Image courtesy of SIEMENS (Malvern, PA)
Regulation of Existing Imaging Technologies
Devices used in medical imaging are considered medical devices under the Federal Food, Drug, and Cosmetic Act (FD&C Act) as amended by the Medical Device Amendments of 1976 and subsequent related amendments.1,2 Most existing medical imaging technologies are considered non-510(k)-exempt Class I and Class II medical devices by FDA, making them eligible for clearance under the 510(k) pathway.
Under this pathway, new medical devices that are substantially equivalent to existing, legally marketed Class I and Class II devices in terms of intended use, indications for use, and technological characteristics may be cleared for the U.S. market. The legally marketed devices with which comparison is made are referred to as predicate devices.
Under the 510(k) pathway, a next-generation imaging technology may be found substantially equivalent. However, the intended use and indications for use of the device must remain essentially unaltered. Intended use generally refers to the overall purpose of a device. The indications for use define the particular conditions or locations to which the device is to be applied. Also, any technological differences in comparison to the predicate device must not raise new questions of safety or effectiveness. However, if a new intended use or indication for use is sought, or if the technology employed raises new questions of safety or effectiveness when compared with the predicate, the new device is designated as Class III and requires a premarket approval (PMA) application. The PMA process requires a demonstration of safety and effectiveness prior to the granting of marketing approval.
Advanced imaging techniques, such as those used in the Gemini GXL from Philips (Andover, MA), are changing healthcare delivery.
Existing imaging devices use a variety of technologies to obtain diagnostic information or guide therapeutic procedures. These devices may be broken down into modalities that use and those that do not use ionizing radiation to form medical images.
Imaging Devices Using Ionizing Radiation. X-rays are the oldest form of medical imaging. They generate images, known as radiographs, by using ionizing radiation. The majority of the equipment used to produce, control, and process x-rays to generate radiographs is Class I or Class II. Actual classification (depending on the type of device) ranges from Class I 510(k)-exempt products such as diagnostic x-ray tube housings, which produce the x-ray beam, to Class II non-510(k)-exempt solid-state x-ray imagers or digital flat panels.
X-ray devices are representative of the product evolution that is possible in the 510(k) pathway. Such devices have become progressively more sophisticated, moving from purely analog imaging methods (i.e., the x-rays pass through the body part being imaged and directly onto a film, which is then processed for interpretation) to modern digital methods (i.e., the x-ray strikes a detector system, which then converts the received x-rays into an image). This evolutionary change was made possible by successive technological improvements in predicate devices over time, all generally cleared via the 510(k) pathway.
A notable exception is digital mammography, for which FDA has required PMA approval. The original designation of digital mammography as Class III is attributed to both its indications for use (detection of breast cancer) and the application of new technology to that indication. Currently, there is serious consideration being given to downclassifying it to Class II.
Fluoroscopy, which is closely related to x-ray imaging, allows both static and dynamic x-ray imaging. Like x-ray, fluoroscopy is a long-established imaging process that has undergone considerable technological advances since 1976, all essentially within the 510(k) pathway. Modern digital fluoroscopic equipment, which allows for image processing and manipulation, would have been technically impossible in 1976. Yet it is now considered Class II and subject to a variety of special controls.
CT scanners existed prior to 1976 and were placed into Class II by the original medical device classification panels. As for x-ray and fluoroscopic technology, the 510(k) path allowed CT technology to progress from fairly simple single x-ray source-and-detector configurations to the sophisticated spiral and multidetector technologies available today. FDA regulations describing what constitutes a CT scanner for regulatory purposes are fairly typical of the general nature of such regulations in medical imaging. A CT scanner is defined as
[A] diagnostic x-ray system intended to produce cross-sectional imaging of the body by computer reconstruction of x-ray transmission data from the same axial plane taken at different angles. This generic type of device may include signal analysis component parts and accessories.3
The FD&C Act is not the only FDA regulation applicable to x-ray, fluoroscopy, CT scanners, and other medical imaging devices that emit ionizing radiation and certain other types of energy. The Radiation Control for Health and Safety Act of 1968 (subsequently incorporated into the FD&C Act at Sections 531–542) addresses such devices. It authorizes FDA to promulgate performance standards for these radiation- and energy-emitting products to minimize unnecessary emission and exposure.4 This legislation applies to any ionizing or nonionizing electromagnetic or particulate radiation; or sonic, infrasonic, or ultrasonic waves emitted from an electronic product as a result of an electronic circuit. Accordingly, any imaging device that emits ionizing radiation marketed in the United States must meet FDA performance standards as well as gain 510(k) clearance.5 The content of the often-complex reports required to satisfy these requirements may be found in various sections of 21 CFR Parts 1000–1050.
