory processes exert steering functions throughout pregnancy. Detecting diversions from this physiologic immune clock may help identify pregnant women at risk for pregnancy-associated complications. We present results from a data-driven selection process to develop a targeted panel of mRNAs that may prove effective in detecting pregnancies diverting from the norm. Based on a de novo dataset from a resource-constrained setting and a dataset from a resource-rich area readily available in the public domain, whole blood gene expression profiles of uneventful pregnancies were captured at multiple time points during pregnancy. BloodGen3, a fixed blood transcriptional module repertoire, was employed to analyze and visualize gene expression patterns in the two datasets. Differentially expressed genes were identified by comparing their abundance to non-pregnant postpartum controls. The selection process for a targeted gene panel considered (i) transcript abundance in whole blood; (ii) degree of correlation with the BloodGen3 module; and (iii) pregnancy biology. We identified 176 transcripts that were complemented with eight housekeeping genes. Changes in transcript abundance were seen in the early stages of pregnancy and similar patterns were observed in both datasets. Functional gene annotation suggested significant changes in the lymphoid, prostaglandin and inflammation-associated compartments, when compared to the postpartum controls. The gene panel presented here holds promise for the development of predictive, targeted, transcriptional profiling assays. Such assays might become useful for monitoring of pregnant women, specifically to detect potential adverse events early. Prospective validation of this targeted assay, in-depth investigation of functional annotations of differentially expressed genes, and assessment of common pregnancy-associated complications with the aim to identify these early in pregnancy to improve pregnancy outcomes are the next steps.

Echocardiographic artificial intelligence for pulmonary hypertension classification

ObjectiveThe classification of pulmonary hypertension (PH) is crucial for determining the appropriate therapeutic strategy. We investigated whether machine learning (ML) algorithms may assist in echocardiographic PH prediction, where current guidelines...

Troponin testing in routine primary care: observations from a dynamic cohort study in the Amsterdam metropolitan area.

Troponin testing is indicated in the diagnostic work-up of acute coronary syndrome (ACS) and incorporated in risk stratification pathways. This study aims to gain insights on the use, outcomes, and diagnostic accuracy of troponin testing in routine primary care; a setting that is understudied. Routine data were used from the academic primary care network in the Amsterdam metropolitan area (968,433 patient records). The study population included adult patients who underwent high-sensitivity troponin I orT (hs-TnI/T) testing between 2011 and 2021. The primary outcome was the reported diagnosis and the secondary outcome was the diagnostic accuracy measured by death or ACS at 30days. 3,184 patients underwent hs-troponin testing, either with hsTNT (n=2,333) or hsTNI (n=851). Median patients' age was 55 (44-65) years, and 62.3 % were female. Predominant symptoms were chest pain and dyspnea (56.7 %). Additional diagnostic laboratory tests were commonly performed (CRP: 47.7 %, natriuretic peptides: 25.6 %, d-dimer: 21.5 %). Most common diagnoses were musculoskeletal symptoms (21.6 %) and coronary heart disease (7.1 %; 1.1 % ACS). Troponin testing showed sensitivity and specificity of 77.8 % (60.9-89.9) and94.3 % (93.5-95.1), respectively. Negative and positive predictive values were 99.7 (99.5-99.9) and 13.5 (11.1-16.4), andpositive and negative likelihood ratios were 13.7 (10.9-17.1) and 0.24 (0.13-0.43). GPs occasionally use troponin testing in very low-risk patients, often as part of a multi-marker rule-out strategy. The diagnostic characteristics of troponin tests, while promising, warrant prospective validation and implementation to facilitate appropriate use.

Conception and prospective multicentric validation of a Robotic Surgery Training Curriculum (RoSTraC) for surgical residents: from simulation via laboratory training to integration into the operation room

