Anyone worried about the escalating cost of questionable medical technology might want to take a close look at proton beam radiation therapy for cancer. There are now six centers in the U.S. that can deliver the precision radiation ion beams. At least five more are either under construction or in development. Pricetag: anywhere from $100 to $225 million each.

And what's the evidence that proton beams are a superior alternative for tumors located in highly sensitive areas where the risk of collateral damage from radiation is high? "No study found that charged particle radiotherapy is significantly better than alternative treatments with respect to patient-relevant clinical outcomes," a new Agency for Healthcare Quality and Research technical brief reports.

Still, that hasn't stopped leading academic medical centers and even some suburban hospitals from plunging ahead and investing in the high-tech facilities. Said William Hartsell, medical director of a $140 million proton therapy center slated to open in Warrenville, Ill. in 2011: "We know there are some tumors where protons are significantly safer and more effective than standard radiation."

Do we? There have been 243 research papers written about proton beam therapy since Loma Linda University Medical Center in Southern California opened the first proton treatment center in 1990. Many involve hard-to-treat head-and-neck cancers.

But the big growth area lately is prostate cancer, which now accounts for 60 percent of all proton beam therapy patients (the hope is that it will spare men from the debilitating treatment side effects of impotence and incontinence, and reduce the risk of radiation-induced secondary cancer). And there the research base is slim. In three decades of research in this new area, no one has definitively shown that it produces better outcomes than standard radiation or other forms of treatment (surgery or watchful waiting).

Indeed, the research base (243 studies) is comprised mostly of retrospective single-arm studies (185) and prospective single-arm studies (35), with a mean patient population of 65. The fact is that most research in this area is little more than glorified case reports.

Despite the presence of these large and expensive machines (they involve building cyclotrons) in some of the nation's leading cancer treatment centers (Harvard-affiliated Massachusetts General Hospital; the M.D. Anderson Cancer Center in Houston; and Indiana University, for instance), there have been only eight randomized and nine unrandomized trials comparing the technology to other forms of therapy. And none used advanced statistical techniques to adjust for confounding variables, thus making the results less than definitive.

"Especially for many common cancers, such as breast, prostate, lung, and pancreatic cancers, it is essential that the theorized advantages of particle beam therapy vs. contemporary alternative interventions are proven in controlled clinical trials, along with concomitant economic evaluations," the AHRQ report concluded.