Posted 01 September 2017
By Michael Mezher
The US Food and Drug Administration (FDA) on Thursday presented its case for regulating 3D-printed, patient-specific anatomical models as medical devices when they are marketed for diagnostic use at a joint meeting with the Radiological Society of North America's 3D Printing Special Interest Group.
Such models are being increasingly used by physicians and surgeons to help visualize or measure patient-specific anatomy to inform treatment decisions. According to FDA, such models are made either by the hospitals and clinics that use them, or by medical device manufacturers or third-party service providers.
During the meeting, officials from FDA's Center for Devices and Radiological Health (CDRH) said the software used to generate such models would be considered a Class II medical device when marketed for creating models for diagnostic use.
According to Jeffrey Ballyns, branch chief for mammography, ultrasound and imaging software in the Division of Radiological Health within CDRH, the agency's current thinking is to regulate the software used to generate the models, rather than the 3D printers or models themselves. This mirrors the agency's approach to regulating medical apps, where the app software is regulated based on its intended use, rather than the platform or device the software runs on.
"Any use that can affect or change making a diagnosis, or patient management or patient treatment is considered a diagnostic use … If there are inaccurate models, they can in fact mislead the physician and result in misdiagnosis, mismanagement of the patient, or delayed treatment," said Nooshin Kiarashi, a lead reviewer at the Division of Radiological Health (DRH) within CDRH.
Who Needs Clearance?
When it comes to getting such software cleared, Kiarashi said that companies should look to demonstrate substantial equivalence to an existing device regulated under product code LLZ, which covers picture archiving and communications systems.
"We believe this is the suitable product code to portray 3D printed anatomic models, because as you can imagine there exists validated performance testing that can show that when we add this functionality to an image analysis software to be able to produce 3D printed anatomic models we can now ensure that there is adequate diagnostic quality, establishing an equivalence between the 3D printed anatomic models and their digital equivalents," she said.
According to Kiarashi, which companies that actually need to get clearance will depend on how the models are marketed and used.
For instance, a 3D printer manufacturer would need to get clearance if the software for their printer is marketed as capable of outputting 3D printed anatomical models for diagnostic use. Similarly, if a software company is marketing software as capable of creating files for making such models, the software would need to be cleared.
In both cases, if the printer or software is being marketed for general use without specific claims for creating anatomical models for diagnostic use, a clearance would not be required.
On the other hand, clinics or other facilities that market the models for diagnostic use would only need to get clearance if the 3D printing software they are using is not cleared by FDA.
And clinics would not need to get clearance for making 3D printed anatomic models for doctors within the clinic, as such use would fall under the practice of medicine, so long as the clinic is not marketing the production of the models as a service to other parties.
Data Requirements and Guidance
To get 510(k) clearance, Kiarashi said the extent of data and validation will depend on the specific diagnostic use a company wants to market their product for, but that any validation will need to address accuracy and reproducibility.
Kiarashi also said the recommended 3D printer model, material and post-processing steps should be detailed in the submission.
For performance testing, Kiarashi said there should be clinically relevant accuracy and precision data for the specific anatomy types the software is intended to create models for, as well as phantom testing to ensure that landmarks from the radiological images are accurately reproduced in the printed model.
Additionally, Kiarashi said that data backing printer compatibility should be provided if the software is designed to work with multiple printers.
As for guidance, James Coburn, senior research engineer and co-chair of FDA's additive manufacturing working group, said to turn to FDA's recently issued draft guidance, Technical Considerations for Additive Manufactured Devices. While the guidance does not specifically touch on 3D printed anatomical models, Coburn said the sections on design, manufacturing, validation and testing are still relevant.
"There are things that you have to think about when you're making anything, especially something that's being made for a patient [and] validation and testing is part of the performance loop that you want to close," Coburn said.