3D Printing and Open Source: A Portable MRI for Under $70,000

What Happened?

The Open Source Imaging Initiative (OSI) has developed the OSI2 ONE, a portable MRI scanner whose core is manufactured with 3D printing and whose design is completely open. According to Tom's Hardware, the total cost of the device is under $70,000, representing less than 7% of the price of a conventional MRI, which starts at $1.1 million. The prototype uses a magnetic field of only 50 mT, far from the 1.5 to 3 T of hospital equipment, which traditionally would imply very low spatial resolution and signal-to-noise ratio (SNR). However, the project is not new: OSI has been working on open scanners since 2016, and the OSI2 ONE has already been replicated multiple times worldwide, according to Tom's Hardware. The key novelty is the integration of artificial intelligence to overcome the physical limitations of low field.

Why Is It Important?

The high cost of MRIs limits their availability in developing countries and in rural or low-resource settings. According to the World Health Organization, approximately two-thirds of the world's population lacks access to diagnostic imaging. A portable, low-cost machine could bring diagnostic imaging to places where it is currently unthinkable. Moreover, being open source, any lab or workshop can replicate, modify, and improve it, fostering decentralized innovation. Technology analyst Brian Roemmele highlighted on X that AI can overcome the limitations of low fields: "Low-field MRI has historically been limited by lower signal-to-noise and greater field inhomogeneity. That is exactly the regime where modern AI thrives." Through neural networks trained on high-field data (1.5T to 8T) or physics-based models, it is possible to reconstruct diagnostic-quality images from poor signals. Roemmele added that AI can adjust sequences in real time, adapting gradients and RF pulses while monitoring signal quality.

Consequences and Impact

If the combination of low-cost hardware and AI software proves clinically useful, it could transform the imaging market. Small hospitals, rural clinics, and NGOs could acquire equipment for a fraction of the current price. However, regulatory barriers remain: in countries with strict health controls, such as the United States or the European Union, obtaining FDA approval or CE marking for an open-source, locally manufactured device will be a challenge. Roemmele responds: "No one can stop us from building in garages." The project initially targets regions with limited access, where regulation is looser or nonexistent, and where any diagnostic capability is better than none. The economic impact could be enormous: if the OSI2 ONE is produced in volume, the cost could drop further, pressuring traditional manufacturers like Siemens, GE, and Philips to reduce prices or innovate. Additionally, the open nature allows collaborative improvements, such as using permanent magnets instead of superconductors, eliminating the need for liquid helium, a scarce and expensive resource.

What Should Readers Know?

• The OSI2 ONE will not replace high-field MRIs for detailed studies, but it may be sufficient for detecting large tumors, hemorrhages, or fractures. Previous research with low-field MRI (such as the Hyperfine Swoop, FDA-approved in 2020) has already demonstrated clinical utility in pediatric neuroimaging and intensive care.
• AI is not an optional add-on: it is the critical component that makes the system viable. Without it, the images would be unusable. Techniques like deep learning-based reconstruction and inhomogeneity correction have been validated in academic settings since 2018.
• The project is completely open: plans, bill of materials, and software are available in public repositories (such as GitHub and GitLab), allowing anyone to build their own unit. OSI also offers simpler educational versions for teaching purposes.
• The $70,000 cost includes materials and electronic components but does not account for labor or certifications. Still, it is orders of magnitude lower than commercial alternatives. For reference, the Hyperfine Swoop costs around $50,000, but it is not open source and requires proprietary components.
• There are precedents: the open-source community has already produced low-cost ultrasound devices (like echOpen), incubators (Incubator Project), and ventilators (like the Respira project in Barcelona). The portable MRI is the next logical step, though technically more complex.
• The 50 mT field is 30-60 times weaker than conventional equipment, but the combination of AI and optimized hardware yields images with sufficient resolution for many applications. Preliminary studies show that with fields as low as 6.5 mT, images of limbs can be obtained.

"Image reconstruction becomes dramatically better when deep networks trained on high-field data or physics-informed models denoise, correct for inhomogeneity, and push resolution beyond the raw acquisition limits." — Brian Roemmele

In summary, the OSI2 ONE is a proof of concept demonstrating that the convergence of 3D printing, open source, and AI can break the price monopoly in medical technology. The path to clinical adoption will be long, but the potential is immense. The open-source community has already shown the ability to iterate quickly: improved variants with higher magnetic fields (perhaps using Halbach arrays) and more powerful AI algorithms are expected in the coming years. Meanwhile, the OSI2 ONE is already available for anyone with access to a 3D printer and basic electronics knowledge to build their own scanner. The democratization of medical imaging has begun.