Imaging Devices Not Employing Ionizing Radiation. A growing number of medical imaging products do not rely on ionizing radiation to produce images. Diagnostic ultrasound technologies also existed before 1976. Regulated as a Class II device, diagnostic ultrasound has incorporated a number of substantial technological advances since 1976. The result is a variety of features (i.e., color Doppler vascular imaging and 3-D imaging) that vastly improve its diagnostic capabilities. Like advances in x-ray, fluoroscopy, and CT scanning, diagnostic ultrasound's considerable evolutionary progress has been possible within the 510(k) process. Notably, its Class II designation includes a variety of imaging transducers for both external and internal use, including intravascular ultrasound catheters used for imaging coronary arteries. In addition to the FD&C Act requirements, ultrasound equipment is also subject to FDA performance standards developed under the Radiation Control for Health and Safety Act, given the emission of ultrasonic waves from these devices.
MRI uses powerful magnetic fields in conjunction with radio-frequency pulses and sophisticated computer algorithms to produce detailed anatomic and functional images. Unlike the previous modalities, clinically usable MRI technology was not developed until the early 1980s and consequently had no predicate device when first brought to market. Because the risk to patients undergoing the then-new procedure was not well characterized, MRI was initially regulated as a Class III device and entered the market through the PMA process. Following the widespread clinical introduction of MRI, considerable clinical data were collected demonstrating the relative safety of the technique. Based on this well-defined risk profile, FDA reclassified MRI, MR spectroscopy, and related coils as Class II devices effective July 28, 1988, allowing them to gain marketing approval via the 510(k) pathway.
Related Imaging Management Equipment. In modern clinical practice, medical imaging devices are only part of the medical imaging enterprise. Images obtained by these devices must be captured, stored, and distributed for both primary interpretation and clinical review. This typically requires a picture archiving and communications system (PACS) together with a variety of related components.
PACS devices and their related components depend on computer workstation technology that was not available in 1976. There also was confusion among manufacturers as to whether these products were medical devices, and if so, how they were to be regulated. Culminating in a final rule issued on April 29, 1998, FDA clarified its treatment of PACS and related devices.6 Under these regulations, medical image storage devices and medical image communications devices are Class I exempt if they do not use irreversible compression. PACS devices, along with medical image digitizers (e.g., film scanners) and medical image hard-copy devices (e.g., film printers), are considered Class II and subject to 510(k) premarket notification. Accordingly, PACS devices, medical image digitizers, and medical imaging hard-copy devices require FDA clearance before they may be legally marketed.
Bringing a New Medical Imaging Device to Market
X-rays are the oldest form of internal medical images.
Given that the majority of medical imaging technology is subject to the 510(k) pathway or is even 510(k) exempt, it might appear that bringing new imaging devices to market is a relatively straightforward process. However, this may not be the case if a new product differs from its potential predicate devices in intended use or indications for use, or if it uses modified or new technology. Even if a new device is considered to present a low risk of patient injury, such changes can have a substantial effect on the product's path to market.
Alterations to an Intended Use or Indications for Use. A new imaging device's intended use is key to determining whether that new product is substantially equivalent to a claimed predicate device.7 If a new device's intended use is determined to be not substantially equivalent (NSE) to a predicate device, the new product is subject to an automatic Class III designation. In practice, the broad intended uses of medical imaging devices generally afford sponsors considerable latitude in identifying an intended use that is reasonably applicable to a new imaging product.
The 510(k) pathway allows sponsors some flexibility with a new device's indications for use. Generally, differences in a new device's indications for use as compared with its predicates' does not preclude a substantially equivalent determination. Of course, those differences must not alter the product's intended diagnostic or therapeutic effect. When examining whether such alterations exist, FDA looks to whether the change affects the device's safety and effectiveness compared with that of its predicates.
