Prospective Validation Cohort Research Articles

Joint Hypermobility, Autonomic Dysfunction, Gastrointestinal Dysfunction and Autoimmune markers (JAG-A): Clinical Associations and Response to Intravenous Immunoglobulin Therapy.

We examined autoimmunity markers (AIM) and autonomic dysfunction in patients with chronic neurogastroenterological symptoms and their relationship to joint hypermobility/hypermobility spectrum disorder (JH/HSD). AIM positivity was defined as a diagnosis of known autoimmune/autoinflammatory disorder (AIDX) with at least one positive seromarker of autoimmunity or at least two positive seromarkers by themselves. Three cohorts were studied: (a) Retrospective (n = 300); (b) Prospective validation cohort (n =133); and (c) Treatment cohort (n=40), administered open-label intravenous immunoglobulin (IVIG). AIM positivity was found in 40% and 29% of the retrospective and prospective cohorts, the majority of whom (71% and 69%, respectively) had AIDX. Significantly more patients with AIM had elevations of C-reactive protein (31% versus 15%, p<0.001) along with an increased proportion of cardiovascular autonomic dysfunction (48% versus 29%; p<.001), small fiber neuropathy (20% versus 9%; p=.002).8) and HLADQ8 positivity (24% versus 13%, p=.01). JH/HSD patients were more likely to have AIM (43% versus 15%, p=.001) along with more severe autonomic and gastrointestinal symptom scores. IVIG treatment was associated with robust improvement in pain, gastrointestinal and autonomic symptoms but adverse events were experienced by 62% patients. Autoimmune markers and autonomic dysfunction are common in patients with unexplained gastrointestinal symptoms, especially in those with JH/HSD. Many patients seem to respond to IVIG treatment but this needs to be confirmed by controlled trials. These results highlight the need for vigilance for autoimmune and autonomic factors and JH/HSD in patients with neurogastroenterological disorders. Clinicaltrials.gov, NCT04859829.

Inter-alpha-trypsin inhibitor heavy chain H3 is a potential biomarker for disease activity in myasthenia gravis

Myasthenia gravis is a chronic antibody-mediated autoimmune disease disrupting neuromuscular synaptic transmission. Informative biomarkers remain an unmet need to stratify patients with active disease requiring intensified monitoring and therapy; their identification is the primary objective of this study. We applied mass spectrometry-based proteomic serum profiling for biomarker discovery. We studied an exploration and a prospective validation cohort consisting of 114 and 140 anti-acetylcholine receptor antibody (AChR-Ab)-positive myasthenia gravis patients, respectively. For downstream analysis, we applied a machine learning approach. Protein expression levels were confirmed by ELISA and compared to other myasthenic cohorts, in addition to myositis and neuropathy patients. Anti-AChR-Ab levels were determined by a radio receptor assay. Immunohistochemistry and immunofluorescence of intercostal muscle biopsies were employed for validation in addition to interactome studies of inter-alpha-trypsin inhibitor heavy chain H3 (ITIH3). Machine learning identified ITIH3 as potential serum biomarker reflective of disease activity. Serum levels correlated with disease activity scores in the exploration and validation cohort and were confirmed by ELISA. Lack of correlation between anti-AChR-Ab levels and clinical scores underlined the need for biomarkers. In a subgroup analysis, ITIH3 was indicative of treatment responses. Immunostaining of muscle specimens from these patients demonstrated ITIH3 localization at the neuromuscular endplates in myasthenia gravis but not in controls, thus providing a structural equivalent for our serological findings. Immunoprecipitation of ITIH3 and subsequent proteomics lead to identification of its interaction partners playing crucial roles in neuromuscular transmission. This study provides data on ITIH3 as a potential pathophysiological-relevant biomarker of disease activity in myasthenia gravis. Future studies are required to facilitate translation into clinical practice.

Placental growth factor promotes neural invasion and predicts disease prognosis in resectable pancreatic cancer

