Lung Reference Set C Application: Rachel Ostroff - SomaLogics (2010)

Abbreviated Name
Lung Ref Set C App: Ostroff (2010)
Lead Investigator
Ostroff, Rachel — SomaLogic
Coordinating Investigator
Zheng, Yingye — Fred Hutchinson Cancer Center
Involved Investigators

Abstract

No abstract availalbe.

Aims

To use the EDRN lung biomarker serum set C to 1. Validate the biomarkers discovered in the train and verification studies 2. Refine the classifier for detection of lung cancer in individuals with indeterminant pulmonary nodules.

Analytic Method

This section describes SomaLogic’s statistical methods for biomarker discovery and classifier construction. These methods were applied to the data presented above, and similar approaches will be used with the EDRN samples. If our application is approved, the EDRN samples would be run on the SLaptamer array (approximately 800 analytes) and analysis completed by June 2009. The samples would then be run on subsequent larger arrays as they are developed. SomaLogic has experienced bioinformatics and statistics experts on staff as well as clinical biostatistics consultants. SomaLogic would nonetheless welcome an opportunity to collaborate with the extensive network of EDRN biostatisticians to learn new ways to identify and characterize potential biomarkers, and develop and optimize classifier algorithms for specific clinical applications. All sample information, including sample receipt, clinical data and assay data will be captured, linked and stored in the SomaLogic laboratory information management LIMS system. Statistical analysis will be conducted using parametric and non-parametric models constructed and implemented in java running in a windows environment. Back-ups of all project data are conducted daily, and all back-up data is stored both locally as well as at a SomaLogic redundant off-site location. There are two primary statistical problems to be solved in the proposed study: (i) identification of biomarkers that discriminate among the sample classes, and (ii) the construction of classifiers that allow for a new sample to be evaluated. The identification of statistically significant differential protein expression between the control samples and the disease samples will be performed using both parametric and nonparametric tests. The Kolmogorov-Smirnov (K-S) test is a non-parametric test that can be applied to marker identification. Similarly, Student’s t-test and an F-test can be used for marker discovery when data are normally distributed. Several classification methods will be applied to the datasets, including a parametric Bayesian classification scheme, linear discriminate analysis, and a nearest-neighbor classification scheme. The sample sets will be divided into train and test sets to assess the quality of the classifiers.

Comments

No comments available.

Outcome

Blood-based proteomic profiles were generated with SOMAscan technology, which measured 1033 proteins. First, preanalytic variability was evaluated with Sample Mapping Vectors (SMV), which are panels of proteins that detect confounders in protein levels related to sample collection. A subset of well collected serum samples not influenced by preanalytic variability was selected for discovery of lung cancer biomarkers. The impact of sample collection variation on these candidate markers was tested in the subset of samples with higher SMV scores so that the most robust markers could be used to create disease classifiers. The discovery sample set (n = 363) was from a multi-center study of 94 non-small cell lung cancer (NSCLC) cases and 269 long-term smokers and benign pulmonary nodule controls. The analysis resulted in a 7-marker panel with an AUC of 0.85 for all cases (68% adenocarcinoma, 32% squamous) and an AUC of 0.93 for squamous cell carcinoma in particular. This panel was validated by making blinded predictions in two independent cohorts (n = 138 in the first validation and n = 135 in the second). The model was recalibrated for a panel format prior to unblinding the second cohort. The AUCs overall were 0.81 and 0.77, and for squamous cell tumors alone were 0.89 and 0.87. The estimated negative predictive value for a 15% disease prevalence was 93% overall and 99% for squamous lung tumors. The proteins in the classifier function in destruction of the extracellular matrix, metabolic homeostasis and inflammation

Secure Outcome

No secure outcome available.

Publications

Biomarkers

Data Collections

  • No data collections available at this time for this protocol.
Start Date
May 1 2010
Estimated Finish Date
Aug 1 2010
Finish Date
Sep 1 2014
Protocol ID
317
Protocol Type
Reference Set
Fields of Research
  • Proteomics
Collaborative Group
Lung and Upper Aerodigestive Cancers Research Group
Cancer Types
  • Malignant neoplasm of bronchus and lung
Phased Status
1

Associated Forms