[Correlogic test with hi-res mass spec]
From the June 2004 issue of Endocrine Related Cancer
>>High-resolution serum proteomic features for ovarian cancer detection
Thomas P Conrads1, Vincent A Fusaro2,3, Sally Ross2, Don Johann2,3, Vinodh Rajapakse2,3, Ben A Hitt4, Seth M Steinberg5, Elise C Kohn3, David A Fishman6, Gordon Whiteley7, J. Carl Barrett8, Lance A Liotta3, Emanuel F Petricoin III2 and Timothy D Veenstra1
1National Cancer Institute Biomedical Proteomics Program, Laboratory of Proteomics and Analytical Technologies, SAIC-Frederick, Inc., National Cancer Institute at Frederick, Frederick, MD 21702
2Food and Drug Administration-National Cancer Institute Clinical Proteomics Program, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD 20892
3Laboratory of Pathology, National Cancer Institute, Center for Cancer Research, Bethesda, MD 20892
4Correlogic Systems Inc., Bethesda, MD 20892
5Biostatistics and Data Management Section, National Cancer Institute, Center for Cancer Research, Bethesda, MD 20892
6National Ovarian Cancer Early Detection Program, Northwestern University, Chicago, IL 60611
7Serum Proteomic Patterns Clinical Reference Laboratory, National Cancer Institute at Frederick, SAIC-Frederick Inc., Frederick, MD 21702)
8Center for Cancer Research, National Cancer Institute, Bethesda, MD 20892
(Requests for offprints should be addressed to Timothy D Veenstra, SAIC-Frederick, National Cancer Institute at Frederick, P.O. Box B, Frederick, MD 21702)
Serum proteomic pattern diagnostics is an emerging paradigm employing low resolution mass spectrometry (MS) to generate a set of biomarker classifiers. In the present study we utilized a well-controlled ovarian cancer serum study set to compare the sensitivity and specificity of serum proteomic diagnostic patterns acquired using a high resolution versus a low resolution MS platform. Using blinded testing sets, the high resolution mass spectral data contained multiple diagnostic signatures that were superior to the lower resolution spectra in terms of sensitivity and specificity (P<0.00001) throughout the range of modeling conditions. Four mass spectral feature set patterns acquired from data obtained exclusively using the high resolution mass spectrometer were 100% specific and sensitive in their diagnosis of serum samples as being acquired from either unaffected patients or those suffering from ovarian cancer. Important to the future of proteomic pattern diagnostics is the ability to statistically recognize inferior spectra, so that those resulting from a specific process error are recognized prior to their potentially incorrect (and damaging) diagnosis. To meet this need we have developed a series of quality assurance and in-process control procedures to a) globally evaluate sources of sample variability, b) identify outlying mass spectra, and c) develop quality control release specifications. Based on these quality assurance and control (QA/QC) specifications, we identified 32 mass spectra out of the total 248 that showed statistically significance differences from the norm. Hence, 216 of the initial 248 high resolution mass spectra were determined to be of high quality and were remodeled using pattern recognition analysis. Again, we obtained four mass spectral feature set patterns that also exhibited 100% sensitivity and specificity in blinded validation tests (68/68 cancer: including 18/18 stage I, and 43/43 healthy). We conclude that a) the use of high resolution MS yields superior classification patterns as compared with those obtained with lower resolution instrumentation, b) though the process error that we discovered did not have a deleterious impact on the present results obtained from proteomic pattern analysis, the major source of spectral variability emanated from mass spectral acquisition and not bias at the clinical collection site, c) this variability can be reduced and monitored through the use of QA/QC statistical procedures and d) multiple and distinct proteomic patterns comprised of low molecular weight biomarkers detected by high resolution MS achieve accuracies surpassing individual biomarkers warranting validation in an large clinical study.<<
Given the incidence rate of about 1.7% of U.S. women, the CIPH test, with its low sensitivity and high specificity, has poor positive predictive power, but excellent negative predictive power. That is we can't be sure if a woman that tests positive really has it, but we can be pretty sure (about 99%) that a woman who test negative really does not have it. In terms of positive predictive power, the sensitivity of CIPH's chosen biomarker set isn't good enough, little better than CA-125 alone.
Plugging in the CIPH numbers, 74% sensitivity, 94% specificity, incidence of 1.7%, I get a positive predictive power of 18.6%. Not even a one in five chance that the woman who test positive actually have ovarian cancer. Negative predictive power is much better; 99.5 times out of a hundred the woman who tests negative does not in fact have ovarian cancer. I use the equations given in the link below, with an arbitrary sample size of 500:
alpha.fdu.edu
Actually has Doesn't really
Ovarian have ovarian
Tests Positive 6.7 29.3 36
Tests Negative 2.3 461.7 464
9 491 500
More women who don't have ovarian but test positive on CIPH's test will have to undergo further testing versus those tested by Correlogic's. However, given the low incidence, this is not huge in absolute terms. CIPH's test, though being developed for SELDI based assays (versus ELISAs), will nevertheless be cheaper than Correlogic's by a factor of about two, I'd guess. One to two hundred dollars versus $200 to $400, because the central labs doing the assaying are actually going to use SELDI instruments and high res mass specs. The high res mass specs cost a lot more, as do the technicians who run them.
My information on the Correlogic test costs (and royalties) comes from this report from last July:
correlogic.com
If CIPH could equal the positive predictive power of the Correlogic test, I'd be a lot happier. I did not realize they were commercializing the Hopkins patterns.
So if I'm reading this right, CIPH commercializes a year ahead, but then Correlogic comes out with a better, though more expensive test. My guess is that anyone who tests positive on either one would be subject to the rest of the confirming process: ultrasound, CT scans, and biopsies. And they might have to run the full gamut, if the CIPH test's false positives are because it detects benign masses on the ovaries, in which case, the ultrasound and CT scans are going to be positive, but the final confirmation -- the biopsy -- will be negative. I would think insurers would prefer the more expensive, more accurate proteomic test, since the false positives of the cheaper one will be outweighed by the cost of confirming them as negative.
If Correlogic's OvaCheck can keep its 100% specificity and sensitivity in large studies, my guess is CIPH's OvaSure will have a short-lived success. Correlogic is also working on prostate cancer, but at this time I don't know how their results stack up against CIPH's.
Anyhow, here are questions I'd want answered by CIPH management:
Can you improve the sensitivity?
If not, will competing on cost be enough?
How much closer to market do you think you are, assuming the FDA views both tests as devices requiring a PMA to be filed?
Has the source of false positives been identified?
Cheers, Tuck |