There is a lack of training curricula and educational concepts for robotic-assisted surgery (RAS). It remains unclear how surgical residents can be trained in this new technology and how robotics can be integrated into surgical residency training. The conception of a training curriculum for RAS addressing surgical residents resulted in a three-step training curriculum including multimodal learning contents: basics and simulation training of RAS (step 1), laboratory training on the institutional robotic system (step 2) and structured on-patient training in the operating room (step 3). For all three steps, learning content and video tutorials are provided via cloud-based access to allow self-contained training of the trainees. A prospective multicentric validation study was conducted including seven surgical residents. Transferability of acquired skills to a RAS procedure were analyzed using the GEARS score. All participants successfully completed RoSTraC within 1 year. Transferability of acquired RAS skills could be demonstrated using a RAS gastroenterostomy on a synthetic biological organ model. GEARS scores concerning this procedure improved significantly after completion of RoSTraC (17.1 (±5.8) vs. 23.1 (±4.9), p < 0.001). In step 3 of RoSTraC, all participants performed a median of 12 (range 5–21) RAS procedures on the console in the operation room. RoSTraC provides a highly standardized and comprehensive training curriculum for RAS for surgical residents. We could demonstrate that participating surgical residents acquired fundamental and advanced RAS skills. Finally, we could confirm that all surgical residents were successfully and safely embedded into the local RAS team.

Designing a Digital Health Solution: A Platform for Automated Surveillance of Fungal Infection.

Surveillance of invasive fungal infection (IFI) requires laborious review of multiple sources of clinical information, while applying complex criteria to effectively identify relevant infections. These processes can be automated using artificial intelligence (AI) methodologies, including applying natural language processing (NLP) to clinical reports. However, developing a practically useful automated IFI surveillance tool requires consideration of the implementation context. We employed the Design Thinking Framework (DTF) to focus on the needs of end users of the tool to ensure sustained user engagement and enable its prospective validation. DTF allowed iterative generation of ideas and refinement of the final digital health solution. We believe this approach is key to increasing the likelihood that the solution will be implemented in clinical practice.

Derivation and validation of novel integrated inpatient mortality prediction score for COVID-19 (IMPACT) using clinical, laboratory, and AI—processed radiological parameter upon admission: a multicentre study

Limited studies explore the use of AI for COVID-19 prognostication. This study investigates the relationship between AI-aided radiographic parameters, clinical and laboratory data, and mortality in hospitalized COVID-19 patients. We conducted a multicentre retrospective study. The derivation and validation cohort comprised of 512 and 137 confirmed COVID-19 patients, respectively. Variable selection for constructing an in-hospital mortality scoring model was performed using the least absolute shrinkage and selection operator, followed by logistic regression. The accuracy of the scoring model was assessed using the area under the receiver operating characteristic curve. The final model included eight variables: anosmia (OR: 0.280; 95%CI 0.095–0.826), dyspnoea (OR: 1.684; 95%CI 1.049–2.705), loss of consciousness (OR: 4.593; 95%CI 1.702–12.396), mean arterial pressure (OR: 0.928; 95%CI 0.900–0.957), peripheral oxygen saturation (OR: 0.981; 95%CI 0.967–0.996), neutrophil % (OR: 1.034; 95%CI 1.013–1.055), serum urea (OR: 1.018; 95%CI 1.010–1.026), affected lung area score (OR: 1.026; 95%CI 1.014–1.038). The Integrated Inpatient Mortality Prediction Score for COVID-19 (IMPACT) demonstrated a predictive value of 0.815 (95% CI 0.774–0.856) in the derivation cohort. Internal validation resulted in an AUROC of 0.770 (95% CI 0.661–0.879). Our study provides valuable evidence of the real-world application of AI in clinical settings. However, it is imperative to conduct prospective validation of our findings, preferably utilizing a control group and extending the application to broader populations.

P454 Systemic free thiols associate with clinical and biochemical disease activity in Inflammatory Bowel Disease: a prospective diagnostic validation study

