HOUSTON (AP) _ Three prominent oncologists contend that flawed research methods make it nearly impossible to determine whether a doctor's controversial alternative cancer treatment works.

``It is scientific nonsense,'' Dr. Howard Ozer, director of the Allegheny Cancer Center in Philadelphia, said Wednesday.

Ozer and two other physicians studied Dr. Stanislaw Burzynski's alternative cancer treatment known as antineoplastons for the Sept. 25 issue of The Cancer Letter, a widely read and well-respected Washington newsletter.

Burzynski has claimed for years that antineoplastons, compounds found in human urine and blood, ``turn off'' cancer genes by interrupting signals that cause the cells to multiply. His treatment involves giving synthetically made antineoplastons to patients orally or intravenously.

Many among Burzynski's fiercely devoted clientele contend they would be dead without the therapy.

But Ozer and the other doctors who reviewed Burzynski's data say the public may never know if his treatment works because his research parameters are ``poorly designed'' and his data is ``not interpretable.''

The other two physicians were Peter Eisenberg, a principal investigator with the Sutter Health West Cancer Research Group, and Henry Friedman, chairman of the brain tumor committee of the Pediatric Oncology Group.

Burzynski said in a faxed, nine-page response that the doctors' findings can be attributed to their lack of complete information, their lack of expertise, or the cancer suffered by the patient.

``The Cancer Letter, unfortunately, is not reporting the truth and only truth,'' the statement said. ``It is also reporting the lies of the oncologists who were selected by The Cancer Letter to review the annual report (submitted to the FDA).''

The three oncologists reviewed an annual report Burzynski submitted to the U.S. Food and Drug Administration in an effort to obtain federal approval of his treatment. The report included data Burzynski gathered from clinical trials of 963 patients who received antineoplastons over a 12-month period.

Statistics released in April and culled by the FDA from Burzynski's preliminary reports show only 36, or 4.3 percent, of 828 patients treated intravenously with antineoplastons responded positively to the drug. Of the 36, 11 died.

Burzynski's lawyer, Rick Jaffe, claimed the FDA's totals were wrong and he chided the agency for releasing preliminary data.