Medical Software with AI: What Rules and Classifications Should You Know?
What Are We Talking About?
Softwares as Medical Devices (SaMD) integrating Artificial Intelligence (AI) are becoming increasingly common in healthcare, which is good news for improving the efficiency of our healthcare system and enhancing patient care.
Currently, the main barrier between healthcare professionals, patients, and these tools is regulatory.
What are the current regulatory requirements for AI-based Software as a Medical Device (SaMD)?
In this article, we offer real-world examples of AI-based Medical Device Software, their classification under MDR 2017/745 or IVDR 2017/746, and guidance on how to achieve compliance.
Examples of AI-based Medical Software, Regulatory Status, and Classification
Among AI-based medical devices, the most common are:
- Diagnostic support algorithms, particularly in radiology (46.9% of AI-based devices approved by the FDA¹), oncology (25%¹), neurology (15.6%¹), dermatology, and genomics.
➡️ Classification (a priori): In vitro diagnostic medical device (IVDMD) under IVDR 2017/746, typically risk class C or D.
- Patient management systems for healthcare professionals that are using risk scores to recommend care, medication dosages, or vaccination reminders.
➡️ Classification (a priori): Medical Device under MDR 2017/745, class IIb.
- Insulin dosing apps
➡️ Classification (a priori): Medical Device under MDR 2017/745, class IIb.
- Prescription assistance software helping doctors draft prescriptions by referencing medication databases.
➡️ Regulatory status (a priori): Considered a Medical Device by HAS under MDR 2017/745, class IIa or IIb.
- Gamified patient applications (DTx) offering rehabilitation or personalized follow-up programs that evolve based on patient profiles and progress.
➡️ Classification (a priori): Medical Device under MDR 2017/745, class I or IIa.
- Self-diagnosis apps, especially for mental health, dermatology, or other conditions.
➡️ Classification (a priori): Medical Device under MDR 2017/745, class IIa.
Some AI-based software remains in a borderline area and is not always CE-marked or FDA-approved.
Examples include:
- Medical assistant chatbots that allow healthcare professionals to consult official recommendations from medical databases.
- Patient apps that guide patients on whether they need to consult a doctor based on medical data and patient criteria.
How to Obtain CE Marking for an AI-based Medical Device
CE marking is mandatory for AI-based Medical Devices when a clinical benefit is claimed.
The evaluation process includes the following steps:
1️⃣ Setting up a Quality Management System (QMS) compliant with ISO 13485 and MDR 2017/745 or IVDR 2017/746.
2️⃣ Compiling a Technical File (TF) demonstrating the safety and performance of the API/module/software, in accordance with European regulatory requirements.
3️⃣ Compliance with the AI Act, with progressive harmonization between the AI Act and MDR 2017/745-IVDR 2017/746
- Articles 47.3 and 48.5 of the AI Act: a single declaration of conformity and combined CE marking are possible, allowing manufacturers to use a single QMS and Technical File covering both regulations.
- Similarities between the AI Act and MDR 2017/745-IVDR 2017/746 include change management and post-market surveillance (PMS), but there are also key differences: training data management, risk management, transparency obligations, and human oversight.
4️⃣ Consult your Notified Body (NB) to determine if they plan to be designated under the AI Act, as none are currently designated.
Need Regulatory Support for Your AI-based Software?
At Sparta Care, we help you with:
- Regulatory strategy for your AI-based Medical Devices (AIaMD)
- Implementation of a Quality Management System compliant with ISO 13485 and MDR 2017/745-IVDR 2017/746
- Compilation of the Technical File for your API/module/software
- Support with Notified Bodies and during audits
Contact us to learn more!
¹ Source: Benjamens, S., Dhunnoo, P. & Meskó, B. The state of artificial intelligence-based FDA-approved medical devices and algorithms. npj Digit. Med. 3, 118 (2020). https://doi.org/10.1038/s41746-020-00324-0
Discutons de vos projets
Ou
contactez nous à hello@sparta.care

