755 Background: The phase II NEONAX trial examined PO and A gemcitabine/nab-paclitaxel (G/nP) efficacy for rPDAC pts with a comprehensive biomarker program. This translational study aimed at identifying liquid biomarker for early prediction of G/nP success in each treatment group (PO and A) and for the combined group using machine learning (ML). Methods: Pts from both study arms were stratified into groups based on Short- or Long-DFS ( n=80 pts total; PO n=42 pts; A n=38 pts). Blood plasma from selected G/nP naïve pts was analyzed by multiplexed ELISA (mELISA, 80-Plex ProcartaPlex Human Immune Response) and mass spectrometry (MS; feature list generation) and clinical data were acquired for each patient. For ML pts were divided into training (80%) and validation (20%) datasets. The Weka-based algorithm WrapperSubsetEval (WSE) with 8 different classifiers was used for feature selection. The best performing (highest accuracy, best ROC-AUC, minimal signature) classifier with the corresponding feature panel was selected by a 10x10-fold cross-validation (CV). Respective panels for all features combined as compared to clinical features and Ca19-9 blood levels were tested for performance via CV and bootstrap aggregating for training and validation datasets (10x with replacement). Results: The feature generation process generated 579 features from G/nP-naive pts (mELISA: 80; clinical data: 99; MS: 400). For response prediction of the whole rPDAC group to G/nP (S-/L-DFS), the ML-based flow identified a panel of 8 proteins from MS (SERPINA1, C1QB, KRT1, C4B, UBB, VCAM1, HPD, LYZ) and 3 proteins from mELISA (Galectin3, IL34, CCL4) combined with a logistic regression classifier. The predictive panel with the best performance for determining PO G/nP success was established using a kernel logistic regression classifier and included 2 proteins from mELISA (CXCL2, IL17A), 4 proteins from MS (C3, IGLV3-25, HLA-B, SHBG), and 2 clinical data (WHO grade, Na/Mg blood level ratio). The best performing biomarker for predicting success in A included 2 clinical data (tumor size at staging, hematocrit) and 1 protein from mELISA (IL22) and was established with a random forest classifier. All panels represented minimal feature signatures (rPDAC: 11; PO: 8; A; 3) and showed a high performance with ROC-AUC (training) > 0.90, ROC-AUC (CV) > 0.85, and ROC-AUC (validation) > 0.90. All panels described were superior compared to similar predictive signatures (with corresponding ML classifiers) for all clinical data or Ca19-9 blood levels. Conclusions: We show that minimal liquid biomarker signatures for early prediction of G/nP success in the NEONAX trial based on mELISA, MS and clinical data can be established by ML. The study shows the potential of ML for biomarker panel development and the value of mELISA/MS for generation of feature lists to use in ML. Clinical trial information: NCT02047513 .
Abstract BACKGROUND Glioblastoma are aggressive and incurable primary tumors of the central nervous system. As patients present with already established tumors, longitudinal imaging by MRI and [18F]FET-PET in human glioblastoma for characterizing temporal and spatial metabolic alterations is only possible to characterize alterations during therapy, but not during tumor evolution. In this regard, genetically engineered preclinical models are very relevant. The replication-competent avian sarcoma-leukosis virus (RCAS)/tumor virus receptor-A (tva) system provides a high grade of spatial and temporal control of somatic gene delivery in vivo. In this study, we used longitudinal MRI and [18F]FET-PET to characterize the evolution and growth of platelet-derived growth factor B (PDGFB)-driven glioma model using the RCAS-tva system. MATERIAL AND METHODS DF-1 cells transfected with PDGFB were implanted into adult 129S.Tg(NES-TVA)-Cdkn2a−/− mice intracranially. Mice received MRI twice and [18F]FET-PET measurements once per week. Longitudinal growth pattern and [18F]FET-PET uptake were assessed and analyzed. We measured symptom-free survival and performed immunohistochemistry on collected tumor tissue samples. RESULTS We observed robust tumor formation with highly reproducible symptom-free survival times of treatment-naïve experimental animals. Contrast-enhanced MRI volumetry revealed exponential growth dynamics starting around day 25 after DF-1 inoculation. Here, the PDGFB-driven gliomas present an escalating [18F]FET-PET uptake as well as increasing BBB permeability over time. These imaging features match the clinical scenario, which underlines the translational relevance of this model. Histology corroborated the imaging findings and an immunogenic glioma-associated microenvironment. CONCLUSION These data provide further valuable insights into this RCAS/tva-based PDGFB-driven murine glioma model and advocates for an imaging-stratified design of preclinical therapeutic interventions.
The Department of Ophthalmology, California Pacific Medical Center, San Francisco, California, USA; The Department of Ophthalmology, Stanford University School of Medicine, Stanford, California, USA; The Francis I. Proctor Foundation, UCSF School of Medicine, San Francisco, California, USA; Division of Ophthalmology, Faculty of Medicine and Health Sciences, Stellenbosch University, Cape Town, South Africa; Centre for Ophthalmology, University Tuebingen, Tuebingen, Germany
Proliferative vitreoretinopathy (PVR) is the major cause for surgical failure after primary rhegmatogenous retinal detachment (RRD). So far, no therapy has been proven to prevent PVR. Promising results for 5-fluorouracil (5-FU) and low-molecular weight heparin (LMWH) in high-risk eyes have been reported previously. The objective of this trial was to examine the effect of adjuvant intravitreal therapy with 5-FU and LMWH compared with placebo on incidence of PVR in high-risk patients with primary RRD.Randomized, double-blind, controlled, multicenter, interventional trial with 1 interim analysis.Patients with RRD who were considered to be at high risk for PVR were included. Risk of PVR was assessed by noninvasive aqueous flare measurement using laser flare photometry.Patients were randomized 1:1 to verum (200 mg/ml 5-FU and 5 IU/ml dalteparin) and placebo (balanced salt solution) intravitreally applied during routine pars plana vitrectomy.Primary end point was the development of PVR grade CP (full-thickness retinal folds or subretinal strands in clock hours located posterior to equator) 1 or higher within 12 weeks after surgery. For grading, an end point committee assessed fundus photographs. Secondary end points included best-corrected visual acuity and redetachment rate. A group sequential design with 1 interim analysis was applied using the O'Brien and Fleming boundaries. Proliferative vitreoretinopathy grade CP incidence was compared using a Mantel-Haenszel test stratified by surgeon.A total of 325 patients in 13 German trial sites had been randomized (verum, n = 163; placebo, n = 162). In study eyes, mean laser flare was 31 ± 26 pc/ms. No significant difference was found in PVR rate. Primary analysis in the modified intention-to-treat population results were: verum 28% vs. placebo 23% (including not assessable cases as failures); odds ratio [OR], 1.25; 95% confidence interval [CI], 0.76-2.08; P = 0.77. Those in the per-protocol population were: 12% vs. 12%; OR, 1.05; 95% CI, 0.47-2.34; P = 0.47. None of the secondary end points showed any significant difference between treatment groups. During the study period, no relevant safety risks were identified.Rate of PVR did not differ between adjuvant therapy with 5-FU and LMWH and placebo treatment in eyes with RRD.
