Alzheimer's disease (AD) can be predicted either by serum or plasma biomarkers, and a combination may increase predictive power, but due to the high complexity of machine learning, it may also incur overfitting problems. In this paper, we investigated whether combining serum and plasma biomarkers with feature selection could improve prediction performance for AD. 150 D patients and 150 normal controls (NCs) were enrolled for a serum test, and 100 patients and 100 NCs were enrolled for the plasma test. Among these, 79 ADs and 65 NCs had serum and plasma samples in common. A 10 times repeated 5-fold cross-validation model and a feature selection method were used to overcome the overfitting problem when serum and plasma biomarkers were combined. First, we tested to see if simply adding serum and plasma biomarkers improved prediction performance but also caused overfitting. Then we employed a feature selection algorithm we developed to overcome the overfitting problem. Lastly, we tested the prediction performance in a 10 times repeated 5-fold cross validation model for training and testing sets. We found that the combined biomarkers improved AD prediction but also caused overfitting. A further feature selection based on the combination of serum and plasma biomarkers solved the problem and produced an even higher prediction performance than either serum or plasma biomarkers on their own. The combined feature-selected serum-plasma biomarkers may have critical implications for understanding the pathophysiology of AD and for developing preventative treatments.
Abstract Background The A/T/N research framework for Alzheimer’s Disease calls for the importance of examining differences in these biomarkers across racial and ethnic groups. This study aims to address this need and examine the correlation and application between A/T/N plasma and neuroimaging biomarkers among a diverse cohort. Method Data were analyzed on Non‐Hispanic White (NHW) (n = 14 A+, n = 117 A‐; n = 4 T+, n = 83 T‐; n = 204 N+, 388 N‐) and Mexican American (MA) (n = 2 A+; n = 75 A‐; n = 3 T+, n = 27 T‐; n = 205 N+, n = 443 N‐) participants from the Health and Aging Brain Study‐ Health Disparities. Participants underwent a blood draw and neuroimaging (MRI, PET Amyloid and Tau). Plasma A/T/N biomarkers of Amyloid Beta (Aβ) 40, 42, Total Tau (T‐Tau), PTau‐181, and Neurofilament Light Chain (Nf‐L) were derived using Single Molecule Array (SIMOA) technology on HD‐X. Pearson correlations were conducted using plasma and neuroimaging A/T/N biomarkers. Due to the limited sample size of A+ and T+, support vector machine models were run in the total sample to examine the application of plasma biomarkers in predicting neuroimaging derived A+/T+/N+. Result In the total sample, Aβ 42 correlated with all PET amyloid ROIs. Ptau‐181 correlated with PET Tau SUVR in the Medial Temporal and Posterior Cingulate and clinical read of Tau +, same with T‐Tau. Nf‐L also correlated with MRI derived N+. Aβ 40 and 42 produced an AUC of 69% in detecting PET Amyloid + while T‐Tau and Ptau‐181 produced an AUC of 86% in detecting PET Tau +. Nf‐L alone produced an AUC of 66% in detecting MRI derived N+. Among MAs, Aβ 40 correlated with PET amyloid Frontal, Anterior Posterior Cingulate, Lateral Parietal, and Global SUVR while Ptau‐181 correlated with clinical read of Tau +. Nf‐L correlated with MRI derived N+ among MAs. Among NHWs, Aβ 40 and 42 correlated with clinical read of Amyloid +, while the later correlated with all PET amyloid ROIs. Neither T‐Tau, Ptau‐181 nor Nf‐L correlated with respective T or N neuroimaging biomarkers among NHWs. Conclusion This work supports prior findings and highlights ethnic specific differences in the interconnection between A/T/N plasma and neuroimaging biomarkers.