Abstract Background In the pathophysiology of inflammatory bowel diseases (IBD), oxidative stress plays an important role. Lower levels of plasma free thiols (FT) reflect a higher oxidative stress burden and correlate with IBD disease activity. However, their clinical significance as biomarker for disease activity remains unclear. This study aimed to prospectively validate the utility of plasma FT levels as biomarker for clinical, biochemical, and endoscopic disease activity in patients with IBD. Methods In an ongoing prospective diagnostic trial aiming to include 100 IBD patients, we conducted an interim analysis of plasma FT levels in 52 patients with IBD (24 Crohn's disease [CD], 21 ulcerative colitis [UC], and 7 patients with undetermined IBD). Clinical disease activity was quantified by the Harvey-Bradshaw Index (HBI) for CD and Simple Clinical Colitis Activity Index (SCCAI) for UC. Biochemical disease activity was determined by measuring blood C-reactive protein [CRP] and faecal calprotectin [fCal]). Endoscopic disease activity was quantified with the Mayo endoscopic subscore for UC and the Simple Endoscopic Score (SES-CD) for CD. Plasma FT levels were assessed for correlations with disease activity parameters using linear and logistic regressions models. Discriminative capacity of biomarkers was evaluated through calculating receiver operating characteristics (ROC) statistics. Results Plasma FT levels were significantly negatively correlated in clinical parameters, like HBI (r = -0.53, p &amp;lt; 0.05) (Fig. 1A). Other inflammatory biomarkers, like CRP, were also negatively correlated with plasma FT levels and were reduced in patients with active IBD (r = -0.43, p &amp;lt; 0.05) (Fig. 1B). Plasma FT levels accurately discriminated between CD patients with clinically quiescent (HBI &amp;lt; 5) and active disease (HBI ≥ 5) (AUC = 0.80, p = 0.02, Fig. 1C), better than fCal (AUC = 0.55, p = 0.83). Plasma FT were not significantly different for biochemically quiescent (CRP &amp;lt; 5 mg/l) and active (CRP ≥ 5mg/l) disease (AUC = 0.69, p = 0.20, Fig. 1D). Both Mayo and SES-CD scores, after adjusting for confounders, also exhibited a trend towards negative association with plasma FT levels (Fig. 1E). Figure 1 Conclusion Further completion of the trial will reveal whether systemic FT levels will reflect endoscopically observed disease activity, and this interim analysis shows strong correlations FT levels and parameters of clinically and biochemically active disease. Measuring systemic FTs in IBD may be a promising, minimally invasive strategy to monitor IBD disease activity. Funding None

P772 Comprehensive disease control in patients with Ulcerative Colitis: refinement of a multi-component endpoint using data from the filgotinib SELECTION trial

Abstract Background Ulcerative colitis (UC) treatment requires a patient-centric efficacy endpoint to assess treatment benefits, enabling a shift towards a personalized approach. Comprehensive disease control (CDC) is an individual multi-component endpoint based on findings from a patient and expert Delphi consensus on remission assessment in UC. CDC was defined as IBDQ remission, partial Mayo Clinic Score (pMCS) remission, biomarker (faecal calprotectin [FCP]) remission and endoscopic improvement. We evaluated different definitions of CDC, and the contribution of different CDC components to the net treatment benefit (NTB) of filgotinib 200 mg (FIL200), an oral, once-daily, Janus kinase 1 inhibitor approved for UC treatment following the SELECTION trial (NCT02914522). Methods CDC was assessed at week 58 (W58) of SELECTION in adult patients with active UC treated with FIL200 or placebo (PBO). The CDC definitions evaluated were: 4-component CDC (CDC4) with FCP &amp;lt;150 µg/g, CDC4-150; CDC4 with FCP &amp;lt;250 µg/g, CDC4-250; 5-component CDC (with histological remission [CDC5]) with FCP &amp;lt;150 µg/g, CDC5-150; and CDC5 with FCP&amp;lt;250 µg/g, CDC5-250. The proportion of patients achieving each combination of components was visualized in UpSet plots. NTB of FIL200 was assessed for each endpoint using generalized pairwise comparisons and is reported here for CDC4-150 (and excluding different components). Results At W58, 21.6%, 26.6%, 19.1% and 22.6% of FIL200-treated patients achieved CDC4-150, CDC4-250, CDC5-150 and CDC5-250, respectively. Among patients who did not achieve CDC4-150, 0.5% of patients did not achieve pMCS remission compared with 5.0%, 10.1% and 6.0% who did not achieve IBDQ remission, biomarker remission and endoscopic improvement, respectively (Figure). Positive overall NTB of FIL200 versus PBO was demonstrated with all CDC definitions, including those with histological remission (CDC5). For CDC4-150 at W58, excluding pMCS remission did not substantially change the NTB (Δ −0.33; Table) versus excluding IBDQ or biomarker remission, or endoscopic improvement (Δ 3.87, 6.93 and 3.20, respectively). Similar NTB results were seen for all CDC definitions. Conclusion CDC is a stringent, patient-level, multi-component endpoint achieved by ~1/5 of FIL200-treated patients at W58. While CDC4-150 could be the optimal operational definition, the addition of histological remission resulted in similar stringency, potentially allowing for the use of different CDC definitions based on context or resources. Additional NTB of FIL200 was identified in patients when IBDQ or biomarker remission or endoscopic improvement were excluded. Overall, CDC appears to be a suitable endpoint to assess individual comprehensive benefit in patients with UC. Prospective validation is needed.

