Quantitas Solutions

🌟 𝗥𝗲𝗳𝗹𝗲𝗰𝘁𝗶𝗻𝗴 𝗼𝗻 𝗔𝗜 𝗮𝗻𝗱 𝗲𝘁𝗵𝗶𝗰𝘀 𝗶𝗻 𝗵𝗲𝗮𝗹𝘁𝗵𝗰𝗮𝗿𝗲 🌟 The team at Quantitas Solutions had the privilege of attending the recent Zorg2025 event, where we explored the evolving role of AI in healthcare, and the ethical dilemmas it brings. The event featured a play, portraying a future where healthcare professionals collaborate with algorithms more than ever. This performance illustrated both the positive and unintended impacts of AI integration in medical practice. We further discussed the challenges around responsible AI, privacy, and fairness, prompting important questions on how we balance innovation with ethical integrity in healthcare. At Quantitas Solutions, we're motivated to contribute to responsible AI in healthcare, ensuring that technology serves not only efficiency but also the well-being and rights of all. Thank you to the organisers of the event. Amsterdam AI, Amsterdam Economic Board, Rabobank, ROM InWest and Sigra

🌟 𝗚𝗿𝗲𝗮𝘁 𝗺𝗼𝗺𝗲𝗻𝘁𝘂𝗺 𝗯𝘂𝘁 𝗮𝗹𝘀𝗼 𝘀𝘁𝗶𝗹𝗹 𝗮 𝗹𝗼𝘁 𝗼𝗳 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗮𝘁 Health-RI! 🌟 Yesterday, the team of Quantitas Solutions attended the Health-RI conference, where we learned more about the ongoing initiatives to create the European Health Data Space. It's encouraging to see how these developments will help advance healthcare technologies across the Netherlands and create a Europe fit for the digital age. The reality however is that for at least the next four years, it will remain challenging for AI developers to work with European medical data for commercial purposes. The fact that we at Quantitas Solutions already offer compliant access to this data makes us extra proud. Together, we can continue to innovate while ensuring that patient privacy and data protection remain a top priority. 🔐 #HealthRI #MedicalData #AI #Innovation #HealthcareTech #QuantitasSolutions

🚀 𝗘𝗺𝗽𝗼𝘄𝗲𝗿𝗶𝗻𝗴 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗥𝗮𝗱𝗶𝗼𝗹𝗼𝗴𝘆 𝘄𝗶𝘁𝗵 contextflow! 🩻 At Quantitas Solutions, we’re thrilled to have collaborated with contextflow on exciting projects that showcase their innovation in radiology: 🫁 Lung Nodule Classification: contextflow’s advanced diagnostic tools are being enhanced by our expert thoracic radiologists, who reviewed and classified lung nodules on CT scans. Their cutting-edge technology is setting new standards in lung cancer detection and management. 👏 We’re proud to have supported a forward-thinking partner and are excited about the potential of their solutions. While we provided contextflow with expert medical annotations for these projects, we also offer both medical annotations and medical data to support medical AI developers. 💡 𝗟𝗲𝘁’𝘀 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗲 𝘁𝗼 𝗱𝗿𝗶𝘃𝗲 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗵𝗲𝗮𝗹𝘁𝗵𝗰𝗮𝗿𝗲 𝘁𝗼𝗴𝗲𝘁𝗵𝗲𝗿! Ester Jiménez Arroyo Julie Sufana Giacomo Deodato Marcel Wassink

𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁 𝗳𝗿𝗼𝗺 𝗮 𝗿𝗮𝗱𝗶𝗼𝗹𝗼𝗴𝗶𝘀𝘁’𝘀 𝗽𝗲𝗿𝘀𝗽𝗲𝗰𝘁𝗶𝘃𝗲 I believe that providers of high-risk AI systems should begin addressing these topics immediately to ensure compliance and prevent future obstacles.… I’m Roderick M. Holewijn, radiologist and co-founder of Quantitas Solutions. My mission is to advance AI adoption, and the AI Act is a significant milestone in this journey (for the full text, check: https://lnkd.in/eUXf6NQk). In this series, I'll explore how this legislation impacts clinical users and our clients. In this post, I'll focus on the steps AI providers should take now to avoid complications with the notified body when applying for MDR and AI Act CE-mark certifications. 𝗡𝗼𝘁𝗶𝗳𝗶𝗲𝗱 𝗕𝗼𝗱𝘆 𝗜𝗻𝘀𝗶𝗴𝗵𝘁𝘀 Recently, several clients have approached us after being informed by notified bodies (NBs) about unforeseen requirements. Here are two important points: 👉 𝗣𝗼𝗽𝘂𝗹𝗮𝘁𝗶𝗼𝗻 𝗥𝗲𝗽𝗿𝗲𝘀𝗲𝗻𝘁𝗮𝘁𝗶𝘃𝗲𝗻𝗲𝘀𝘀: the NB requires that your AI system’s data must include a significant portion from EU populations if you want to market your product in the EU. 👉 𝗠𝗲𝗱𝗶𝗰𝗮𝗹 𝘀𝗽𝗲𝗰𝗶𝗮𝗹𝗶𝘀𝘁 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱𝘀: the NB scrutinizes the credentials of medical specialists who annotate data. This includes their years of experience, training locations, and subspecialties. 𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻 𝗳𝗼𝗿 𝘆𝗼𝘂𝗿 𝗔𝗜 𝗽𝗿𝗼𝗱𝘂𝗰𝘁? To ensure compliance, you need to ask: “Can I demonstrate two years from now that my AI system was developed in accordance with the standards from art. 10 AI Act?”. Our recommendations:   👉 𝗦𝗸𝗶𝗹𝗹𝗲𝗱 𝗮𝗻𝗻𝗼𝘁𝗮𝘁𝗶𝗼𝗻 𝗽𝗿𝗼𝗳𝗲𝘀𝘀𝗶𝗼𝗻𝗮𝗹𝘀: ensure that your annotators have a proven track-record and align with the target population. Using medical specialists from outside the EU or those not actively working in clinical settings may lead to complications and eventually higher costs.   👉 𝗛𝗶𝗴𝗵-𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝗿𝗲𝗽𝗿𝗲𝘀𝗲𝗻𝘁𝗮𝘁𝗶𝘃𝗲 𝗱𝗮𝘁𝗮: you must be able to trace the source of your data. Relying on cheap bulk data without knowing its full history and context can create issues later on. 𝗛𝗼𝘄 𝗰𝗮𝗻 𝘄𝗲 𝗵𝗲𝗹𝗽? At Quantitas Solutions, we address these challenges by providing high-quality European data and annotation services from Western European specialists, all of whom are actively practicing. I didn’t anticipate the AI Act when I founded the company - I’m not a lawyer - but I knew these topics were essential for AI to function effectively in clinical practice. 𝗡𝗲𝘅𝘁 𝘂𝗽: Stay tuned for my upcoming post on the user perspective of medical AI systems.   👨⚕️⚕At Quantitas Solutions, we excel in helping AI developers bring their products to market. We provide annotation services by European clinicians and high-quality European data to ensure compliance and efficacy. ➡️ Interested in hearing more about Roderick M. Holewijn his vision regarding the AI Act? Send us a DM.

𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁 𝗳𝗿𝗼𝗺 𝗮 𝗿𝗮𝗱𝗶𝗼𝗹𝗼𝗴𝗶𝘀𝘁’𝘀 𝗽𝗲𝗿𝘀𝗽𝗲𝗰𝘁𝗶𝘃𝗲 AI Act obligations for the majority of health AI developers take effect per 2027. Time to relax until that date? (Un)fortunately not,… I’m Roderick M. Holewijn, radiologist and co-founder of Quantitas Solutions. My mission is to advance AI adoption, and the AI Act is a significant milestone in this journey (for the full text, check: https://lnkd.in/eUXf6NQk). In this series, I'll explore how this legislation impacts clinical users and our clients. Here, I’ll discuss the AI Act's timelines and why action is needed now. 𝗔𝗜 𝗔𝗰𝘁 𝘁𝗶𝗺𝗲𝗹𝗶𝗻𝗲 The AI Act came into force on August 1, 2024, but compliance deadlines vary by AI risk classification: 🕚 𝗙𝗲𝗯𝗿𝘂𝗮𝗿𝘆 𝟮, 𝟮𝟬𝟮𝟱: Prohibitions begin for ‘Unacceptable Risk AI’. 🕚 𝗔𝘂𝗴𝘂𝘀𝘁 𝟮, 𝟮𝟬𝟮𝟲: Obligations start for ‘High-Risk AI’ systems in Annex III. 🕚 𝗔𝘂𝗴𝘂𝘀𝘁 𝟮, 𝟮𝟬𝟮𝟳: Obligations apply for ‘High-Risk AI’ systems requiring third-party assessment. 𝗪𝗵𝘆 𝘆𝗼𝘂 𝗰𝗮𝗻’𝘁 𝘄𝗮𝗶𝘁 𝘂𝗻𝘁𝗶𝗹 𝟮𝟬𝟮𝟳: 𝗠𝗲𝗱𝗶𝗰𝗮𝗹 𝗗𝗲𝘃𝗶𝗰𝗲 𝗥𝗲𝗴𝘂𝗹𝗮𝘁𝗶𝗼𝗻 (𝗠𝗗𝗥) 𝗮𝗻𝗱 𝗔𝗜 𝗔𝗰𝘁 𝗼𝘃𝗲𝗿𝗹𝗮𝗽 Most AI systems in healthcare will be classified as high-risk (see my previous post), meaning obligations will start by August 2, 2027. But, you can’t wait until then. The MDR already demands ‘state-of-the-art practices’, aligning with the AI Act. In my opinion, this means AI Act requirements should already be part of your quality management system. 𝗪𝗵𝗮𝘁 𝘀𝗵𝗼𝘂𝗹𝗱 𝘆𝗼𝘂 𝗱𝗼 𝗻𝗼𝘄? 👉 Familiarize yourself with best practices and relevant ISO standards, and integrate these into your product development and validation process . 👉 Pay particular attention to quality aspects during the development phase that cannot be easily ‘fixed’ later on. 𝗡𝗲𝘅𝘁 𝘂𝗽: stay tuned for my next update on specific requirements for high-risk AI providers and immediate steps to take. 👨⚕️⚕At Quantitas Solutions, we excel in helping AI developers bring their products to market. We provide expert level annotation services by European clinicians and high-quality European data to ensure compliance and efficacy. ➡️ Interested in hearing more about Roderick his vision regarding the AI Act? Feel free to send us a DM.

𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁 𝗳𝗿𝗼𝗺 𝗮 𝗿𝗮𝗱𝗶𝗼𝗹𝗼𝗴𝗶𝘀𝘁’𝘀 𝗽𝗲𝗿𝘀𝗽𝗲𝗰𝘁𝗶𝘃𝗲 Why do almost all of the healthcare AI systems qualify as ‘high-risk’ under the AI Act? And why is this crucial to define the requirements you must adhere to...? My name is Roderick M. Holewijn, radiologist and founder of Quantitas Solutions. My mission is to advance the adoption of AI in healthcare. The recently enacted AI Act is a big step forward in achieving these goals (for the full text, check: https://lnkd.in/eUXf6NQk). In this series of posts, I'll look into the specifics of how this legislation impacts my work as a clinical user and the broader implications for our clients. In this post, I’ll talk about the relation between the Medical Device Regulation and the AI Act. 𝗜𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝗰𝗲 𝗼𝗳 𝘁𝗵𝗲 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗔𝗜 𝗔𝗰𝘁 𝗿𝗶𝘀𝗸 𝗰𝗹𝗮𝘀𝘀𝗲𝘀 The AI Act divides AI systems into four risk classes: unacceptable risk, high risk, limited risk and low / minimal risk. Based on the risk class, the AI Act provides a set of requirements that must be met in order for an AI system to be considered safe and legitimately offered on the EU market. It is therefore crucial for our clients (‘Providers’) and myself as a user of an AI systems to understand the risk class. 𝗛𝗼𝘄 𝘄𝗲 𝗰𝗮𝗻 𝘂𝘀𝗲 𝘁𝗵𝗲 𝗠𝗲𝗱𝗶𝗰𝗮𝗹 𝗗𝗲𝘃𝗶𝗰𝗲 𝗥𝗲𝗴𝘂𝗹𝗮𝘁𝗶𝗼𝗻 (‘𝗠𝗗𝗥’) 𝘁𝗼 𝗶𝗱𝗲𝗻𝘁𝗶𝗳𝘆 𝘁𝗵𝗲 𝗿𝗶𝘀𝗸 𝗰𝗹𝗮𝘀𝘀? 👉 The AI Act: All AI systems falling under the MDR and simultaneously, according to the MDR, requiring a third-party conformity assessment are considered to be high-risk AI systems under the AI Act. 👉 The MDR: Software used for diagnosis, making therapeutic decisions, monitoring physiological processes, or contraception, fall in Class IIa or higher. For Class IIa, Class IIb and Class III devices are subject to a conformity assessment by third-party authorities. 95% of the products our clients develop fall into these categories. 👉  Conclusion: The interplay of the AI Act and the MDR results in medical devices that are recognised as Class IIa or higher under the MDR to be defined as high-risk under the AI Act. Stay tuned for more in-depth insights in my upcoming posts! 𝗡𝗲𝘅𝘁 𝘂𝗽: When will you as a Provider need to adhere to the AI Act requirements for high-risk AI systems? 👨⚕️⚕At Quantitas Solutions, we excel in helping AI developers bring their products to market. We provide annotation services by European clinicians and high-quality European data to ensure compliance and efficacy. ➡️ Interested in hearing more about Roderick M. Holewijn his vision regarding the AI Act? Feel free to send us a DM.

𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁 𝗳𝗿𝗼𝗺 𝗮 𝗿𝗮𝗱𝗶𝗼𝗹𝗼𝗴𝗶𝘀𝘁’𝘀 𝗽𝗲𝗿𝘀𝗽𝗲𝗰𝘁𝗶𝘃𝗲 As a radiologist, I'm thrilled about the implementation of the AI Act. Why?.. My name is Roderick M. Holewijn, radiologist and founder of Quantitas Solutions. My mission is to advance the adoption of AI in healthcare, while ensuring these systems are of the highest quality, safety, and trustworthiness. The recently enacted AI Act is a significant step forward in achieving these goals, and I’m thrilled about the potential it brings (for the full text, check: https://lnkd.in/eUXf6NQk). In upcoming posts, I'll look into the specifics of how this legislation impacts my work as a clinical user and the broader implications for our clients. But first, let me elaborate why the AI Act will serve as a business enabler. 𝗪𝗵𝗮𝘁 𝗶𝘀 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁? Based on your role under the AI Act (in my case ‘User’ and in the case of our AI developing clients ‘Provider’) and the risk type of the AI system, the AI Act establishes guidelines to ensure the safe and ethical deployment of AI technologies. 𝗪𝗵𝘆 𝗜 𝘄𝗲𝗹𝗰𝗼𝗺𝗲 𝘁𝗵𝗲 𝗔𝗜 𝗔𝗰𝘁: 👉 Improved adoption of AI systems: This leads to enhanced hospital workflow, greater accuracy, and reduced diagnostic errors. 👉 Increased patient trust: Ensuring AI systems are safe, ethical, and transparent will build patient confidence. 👉 Healthcare accessibility: The Act helps guarantee that (AI-driven) healthcare remains accessible to all patients now and in the future. 𝗪𝗵𝗼 𝗶𝘀 𝗮𝗳𝗳𝗲𝗰𝘁𝗲𝗱? This regulation is crucial for: 👉 AI developers (‘Provider’): Those integrating AI into products and services. 👉 Organizations using AI (‘User’): Including those utilizing AI systems in sectors like healthcare and use these tools in daily clinical practice. Stay tuned for more in-depth insights in my upcoming posts! 𝗡𝗲𝘅𝘁 𝘂𝗽: why almost all of the AI systems that are used in healthcare qualify as ‘high-risk’ AI and what this means for Providers and Users. 👨⚕️⚕At Quantitas Solutions, we excel in helping AI developers bring their products to market. We provide annotation services by European clinicians and high-quality European data to ensure compliance and efficacy. ➡️ Interested in hearing more about Roderick M. Holewijn his vision regarding the AI Act? Feel free to send us a DM.