Background: Blood biomarkers have the potential to transform Alzheimer’s disease (AD) diagnosis and monitoring, yet their integration with common medical comorbidities remains insufficiently explored. Objective: This study aims to enhance blood biomarkers’ sensitivity, specificity, and predictive performance by incorporating comorbidities. We assess this integration’s efficacy in diagnostic classification using machine learning, hypothesizing that it can identify a confident set of predictive features. Methods: We analyzed data from 1,705 participants in the Health and Aging Brain Study-Health Disparities, including 116 AD patients, 261 with mild cognitive impairment, and 1,328 cognitively normal controls. Blood samples were assayed using electrochemiluminescence and single molecule array technology, alongside comorbidity data gathered through clinical interviews and medical records. We visually explored blood biomarker and comorbidity characteristics, developed a Feature Importance and SVM-based Leave-One-Out Recursive Feature Elimination (FI-SVM-RFE-LOO) method to optimize feature selection, and compared four models: Biomarker Only, Comorbidity Only, Biomarker and Comorbidity, and Feature-Selected Biomarker and Comorbidity. Results: The combination model incorporating 17 blood biomarkers and 12 comorbidity variables outperformed single-modal models, with NPV12 at 92.78%, AUC at 67.59%, and Sensitivity at 65.70%. Feature selection led to 22 chosen features, resulting in the highest performance, with NPV12 at 93.76%, AUC at 69.22%, and Sensitivity at 70.69%. Additionally, interpretative machine learning highlighted factors contributing to improved prediction performance. Conclusions: In conclusion, combining feature-selected biomarkers and comorbidities enhances prediction performance, while feature selection optimizes their integration. These findings hold promise for understanding AD pathophysiology and advancing preventive treatments.
Abstract Cerebrovascular disease is associated with symptoms and pathogenesis of Alzheimer's disease (AD) among adults with Down syndrome (DS). The cause of increased dementia‐related cerebrovascular disease in DS is unknown. We explored whether protein markers of neuroinflammation are associated with markers of cerebrovascular disease among adults with DS. Participants from the Alzheimer's disease in Down syndrome (ADDS) study with magnetic resonance imaging (MRI) scans and blood biomarker data were included. Support vector machine (SVM) analyses examined the relationship of blood‐based proteomic biomarkers with MRI‐defined cerebrovascular disease among participants characterized as having cognitive decline (n = 36, mean age ± SD = 53 ± 6.2) and as being cognitively stable (n = 78, mean age = 49 ± 6.4). Inflammatory and AD markers were associated with cerebrovascular disease, particularly among symptomatic individuals. The pattern suggested relatively greater inflammatory involvement among cognitively stable individuals and greater AD involvement among those with cognitively decline. The findings help to generate hypotheses that both inflammatory and AD markers are implicated in cerebrovascular disease among those with DS and point to potential mechanistic pathways for further examination.
Abstract Extrachromosomal circular DNAs (eccDNAs) have been reported in most eukaryotes. However, little is known about the cell-free eccDNA profiles in circulating system such as blood. To characterize plasma cell-free eccDNAs, we performed sequencing analysis in 26 libraries from three blood donors and negative controls. We identified thousands of unique plasma eccDNAs in the three subjects. We observed proportional eccDNA increase with initial DNA input. The detected eccDNAs were also associated with circular DNA enrichment efficiency. Increasing the sequencing depth in an additional sample identified many more eccDNAs with highly heterogenous molecular structure. Size distribution of eccDNAs varied significantly from 31 bp to 19,989 bp. We found significantly higher GC content in smaller eccDNAs (<500 bp) than the larger ones (>500 bp) (p < 0.01). We also found an enrichment of eccDNAs at exons and 3′UTR (enrichment folds from 1.36 to 3.1) as well as the DNase hypersensitive sites (1.58–2.42 fold), H3K4Me1 (1.23–1.42 fold) and H3K27Ac (1.33–1.62 fold) marks. Junction sequence analysis suggested fundamental role of nonhomologous end joining mechanism during eccDNA formation. Further characterization of the extracellular eccDNAs in peripheral blood will facilitate understanding of their molecular mechanisms and potential clinical utilities.
Abstract Cancers that are histologically defined as the same type of cancer often need a distinct therapy based on underlying heterogeneity; likewise, histologically disparate cancers can require similar treatment approaches due to intrinsic similarities. A comprehensive analysis integrated with drug response data and genomic alterations, particularly to reveal therapeutic concordance mechanisms across histologically disparate tumour subtypes, has not yet been fully exploited. In this study, we used pharmacogenomic profiling data provided from the Cancer Genome Project (CGP) in a systematic in silico investigation of the pharmacological subtypes of cancers and the intrinsic concordance of molecular mechanisms leading to similar therapeutic responses across histologically disparate tumour subtypes. We further developed a novel approach to redefine cell-to-cell similarity and drug-to-drug similarity from the therapeutic concordance, providing a new point of view to study cancer heterogeneity. Our study identified that histologically different tumours, such as malignant melanoma and colorectal adenocarcinoma, could belong to the same pharmacological subtype regarding drug sensitivity to MEK inhibitors, which was determined by their genomic alterations, high occurrence of BRAF or KRAS mutations. Therapeutic concordance for chemotherapy drugs was identified across histologically disparate hematological tumors mainly due to the extraordinary activation of the cell cycle in blood cancers. A subcluster of SCLC had a more similar profile with hematological tumors, and was associated with the malignant phenotype, with a higher level of MYC expression. We developed a website to store and visualize the pharmacological subtypes of drugs, as well as their connected genomic and expression alterations.
