Alejandro Abarca Blanco
Cytocatch Imaging System
This project builds upon a previous one I undertook in collaboration with Delee Corp, where I developed an automated blood sample preparation device named "Cytocatch Sample Prep Device" specifically for liquid biopsy applications.
​
Now, as a part of Delee Corp's ongoing commitment to innovation and enhancement, they've asked me to design a fluorescence microscope. This advanced microscope is intended to scan and analyze samples with immunostained Circulating Tumor Cells autonomously, with the goal of boosting both the sensitivity and specificity of their tests.
1. Basic Research
The Cytocatch Imaging System is a microscope specialized in scanning non-planar samples. This microscope boasts advanced artificial intelligence and algorithms that enable constant focusing on immunostained samples, capturing images, and classifying cells based on their morphology and color.
​
The microscope was developed to address an issue with current microscope equipment. These existing systems often exhibit hardware and software variability, leading to inconsistencies and human operator errors, which in turn impedes the widespread adoption of liquid biopsy technology.
For further technical information of the scientific principles, please read:
2. Planning & Prototyping
Our starting point was a comprehensive understanding of current microscopy limitations, achieved through extensive literature reviews, user interviews, and direct observation of existing systems in operation. With a clear problem definition, our multidisciplinary team brainstormed and generated initial design concepts, utilizing advanced CAD tools for visualization. Embracing a user-centric approach, we consulted with experienced microscopy technicians and researchers to refine our designs for ergonomics, functionality, and efficiency. Leveraging rapid prototyping techniques, such as 3D printing and CNC machining, we translated our digital designs into physical, working prototypes. These initial models were subjected to a series of tests: from optical quality checks and software interface usability tests to mechanical stability assessments.
3. Intellectual Property
Throughout our medical device development journey, safeguarding our intellectual property (IP) was of paramount importance. Early in the concept phase, we undertook a thorough patent landscape analysis to ensure our innovations didn't infringe upon existing patents, while also identifying unique aspects of our device for potential patenting. As we refined our designs, we liaised with IP attorneys to file provisional patents, subsequently leading to full patent applications, trademarks, and copyrights as relevant.
4. Software Engineering & AI
We implemented an automated cell recognition and counting software which can distinguish tumor cells from contaminating peripheral blood cells trapped on a filtering membrane. The recognition algorithm utilizes a machine learning model trained to classify cells in microscope slides. We supplied a labeled training database prepared by professional pathologists, where cell features such as shape, size, and staining intensity are fed into a supervised machine learning model, like a convolutional neural network, which is then used to classify cells in new, unseen slides and cells.
5. Design for Manufacturing
Understanding the nuances of both optical precision and production scalability, the design process initiated with detailed sketches and CAD simulations, allowing for thorough evaluation of assembly complexities and component interactions. Collaboration was key: by forging early partnerships with potential manufacturers, I was able to gain insights into fabrication techniques, materials, and common pitfalls. This feedback loop enabled iterative refinements, optimizing the design for ease of assembly, component standardization, and minimizing potential defects.
6. Quality and Regulatory
The original intention of this microscope was to be released to the public as a Research Use Only (RUO) device, but eventually the company plast to release it for diagnostic objectives. Therefore we are developing a plan to present the microscope with a ML algorithm for FDA approval.
Iit's pivotal to engage in pre-submission consultations with the FDA. The device's classification will dictate its regulatory pathway. Documentation should provide a detailed software description, development process, risk analysis, and clinical validation. Hardware specifics and how they interface with the ML component are crucial. Clinical trials or studies are essential to demonstrate real-world reliability.
A post-market surveillance plan may be necessary, especially if the model undergoes periodic updates. Compliance with Quality System Regulation, ensuring data security, and providing user training are also key components. Collaborating with regulatory experts can facilitate navigating this complex process.
7. Market Transition
We established a close collaboration with manufacturing partners early on, enabling us to preemptively address potential scale-up challenges and maintain production quality. By utilizing modular and scalable designs, we ensured that our device could adapt to evolving market demands without necessitating a complete redesign. Continuous training and skill development programs were instituted for our technical team, enabling them to swiftly respond to post-launch feedback and potential product enhancements. Concurrently, we established strong communication channels with regulatory bodies, ensuring our device always met or exceeded compliance standards.