Onyx Pharmaceuticals


is down 28% Monday, only partly because investors are disappointed that the biotech firm will delay the U.S. filing of its kidney cancer drug BAY 43-9006, which it is developing with German drugmaker



. In addition to the delay, Onyx is taking a hit today because confidence in the fundamental bull story around BAY 43-9006 has been undermined.

Onyx has been a great biotech story: a relatively unknown, single-digit stock in late 2002 that soared to a high of $60 per share in mid-2004 because investors believed that BAY 43-9006 would be the next big thing in cancer drugs. Its cult following grew over the years, but reality slapped a lot of people in the face on Monday.

Onyx and Bayer are putting a positive spin on the data from a phase II study of BAY 43-9006, and with some justification. But at the same time, there are aspects of this data that suggest this drug may not be as potent as previously believed. Add these concerns to the fact that BAY 43-9006 may get beat to market by


(PFE) - Get Report

rival kidney cancer drug SU-11248 and you've got a recipe for some nervous investors.

Tracing the Timeline

First, let me address the timeline for a BAY 43-9006 filing with the Food and Drug Administration. Onyx and Bayer say they expect U.S. approval in 2006, but the companies won't be any more specific, citing competitive reasons. If you assume, very optimistically, that BAY 43-9006 receives approval in January 2006, this implies that the drug would need to be filed to the FDA by the middle of 2005 so that the agency has six months to review.

Hard-core Onyx bulls were hoping for an early 2005 FDA filing with an approval in mid- or late-2005. The filing delay disclosed Monday is obviously a disappointment, but given the fact that several analysts and knowledgeable Onyx investors already were anticipating a later filing, the stock's steep slide this morning doesn't make sense.

That is, the slide doesn't make sense until you look at the drug's efficacy, which is also taking some hits. According to data from a phase II study released Monday, BAY 43-9006 wasn't very effective in shrinking tumors in kidney cancer patients.

Only 4% of patients in the study saw their tumors shrink by 50% or more. This 4% "partial response" rate is down from a 15% rate reported last June in a preliminary fashion at the annual meeting of the American Society of Clinical Oncology. In other words, as more patient data were analyzed, the partial response rate of BAY 43-9006 decreased, and by a fairly significant amount.

On its conference call Monday morning, Onyx executives downplayed these results, stating that the phase II study was not designed to measure tumor shrinkage. Instead, the phase II study was intended to determine whether BAY 43-9006 could slow or stop tumor growth, measured by the time it takes for a patient's tumor to progress.

Here, Onyx says the phase II study was positive. A higher percentage of patients taking BAY 43-9006 didn't see their tumors progress, or start growing again, at 12 weeks, compared with similar patients taking a placebo. Overall, patients taking the drug and a placebo had a "time to progression" (the time it takes a tumor to start growing again) of 169 days. Onyx says a historical search of other kidney cancer studies shows that patients given a placebo typically see their tumors progress after 60 to 90 days.

Onyx executives believe this data suggest that BAY 43-9006, by slowing or stopping tumors from growing, will lead kidney cancer patients to live longer. Prolonged survival is the primary endpoint of the ongoing phase III study, on which the companies expect to seek FDA approval.

But in order to be entirely comfortable with this phase II time-to- progression data and be confident that it will lead to a positive phase III study, it's important to know how much better patients taking BAY 43-9006 performed compared wiht patients taking placebos. At this point, Onyx and Bayer won't say, choosing to keep the data under wraps until they can be presented at a scientific meeting.

Patients in this study had their tumors scanned by doctors every 12 weeks. This is a relatively wide time interval; in many studies, tumors are scanned every four to six weeks. More frequent scans produce more accurate data and reduce the risk that results are overestimated. (A problem with "assessment bias" of time-to-progression data helped trip up Genta and its cancer drug Genasense earlier this year.)

BAY 43-9006 may indeed slow or stop the growth of kidney cancer tumors and hopefully boost patient survival, as Onyx believes. But then again, there are legitimate questions about that claim that won't be resolved until we see data from the phase III study.

A lot of this debate may seem too detailed for the average investor to worry about, but the minutiae is important because kidney cancer is a relatively rare cancer type and does not represent a very large commercial market -- $400 million to $600 million in the U.S. and Europe.

Recall that Onyx has to share sales and profits of BAY 43-9006 with Bayer. If Pfizer and its kidney cancer drug SU-11248 beat BAY 43-9006 to approval and launch, Onyx could be squeezed even further. The story gets more troublesome if Pfizer's drug proves itself to be clinically stronger than BAY 43-9006 --something some observers believe, especially after seeing data on the drug last spring.

Drug development is hard work. Too often, investors fall in love with drugs that show early promise, only to be disappointed later when reality can't keep pace with expectations. On Monday, Onyx investors learned this lesson the hard way.

Adam Feuerstein writes regularly for RealMoney.com. In keeping with TSC's editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet.com. He also doesn't invest in hedge funds or other private investment partnerships. He invites you to send your feedback to