A major hurdle for blood-based proteomic diagnostics is efficient transport of specimens from the collection site to the testing laboratory. Dried blood spots have shown utility for diagnostic applications, specifically those where red blood cell hemolysis and contamination of specimens with hemoglobin is not confounding. Conversely, applications that are sensitive to the presence of the hemoglobin subunits require blood separation, which relies on centrifugation to collect plasma/serum, and then cold-chain custody during shipping. All these factors introduce complexities and potentially increased costs. Here we report on a novel whole blood-collection device (BCD) that efficiently separates the liquid from cellular components, minimizes hemolysis in the plasma fraction, and maintains protein integrity during ambient transport. The simplicity of the design makes the device ideal for field use. Whole blood is acquired through venipuncture and applied to the device with an exact volume pipette. The BCD design was based on lateral-flow principles in which whole blood was applied to a defined area, allowing two minutes for blood absorption into the separation membrane, then closed for shipment. The diagnostic utility of the device was further demonstrated with shipments from multiple sites (n = 33) across the U.S. sent to two different centralized laboratories for analyses using liquid chromatography/mass spectrometry (LC/MS/MS) and matrix assisted laser desorption/ionization-time of flight (MALDI-ToF) commercial assays. Specimens showed high levels of result label concordance for the LC/MS/MS assay (Negative Predictive Value = 98%) and MALDI-ToF assay (100% result concordance). The overall goal of the device is to simplify specimen transport to the laboratory and produce clinical test results equivalent to established collection methods.

Clinical oncology assays using mass spectrometry have exclusively been run on instrument platforms designed for use in research laboratories. Here, we report on the use of a bench-top MALDI-TOF mass spectrometer (MS) to run a serum proteomic Laboratory Developed Test (LDT), VeriStrat®. A number of previous studies have shown that this assay measures a serum proteomic fingerprint using MALDI-TOF MS which, following spectral acquisition and data processing, classifies advanced non-small cell lung cancer (NSCLC) patients into treatment categories. An extension of VeriStrat, BDX004, is currently under development as part of a global, randomized, double-blind Phase 2 clinical study, FOCAL. BDX004 will be used prospectively to select previously untreated, EGFR mutation-positive patients with advanced NSCLC to receive one of two treatments: (1) the combination of ficlatuzumab (a humanized monoclonal antibody against hepatocyte growth factor) and an EGFR tyrosine kinase inhibitor (TKI), or (2) an EGFR TKI with placebo. For the development of BDX004 as a companion diagnostic (CDx), we have included the microflex™ LT (LT), a component of Bruker's MALDI Biotyper CA system, which was granted FDA clearance for the identification of Gram negative bacteria. In the present study, we evaluated the LT platform through comparative studies with other MALDI-TOF instruments (autoflex™ III and autoflex SPEED) utilizing samples from patients previously diagnosed with NSCLC. In summary, we found that under common laboratory procedures the LT produces results equivalent to research grade MALDI-TOF MS instrumentation. In initial studies, the microflex LT showed similar resolution (R > 600 at 6632.1 m/z) and sensitivity (S/N ≥ 50 with 500 fmoles BSA in 100 laser shots) performance parameters as the autoflex series research instruments (R > 800 at 6632.1 m/z and S/N ≥ 50 with 500 fmoles BSA in 250 laser shots). The LT was qualified with several independent sample sets [n = 4 (8 replicates; 2000 spectra/replicate); n = 67 (3 replicates; 2000 spectra/replicate) and n = 20 (3 replicates; 2000 spectra/replicate)]. In these studies, the LT achieved 100%, >97%, and 100% concordance with reference data sets previously acquired on the autoflex platforms. As expected, due to the lower laser repetition rate of the LT's nitrogen laser, the acquisition speed was slower on the LT as compared to the autoflex instruments by a factor of two. However, the slower acquisition speed is partially compensated for with increased sensitivity, which is particularly evident in the reference sample mass spectra which have signal-to-noise ratios ∼1.5 greater than those acquired on the autoflex platform. Based on these results, the microflex LT appears to be suitable for measurement of serum proteomic profiles in NSCLC and suggests that the inclusion of the microflex LT in CDx test development has the potential to expand the utility of MALDI-based testing in the clinic. Citation Format: Nicholas Dupuis, Jamie Chang, Maximillian Steers, Zhigang Weng, Jeno Gyuris, Gary Pestano. Feasibility of serum proteomic companion diagnostic (CDx) test development on the microflex LT platform. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2011. doi:10.1158/1538-7445.AM2015-2011

BDX004 is a multivariate serum proteomics companion diagnostic (CDx) test currently under development as part of a global, randomized, double-blind Phase 2 clinical study, FOCAL. The test will be used prospectively to select previously untreated, EGFR mut-positive patients with advanced NSCLC to receive one of two treatments: (1) the combination of ficlatuzumab (a humanized mAb against hepatocyte growth factor) and an EGFR tyrosine kinase inhibitor (TKI), or (2) an EGFR TKI with placebo. BDX004 measures the abundance of eight serum protein variants using MALDI-TOF mass spectrometry, the data from which is coupled with pre-processing steps and a k-nearest neighbors (kNN) algorithm to classify subjects into positive or negative groups. To date, clinical oncology assays using mass spectrometry, including those to measure serum proteomic profiles in NSCLC, have exclusively been run on instruments designed for use as research use only. Here, we describe a comparative assessment of three MALDI-TOF platforms (microflex™, autoflex™ and ultraflex™). The microflex LT we used is an RUO version of Bruker Daltonics' FDA cleared MALDI Biotyper CA system. All three instrument platforms were qualified with at least two independent sample sets and achieved >97% classification label concordance (Positive and Negative) with a gold standard data set. In summary, the microflex produced results equivalent to the other, higher performance MALDI-TOF MS instruments. In initial studies, the microflex showed similar resolution (R > 600 at 6632.1 m/z) and sensitivity (S/N ≥ 50 with 500 fmoles BSA in 100 laser shots) performance parameters as the autoflex and ultraflex instruments. The microflex was further evaluated with several additional independent sample sets [n = 4 (8 replicates; 2000 spectra/replicate), n = 67 (3 replicates; 2000 spectra/replicate), and n = 20 (3 replicates; 2000 spectra/replicate)] and achieved 100%, >97%, and 100% classification label concordance against reference data sets. At a quantitative level we also measured the feature value (protein peak) CVs that underlie the classification labels. These were found to be highly comparable across all three platforms at 5.7%, 6.4% and 4.8% for the microflex, autoflex and ultraflex, respectively. We conclude that all three of the platforms tested demonstrated equivalent performance, and the microflex LT is well suited for the development of novel MALDI-TOF CDx tests. Citation Format: Nicholas F. Dupuis, Maximillian Steers, Amanda Weaver, Zhigang Weng, Francis C. Payumo, Michael N. Needle, Gary A. Pestano. Comparative assessment of MALDI-TOF platforms for utility in CDx development of a proteomic signature for ficlatuzumab. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr B128.