P267 Prospective validation of the IBD - Distribution, Chronicity, Activity (DCA) histological score for Ulcerative Colitis and correlation with UCEIS and MES

Abstract Background Histological assessment is not currently recognised as an independent parameter to guide personalised treatment for patients with ulcerative colitis (UC). The development of the Inflammatory Bowel Disease – Distribution, Chronicity, Activity (IBD-DCA) score allows a pragmatic approach to quantitatively evaluate mucosal injury and healing in UC, compared with other pre-existing histological scores (Lang-Schwarz et al., J Crohns Colitis, 2021). This scoring system facilitates a template-based reporting to produce consistent reports with information representing UC severity. Each domain of the IBD-DCA is scored as one of 0 (normal), 1 (mild), or 2 (moderate to severe) based on the presence of established histological UC findings (Figure 1). This project aims to prospectively validate the IBD-DCA score for routine clinical deployment and correlate with endoscopic scores – UC Endoscopic Index of Severity (UCEIS) and Mayo Endoscopic Score (MES). Methods Colonoscopic biopsies from patients with a previously confirmed diagnosis of UC attending one of four hospitals in a single healthcare group from August 2022 to January 2023 were reported using the IBD-DCA histological score, and the corresponding, clinician-reported UCEIS and MES scores were recorded. The histopathology reports were assessed based on the completeness of information in their respective D, C, and A domains. The proportion of reports with information fully represented by the IBD-DCA reporting template and the distribution of biopsy sites were described. The correlation between IBD-DCA with UCEIS and MES was quantified via Kendall’s tau (τ) using the R statistical package v4.3.2. Results Of the 151 cases reported, 70 (46.4%) patients were female and 81 (53.6%) were male. 12 patients (7.9%) had full colonic biopsies, while the remaining cases were assessed with selective biopsies. Rectal biopsy was most often performed (83.4%). The IBD-DCA components were reported for every biopsied sites in 133 (88.1%) cases. Notable correlations included UCEIS vascular pattern with histological chronicity (τ = 0.3891), UCEIS erosions/ulceration with histological activity (τ = 0.3252), and histological activity with the total UCEIS score (τ = 0.3450). MES showed the strongest correlation with histological disease distribution (τ = 0.3450)(Table 1). All p-values &amp;lt; 0.001. Conclusion These data show encouraging uptake of this simple histological reporting template in UC in daily routine use. The findings suggest that the IBD-DCA score may be a useful addition in the clinical management in patients with UC.

P213 ArgesMES: A high-performance, generalizable AI model for scoring Ulcerative Colitis severity from endoscopy videos

