Team Project

Lung Reference Set A Application: LaszloTakacs - Biosystems (2010)

Lung Ref Set A App: Takacs (2010)
Feng, ZidingFred Hutchinson Cancer Research Center
mAb candidates
No design specified.
Lung and Upper Aerodigestive Cancers Research Group

We would like to access the NCI lung cancer Combined Pre-Validation Reference Set A in order to further validate a lung cancer diagnostic test candidate. Our test is based on a panel of antibodies which have been tested on 4 different cohorts (see below, paragraph “Preliminary Data and Methods”). This Reference Set A, whose clinical setting is “Diagnosis of lung cancer”, will be used to validate the panel of monoclonal antibodies which have been demonstrated by extensive data analysis to provide the best discrimination between controls and Lung Cancer patient plasma samples, sensitivity and specificity values from ROC analyses are superior than 85 %.

Validate the panel of monoclonal antibodies
To complete the research phase of the project, we refine our algorithm. From this experiment, we will select the final panel of antibodies which will be used to develop the final diagnostic test. This final panel of antibodies used is selected as the panel showing the highest discrimination between the two groups (control vs. lung cancer). A learning algorithm based on LDA (Linear Discrimination Analysis) model is applied on each potential panel (called classifiers) and the performance of each classifier is evaluated calculating the AUC (area under the receiver operating curve) by Leave-One-Out-Cross validation (LOOCV). With the refined algorithm, we would like to test the Combined Pre-Validation Set A of the NCI lung cancer reference set.

No datasets are currently associated with this protocol.

Version 5.1.0