It is becoming clear that no one biomarker has the power necessary to identify individual patients likely to respond to particular oncology drugs. Even for drugs for which the target molecule is clearly defined, the mere presence of that target is insufficient for adequate classification of response.  Even for those breast carcinoma cases that express high levels of the tyrosine kinase receptor, HER-2, there is still only a 50% chance of response to antibodies that target this molecule.  Clearly, either the tests are imperfect or there are other factors involved in determining response to these drugs (or both).

Better success has been achieved through use of more than one analyte for classifying patients into drug-responsive subgroups. The diagnostic community is beginning to utilize complex analysis of tumor phenotypes using gene array technologies to segregate drug responsive groups. While promising, these technologies are extremely complex and fraught with error.

Pathogenesys has elected to define simple algorithms using only few biomarkers to define drug responsiveness. To reach this goal, a more complex database containing the semi-quantitative results one or more Immunohistochemical biomarkers, morphologic features obtained from histologic examination of the tumor, and selected demographic patient information is analyzed using serial Discriminant Analysis or other statistical algorithms.  The results are often a simple formula that can be used to better define the resistant or responsive subgroup.  Efforts are currently underway to define algorithms and validate their utility in the conduct of clinical trials.

Ideally, data that includes the identification of the presence of a molecular target, some measure of relative resistance to apoptosis, a measurement of cell cycle activity, and some measure of downstream signal transduction activity would be useful for identifying adequate predictive algorithms. Other useful biomarkers could include the identification of specific mutations, such as p53, KRAS, or PTEN mutations. A future focus of Pathogenesys will be to use the patterns of typical gene amplifications and deletions found frequently in tumors to segregate patients into various responsive subgroups.  Regardless of the type or number of biomarkers available for analysis, a largely empirical approach towards the discovery of a useful algorithm will likely include as many variables as are available

To augment the biomarker database, Pathogenesys routinely performs a comprehensive morphologic assessment of each tumor which yields as many as 16 morphologic variables. These variables are compatible with standards of practice in Pathology and with CAP and AJCC guidelines.

All interested groups are encouraged to call Dr. James Thompson for more information on this approach at 949-258-0318 or e-mail him at jthompson@pathogenesys.com.

[Home] [About Us] [Mission Statement] [Services] [Clinical Trial Biomarker Algorithms] [Pharama Test Menu] [CLIA Test Menu] [Assay Validation] [Proprietary Technologies] [Investment Opportunities] [FAQ] [News] [References] [Contractual Relationship] [Events] [Study Reporting]