In practice, minor alterations in the indications for use for an existing imaging technology are often not viewed by FDA as altering a product's diagnostic or therapeutic effect. An example is limiting the application of a cleared imaging technology to a specific area of the body.
More-pronounced changes to a device's indications for use, such as incorporating disease-specific claims (i.e., cancer detection) where the predicate products made only general imaging claims, may cause difficulties in establishing substantial equivalence. In such instances, FDA could easily see the modified indication as affecting the device's diagnostic effect. A relevant example is FDA's treatment of digital mammography. The addition of a breast cancer–screening claim to a breast-specific refinement of digital imaging technology was a major factor in FDA's classification decision. In that case, FDA did require a PMA, even though similar, existing digital imaging systems without a specific cancer-screening indication had been subject to 510(k) clearance. The decision to require a PMA for digital mammography was difficult for many in the imaging community to accept. Nonmammography digital imaging systems with broad indications for use were routinely used in the diagnosis and management of cancer, despite the lack of an explicit cancer-related indication for use. Simply put, even the most established imaging technology may be subject to the PMA pathway if a new indication significantly alters that technology's established diagnostic effects.
Alterations to Device Technology. Assuming that the intended use of a new device is unchanged from its predicates and the indications for use do not alter its diagnostic or therapeutic effect, the substantial equivalence analysis shifts to the technology used to create the images. Should that technology be substantially similar to that of the new product's predicate devices, substantial equivalence is established.
However, if a new product's technology differs from that of the predicate devices, the question becomes whether the change could significantly affect safety or effectiveness. If the answer is no, as might be the case with a minor alteration to image manipulation capabilities or the algorithm by which those images were generated, the product is considered substantially equivalent. However, if the change in technology might significantly affect safety or effectiveness, the analysis moves to whether that change does in fact raise new questions of safety or effectiveness.
The analysis of whether new technological characteristics raise new concerns of safety or effectiveness is often the key question with innovative imaging products. Such products often use completely new methods to obtain diagnostic information. Even if such technology is unquestionably safe, the novel nature of a new device's technology may raise questions of effectiveness. For example, if the imaging technology does not impart energy of any kind into the patient or uses a type of energy whose risks are well characterized, FDA may question the technology's effectiveness. Whenever a new device's technological characteristics do raise new questions, the agency's substantial equivalence analysis ends with an NSE determination.
Consider instances where FDA accepts that no new questions of safety and effectiveness have been raised by a new imaging device's technology. A sponsor must still demonstrate that accepted scientific methods exist for assessing the new technological characteristics and present performance data obtained using those methods to prove substantial equivalence. In practice, clinical data are frequently necessary to establish that the technology has safety and efficacy comparable to its predicates. Designing and conducting clinical studies to demonstrate comparable safety and effectiveness may be difficult. Typically, it requires demonstration of the new product's clinical utility. This, in turn, generally requires the use of the new device to evaluate a specific clinical condition whose existence can be confirmed with a reasonable degree of confidence. This confirmation allows the sensitivity and specificity of the new imaging product to be compared with its predicate devices in a clinical setting in which the new device is to be used.
Approaches to a Not Substantially Equivalent Determination. Should FDA determine that the intended use of a new imaging device differs from that of its predicates, that its indications for use change its effect, or that it raises new questions of safety or effectiveness, agency policy requires that the new device be found NSE.
There are certain indications from FDA that an NSE determination is likely. These include a letter requesting additional information and containing questions that the sponsor cannot address with data that are either on hand or reasonably available.
There are several strategies for dealing with an NSE decision, depending on the underlying basis for the determination.
One potential reason for an NSE decision could be that the intended use has gone beyond the predicate device's intended use. For example, a particular indication for use, such as a claim that an imaging device can either screen for or characterize certain forms of cancer, may require the submission of extensive clinical data or a PMA prior to receiving marketing approval.8 Abandoning or modifying that particular indication, assuming that there were no agency questions with regard to the device's underlying technology, may allow the product to be cleared via the 510(k) pathway.
