Stat Bite: Persons Living With Major Cancers in the United States, 1998

approach of generating monoclonal antibodies from mice immunized with human tumor cells."We had to use high cut-offs in order not to overlap with normal samples [which also had measurable levels of these proteins]," noted Kenneth Kinzler, Ph.D., senior author on the study, referring to the combined assay."But what's exciting is it's a totally molecular thing.You can see [a protein is] elevated in the cancer, make the prediction that it might be elevated in the serum and then demonstrate that it is."However, the TIMP-1 experiments were not a slam dunk.Although serum levels of the protein were on average significantly higher in pancreatic cancer patients compared to the other groups, Zhou et al. nevertheless ran into problems with TIMP-1 alone as a tumor marker.They found that serum levels of the protein varied widely among pancreatic cancer patients, increasing the likelihood that testing only for TIMP-1 would lead to a relatively high rate of false positivity.Because pancreatic cancer is rare, even as low as a 5% rate of false positivity may be unacceptable for clinical medicine.