High-grade astrocytomas are some of the most common and aggressive brain cancers, whose signs and symptoms are initially non-specific. Up to the present date, there are no diagnostic tools to observe the early onset of the disease. Here, we analyzed the combination of blood serum proteins, which may play key roles in the tumorigenesis and the progression of glial tumors. Fifty-nine astrocytoma patients and 43 control serums were analyzed using Custom Human Protein Antibody Arrays, including ten targets: ANGPT1, AREG, IGF1, IP10, MMP2, NCAM1, OPN, PAI1, TGFβ1, and TIMP1. The decision tree analysis indicates that serums ANGPT1, TIMP1, IP10, and TGFβ1 are promising combinations of targets for glioma diagnostic applications. The accuracy of the decision tree algorithm was 73.5% (75/102), which correctly classified 79.7% (47/59) astrocytomas and 65.1% (28/43) healthy controls. The analysis revealed that the relative value of osteopontin (OPN) protein level alone predicted the 12-month survival of glioblastoma (GBM) patients with the specificity of 84%, while the inclusion of the IP10 protein increased model predictability to 92.3%. In conclusion, the serum protein profiles of ANGPT1, TIMP1, IP10, and TGFβ1 were associated with the presence of astrocytoma independent of its malignancy grade, while OPN and IP10 were associated with GBM patient survival.
In this study, the applicability of commonly used Damage Frequency Method (DFM) is addressed in the context of Laser-Induced Damage Threshold (LIDT) testing with pulsed lasers. A simplified computer model representing the statistical interaction between laser irradiation and randomly distributed damage precursors is applied for Monte Carlo experiments. The reproducibility of LIDT predicted from DFM is examined under both idealized and realistic laser irradiation conditions by performing numerical 1-on-1 tests. A widely accepted linear fitting resulted in systematic errors when estimating LIDT and its error bars. For the same purpose, a Bayesian approach was proposed. A novel concept of parametric regression based on varying kernel and maximum likelihood fitting technique is introduced and studied. Such approach exhibited clear advantages over conventional linear fitting and led to more reproducible LIDT evaluation. Furthermore, LIDT error bars are obtained as a natural outcome of parametric fitting which exhibit realistic values. The proposed technique has been validated on two conventionally polished fused silica samples (355 nm, 5.7 ns).
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.