BackgroundSurgery represents the only curative treatment option for pancreatic ductal adenocarcinoma (PDAC), but recurrence in more than 85% of patients limits the success of curative-intent tumor resection. Neural invasion (NI), particularly the spread of tumor cells along nerves into extratumoral regions of the pancreas, constitutes a well-recognized risk factor for recurrence. Hence, monitoring and therapeutic targeting of NI offer the potential to stratify recurrence risk and improve recurrence-free survival. Based on the evolutionary conserved dual function of axon and vessel guidance molecules, we hypothesize that the proangiogenic vessel guidance factor placental growth factor (PlGF) fosters NI. To test this hypothesis, we correlated PlGF with NI in PDAC patient samples and functionally assessed its role for the interaction of tumor cells with nerves.MethodsSerum levels of PlGF and its soluble receptor sFlt1, and expression of PlGF mRNA transcripts in tumor tissues were determined by ELISA or qPCR in a retrospective discovery and a prospective validation cohort. Free circulating PlGF was calculated from the ratio PlGF/sFlt1. Incidence and extent of NI were quantified based on histomorphometric measurements and separately assessed for intratumoral and extratumoral nerves. PlGF function on reciprocal chemoattraction and directed neurite outgrowth was evaluated in co-cultures of PDAC cells with primary dorsal-root-ganglia neurons or Schwann cells using blocking anti-PlGF antibodies.ResultsElevated circulating levels of free PlGF correlated with NI and shorter overall survival in patients with PDAC qualifying for curative-intent surgery. Furthermore, high tissue PlGF mRNA transcript levels in patients undergoing curative-intent surgery correlated with a higher incidence and greater extent of NI spreading to tumor-distant extratumoral nerves. In turn, more abundant extratumoral NI predicted shorter disease-free and overall survival. Experimentally, PlGF facilitated directional and dynamic changes in neurite outgrowth of primary dorsal-root-ganglia neurons upon exposure to PDAC derived guidance and growth factors and supported mutual chemoattraction of tumor cells with neurons and Schwann cells.ConclusionOur translational results highlight PlGF as an axon guidance factor, which fosters neurite outgrowth and attracts tumor cells towards nerves. Hence, PlGF represents a promising circulating biomarker of NI and potential therapeutic target to improve the clinical outcome for patients with resectable PDAC.Graphical

MRI-based radiomics feature combined with tumor markers to predict TN staging of rectal cancer.

The aim of this study is to evaluate the predictive ability of MRI-based radiomics combined with tumor markers for TN staging in patients with rectal cancer and to develop a prediction model for TN staging. A total of 190 patients with rectal adenocarcinoma who underwent total mesorectal excision at the First Affiliated Hospital of the Air Force Medical University between January 2016 and December 2020 were included in the study. An additional 54 patients from a prospective validation cohort were included between August 2022 and August 2023. Preoperative tumor markers and MRI imaging data were collected from all enrolled patients. The 190 patients were divided into a training cohort (n = 133) and a validation cohort (n = 57). Radiomics features were extracted by outlining the region of interest (ROI) on T2WI sequence images. Feature selection and radiomics score (Rad-score) construction were performed using least absolute shrinkage and selection operator regression analysis (LASSO). The postoperative pathology TNM stage was used to differentiate locally advanced rectal cancer (T3/4 or N1/2) from locally early rectal cancer (T1/2, N0). Logistic regression was used to construct separate prediction models for T stage and N stage. The models' predictive performance was evaluated using DCA curves and calibration curves. The T staging model showed that Rad-score, based on 8 radiomics features, was an independent predictor of T staging. When combined with CEA, tumor diameter, mesoretal fascia (MRF), and extramural venous invasion (EMVI), it effectively differentiated between T1/2 and T3/4 stage rectal cancers in the training cohort (AUC 0.87 [95% CI: 0.81-0.93]). The N-staging model found that Rad-score, based on 10 radiomics features, was an independent predictor of N-staging. When combined with CA19.9, degree of differentiation, and EMVI, it effectively differentiated between N0 and N1/2 stage rectal cancers. The training cohort had an AUC of 0.84 (95% CI: 0.77-0.91). The calibration curves demonstrated good precision between the predicted and actual results. The DCA curves indicated that both sets of predictive models could provide net clinical benefits for diagnosis. MRI-based radiomics features are independent predictors of T staging and N staging. When combined with tumor markers, they have good predictive efficacy for TN staging of rectal cancer.

Classification of multiple primary lung cancer in patients with multifocal lung cancer: assessment of a machine learning approach using multidimensional genomic data.

Multiple primary lung cancer (MPLC) is an increasingly well-known clinical phenomenon. However, its molecular characterizations are poorly understood, and still lacks of effective method to distinguish it from intrapulmonary metastasis (IM). Herein, we propose an identification model based on molecular multidimensional analysis in order to accurately optimize treatment. A total of 112 Chinese lung cancers harboring at least two tumors (n = 270) were enrolled. We retrospectively selected 74 patients with 121 tumor pairs and randomly divided the tumor pairs into a training cohort and a test cohort in a 7:3 ratio. A novel model was established in training cohort, optimized for MPLC identification using comprehensive genomic profiling analyzed by a broad panel with 808 cancer-related genes, and evaluated in the test cohort and a prospective validation cohort of 38 patients with 112 tumors. We found differences in molecular characterizations between the two diseases and rigorously selected the characterizations to build an identification model. We evaluated the performance of the classifier using the test cohort data and observed an 89.5% percent agreement (PA) for MPLC and a 100.0% percent agreement for IM. The model showed an excellent area under the curve (AUC) of 0.947 and a 91.3% overall accuracy. Similarly, the assay achieved a considerable performance in the independent validation set with an AUC of 0.938 and an MPLC predictive value of 100%. More importantly, the MPLC predictive value of the classification achieved 100% in both the test set and validation cohort. Compared to our previous mutation-based method, the classifier showed better κ consistencies with clinical classification among all 112 patients (0.84 vs. 0.65, p <.01). These data provide novel evidence of MPLC-specific genomic characteristics and demonstrate that our one-step molecular classifier can accurately classify multifocal lung tumors as MPLC or IM, which suggested that broad panel NGS may be a useful tool for assisting with differential diagnoses.

