Oncotech Inc. has developed a testing system using humantumor samples that the company said can accurately identifythe specific cancers most likely to respond to new drugs.

There have been no reliable methods of predicting how wellnew cancer drugs will actually work in humans.

As a result, said Dr. Larry Weisenthal, vice president forscientific affairs and inventor of Oncotech's system, fewer than10 percent of the 100 or so drugs tested against cancerbetween 1960 and 1980 won marketing approval.

"We put our money where our mouth is by making predictionsabout two drugs that haven't been tested yet," Weisenthal toldBioWorld.

He predicted that enloplatin, a new anti-cancer drug justentering clinical trials, would have greatest activity againstnon-Hodgkin's lymphoma. His second prediction is thatcytarabine, a widely used drug, will be very effective againstchronic lymphocytic leukemia, a disease it has never beentested against.

It will take about a year to see if he's right.

Weisenthal made the predictions Friday at the annual meetingof the American Association of Cancer Research in Houston. Healso presented data showing that the system accuratelymodeled the known clinical activity of existing anti-cancerdrugs.

Previous tests failed because they used cell cultures containingimmortal lines of rapidly dividing cells, said Weisenthal. Inthese cell lines, all cancers tend to look pretty similar, and it'shard to tell if a drug works better on one cancer than another.

In actual tumors used in the Oncotech method, only a smallproportion of cells are actively dividing. "If you look at drugeffects on the total cell population, then drug effects aremagnified tremendously," said Weisenthal.

Oncotech, based in Irvine, Calif., has selected about 50 of thedrugs tested between 1960 and 1980, all of which are offpatent, and plans to test them using its system. If any of thedrugs look promising, the company will ask for orphan drugdesignation and enter Phase II clinicals.

Privately held Oncotech developed the system about 10 yearsago and has been using it to predict which drugs won't work inpatients receiving chemotherapy. -- Karen Bernstein

(c) 1997 American Health Consultants. All rights reserved.