Novartis' (NYSE: NVS) cancer drug patupilone doesn't work well. At least, not yet.

The company said yesterday that a phase 3 trial in ovarian cancer patients showed that patupilone was unable to increase survival compared to Johnson & Johnson's (NYSE: JNJ) Doxil, which is called Caelyx outside the U.S. and sold by Merck (NYSE: MRK) in most ex-U.S. markets.

It is possible that patupilone actually works as well as Doxil/Caelyx, but not better than it. Novartis' exact quote was that the drug "did not show a significant overall survival advantage." Until the company releases the full data set, investors can hold out hope that the drug works, just not enough to justify bringing it to market for ovarian cancer patients.

Novartis is also testing patupilone against other cancers, including metastatic colorectal cancer, metastatic lung cancer, and prostate cancer, and I wouldn't discount its chances of success based on this one trial.

First, there are many examples of cancer drugs working for one tumor type but not another. Pfizer's (NYSE: PFE) Sutent works for kidney, gastrointestinal, and pancreatic cancers, but not against liver or breast cancer, for instance.

Second, the patients in this trial were pretty sick. They could have failed up to three previous chemotherapy regimens, including one of the standard first-line treatments of Bristol-Myers Squibb's (NYSE: BMY) Taxol or sanofi-aventis' (NYSE: SNY) Taxotere. Testing a drug against advanced patients is a good way to get on the market, but it's often a long shot, too.

It looks like patupilone went wide of goal this time; fortunately Novartis still has a few more shots.

Matt Hoffman explains how soccer is like investing without using shots on goal metaphors.