Internal jugular vein pressure monitoring guided venous reconstruction could improve perioperative safety after superior vena cava resection for mediastinal tumors: a cohort study.

After superior vena cava (SVC) resection, the decision on unilateral or bilateral reconstruction was mostly based on the expertise of surgeons without objective measurements. This study explored the use of internal jugular vein pressure (IJVP) monitoring to guide the SVC reconstruction strategy. In a retrospective cohort, perioperative outcomes of unilateral and bilateral reconstruction based on surgeons' experience were compared. Then, IJVP threshold was measured when temporarily clamping the left innominate vein in a testing cohort. Venous reconstruction according to IJVP monitoring was performed in a prospective validation cohort afterward. Perioperative outcomes were compared between the prospective and the retrospective cohorts. For some interested variables, intuitive explanations would be given using Bayesian methods. Potential risk factors for postoperative complications were investigated by multivariable analysis. From March 2009 to September 2022, 57 patients underwent SVC reconstruction based on surgeons' experience. Bayesian analysis indicated a posterior probability of 80.49% that unilateral reconstruction had less blood loss than bilateral reconstruction (median 550ml vs. 1200ml). Cerebral edema occurred in two patients after unilateral reconstruction. In the testing cohort, median IJVP was 22.7 (18-27) cmH 2 O after temporary left innominate vein clamping in 10 patients. In the prospective cohort, unilateral reconstruction only was performed if the contralateral IJVP was <30cmH 2 O in 16 patients. Bilateral reconstruction was performed if IJVP was ≥30cmH 2 O after unilateral bypass in nine patients. No cerebral edema occurred in the prospective cohort. Less postoperative complications occurred in the prospective cohort than the retrospective cohort (12.0 vs. 38.6%, P =0.016). Upon multivariable analysis, IJVP-monitoring guided SVC reconstruction was associated with significantly less postoperative complications ( P =0.033). Intraoperative IJVP-monitoring is a useful strategy for selection of unilateral or bilateral SVC reconstruction and improving perioperative safety in patients with mediastinal tumors.

Machine Learning-Based Prediction of Malnutrition in Surgical In-Patients: A Validation Pilot Study.

Malnutrition in hospitalised patients can lead to serious complications, worse patient outcomes and longer hospital stays. State-of-the-art screening methods rely on scores, which need additional manual assessments causing higher workload. The aim of this prospective study was to validate a machine learning (ML)-based approach for an automated prediction of malnutrition in hospitalised patients. For 159 surgical in-patients, an assessment of malnutrition by dieticians was compared to the ML-based prediction conducted in the evening of admission. The model achieved an accuracy of 83.0% and an AUROC of 0.833 in the prospective validation cohort. The results of this pilot study indicate that an automated malnutrition screening could replace manual screening tools in hospitals.

Abstract 1167: Prospective analysis of pretreatment circulating CD4 memory T cell features associated with severe immunotherapy toxicity

Abstract Introduction: Immune checkpoint inhibitors (ICIs) have revolutionized melanoma treatment, significantly improving outcomes. Nevertheless, too many melanoma patients undergoing ICI treatment experience severe and often debilitating immune-related adverse events (irAEs). In a retrospective melanoma cohort treated with ICIs, we previously found that elevated baseline levels of circulating CD4 effector memory T (TEM) cells were associated with the development of severe irAEs, irrespective of the affected organ system (Lozano et al., Nature Medicine, 2022). Here, we present data from an extended prospective validation cohort from 47 patients across two academic medical centers. Methods: Of 47 patients, 26 received combination anti-PD1/anti-CTLA4 ICIs, 3 combination anti-PD1/anti-LAG3 ICIs, 17 anti-PD1 ICI, and 1 anti-CTLA4 ICI. Peripheral blood mononuclear cells (PBMCs) were prepared from pretreatment (Cycle 1 Day 1) blood samples obtained from all patients. We performed mass cytometry (CyTOF) on PBMCs using a 38-marker panel to deeply profile immune subsets. We further leveraged bulk RNA-seq from 24 patients to (i) deconvolve CD4 memory T cell abundance using CIBERSORTx and (ii) determine T cell receptor (TCR) diversity using Shannon entropy following MiXCR clonotype assembly. irAEs were graded on a 5-point scale using the Common Terminology Criteria for Adverse Events v5. Results: After treatment initiation, 14 patients experienced severe irAEs (grade 3+) at a median of 12.1 weeks (range 4-54), including 4 with life-threatening (grade 4) irAEs. CyTOF revealed elevated circulating CD4 TEM cells in patients that developed severe irAEs (P&amp;lt;0.0001; AUC=0.87), including those treated with combination ICIs (P=0.0001; AUC=0.89). We further observed a continuous relationship between increasing CD4 TEM levels and irAE grade (P=3e-6, Jonckheere-Terpstra test). We then combined CD4 memory T cell abundance and TCR diversity derived from bulk RNA-seq on 24 of these patients using logistic regression (trained on data from Lozano et al.) to yield a composite model score. The pretreatment composite model score was strongly associated with severe irAE development (P=0.0009, AUC=0.88), and in a median split, identified patients with shorter time to severe irAE development (3 months vs. not reached; P=0.006; HR=10.5). Notably, composite model features and the overall score had no significant association with immunotherapy benefit. Conclusion: In this prospective study, both circulating CD4 TEM levels determined by CyTOF and a composite model score measured by RNA-seq were strongly associated with severe irAE development, independent of response status. These findings hold promise for future pretreatment risk stratification of ICI therapy. Citation Format: Abul Usmani, Noah Earland, Wubing Zhang, Peter K. Harris, Antonietta Bacchiocchi, Chloe M. Sachs, Aishwarya Nene, David Y. Chen, Mario Sznol, Ruth Halaban, Aaron M. Newman, Aadel A. Chaudhuri. Prospective analysis of pretreatment circulating CD4 memory T cell features associated with severe immunotherapy toxicity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1167.