Abstract Background The Mayo endoscopic subscore (MES) is widely used to assess endoscopic disease severity assigned by human readers in Ulcerative Colitis (UC) clinical trials. AI-based automation of the MES could reduce inter-rater variability and allow for the development of more sensitive endoscopic measures. This report assesses whether a previously trained (locked) algorithm is suitable for automating full colon or segment-level MES scoring on a prospective UC clinical trial. Methods Endoscopy videos from two UC clinical trials (UNIFI: NCT02407236, Phase 3, 965 subjects, 3128 videos; and JAK-UC: NCT01959282, Phase 2, 211 subjects, 448 videos) were used to train an AI-based MES classifier, where 20% of the total data was retained as a holdout set. This AI model training had two steps: 1) training a feature extraction module using self-supervised learning (SSL), and 2) supervised training of a small transformer network with an attention-based classifier using SSL features to estimate full colon MES. Videos from an independent, prospective UC trial (QUASAR: NCT04033445, Phase 2b induction study, 313 subjects, 615 videos) were used to validate the locked AI model. MES scoring in QUASAR included full colon MES values and additional MES values for three left colon segments: descending colon, sigmoid colon, and rectum. Comparisons between AI-model and human reader scores were performed using AUC, Accuracy, F1 score, and Fleiss kappa. A non-inferiority test was also conducted to determine interchangeability between AI- and human-derived full colon MES values. Results Full colon MES on the QUASAR data showed AUC, Accuracy, and F1 scores of 0.810, 0.687, and 0.693, respectively, comparable to results obtained on the UNIFI holdout data (0.803, 0.645, and 0.647). The Fleiss kappa score was 0.682, comparable to the inter-rater agreement between two human readers-local and central readers (Fleiss kappa = 0.712). The non-inferiority test (p-value &amp;lt; 0.05) indicated that the AI-computed full colon MES readout was interchangeable to that of human readers. Similar performance was observed for the AI-computed segment-level MES: descending colon, sigmoid colon, and rectum as shown in Table 1. This result demonstrates the model's effectiveness at scoring the segment-level MES despite not being trained with segment level ground truth. Conclusion ArgesMES, an AI-based model, underwent successful prospective validation, demonstrating proficiency in automating full colon and segment-level MES scores. ArgesMES has the potential to facilitate rapid, reliable, and reproducible MES scoring at full colon and segment levels in prospective clinical trials.

The construct validity, responsiveness, reliability and interpretability of the Recap of atopic eczema questionnaire (RECAP) in children.

The Recap of atopic eczema questionnaire (RECAP) was developed to measure eczema control in patients with atopic dermatitis (AD). The measurement properties of RECAP have not yet been validated in caregivers of children with AD. To assess the construct validity, responsiveness, reliability and interpretability of the Dutch proxy version of RECAP. A prospective validation study was conducted in children (aged < 12 years) with AD and their caregivers (in a Dutch tertiary hospital). At three timepoints (T0 = baseline; T1 = after 1-7 days; T2 = after 4-8 weeks) RECAP and multiple reference instruments were completed by caregivers of child patients. Single- and change-score validity (responsiveness) were tested with a priori hypotheses on correlations with reference instruments. Intraclass correlation coefficients (ICCagreement) and standard error of agreement (SEMagreement) were reported. Bands for perceived eczema control were proposed. The smallest detectable change (SDC) and minimally important change (MIC) were determined. Two anchor-based methods based on receiver operating characteristic curve (ROC) and predictive modelling were used to determine the MIC. A total of 231 children with AD and their caregivers participated. Of our a priori hypotheses for single-score and change-score validity, 77% and 80% were confirmed, respectively. A stronger correlation than hypothesized was found for all rejected hypotheses.Excellent reliability was found (ICCagreement = 0.94, 95% confidence interval 0.90-0.96). The SEMagreement was 1.9 points. The final banding was 0-1 (completely controlled), 2-7 (mostly controlled), 8-12 (moderately controlled), 13-18 (a little controlled) and 19-28 (not at all controlled). A cutoff point of ≥ 8 was selected to identify children whose AD is not under control. The SDC was 5.3 and the MIC values were 1.5 and 3.6 for the ROC and predictive modelling approaches, respectively. No floor or ceiling effects were observed. The proxy version of RECAP is a valid, reliable and responsive measurement instrument for measuring eczema control in children with AD. An improvement of ≥ 6 points can be regarded as a real and important change in children with AD.

Use of tumor-informed ctDNA across gastrointestinal cancers.