Similarly, it is not unusual for FDA to object to particular aspects of a new device's technology, such as a specific pulse sequence or imaging algorithm. Removing the portion of the new product's technology that has raised the agency's concern may often allow a substantial equivalence determination to be made on a modified, although less-capable, version of the device.
In instances where a novel new imaging device is low risk but NSE to a predicate device, de novo downclassification under Section 513(f)(2) of the FD&C Act may be an option. Under this pathway, the new device is downclassified into Class I or Class II and subsequently cleared by the agency. FDA will note that a particular device may be eligible for the de novo pathway in the letter notifying the sponsor of the agency's NSE decision. However, if the de novo pathway is being considered, the downclassification should be discussed well before any 510(k) notice is submitted.
To qualify for de novo downclassification, the new imaging device must meet the statutory requirements for inclusion in either Class I or Class II. Administratively, the sponsor must submit a written request for downclassification within 30 days of receiving an NSE letter. The request should include a description of the device, the rationale for the downclassification, and information to support the request. This submission should also discuss proposed general or special controls for the new product that can provide reasonable assurance of its safety and effectiveness.
In practice, clinical data are often necessary, and the downclassification request is typically prenegotiated with FDA. Submission of that request is usually made immediately following the agency's pro forma issuance of an NSE determination. These considerations notwithstanding, FDA has 60 days following the submission of the written request to downclassify the new device. If the agency finds that the product is properly classified in Class I or Class II, the device may be commercially distributed based on the information submitted as part of the de novo downclassification request. However, if FDA finds that the new device is appropriately in Class III, the new device is subject to the PMA process. Although attractive in concept, the de novo process is still used infrequently.
In instances where FDA will not consider either the 510(k) or de novo pathways for a medical imaging device, the PMA pathway is the only option to bring a new imaging device to market. Under the PMA paradigm, clinical trials are necessary to demonstrate that the device is reasonably safe and effective for its intended use and indications for use. As noted earlier, designing clinical studies for imaging products can be challenging, particularly demonstrating a new device's effectiveness and clinical utility. Depending on the particulars of the technology, these studies may have to demonstrate not only that the device can successfully detect a disease or condition with a certain sensitivity and specificity, but also that physicians and other healthcare professionals can use the information provided in a clinically useful manner. Even in instances where the clinical trials for a PMA are comparable in size and complexity with those needed for a 510(k) notice or de novo downclassification, sponsors of a PMA-pathway device are typically subject to substantially higher user fees and a preapproval manufacturing inspection.
FDA regulation of medical imaging devices may appear to some as deceptively simple, because the majority of imaging products are subject to the 510(k) pathway. However, when a sponsor is contemplating new or modified indications for use for an established imaging technology, a variation of established technology, or a totally new technology, the sponsor should fully and realistically explore the potential regulatory consequences.
When developing a substantial equivalence argument for such a product, it is often prudent to consider alternatives to the planned device in terms of intended use and technology, in case FDA initially rejects the claim of substantial equivalence. In considering these options, sponsors should also perform a realistic assessment of the resources available to design and develop clinical trials should clinical data be necessary for 510(k) clearance or should de novo downclassification or a PMA prove necessary.
John J. Smith is counsel at Hogan & Hartson LLP (Washington, DC) where he specializes in regulatory and scientific issues related to FDA marketing approval for medical devices. Smith can be contacted at [email protected].
1. Federal Food, Drug, and Cosmetic Act, Public Law 75-717, 52 Stat. 1040 (1938), 21 USC Sections 301–394.
2. Medical Device Amendments of 1976, Public Law 94-295, 90 Stat. 539 (1976; codified as amended in scattered sections of 21 USC).
3. Code of Federal Regulations, 21 CFR 892.1750.
4. Radiation Control for Health and Safety Act of 1968, Public Law 90-602, 82 Stat. 1171 (1968).
5. Code of Federal Regulations, 21 CFR 1040.
6. Federal Register, 63 FR:23385, April 29, 1998.
7. “Guidance on the CDRH Premarket Notification Review Program 6/30/86 (K86-3),” Blue Book Memorandum (Rockville, MD: FDA, CDRH, 1986).
8. “Guidance for Industry: General/Specific Intended Use” (Rockville, MD: FDA, CDRH, 1998).