Prospective validation of the role of PET/CT in detecting disease after neoadjuvant chemotherapy in advanced ovarian cancer.

The study aimed to compare the diagnostic accuracies of 2-[18F]FDG PET/CT and contrast-enhanced CT (ceCT) after neoadjuvant chemotherapy (NACT) in advanced ovarian cancer (OC). This study consisted historical observational cohort and prospective validation cohort. Patients with newly diagnosed stage III-IV OC scheduled for NACT were recruited, with imaging performed after three to six cycles of NACT before interval debulking surgery. Nineteen regions in the abdominopelvic cavity were scored for the presence and absence of disease, referenced to the intra-operative findings or histological specimens. Diagnostic metrics were compared using McNemar's test. In the historical cohort (23 patients, age 58 ± 13), 2-[18F]FDG PET had an overall accuracy (Acc) 82%, sensitivity (Sen) 38%, specificity (Spe) 97%, positive predictive value (PPV) 79% and negative predictive value (NPV) 82%; ceCT had an overall Acc 86%, Sen 64%, Spe 93%, PPV 75% and NPV 89%. In the prospective cohort (46 patients, age 59 ± 9), 2-[18F] FDG PET had an overall Acc 87%, Sen 48%, Spe 98%, PPV 84% and NPV 88%; ceCT had an overall Acc 89%, Sen 66%, Spe 95%, PPV 77% and NPV 91%. No significant difference was demonstrated between the two imaging modalities (p > 0.05). High false-negative rates were observed in the right subdiaphragmatic space, omentum, bowel mesentery and serosa. High omental metabolic uptake after NACT was associated with histological non-responders (p < 0.05). 2-[18F]FDG PET/CT had no additional value over ceCT with comparable diagnostic accuracy in detecting disease after NACT in advanced OC. 2-[18F]FDG PET/CT is not superior to contrast-enhanced CT in determining disease after neoadjuvant chemotherapy in advanced ovarian cancer; contrast-enhanced CT should be suffice for surgical planning before interval debulking surgery. • Additional value of 2-[18F]FDG PET/CT over contrast-enhanced CT is undefined in detecting disease after neoadjuvant chemotherapy. • 2-[18F]FDG PET/CT has comparable diagnostic accuracy compared to contrast-enhanced CT. • Contrast-enhanced CT will be suffice for surgical planning after neoadjuvant chemotherapy.

Gut microbiome features and metabolites in non-alcoholic fatty liver disease among community-dwelling middle-aged and older adults

BackgroundThe specific microbiota and associated metabolites linked to non-alcoholic fatty liver disease (NAFLD) are still controversial. Thus, we aimed to understand how the core gut microbiota and metabolites impact NAFLD.MethodsThe data for the discovery cohort were collected from the Guangzhou Nutrition and Health Study (GNHS) follow-up conducted between 2014 and 2018. We collected 272 metadata points from 1546 individuals. The metadata were input into four interpretable machine learning models to identify important gut microbiota associated with NAFLD. These models were subsequently applied to two validation cohorts [the internal validation cohort (n = 377), and the prospective validation cohort (n = 749)] to assess generalizability. We constructed an individual microbiome risk score (MRS) based on the identified gut microbiota and conducted animal faecal microbiome transplantation experiment using faecal samples from individuals with different levels of MRS to determine the relationship between MRS and NAFLD. Additionally, we conducted targeted metabolomic sequencing of faecal samples to analyse potential metabolites.ResultsAmong the four machine learning models used, the lightGBM algorithm achieved the best performance. A total of 12 taxa-related features of the microbiota were selected by the lightGBM algorithm and further used to calculate the MRS. Increased MRS was positively associated with the presence of NAFLD, with odds ratio (OR) of 1.86 (1.72, 2.02) per 1-unit increase in MRS. An elevated abundance of the faecal microbiota (f__veillonellaceae) was associated with increased NAFLD risk, whereas f__rikenellaceae, f__barnesiellaceae, and s__adolescentis were associated with a decreased presence of NAFLD. Higher levels of specific gut microbiota-derived metabolites of bile acids (taurocholic acid) might be positively associated with both a higher MRS and NAFLD risk. FMT in mice further confirmed a causal association between a higher MRS and the development of NAFLD.ConclusionsWe confirmed that an alteration in the composition of the core gut microbiota might be biologically relevant to NAFLD development. Our work demonstrated the role of the microbiota in the development of NAFLD.

