WEDNESDAY, Oct. 29 (HealthDay News) -- What if a blood test or biopsy could predict if a cancer therapy will help cure you, or only make you feel worse?

Tests like these, based on genes, proteins or other "molecular markers" may someday do just that for people battling colon, lung and pancreatic tumors, scientists reported at a news conference Tuesday.

"The ultimate goal is to bring personalized medicine to reality, to identify characteristics of tumors or patients where we can make a relatively dramatic impact using targeted agents," said Dr. Bruce Johnson of the Dana-Farber Cancer Institute and Harvard Medical School, in Boston.

Johnson moderated the teleconference, sponsored by the American Society of Clinical Oncology (ASCO). The briefing focused on research being presented at the second annual meeting on Molecular Markers in Cancer, which will take place Oct. 30 through Nov. 1 in Hollywood, Fla. The meeting is co-sponsored by ASCO, the U.S. National Cancer Institute and the European Organization for Research and Treatment of Cancer.

First up, colon cancer. Researchers pooled data from four clinical trials involving 715 patients. They confirmed that the cancer drug Vectibix (panitumumab) was only successful in treating advanced colorectal cancer in patients with the normal ("wild type") form of the KRAS gene -- not a mutated version.

Vectibix blocks the epidermal growth factor receptor (EGFR) on cancer cells.

The patient response rate to Vectibix was 14 percent if they carried the normal KRAS gene, but that rate sank to zero if the patient had a mutated form of the gene, said study lead author Daniel Freeman, a principal scientist in oncology research at Amgen Inc., which makes the drug.

Progression-free survival (3.3 months versus 1.7 months) and overall survival (8.3 months versus 5.7 months) were also better in patients with the normal KRAS gene.

This isn't the first time the KRAS gene has dictated just how well a cancer therapy might work. Just last week, researchers reported in the New England Journal of Medicine that colon cancer patients with a KRAS mutation would not respond to the drug Erbitux (cetuximab).

"The data is quite consistent that KRAS is an important negative predictor of patients unlikely to get a benefit from EGFR antibodies," Johnson said.

A second study discussed Tuesday focused on the predictive value of a marker for lung cancer, the No. 1 cancer killer in the United States.

Spanish researchers found that when both blood and tumor samples from patients with metastatic non-small cell lung cancer had a mutated form of the EGFR gene, the patients had a more aggressive cancer, with less time to live -- even after taking Tarceva (erlotinib), another targeted therapy that blocks EGFR.

It could be that more aggressive tumors lead to more mutant DNA in the blood, according to the researchers, from the Catalan Institute of Oncology in Barcelona.

The finding has practical implications for lung cancer patients, they noted. That's because blood samples could yield the same critical information as hard-to-obtain tissue samples, circumventing the need for biopsy. In fact, about a third of patients with this type of lung cancer don't yield enough tumor tissue to analyze for the mutation, the study authors said.

Finally, a third study looked at a particularly deadly malignancy, pancreatic cancer. Researchers found that patients whose pancreatic cancer tumors do not contain the S100A2 protein have double the life expectancy of patients whose tumors show high levels of the protein.

Testing for the presence of the protein beforehand could help select which patients would benefit from surgery. Such surgeries are difficult to perform and available to only a minority of patients with this type of cancer.

The protein may also provide a target for new treatments for this cancer killer, the team said.

The goal of all three studies are tests that could pinpoint which patients will reap the biggest benefit from a particular cancer therapy, Johnson said.

"Most of the pharmaceutical industry has been focused on making 100 percent of patients 20 percent better," he reasoned. "Today, we're talking about markers that can help 20 percent of patients get 100 percent better."

More information

Visit Cancer.net for more on different types of cancer.