Panel members concluded Iressa, made by AstraZeneca, has some ability to shrink tumors and is likely to be of some benefit to patients who take it. The benefits, however, are modest: Tumors shrink, but patient survival is not improved.

The panel agreed to recommend the drug for ''fast-track'' approval even though the company could not support claims about symptom improvement. The company presented findings from a study that combined Iressa with chemotherapy but did not compare those patients with patients who received chemotherapy alone.

The panel also faced an unusual situation: In August, AstraZeneca reported results of studies in which Iressa failed to improve survival of patients who took the drug alone, without chemotherapy. That data was not presented to the panel because AstraZeneca is seeking fast-track approval for Iressa as a last-ditch treatment after chemotherapy has failed, not as an initial treatment. Fast-track approval is less rigorous than standard approval, but companies must provide additional data later.

The panel said that the tumor shrinkage gave them confidence that patients who have chemotherapy will have at least some limited response. FDA spokeswoman Kathleen Kolar said final approval is likely to come in six months.

One reason the panel recommended approval without more positive data is the lack of effective alternatives. Standard chemotherapy provides only modest improvements in survival for patients with the most advanced cases of lung cancer -- a two-month median improvement in survival over patients who are not treated.

Iressa, which is already approved in Japan, belongs to a new class of drugs called tyrosine kinase inhibitors that are supposed to work more directly on tumor cells than chemotherapy. They also appear to be much less toxic. Gleevac, which was approved for a form of leukemia, is in this class and is considered to be a breakthrough.

Genetech Inc. and OSI Pharmaceuticals are testing a tyrosine kinase inhibitor called Tarceva for lung cancer. Preliminary data appear positive.

''The importance of this class of drugs in cancer is quite well-established,'' says Robert Mass of Genentech. ''There are many different approaches that companies are taking.''

The experimental drug Erbitux is in the same class, but the FDA declined last December to approve it based on data presented by Imclone Systems and its partner, Bristol Meyers Squibb.

© Copyright 2002 USA TODAY, a division of Gannett Co. Inc.