Exploring the potential of fully automated LUMIPULSE G plasma assays for detecting Alzheimer’s disease pathology

BackgroundLUMIPULSE G-automated immunoassays represent a widely used method for the quantification of Alzheimer’s disease (AD) biomarkers in the cerebrospinal fluid (CSF). Less invasive blood-based markers confer a promising tool for AD diagnosis at prodromal stages (mild cognitive impairment (MCI)). Highly sensitive assays for the quantification of amyloid-beta (Aβ) and phosphorylated Tau-181 (p-Tau181) in the blood are showing promising results. In this study, we evaluated the clinical performance of the recently available fully automated LUMIPULSE plasma marker assays for detecting brain AD pathology and for predicting progression from MCI to AD dementia stage.MethodsA retrospective exploratory cohort of 138 individuals (22 neurological controls [NC], 72 MCI, and 44 AD dementia patients) was included. Data regarding baseline CSF concentrations of Aβ42, Aβ40, t-Tau, and p-Tau181 was available and used to establish the presence of AD brain pathology. Baseline Aβ42, Aβ40, and p-Tau181 concentrations were determined in stored plasma samples using high-throughput fully automated LUMIPULSE assays. Progression from MCI to AD dementia was evaluated during follow-up (mean 6.4 ± 2.5 years). Moreover, a prospective validation cohort of 72 individuals with memory complaints underwent AD biomarker quantification, closely mirroring typical clinical practice. This cohort aimed to confirm the study’s main findings.ResultsIn the exploratory cohort, correlations between CSF and plasma were moderate for p-Tau181 (ρ = 0.61, p < 0.001) and weak for Aβ42/Aβ40 ratio (ρ = 0.39, p < 0.001). Plasma p-Tau181 and p-Tau181/Aβ42 concentrations were significantly increased while Aβ42/Aβ40 was significantly decreased (p < 0.001) in patients with AD dementia and prodromal AD, as well as in individuals with CSF abnormal amyloid concentrations (A +). Plasma p-Tau181 showed a robust performance in differentiating patients clinically diagnosed as AD (AUC = 0.89; 95% CI 0.83–0.94); A + vs. A − (AUC = 0.84, 95% CI 0.77–0.91) and also in predicting conversion to AD dementia in MCI patients (AUC = 0.89, 95% CI 0.81–0.96). When tested in the validation cohort, plasma p-Tau181 displayed 83.3% of the overall percentage of agreement according to amyloid status.ConclusionsOur results show that the measurement of p-Tau181 in plasma has great potential as a non-invasive prognostic screening tool for implementation in a clinical setting.

Exploring metformin monotherapy response in Type-2 diabetes: Computational insights through clinical, genomic, and proteomic markers using machine learning algorithms

BackgroundIn 2016, the UK had 4.5 million people with diabetes, predominantly Type-2 Diabetes Mellitus (T2DM). The NHS allocates £10 billion (9% of its budget) to manage diabetes. Metformin is the primary treatment for T2DM, but 35% of patients don't benefit from it, leading to complications. This study aims to delve into metformin's efficacy using clinical, genomic, and proteomic data to uncover new biomarkers and build a Machine Learning predictor for early metformin response detection. MethodsHere we report analysis from a T2DM dataset of individuals prescribed metformin monotherapy from the Diastrat cohort recruited at the Altnagelvin Area Hospital, Northern Ireland. ResultsIn the clinical data analysis, comparing responders (those achieving HbA1c ≤ 48 mmol/mol) to non-responders (with HbA1c > 48 mmol/mol), we identified that creatinine levels and bodyweight were more negatively correlated with response than non-response. In genomic analysis, we identified statistically significant (p-value <0.05) variants rs6551649 (LPHN3), rs6551654 (LPHN3), rs4495065 (LPHN3) and rs7940817 (TRPC6) which appear to differentiate the responders and non-responders. In proteomic analysis, we identified 15 statistically significant (p-value <0.05, q-value <0.05) proteomic markers that differentiate controls, responders, non-responders and treatment groups, out of which the most significant were HAOX1, CCL17 and PAI that had fold change ∼2. A machine learning model was build; the best model predicted non-responders with 83% classification accuracy. ConclusionFurther testing in prospective validation cohorts is required to determine the clinical utility of the proposed model.