758 Background: In colorectal cancer (CRC), the use of ctDNA for the purpose of early recurrence detection and prognostication is established, and multiple studies are testing its utility for guiding adjuvant therapy decisions. Tumor informed testing can be performed in clinical situations where adequate tissue is available for next-generation sequencing (NGS). Increasingly, ctDNA testing is being used clinically for colorectal cancer patients and other gastrointestinal (GI) cancers when there is an appropriate clinical scenario and adequate tumor tissue is available for sequencing. Methods: We identified GI cancer patients treated at our institution where Signatera (Natera) testing was ordered and performed between 2020 – 2023. We included all patients for whom testing was ordered. Then, among patients with any result further analyzed results by disease type, stage, and results. We examined clinical rationale for testing and whether results informed subsequent decisions. Results: A total of 99 GI cancer patients for whom ctDNA testing was ordered were identified. 80 patients had at least 1 ctDNA result including 68 colorectal patients and 12 other GI cancers (anal, bile duct, esophagogastric, pancreatic, and small bowel). Among patients with measurable ctDNA, at the most recent testing timepoint 16 had levels that met criteria for positive detection, while 64 were undetectable. Most commonly, testing was ordered in stage II/III CRC patients for post-operative surveillance. Other clinical use scenarios included 1) surveillance/treatment monitoring in oligometastatic CRC, 2) surveillance in non-metastatic esophagogastric cancer, 3) treatment monitoring in MSI-H/immunotherapy sensitive GI cancers. For at least 18 patients, results were a factor in shared clinical decision making between patients and providers. Conclusions: We found ctDNA monitoring is most commonly used in the adjuvant setting for CRC patients which is best supported by existing studies. Our data demonstrates that tumor informed testing can be successful in other GI cancers for both surveillance and therapy monitoring. Given the broad and rapid adoption of this technology, there needs to be continued efforts for both prospective validation studies and retrospective real world studies to guide clinical practice.

Malnutrition prevalence according to GLIM and its feasibility in geriatric patients: a prospective cross-sectional study.

In 2019, the Global Leadership Initiative on Malnutrition (GLIM) suggested a 2-step diagnostic format for malnutrition including screening and diagnosis. Prospective validation and feasibility studies, using the complete set of the five GLIM criteria, are needed. The aims of this study were to determine the prevalence of malnutrition, and investigate how the prevalence varied with mode of screening. Furthermore, we assessed the feasibility of GLIM in geriatric patients. Consecutive patients from two acute geriatric wards were included. For screening risk of malnutrition, the Mini Nutritional Assessment-Short Form (MNA-SF) or Malnutrition Screening Tool (MST) were used. In accordance with GLIM, a combination of phenotypic and etiologic criteria were required for the diagnosis of malnutrition. Feasibility was determined based on % data completeness, and above 80% completeness was considered feasible. One hundred patients (mean age 82years, 58% women) were included. After screening with MNA-SF malnutrition was confirmed by GLIM in 51%, as compared with 35% after screening with MST (p = 0.039). Corresponding prevalence was 58% with no prior screening. Using hand grip strength as a supportive measure for reduced muscle mass, 69% of the patients were malnourished. Feasibility varied between 70 and 100% for the different GLIM criteria, with calf circumference as a proxy for reduced muscle mass having the lowest feasibility. In acute geriatric patients, the prevalence of malnutrition according to GLIM varied depending on the screening tool used. In this setting, GLIM appears feasible, besides for the criterion of reduced muscle mass.

Predicting systemic diseases in fundus images: systematic review of setting, reporting, bias, and models' clinical availability in deep learning studies.

Analyzing fundus images with deep learning techniques is promising for screening systematic diseases. However, the quality of the rapidly increasing number of studies was variable and lacked systematic evaluation. To systematically review all the articles that aimed to predict systemic parameters and conditions using fundus image and deep learning, assessing their performance, and providing suggestions that would enable translation into clinical practice. Two major electronic databases (MEDLINE and EMBASE) were searched until August 22, 2023, with keywords 'deep learning' and 'fundus'. Studies using deep learning and fundus images to predict systematic parameters were included, and assessed in four aspects: study characteristics, transparent reporting, risk of bias, and clinical availability. Transparent reporting was assessed by the TRIPOD statement, while the risk of bias was assessed by PROBAST. 4969 articles were identified through systematic research. Thirty-one articles were included in the review. A variety of vascular and non-vascular diseases can be predicted by fundus images, including diabetes and related diseases (19%), sex (22%) and age (19%). Most of the studies focused on developed countries. The models' reporting was insufficient in determining sample size and missing data treatment according to the TRIPOD. Full access to datasets and code was also under-reported. 1/31(3.2%) study was classified as having a low risk of bias overall, whereas 30/31(96.8%) were classified as having a high risk of bias according to the PROBAST. 5/31(16.1%) of studies used prospective external validation cohorts. Only two (6.4%) described the study's calibration. The number of publications by year increased significantly from 2018 to 2023. However, only two models (6.5%) were applied to the device, and no model has been applied in clinical. Deep learning fundus images have shown great potential in predicting systematic conditions in clinical situations. Further work needs to be done to improve the methodology and clinical application.