Detecting breast cancer at early stages can be challenging. Traditional mammography and tissue microarray that have been studied for early breast cancer detection and prediction have many drawbacks. Therefore, there is a need for more reliable diagnostic tools for early detection of breast cancer due to a number of factors and challenges. In the paper, we presented a five-marker panel approach based on SVM for early detection of breast cancer in peripheral blood and show how to use SVM to model the classification and prediction problem of early detection of breast cancer in peripheral blood. We found that the five-marker panel can improve the prediction performance (area under curve) in the testing data set from 0.5826 to 0.7879. Further pathway analysis showed that the top four five-marker panels are associated with signaling, steroid hormones, metabolism, immune system, and hemostasis, which are consistent with previous findings. Our prediction model can serve as a general model for multibiomarker panel discovery in early detection of other cancers.
Abstract The phosphatidylinositol 3 kinase (PI3K) pathway is the most frequently activated pathway across multiple tumor lineages and this is a potential therapeutic target. As a critical step to accelerate therapeutic development in endometrial cancer, we performed a comprehensive analysis of mutation and function of PI3K pathway members in a set of 243 highly characterized endometrial tumors. Whole gene resequencing revealed the highest frequencies of mutation in the PI3K pathway of any tumor lineage including PTEN (44%), PIK3CA (40%; encoding the p110α catalytic subunit of PI3K) and PIK3R1 (20%; encoding the p85a regulatory subunit). Indeed, when complete protein loss is considered, almost 80% of endometrioid endometrial cancers demonstrate an aberration in the PI3K pathway. Remarkably, mutations in the PI3K pathway were not mutually exclusive and indeed co-existence of PIK3CA or PIK3R1 mutation with heterozygous PTEN mutation occurred at frequencies higher than predicted by the frequency of each lesion alone. High-throughput reverse-phase protein array suggested that the dominant signaling effects occurred with PTEN protein loss and that PIK3CA or PIK3R1 mutation functionally mimic PTEN protein loss as indicated by protein markers including phosphorylated Akt and stathmin. Thus it appears likely that co-mutations in the PI3K pathway are selected to compensate haploinsufficiency of PTEN due to PTEN heterozygous mutation. To determine the functional consequences of mutations in PIK3R1, we demonstrated that several somatic PIK3R1 mutations tested conferred cytokine-independent growth on interleukin-3-dependent Ba/F3 cells and induced Akt phosphorylation in endometrial cancer cell lines indicating that they were gain of function mutations likely acting as oncogenes. Strikingly two of the gain of function mutations (E160* and R348*) lacked the ability to bind to the p110 catalytic subunit and E160* failed to bind to both p110 and PTEN. We demonstrate that E160* disrupts a novel mechanism of pathway regulation wherein p85 binds and stabilizes PTEN. Together, the data indicate that the PI3K pathway is targeted in the vast majority of endometrioid endometrial cancers representing a novel opportunity for implementation of targeted therapy. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr LB-247. doi:10.1158/1538-7445.AM2011-LB-247
Background: Despite the diagnostic accuracy of advanced neurodiagnostic procedures, the detection of Alzheimer’s disease (AD) remains poor in primary care. There is an urgent need for screening tools to aid in the detection of early AD. Objective: This study examines the predictive ability of structural retinal biomarkers in detecting cognitive impairment in a primary care setting. Methods: Participants were recruited from Alzheimer’s Disease in Primary Care (ADPC) study. As part of the ADPC Retinal Biomarker Study (ADPC RBS), visual acuity, an ocular history questionnaire, eye pressure, optical coherence tomography (OCT) imaging, and fundus imaging was performed. Results: Data were examined on n = 91 participants. The top biomarkers for predicting cognitive impairment included the inferior quadrant of the outer retinal layers, all four quadrants of the peripapillary retinal nerve fiber layer, and the inferior quadrant of the macular retinal nerve fiber layer. Conclusion: The current data provides strong support for continued investigation into structural retinal biomarkers, particularly the retinal nerve fiber layer, as screening tools for AD.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)