Study Data
Validation of tests through proper study design
When designing clinical studies, it is important to include the appropriate cases and control groups. We place an emphasis on multi- institutional studies from both referral and first opinion practices to ensure that our tests and sample shipping logistics are fully compatible with the diverse range of veterinary practice. Close attention is paid to the inclusion of clinically relevant controls from both diseased and healthy individuals (including inflammatory conditions resulting in non malignant lymphadenopathy). The process is described in detail in a paper which has been published in the Journal of Veterinary and Comparative Oncology*. The initial test data comprised serum samples from 92 non lymphoma dogs and 87 dogs with lymphoma obtained from 14 separate clinical centres. Following this the test performance was verified using 96 individual blind samples from first opinion practices in both the UK and USA. To date, PetScreen has analyzed more than 1,500 serum samples using the lymphoma blood test from first opinion and referral practices.
Blinded trials are essential to any good study. We recently completed a retrospective, blind analysis of 295 samples obtained from first opinion practices. This work demonstrated the following test performance:
Sensitivity = 84%
Specificity = 86%
In another blinded study of 148 pre-treatment confirmed lymphoma cases obtained from referral practices, a sensitivity of 92% was observed.
* Ratcliffe et al. Proteomic identification and profiling of canine lymphoma patients. Journal of Veterinary and Comparative Oncology. June 2009 Volume 7 Issue 2 p92-105
Back...