Soluble lymphocyte activation gene-3 (sLAG3) and CD4/CD8 ratio dynamics as predictive biomarkers in patients undergoing immune checkpoint blockade for solid malignancies

BackgroundThe search for biomarkers to identify suitable candidates for immune checkpoint inhibitor (ICI) therapy remains ongoing. We evaluate how soluble levels of the next generation immune checkpoint Lymphocyte Activation Gene-3 (sLAG-3) and its association with circulating T lymphocyte subsets could pose as a novel biomarker to predict outcome to ICI therapy.MethodsCirculating levels of sLAG3 were analyzed using multiplex immunoassay in n = 84 patients undergoing ICI therapy for advanced solid cancer, accompanied by flow cytometry analyses of peripheral blood mononuclear cells (PBMCs).ResultsUni- and multivariate analysis shows that patients with higher sLAG3 concentrations before ICI therapy had a significantly impaired progression-free (PFS) and overall survival (OS) (HRPFS: 1.005 [95%CI: 1.000–1.009], p = 0.039; HROS: 1.006 [95%CI: 1.001–1.011], p = 0.015). The CD4/CD8 cell ratio and its dynamics during therapy were strong predictors of PFS and OS with patients with a decreasing ratio between baseline and after 1–2 cycles having an improved median OS compared to patients with increasing values (p = 0.012, HR: 3.32). An immunological score combining sLAG3 and the CD4/CD8 ratio showed the highest predictive potential (HROS: 10.3).ConclusionPending prospective validation, sLAG3 and correlating circulating T-cell subsets can be used as a non-invasive predictive marker to predict outcome to ICI therapy to help identifying ideal ICI candidates in the future.

Abstract A044: Tumor-infiltrating lymphocyte density correlates with survival in pancreatic adenosquamous carcinoma

Abstract Background: Pancreatic adenosquamous carcinoma (PASC) is a rare and aggressive form of pancreatic cancer, representing only 1-5% of cases. Histologically, it is defined by the presence of at least 30% squamous differentiation mixed with ductal adenocarcinoma. No standard therapies or prognostic biomarkers have been validated to date in this understudied disease. Methods: All patients diagnosed with pancreatic cancer in the Johns Hopkins Cancer Registry from 2013 to 2018 were queried for adenosquamous histology. Archived slides of available tumor specimens were examined to quantify the percentage of squamous differentiation and tumor-infiltrating lymphocyte (TIL) density. Descriptive statistics were used to summarize clinical, pathologic, and molecular characteristics. Kaplan-Meier estimates and log-rank tests were used for survival analyses. Results: Among a total of 8,908 patients queried, 155 patients with adenosquamous histology were identified. Median overall survival (OS) was 14.2 months and 7.5 months for resectable and unresectable/metastatic disease, respectively. Molecular tumor testing revealed frequent somatic mutations in KRAS, TP53, and SMAD4 in addition to multiple DNA damage repair genes. The percentage of squamous differentiation and TIL density was correlated with survival outcomes in a subset of 28 patients with evaluable tumor specimens. No association between the percentage of squamous differentiation and OS was identified. Of note, compared to those with low TIL density (0-10%), patients with an intermediate/high TIL density (&amp;gt;10%) exhibited significantly improved survival (median OS 36.3 vs 14.2 months, p = 0.011). Conclusions: For the first time, intermediate/high TIL density is associated with prolonged OS in a single-institution cohort of PASC patients. Prospective validation in larger cohorts is needed to establish prognostic biomarkers that can guide therapeutic decision-making in this lethal and under-examined disease subtype. Citation Format: Saurav D. Haldar, Christopher D. Jakubowski, Su Jin Lim, Hao Wang, Richard A. Burkhart, Kelly Lafaro, William R. Burns III, Jin He, Daniel A. Laheru, Elizabeth D. Thompson, Nilofer S. Azad. Tumor-infiltrating lymphocyte density correlates with survival in pancreatic adenosquamous carcinoma [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr A044.

Intra-tumoral T cells in pediatric brain tumors display clonal expansion and effector properties.