Deep learning combining mammography and ultrasound images to predict the malignancy of BI-RADS US 4A lesions in women with dense breasts: a diagnostic study.

The authors aimed to assess the performance of a deep learning (DL) model, based on a combination of ultrasound (US) and mammography (MG) images, for predicting malignancy in breast lesions categorized as Breast Imaging Reporting and Data System (BI-RADS) US 4A in diagnostic patients with dense breasts. A total of 992 patients were randomly allocated into the training cohort and the test cohort at a proportion of 4:1. Another, 218 patients were enrolled to form a prospective validation cohort. The DL model was developed by incorporating both US and MG images. The predictive performance of the combined DL model for malignancy was evaluated by sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). The combined DL model was then compared to a clinical nomogram model and to the DL model trained using US image only and to that trained MG image only. The combined DL model showed satisfactory diagnostic performance for predicting malignancy in breast lesions, with an AUC of 0.940 (95% CI: 0.874-1.000) in the test cohort, and an AUC of 0.906 (95% CI: 0.817-0.995) in the validation cohort, which was significantly higher than the clinical nomogram model, and the DL model for US or MG alone ( P <0.05). The study developed an objective DL model combining both US and MG imaging features, which was proven to be more accurate for predicting malignancy in the BI-RADS US 4A breast lesions of patients with dense breasts. This model may then be used to more accurately guide clinicians' choices about whether performing biopsies in breast cancer diagnosis.

Assessing the clinical utility of rapid post-therapy whole-body digital SPECT/CT in evaluating early treatment response of 177Lu-PSMA-617 treatment.

32 Background: Multi-detector cadmium-zinc-telluride (CZT) based whole-body digital SPECT/CT is a new generation of imaging system. We aim to assess the clinical utility of rapid post-therapy whole-body SPECT/CT scan in evaluating early treatment response of 177Lu-PSMA-617 treatment. Methods: We retrospectively reviewed patients with progressive metastatic castration resistant prostate cancer (mCRPC) who were treated with at least two cycles of 177Lu-PSMA-617 at our institution from June 2022 to June 2023. Post-therapy whole-body digital SPECT/CT (GE StarGuide) was performed from vertex to mid-thigh at 1-2 hours after 177Lu-PSMA-617 infusion. Post-therapy SPECT/CT images were quantified with MIM software. Lu177-PSMA positive lesions were delineated with liver parenchyma uptake as cut-off. Lu-PSMA positive total tumor volume (Lu-TTV), tumor SUVmax and SUVmean were obtained. Post-therapy SPECT/CT image after cycle 1 was used as baseline for comparison with SPECT/CT after cycle 2 and 3. Overall survival, PSA progression free survival (PSA-PFS) as defined by PCWG3 and PSA decline &gt; 50% from baseline (PSA50) at any time after treatment were measured. Changes in post-therapy SPECT/CT were correlated with clinical outcomes to assess quantitative SPECT/CT as a tool for early treatment response. Analyses were performed with SPSS. Results: A total of 56 patients (76±8, mean ± SD, range 60-93 years old) who were imaged with at least 2 post-therapy SPECT/CT were included in the analysis. Post-therapy whole body SPECT/CT was acquired in ~12 mins. All scanned patients tolerated the rapid whole-body SPECT/CT very well. With median follow up of 10 months, median PSA-PFS was 5.0 months (range 1.0-15 months), 33 of 56 patients (58.9%) achieved PSA50 at any time after treatment, and 42 of 56 patients (75%) were alive at data cutoff. Quantitative analysis of SPECT/CT images showed that 36 of 56 patients (64%) had a &gt;30% decrease in Lu-TTV on early follow-up post-therapy SPECT/CT after cycle 2 or 3. Kaplan-Meier survival analysis showed that a &gt;30% decrease in Lu-TTV was associated with longer overall survival (median not reached vs 6 months, P = 0.008) and longer PSA-PFS (6 months vs 1 months, P &lt; 0.001). Decrease in SUVmax or SUVmean, however, was not associated with PSA-PFS or overall survival. Conclusions: Rapid post-therapy whole-body digital SPECT/CT was well tolerated in post-therapy SPECT/CT imaging following 177Lu-PSMA-617 treatment. Quantitation of post-therapy SPECT/CT may be useful in evaluating early treatment response of 177Lu-PSMA-617. Prospective validation and larger patient cohorts are needed to further elucidate the role of post-therapy SPECT/CT imaging in guiding patient management in 177Lu-PSMA-617 therapy.

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.

