FDA decision on digital mammography may add years to commercialization

January 18, 1995

FDA to regulate full-FOV systems as class-three devicesThe much anticipated market introduction of full-field-of-viewdigital mammography systems for screening and diagnosis couldbe postponed for years as a result of a Food and Drug

FDA to regulate full-FOV systems as class-three devices

The much anticipated market introduction of full-field-of-viewdigital mammography systems for screening and diagnosis couldbe postponed for years as a result of a Food and Drug Administrationdecision to place the systems in the agency's most restrictiveproduct regulatory class.

FDA officials have decided to regulate full-field-of-view digitalsystems as class-three devices, a move that will likely requirevendors to conduct comprehensive clinical studies and file premarketapproval applications (PMAs) rather than 510(k) submissions. Notonly will performing these studies slow vendors' planned productintroductions, but the PMA review process can drag on for years,delaying commercialization even further.

Not surprisingly, mammography vendors are taking issue withthe FDA's position. The decision suggests that radiologists areincapable of determining whether a technology is beneficial, saidMorgan Nields, chairman and CEO of Fischer Imaging in Denver.

"Radiologists, through their own clinical studies, aregoing to discern whether the technology is going to add valueor confusion," Nields said. "The FDA is simply goingto slow down the dissemination of the technology and make it alonger time before digital mammography is available to women whocould benefit from it."

FDA officials recognize industry concerns regarding the timelyrelease of new technology, but say the cautious approach to full-field-of-viewdigital mammography is warranted.

"We don't want to put a tremendous obstacle in anyone'sway, but we need to have an appropriate amount of data,"said Dr. Susan Alpert, director of the Office of Device Evaluation(ODE).

The review process for full-field-of-view digital mammographywill require a variety of information, Alpert said. For starters,the agency will want nonclinical data about the equipment. Butmost important will be clinical data showing the safety and effectivenessof digital mammography as a screening tool in cancer diagnosis.

NEMA prepares a response. The National Electrical ManufacturersAssociation in Washington, DC, responded during the first weekof January by polling member companies that are developing digitalmammography systems to determine whether to contest the FDA'sstance, said Robert Britain, NEMA vice president of medical products.

"With all the hullabaloo about mammography last year,(the decision) doesn't make sense," Britain said. "TheFDA has taken a step backwards."

Having a prospective technology placed in the class-three categoryis a vendor's worst nightmare. As of fiscal 1994, the typicalPMA review, which is required for all class-three devices, tookabout 800 days to complete.

"We are not proud of those numbers," said CDRH directorDr. Bruce Burlington. "It's the same amount of time, or actuallya little longer, that the FDA center for drugs needs to reviewNDAs (new drug applications)."

In fact, many class-three devices require much longer -- threeyears or more -- than this typical period.

In contrast, the average review for class-two devices, whichgo through the less demanding 510(k) process, is now at about200 days (see story, page 2). The review time for 510(k) applicationsis shrinking and could drop to 100 days by summer, but the PMAprocess shows no such improvement.

At least one mammography vendor, Lorad of Danbury, CT, plansto file for fast-track PMA review of its full-field-of-view productif the FDA goes through with its plan to require PMAs, accordingto Lorad president Hal Kirshner. The FDA implemented fast-trackreview in 1993 as part of a package of reforms to improve productreview times (SCAN 7/14/93).

All eyes are on ultrasound vendor ATL, which got the go-aheadfor fast-track PMA review last year for its method of using ultrasoundto differentiate breast lesions (SCAN 1/19/94 and 5/18/94). Thevendor had hoped to receive FDA clearance by the end of 1994,but as of this month, the agency had not yet cleared the application.

NEMA's Britain believes that digital mammography systems wouldbe more appropriately placed in the class-two category, as aredigital spot mammography devices for guiding stereotactic biopsy.Such devices are being marketed by Lorad, Fischer and BennettX-Ray.

"Digital mammography is not a new science," he said.

Radiologists who use these systems thoroughly evaluate thembefore they are introduced into routine clinical use, said Nieldsof Fischer Imaging, which markets the MammoVision digital spotmammography device.

