Clarification in the Point/Counterpoint Discussion Related to Surface-Enhanced Laser Desorption/Ionization Time-of-Flight Mass Spectrometric Identification of Patients with Adenocarcinomas of the Prostate

摘要

Articles on the detection of prostate cancer by surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) (1)(2) address important issues concerning this technology proposed for the early detection of disease. Dr. Diamandis (1) suggests that the two laboratories studying the early detection of prostate cancer by SELDI-TOF-MS (3)(4) should have identified some of the same peaks unless thousands of peaks separated disease from nondisease. Statistically, methodologically, and biologically identifying the same peaks in these two studies is actually not likely.Different peaks were identified because of different methods (e.g., different protein chips using completely different chemistries) and because of different approaches to analysis (5). Adam et al. (3) identified peaks at an initial mass-to-charge ratio ( m/z ) and matched peaks by considering peaks within ± 0.2% of molecular weight to be the same peak. Peaks with amplitudes that separated patients with prostate cancer from individuals without disease [prostate-specific antigen (PSA) ≤4 μg/L] or from those with benign prostatic hyperplasia (i.e., 4 μg/L PSA 10 μg/L) were put through a training analysis to separate optimally these three conditions by identifying a group of peaks with the best sensitivity and specificity. This optimized set was evaluated in a separate test set of patients. In contrast, Petricoin et al. (4) did not rely on peak identification and matching, but considered each m/z ratio as a separate variable with an associated amplitude. Using a training set, they mapped every m/z and its amplitude into n dimensional space to group patients … [?][1]aAddress correspondence to Dr. Ptericoin at: CBER/FDA, Bldg. 29A/2D12, 8800 Rockville Pike, Bethesda, MD 20892. Fax 301-480-5005; e-mail petricoin{at}cber.fda.gov; or Dr. Liotta at: CCR/NCI/NIH, 10 Center Drive, Bethesda, MD 20892; e-mail liottal{at}mail.nih.gov. [1]: #xref-corresp-2-1