Highly Informative Analytical Platforms for Rapid, Non-Invasive Diagnosis and Stratification of Patients with Cancer

As the tissue that contains the largest representation of the human proteome, blood is the most important fluid for clinical diagnostics. However, although changes of plasma protein profiles reflect physiological or pathological conditions associated with many human diseases, only a handful of plasma proteins are routinely used in clinical tests. Reasons for this include the intrinsic complexity of the plasma proteome, the heterogeneity of human diseases and the rapid degradation of proteins in sampled blood. The first part of this thesis reports an integrated microfluidic system, the integrated blood barcode chip (IBBC) that can sensitively sample a large panel of protein biomarkers over broad concentration ranges and within 10 minutes of sample collection. It enables on-chip blood separation and rapid measurement of a panel of plasma proteins from quantities of whole blood as small as those obtained by a finger prick. The device holds potential for inexpensive, noninvasive and informative clinical diagnoses, particularly in point-of-care settings.;Proteomic approaches, on which the IBBC platform is based, have shown great promise in recent years for correctly classifying and diagnosing cancer patients. However, no large antibody-based microarray studies have yet been conducted to evaluate and validate plasma molecular signatures for detection of glioblastoma and monitoring of its response to therapy. In the second part of this thesis, plasma samples from 46 glioblastoma patients (72 total samples) are compared with those of 47 healthy controls with respect to the plasma levels of 35 different proteins known to be generally associated with tumor growth, survival, invasion, migration, and immune regulation. Average-linkage hierarchical clustering of the patient data stratified the two groups effectively, permitting accurate assignment of test samples into either GBM or healthy control groups with a sensitivity and specificity as high as 90% and 94%, respectively (when test samples within unbiased clusters were removed). The accuracy of these assignments improved (sensitivity and specificity as high as 94% and 96%, respectively) when the cluster analysis was repeated on increasingly trimmed sets of proteins that exhibited the most statistically significant (p < 0.05) differential expression. The diagnostic accuracy was also higher for test samples that fell into more homogeneous clusters. Intriguingly, test samples that fell within perfectly homogeneous clusters (all members belonging to the same group) could be diagnosed with 100% accuracy. Using the same 35-protein panel, we then analyzed plasma samples from GBM patients who were treated with the chemotherapeutic drug Avastin (Bevacizumab) in an effort to stratify patients based on treatment-responsiveness. Specifically, we compared 52 samples from (25) patients who exhibited tumor recurrence with 51 samples from (21) patients who did not exhibit recurrence. Again, several proteins were highly differentially expressed and cluster analysis provided effective stratification of patients between these two groups (sensitivity and specificity of 90% and 96%, respectively).