Artificial intelligence (AI) is a broad discipline that aims to understand and design systems that display properties of intelligence. Machine learning (ML) is a subset of AI that describes how algorithms and models can assist computer systems in progressively improving their performance. In health care, an increasingly common application of AI/ML is software as a medical device (SaMD), which has the intention to diagnose, treat, cure, mitigate, or prevent disease. AI/ML includes either “locked” or “continuous learning” algorithms. Locked algorithms consistently provide the same output for a particular input. Conversely, continuous learning algorithms, in their infancy in terms of SaMD, modify in real-time based on incoming real-world data, without controlled software version releases. This continuous learning has the potential to better handle local population characteristics, but with the risk of reinforcing existing structural biases. Continuous learning algorithms pose the greatest regulatory complexity, requiring seemingly continuous oversight in the form of special controls to ensure ongoing safety and effectiveness. We describe the challenges of continuous learning algorithms, then highlight the new evidence standards and frameworks under development, and discuss the need for stakeholder engagement. The paper concludes with 2 key steps that regulators need to address in order to optimize and realize the benefits of SaMD: first, international standards and guiding principles addressing the uniqueness of SaMD with a continuous learning algorithm are required and second, throughout the product life cycle and appropriate to the SaMD risk classification, there needs to be continuous communication between regulators, developers, and SaMD end users to ensure vigilance and an accurate understanding of the technology.

BackgroundAutoimmune hepatitis (AIH) is a rare chronic progressive liver disease, managed with corticosteroids and immunosuppressants and monitored using a combination of liver biochemistry and histology. Liver biopsy (gold standard) is invasive, costly and has risk of complications. Non-invasive imaging using multiparametric magnetic resonance (mpMR) can detect the presence and extent of hepatic fibroinflammation in a risk-free manner.ObjectiveTo conduct early economic modelling to assess the affordability of using mpMR as an alternative to liver biopsy.MethodsMedical test costs associated with following 100 patients over a 5-year time horizon were assessed from a National Health Service payor perspective using tariff costs and average biopsy-related adverse events costs. Sensitivity analyses modelling the cost consequences of increasing the frequency of mpMR monitoring within the fixed cost of liver biopsy were performed.ResultsPer 100 moderate/severe AIH patients receiving an annual mpMR scan (in place of biopsy), early economic modelling showed minimum cost savings of £232 333. Per 100 mild/moderate AIH patients receiving three mpMR scans over 5 years estimated minimum cost savings were £139 400. One-way sensitivity analyses showed increasing the frequency of mpMR scans from 5 to 10 over 5 years in moderate/severe AIH patients results in a cost saving of £121 926.20. In patients with mild/moderate AIH, an increase from 3 to 6 mpMR scans over 5 years could save £73 155.72. In a minimalistic approach, the use of 5 mpMR scans was still cost saving (£5770.48) if they were to replace two biopsies over the 5-year period for all patients with moderate/severe or mild/moderate AIH.ConclusionsIntegration of mpMR scans in AIH patient pathways leads to significant cost savings when liver biopsy frequency is either reduced or eliminated, in addition to improved patient experience and clinician acceptability as well as providing detailed phenotyping to improve patient outcomes.Trial registrationNCT03979053.

UNSTRUCTURED Artificial intelligence (AI) is a broad discipline that aims to understand and design systems that display properties of intelligence. Machine learning (ML) is a subset of AI that describes how algorithms and models can assist computer systems in progressively improving their performance. In health care, an increasingly common application of AI/ML is software as a medical device (SaMD), which has the intention to diagnose, treat, cure, mitigate, or prevent disease. AI/ML includes either “locked” or “continuous learning” algorithms. Locked algorithms consistently provide the same output for a particular input. Conversely, continuous learning algorithms, in their infancy in terms of SaMD, modify in real-time based on incoming real-world data, without controlled software version releases. This continuous learning has the potential to better handle local population characteristics, but with the risk of reinforcing existing structural biases. Continuous learning algorithms pose the greatest regulatory complexity, requiring seemingly continuous oversight in the form of special controls to ensure ongoing safety and effectiveness. We describe the challenges of continuous learning algorithms, then highlight the new evidence standards and frameworks under development, and discuss the need for stakeholder engagement. The paper concludes with 2 key steps that regulators need to address in order to optimize and realize the benefits of SaMD: first, international standards and guiding principles addressing the uniqueness of SaMD with a continuous learning algorithm are required and second, throughout the product life cycle and appropriate to the SaMD risk classification, there needs to be continuous communication between regulators, developers, and SaMD end users to ensure vigilance and an accurate understanding of the technology.