Development of an Affordable Skeletal Muscle MRI Phantom; Systemic Failures in MRI Calibration: Distributed Responsibility and the Ethics of Medical Misdiagnoses

The technical and STS research components of my thesis are interconnected as they both recognize the importance of reliability and interpretation of MRI data. The technical capstone project involves the design of a skeletal MRI phantom, which is an artificial calibration tool used to simulate human tissue properties for imaging quality control. The phantom is intended to be an affordable, reproducible standard for ensuring image fidelity and quality across research environments. This project addresses the problem regarding calibration accessibility and aims to design solutions. It focuses on the need for physical testing materials that can support structural MRI protocols and ensuring that MRI will continue to be the gold standard for imaging.
In parallel, the STS research investigates systemic failures in MRI calibration that contribute to medical misdiagnosis. It uses Actor-Network Theory (ANT) to analyze the distributed responsibilities among a complex network of human and non-human actors involved in the whole MRI system. The STS portion of my thesis also recognizes the issue of calibration error in MRI affecting health outcomes and the importance of understanding these issues to reduce potential public harm. Both projects are motivated by a shared concern for the accuracy and trustworthiness of MRI as a diagnostic technology. The technical work aims to provide a solution, while the sociotechnical analysis offers a critique of the current system that fails to prioritize calibration, maintenance, and distributed responsibility. Together, the technical capstone project and STS research recognize the importance of technical innovation and the necessity of examining the social systems that shape how technologies are used, regulated, and trusted.

The goal of our capstone project is to develop and validate a more affordable lower-limb skeletal muscle MRI phantom that mimics the imaging properties and orientations between tissue and muscle for biomedical researchers working with musculoskeletal modeling and MRI application. MRI phantoms are essential tools used to assess scanner performance, test imaging sequences, and ensure consistency across different MRI machines. However, a full body skeletal MRI phantom can cost up to $37,000 and varying degrees or muscle fiber orientation are rarely involved in musculoskeletal phantom designs.
To create the phantom design, a range of materials were researched and tested for their ability to mimic the imaging characteristics of real human tissue. A cost-benefit analysis for each material was completed to ensure an optimal balance of high-quality and efficient phantom for a broad range of users.
Combinations of different materials representing muscle, fat, and bone were tested in an MRI scan and their signal intensities normalized in a water image, percent uniformity, and signal-to-noise ratio were recorded. The data from each material was compared to a human MRI scan using the same protocols, as well as thresholds determined by previous literature. It was found that cement base, flexible epoxy resin, and oil are low-cost materials that can be used to accurately represent the imaging properties of bone, muscle, and fat, respectively. Overall, the phantom can be used by biomedical engineers to validate an MRI in terms of spatial accuracy. Future research will include analyzing the phantom’s ability to create pennation angle as well as include more complex anatomical features like fluid for DTI imaging. Additionally, long-term durability of the phantom has not been tested.

The sociotechnical thesis paper investigates how failures in MRI calibration contribute to medical misdiagnosis, framing these issues as systemic problems rather than isolated human or technical errors. Actor-Network Theory is implemented as an analytical framework to argue that the calibration failures stem from a systemic breakdown involving a complex network of actors, including, radiologists, MRI technicians, engineers, regulatory agencies, hospital administration, calibration phantoms, and the patients themselves. There has been a prevailing tendency to blame medical professionals and healthcare practitioners but this framework reveals how accuracy of diagnoses relies on the fully coordinated function of the human and non-human actors within the healthcare system.
MRI is a crucial tool in modern diagnostics that has become the gold standard in imaging due to its high-resolution and non-invasive imaging capabilities. However, inaccuracies in MRI output can be caused by calibration mistakes and can lead to delayed or incorrect diagnoses with significant consequences for patient outcomes. Despite its importance, calibration is often under prioritized due to institutional pressures, limited regulatory enforcement, and various cost-cutting measures. This paper uses a literature review and case study on previous malpractice cases of calibration eros to highlight how the systemic negligence often places undue liability on individual clinicians while ignoring deeper organizational failures.
The thesis concludes by calling for a more holistic approach to MRI quality control that includes routine maintenance checks, improved commitment to engineering design improvement, and clearer regulatory standards for calibration. Interdisciplinary collaboration will be crucial for system-wide accountability and will widen our view going forward about ethical responsibility in biomedical engineering and healthcare.

I thought that working on both of these projects simultaneously introduced interesting perspectives for each side. My STS research on calibration failures helped recognize that technical innovations can sometimes not be enough for solving an engineering problem completely and that the design must be supported by a full system of training, accountability, and regulation. It allowed me to think critically about the social, ethical, and institutional dimensions of my design. Additionally, as I developed my phantom, I began to think more about how decisions including materials, cost, and signal properties have broader implications in accessibility and trust. As an engineer, it is important to think deeply about the users and any groups that are influenced by the design. I thought more about how the phantom would be used, regulated, and maintained, and how it could fit into the existing system from development to the delivery of results. Overall, the combinations of projects reinforced my understanding of biomedical engineering as a field. It is not just a technical field, but one that is embedded in systems of responsibility, policy, and human impact. I have learned a lot about the integration of the technical and socio technical perspectives that I have held throughout this past year and I will continue to apply this knowledge to my research in graduate school in the coming years.
In parallel, the STS research investigates systemic failures in MRI calibration that contribute to medical misdiagnosis. It uses Actor-Network Theory (ANT) to analyze the distributed responsibilities among a complex network of human and non-human actors involved in the whole MRI system. The STS portion of my thesis also recognizes the issue of calibration error in MRI affecting health outcomes and the importance of understanding these issues to reduce potential public harm. Both projects are motivated by a shared concern for the accuracy and trustworthiness of MRI as a diagnostic technology. The technical work aims to provide a solution, while the sociotechnical analysis offers a critique of the current system that fails to prioritize calibration, maintenance, and distributed responsibility. Together, the technical capstone project and STS research recognize the importance of technical innovation and the necessity of examining the social systems that shape how technologies are used, regulated, and trusted.

School of Engineering and Applied Science

Bachelor of Science in Biomedical Engineering

Technical Advisor: Silvia Blemker, Mario Garcia

STS Advisor: MC Forelle

Technical Team Members: Quentin Shin, Lydia Francis