Brain tumors in children are a devastating disease in a high proportion of patients. Owing to inconsistent results in clinical trials in unstratified patients, the role of immunotherapy remains unclear. We performed an in-depth survey of the single-cell transcriptomes and clonal relationship of intra-tumoral T cells from children with brain tumors. Our results demonstrate that a large fraction of T cells in the tumor tissue are clonally expanded with the potential to recognize tumor antigens. Such clonally expanded T cells display enrichment of transcripts linked to effector function, tissue residency, immune checkpoints and signatures of neoantigen-specific T cells and immunotherapy response. We identify neoantigens in pediatric brain tumors and show that neoantigen-specific T cell gene signatures are linked to better survival outcomes. Notably, among the patients in our cohort, we observe substantial heterogeneity in the degree of clonal expansion and magnitude of T cell response. Our findings suggest that characterization of intra-tumoral T cell responses may enable selection of patients for immunotherapy, an approach that requires prospective validation in clinical trials.

Translation into Clinical Practice of the G1-G7 Molecular Subgroup Classification of Glioblastoma: Comprehensive Demographic and Molecular Pathway Profiling.

Glioblastoma is the most frequent and malignant primary neoplasm of the central nervous system. In a recent breakthrough study on a prospective Discovery cohort, I proposed the first all-inclusive molecular classification of glioblastoma into seven subgroups, G1-G7, based on MAPK pathway activation. New data from a WHO-grade-4 diffuse glioma prospective Validation cohort offers, in this study, an integrated demographic-molecular analysis of a 213-patient Combined cohort. Despite cohort differences in the median age and molecular subgroup distribution, all the prospectively-acquired cases from the Validation cohort mapped into one of the G1-G7 subgroups defined in the Discovery cohort. A younger age of onset, higher tumor mutation burden and expanded G1/EGFR-mutant and G3/NF1 glioblastoma subgroups characterized the glioblastomas from African American/Black relative to Caucasian/White patients. The three largest molecular subgroups were G1/EGFR, G3/NF1 and G7/Other. The fourth largest subgroup, G6/Multi-RTK, was detailed by describing a novel gene fusion ST7-MET, rare PTPRZ1-MET, LMNA-NTRK1 and GOPC-ROS1 fusions and their overexpression mechanisms in glioblastoma. The correlations between the MAPK pathway G1-G7 subgroups and the PI3-kinase/PTEN, TERT, cell cycle G1 phase and p53 pathways defined characteristic subgroup pathway profiles amenable to personalized targeted therapy. This analysis validated the first all-inclusive molecular classification of glioblastoma, showed significant demographic and molecular differences between subgroups, and provided the first ethnic molecular comparison of glioblastoma.

Computer-aided diagnosis of distal metastasis in non-small cell lung cancer by low-dose CT based radiomics and deep learning signatures.

This study aimed to develop and validate radiomics and deep learning (DL) signatures for predicting distal metastasis (DM) of non-small cell lung cancer (NSCLC) in low-dose computed tomography (LDCT). Images and clinical data were retrospectively collected for 381 NSCLC patients and prospectively collected for 114 patients at the Fifth Affiliated Hospital of Sun Yat-Sen University. Additionally, we enrolled 179 patients from the Jiangmen Central Hospital to externally validate the signatures. Machine-learning algorithms were employed to develop radiomics signature while the DL signature was developed using neural architecture search. The diagnostic efficiency was primarily quantified with the area under receiver operating characteristic curve (AUC). We interpreted the reasoning process of the radiomics signature and DL signature by radiomics voxel mapping and attention weight tracking. A total of 674 patients with pathologically-confirmed NSCLC were included from two institutions, with 143 of them having DM. The radiomics signature achieved AUCs of 0.885, 0.854, and 0.733 in the internal validation, prospective validation, and external validation while those for DL signature were 0.893, 0.786, and 0.780. The proposed signatures achieved a promising performance in predicting the DM of NSCLC and outperformed the approaches proposed in previous studies. Interpretability analysis revealed that both radiomics and DL signatures could detect the variations among voxels inside tumors, which helped in identifying the DM of NSCLC. Our study demonstrates the potential of LDCT-based radiomics and DL signatures for predicting DM in NSCLC. These signatures could help improve lung cancer screening regarding further diagnostic tests and treatment strategies.