Hopkins researchers develop test for cancer

Scientists at the Johns Hopkins Medical Institutions say that they have developed a simple and inexpensive genetic test capable of detecting early cancers in samples of bodily fluids, making some biopsies, X-rays or other invasive techniques unnecessary.

The doctors, reporting in today's issue of Proceedings of the National Academy of Sciences, said the test could revolutionize cancer detection by flagging tumors -- including microscopic tumors -- that are far too small to find with conventional means.


Dr. David Sidransky, a cancer and ear-nose-throat specialist, cautioned yesterday that the test must undergo further refinement. But he said that it may someday offer community doctors the means to test not just patients at high risk for particular cancers but also to screen broad populations for common cancers such as those of the lung, breast, colon and prostate.

"What makes this so easy is that it's just a very simple technology," Dr. Sidransky said. "It's a one-step thing. The whole thing can be done in one day."


The test should give doctors the ability to detect bladder cancer in urine samples, colon cancer in stool samples, lung cancer and cancers of the throat and mouth in sputum, and prostate cancer in urine and ejaculate. Breast cancer could be detected in secretions.

Scientists recently tested the new method on 100 people with known cancers of the lung, bladder, kidney, and head and neck. They were able to find genetic evidence of the cancers in 25

percent of the cases, but Dr. Sidransky said that he is confident that the test's accuracy will be refined to the point where it approaches 100 percent.

Next, they plan to use the technique on a much larger number of people who are thought to be at risk for bladder or cervical cancer but who have no diagnosed disease. Also, they plan to try it on cancer patients whose tumors have been surgically removed in an effort to see if the patients have any residual cancer cells.

"The one major caveat is that like anything else, despite the fact it is so promising and potentially cost-effective, we need to do the large populations," Dr. Sidransky said.

"But the reason we're so confident is that even if you only detected half of all cancers -- even if we fell short of expectations -- you'd be detecting a lot more cancers than you are now."

In the past few years, Dr. Bert Vogelstein of the Johns Hopkins Oncology Center has unveiled tests for finding the genetic mistakes responsible for some types of colon cancer. And recently, researchers at the University of Utah found the mutation that causes some breast cancers that run in families.

But Dr. Sidransky said that a major limitation of genetic tests for human cancers is that no single mutation is responsible for any particular cancer.


"We know that a given colon cancer has nine or 10 mutations, and we've discovered two or three of them," he said. "We don't have all the genes that cause cancer."

Instead of looking for mutations, the new test scans fluid samples for errors in "junk DNA" -- harmless sequences along the genetic code that are changed no matter what the particular mutation is. These genetic changes don't actually cause cancer, but serve as telltale markers of cells that have gone awry.

Cancer is nothing more than the uncontrolled growth of cells that progressively invade surrounding tissue. And as cancer cells reproduce, errors in the "junk DNA" are repeated in each cell. These sequences are found by comparing strands of DNA in fluid samples with those found in normal blood cells. When the sequences differ, scientists presumably have detected cancer.

Testing one patient for a given disease should cost about $50, compared with $1,000 for the genetic tests that look for specific mutations.

Around the nation, scientists reviewing the paper greeted it with mixed enthusiasm and caution.

"I've been saying to my colleagues for a couple of years now that by the turn of the century, which is not very far away, the diagnosis of cancer is going to be very different from what it has been in the past," said Dr. John Laszlo, national vice president for research of the American Cancer Society. "We're going to use a lot more genetic markers."


But he said that the test may pose practical problems. First, even at its low price, the total societal cost would soar if it is ultimately regarded as a screening tool sought by millions of healthy Americans annually.

Also, finding evidence of microscopic cancers in laboratory samples doesn't mean that a doctor can find the corresponding tumor in the human body.

"Think about the implications," said Dr. Laszlo. "If you can't find the cancer, then you sit and wait until you can see something, at which point it's too late to cure him."