Abstract A112: Liquid biopsy shows distinctive proteome of long-term recurrence free survivors after PDAC resection

Abstract Pancreatic ductal adenocarcinoma (PDAC)patients have a 5-year survival rate of less than 10%, partly due to 80% developing recurrence within 2 years after curative intent surgery. As surgical resectability criteria expand and techniques progress, better understanding of tumor biology regarding post-resection prognosis is needed to improve patient outcomes. Our study therefore investigated usefulness of liquid biopsy and proteomics analysis for stratification of post-resection risk groups. Patients undergoing standardized follow-up after resection of PDAC at the European Pancreas Center (Heidelberg, Germany) donated blood samples and clinical data after giving informed consent. Outcomes were extracted from the institutional database or Electronic Health Records and documented retrospectively. From all included patients (n=101) 7mL serum was drawn at different follow-up time points. Overall, 139 samples were collected. Protein expression values of 2953 proteins were determined via antibody microarrays. For analysis and statistics, GenePix 6.0, R and GraphPad Prism were used. First, a principal component analysis was performed on the complete data set. Then the data set was split in a training and test set and A subgroup of 43 patients experienced recurrence of PDAC within 24 months post resection (median: 11.39 months), classified as “early recurrence” (ER). Furthermore, 23 patients experienced recurrence between 26 and 144 months and were termed “late recurrence” (LR). 32 patients had a follow up &amp;gt; 26 months after resection and 6 months after sampling and remained recurrence-free; those were termed “long-term recurrence free survivors” (LT-RFS). Principal component analysis of protein microarray data revealed distinct clustering of LT-RFS together along the primary axis. Contrarily, all of the samples taken at time of recurrence were exclusively found in other quadrants, as were all samples but one of the ER subgroup preceeding diagnosis of recurrence. Samples of the LR subgroup appeared to shift from clustering with LT-RFS samples towards samples from ER and post-recurrence. Lasso regression yielded a panel of 10 proteins, with a specificity of 98.0%/90.9% and a sensitivity of 89.4%/65.0% (training/test set), for identification of samples from LT-RFS. The findings of our protein microarray analysis on PDAC recurrence monitoring suggest a distinct proteome of patients with RFS of &amp;gt;2a post resection. Furthermore, we found hints that the post-resection proteome undergoes a dynamic shift from recurrence-free phenotype towards a phenotype of recurrence. These data need validation in prospective cohorts to determine clinical relevance, however, they point towards opportunities offered by liquid biopsy for sophisticated disease monitoring post resection and for novel insights into mechanisms of disease in PDAC. A protein panel predictive of long-term RFS might enable personalized follow-up pathways and have a positive impact on frequency and invasiveness of post-resection diagnostic testing for PDAC patients. Citation Format: Teresa Colbatzky, Fawaz N. Al-Shaheri, Andrea S. Bauer, Nathalia Giese, Jörg D. Hoheisel, Ulrike Heger. Liquid biopsy shows distinctive proteome of long-term recurrence free survivors after PDAC resection [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 A112.

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.

Development of a Novel Biomarker for the Progression of Idiopathic Pulmonary Fibrosis.

The progression of idiopathic pulmonary fibrosis (IPF) is diverse and unpredictable. We identified and validated a new biomarker for IPF progression. To identify a candidate gene to predict progression, we assessed differentially expressed genes in patients with advanced IPF compared with early IPF and controls in three lung sample cohorts. Candidate gene expression was confirmed using immunohistochemistry and Western blotting of lung tissue samples from an independent IPF clinical cohort. Biomarker potential was assessed using an enzyme-linked immunosorbent assay of serum samples from the retrospective validation cohort. We verified that the final candidate gene reflected the progression of IPF in a prospective validation cohort. In the RNA-seq comparative analysis of lung tissues, CD276, COL7A1, CTSB, GLI2, PIK3R2, PRAF2, IGF2BP3, and NUPR1 were up-regulated, and ADAMTS8 was down-regulated in the samples of advanced IPF. Only CTSB showed significant differences in expression based on Western blotting (n = 12; p < 0.001) and immunohistochemistry between the three groups of the independent IPF cohort. In the retrospective validation cohort (n = 78), serum CTSB levels were higher in the progressive group (n = 25) than in the control (n = 29, mean 7.37 ng/mL vs. 2.70 ng/mL, p < 0.001) and nonprogressive groups (n = 24, mean 7.37 ng/mL vs. 2.56 ng/mL, p < 0.001). In the prospective validation cohort (n = 129), serum CTSB levels were higher in the progressive group than in the nonprogressive group (mean 8.30 ng/mL vs. 3.00 ng/mL, p < 0.001). After adjusting for baseline FVC, we found that CTSB was independently associated with IPF progression (adjusted OR = 2.61, p < 0.001). Serum CTSB levels significantly predicted IPF progression (AUC = 0.944, p < 0.001). Serum CTSB level significantly distinguished the progression of IPF from the non-progression of IPF or healthy control.