"Because of medico-legal concerns, radiologists are notgoing to launch into a new technology without thorough evaluationthemselves, no matter what the FDA or anybody else says aboutit," Nields said. "Therefore, digital mammography isthe same thing (as digital spot mammography)."

But the precedent established by digital stereotactic biopsysystems does not apply, said Harvey Rudolph, acting director ofthe CDRH Office of Science and Technology.

"There is more than just hardware involved," Rudolphsaid. "It is the intended use of the device that matters."

Ironically, one of the most promising aspects of digital technology-- that the enhanced resolving power of CCDs (charge-coupled devices)might allow the identification of smaller lesions -- has led tothe current dilemma. According to ODE director Alpert, her officeis putting these devices into class three because she and othersin the FDA fear that the new digital systems might actually betoo sensitive, resulting in unnecessary referrals for surgicalbiopsy.

"Screening tools have to be sensitive enough to find thelesions, but there is a real issue about being too sensitive (and)identifying too many people for additional procedures," shesaid.

The classification being given to digital mammography, therefore,is appropriate because class three includes devices that havenot been assessed for potential patient risks. The issue revolvesaround digital mammography as a screening tool in place of theconventional mammogram, according to Alpert.

"We think there is a public health issue here becausethis technology will be used to determine whether women go onfor additional procedures for the diagnosis of breast cancer,"Alpert said. "That is not a trivial issue."

Vendors will need clinical data. If the contingent of digitalmammography vendors led by NEMA decides to fight the FDA's decision,Britain said they will not wait until the next scheduled meetingof the FDA panel that reviews radiology matters, set for March6. Instead, they would go directly to the FDA as soon as possible.

But if the industry approaches the FDA, it had better bringclinical data. Those data, according to Alpert, will have to showhow effective the technology is in spotting cancerous lesions,with explicit information regarding false negatives and falsepositives.

"We'll need to know how many biopsies are performed (asthe result of a digital mammogram) and how many of these wereconfirmed as cancer," she said.

There is, in fact, no way vendors will be able to avoid performingclinical studies. Even if the agency grants class-two status tofull-field-of-view digital mammography, clinical studies willbe a part of the 510(k) review process, said Lillian Yin, directorof the Division of Reproductive, Abdominal, ENT and RadiologicalDevices, the segment within the ODE that reviews medical imagingdevices.

Yin's group is already developing a draft clinical protocolfor full-field-of-view digital mammography systems. The draft,scheduled for completion Feb. 1, will be released to the medicalimaging community about three weeks prior to the March 6 panelmeeting.

"In the meantime we are still welcoming companies to sharetheir own clinical protocols," Yin said. "It would begood to share them with us ahead of time. We will try to integratethem into our draft protocol, if that is feasible."

Entrepreneurs move forward. Although the goal of the FDA is toprotect the public health, an unintended result of its recentactions and decisions might be to discourage entrepreneurs fromdeveloping novel technologies for digital mammography. PrimeXGeneral Imaging (PGI), a sensor technology company in Carlsbad,CA, will not be deterred, according to Ronald B. Schilling, chairmanand CEO. PGI has adapted technology originally devised for nationaldefense applications to digital mammography.

The new technology is silicon-based, as is CCD technology.But rather than converting x-rays to light, the foundation forCCD-based systems, the PGI approach is to convert x-rays directlyto electronic signals that are then reconstructed into images.Schilling wants to partner with an OEM to integrate PGI's digitaldetector into a mammography system.

Schilling is concerned about the FDA's move to make full-field-of-viewdigital mammography a class-three device, but believes it willnot be a hardship for his company. One reason is that PGI is exploringother, nonradiological markets for its detector technology. Asecond reason is more philosophical.

"Our strategy is based on what diagnostic imaging hasbecome, not what it was," Schilling said. "Manufacturershave to stop talking about technical performance and demonstratediagnostic and therapeutic impact, which includes patient outcomes."

That, said Schilling, means clinical research. Obviously, theFDA agrees.

© 1996

Miller Freeman, Inc.

All rights reserved.