Urine Proteomic Analysis Between Calcium Stone Formers and Non-Stone Formers
Gregory Bernstein1, Stephen Brassell1, Lionel Banez*2, Premkala Prasanna*2, Shiv Srivastava*2, Thomas Novak*1, Noah Schenkman1
1Walter Reed Army Medical Center, Washington, DC;2Center for Prostate Disease Research, Rockville, MD
Introduction: Our goal was to detect differences between the urine proteomic patterns in stone forming patients and non-stone forming patients using surface enhanced laser desorption/ionization time of flight (SELDI-TOF) mass spectrometry.
Methods: The test group consisted of 50 male and 50 female patients, all with a diagnosis of calcium oxalate stones. The control group consisted of 50 male and 50 female patients without history of stone disease. Each sample was processed using a protocol of SELDI-TOF-MS analysis with WCX2 chips. Data analysis was carried out using Biomarker Wizard for peak clustering and Biomarkers Patterns Software (BPS), Support Vector Machine (SVM) and Artificial Neural Networks (ANN) for pattern detection.
Results: 163 samples were processed and analyzed. The samples were divided into a training set and a blinded validation set. The training set consisted of 42 stone-formers and 42 controls. The blinded validation set consisted of 37 controls and 42 stone-formers. With the training set; BPS, SVM and ANN were used to build classification algorithms to discriminate the two groups. The blinded validation set was then used to challenge the algorithms. Algorithms generated from combined SVM and ANN analysis gave a sensitivity of 83.3%, and specificity of 67.6% in detecting stone formers.
Conclusions:We present the initial experience of examining SELDI-TOF protein biomarker patterns to predict urinary stone formation. Urine proteomic analysis appears to have potential in defining the protein biomarkers that are characteristic of stone forming patients, however, refinement of techniques are needed.
Back to Final Program