Proteomics-driven noninvasive screening of circulating serum protein panels for the early diagnosis of hepatocellular carcinoma

Early diagnosis of hepatocellular carcinoma (HCC) lacks highly sensitive and specific protein biomarkers. Here, we describe a staged mass spectrometry (MS)-based discovery-verification-validation proteomics workflow to explore serum proteomic biomarkers for HCC early diagnosis in 1002 individuals. Machine learning model determined as P4 panel (HABP2, CD163, AFP and PIVKA-II) clearly distinguish HCC from liver cirrhosis (LC, AUC 0.979, sensitivity 0.925, specificity 0.915) and healthy individuals (HC, AUC 0.992, sensitivity 0.975, specificity 1.000) in an independent validation cohort, outperforming existing clinical prediction strategies. Furthermore, the P4 panel can accurately predict LC to HCC conversion (AUC 0.890, sensitivity 0.909, specificity 0.877) with predicting HCC at a median of 11.4 months prior to imaging in prospective external validation cohorts (No.: Keshen 2018_005_02 and NCT03588442). These results suggest that proteomics-driven serum biomarker discovery provides a valuable reference for the liquid biopsy, and has great potential to improve early diagnosis of HCC.

Co-occurring BRCA2/SPOP Mutations Predict Exceptional Poly (ADP-ribose) Polymerase Inhibitor Sensitivity in Metastatic Castration-Resistant Prostate Cancer

Background and objectiveBRCA2 mutations in metastatic castration-resistant prostate cancer (mCRPC) confer sensitivity to poly (ADP-ribose) polymerase (PARP) inhibitors. However, additional factors predicting PARP inhibitor efficacy in mCRPC are needed. Preclinical studies support a relationship between speckle-type POZ protein (SPOP) inactivation and PARP inhibitor sensitivity. We hypothesized that SPOP mutations may predict enhanced PARP inhibitor response in BRCA2-altered mCRPC. MethodsWe conducted a multicenter retrospective study involving 13 sites. We identified 131 patients with BRCA2-altered mCRPC treated with PARP inhibitors, 14 of which also carried concurrent SPOP mutations. The primary efficacy endpoint was prostate-specific antigen (PSA) response rate (≥50% PSA decline). The secondary endpoints were biochemical progression-free survival (PSA-PFS), clinical/radiographic progression-free survival (PFS), and overall survival (OS). These were compared by multivariable Cox proportional hazard models adjusting for age, tumor stage, baseline PSA level, Gleason sum, prior therapies, BRCA2 alteration types, and co-occurring mutations. Key findings and limitationsBaseline characteristics were similar between groups. PSA responses were observed in 60% (70/117) of patients with BRCA2mut/SPOPwt disease and in 86% (12/14) of patients with BRCA2mut/SPOPmut disease (p = 0.06). The median time on PARP inhibitor treatment was 24.0 mo (95% confidence interval [CI] 19.2 mo to not reached) in this group versus 8.0 mo (95% CI 6.1–10.9 mo) in patients with BRCA2 mutation alone (p = 0.05). In an unadjusted analysis, patients with BRCA2mut/SPOPmut disease experienced longer PSA-PFS (hazard ratio [HR] 0.33 [95% CI 0.15–0.72], p = 0.005) and clinical/radiographic PFS (HR 0.4 [95% CI 0.18–0.86], p = 0.02), and numerically longer OS (HR 0.4 [95% CI 0.15–1.12], p = 0.08). In a multivariable analysis including histology, Gleason sum, prior taxane, prior androgen receptor pathway inhibitor, stage, PSA, BRCA2 alteration characteristics, and other co-mutations, patients with BRCA2mut/SPOPmut disease experienced longer PSA-PFS (HR 0.16 [95% CI 0.05–0.47], adjusted p = 0.001), clinical/radiographic PFS (HR 0.28 [95% CI 0.1–0.81], adjusted p = 0.019), and OS (HR 0.19 [95% CI 0.05–0.69], adjusted p = 0.012). In a separate cohort of patients not treated with a PARP inhibitor, there was no difference in OS between patients with BRCA2mut/SPOPmut versus BRCA2mut/SPOPwt disease (HR 0.97 [95% CI 0.40–2.4], p = 0.94). In a genomic signature analysis, Catalog of Somatic Mutations in Cancer (COSMIC) SBS3 scores predictive of homologous recombination repair (HRR) defects were higher for BRCA2mut/SPOPmut than for BRCA2mut/SPOPwt disease (p = 0.04). This was a retrospective study, and additional prospective validation cohorts are needed. Conclusions and clinical implicationsIn this retrospective analysis, PARP inhibitors appeared more effective in patients with BRCA2mut/SPOPmut than in patients with BRCA2mut/SPOPwt mCRPC. This may be related to an increase in HRR defects in coaltered disease. Patient summaryIn this study, we demonstrate that co-alteration of both BRCA2 and SPOP predicts superior clinical outcomes to treatment with poly (ADP-ribose) polymerase (PARP) inhibitors than BRCA2 alteration without SPOP mutation.