(Guest Post by Evelyn McLean) Case Study: The Clinical Translation of MRI

This is a guest post by Evelyn McLean

Key Lessons:

• The relatively less-known EMI research lab deserves more attention in the discourse around nontraditional research institutions like Bell Labs and PARC
• Obtaining the first MRI images required Lauterbur to move back-and-forth between academia and industry in a way that would be unusual today

Medical Imaging by The Beatles

In 1962 the British company Electrical & Musical Industries Ltd. (EMI) signed The Beatles to their label, making music history and ushering in an era of unprecedented success for any musical artist in history. A decade later EMI introduced the Computed Tomography (CT) scanner (then called the EMI scanner) to the world, ushering in an era of unprecedented advancements in medical imaging, and ultimately leading to the development of what would become the “gold standard” of medical imaging, Magnetic Resonance Imaging (MRI).

The 1950s and 60s were good times for EMI. In addition to running a world-renowned record label, the company manufactured a wide range of electronics, from television cameras and tape recorders to guided missiles and radar. In the 1950s an electrical engineer in EMI’s computer department, Godfrey Hounsfield, led the team that built the UK’s first commercially available all-transistor computer. EMI eventually shuttered the department and transferred Hounsfield to the central research lab, where his previous work in imaging for radar and television got him thinking about whether it was possible to use computerized imaging with x-ray. EMI worked out a partnership with the Department of Health and Social Security (DHSS) to fund Hounsfield’s prototype scanner, likely sourcing some of their portion of the funds from the tidal wave of money flowing in from the release of the first few Beatles’ albums.¹ The analogy to Bell Labs, PARC, or a FAANG research lab is apparent.

Immediately upon release the CT scanner transformed medical imaging. CT was the first technology to explore the possibilities of using computerized images in medicine, supplanting direct imaging techniques used in conventional radiography. It improved upon the revelation of x-ray vision in two main ways: one, by producing three-dimensional images, an improvement on the superimposition of structures in x-ray images; and two, by depicting soft tissue. Doctors could use CT to identify the precise depth of tumors, see “behind” bones, and (miraculously!) see inside the brain without having to cut open a patient’s skull.

The fanfare from the medical community was loud, but CT was far from perfect. The images from CT scans were better than the images from x-rays, but still left a lot to be desired, and the dyes that had to be injected into patients for contrast could have unpleasant side effects. More importantly, CT still relied on x-ray technology, and exposed patients to ten times the levels of ionizing radiation as x-rays.

Still, the medical industry raced to bring CT to the world. Within a year of releasing their first model EMI had orders from hospitals in the US and the UK, and within 3 years 10 other companies were advertising their own models, including medical-industry heavyweights GE and Siemens. EMI declined to license their patents to their competitors, upending the common practice within the medical industry of avoiding court battles over patents through extensive licensing. (See Appendix for some interesting notes on this.) In the next few years increases in computer processing power and the application of more sophisticated reconstruction algorithms rolled out at breakneck speed. By the late 1970s the R&D departments at EMI’s bigger competitors had made their first model obsolete several times over; GE took most of the market, followed by Siemens and Johnson & Johnson. Despite EMI’s efforts to hold onto their edge, the medical industry giants won out.

But the story of MRI began before EMI revealed the CT scanner to the world. That same year, 1972, a physician and faculty member at SUNY Downstate Medical Center named Raymond Damadian filed a patent for a novel application of Nuclear Magnetic Resonance (NMR) to look inside the body. Though a few of the key details about the machine outlined in his patent application changed in the interceding four years before he built it, his "Apparatus and Method for Detecting Cancer Tissue" was the blueprint for the first MRI scanner.²

Biological Signals

The field of NMR took off with a few big breakthroughs in the 1940s. Scientists devised methods to measure differences in the magnetic moments (spins) of the nuclei of different atoms within a strong magnetic field, and for the next 25 years their findings, and the scientific field that sprang up to devise new applications for these discoveries, remained within the purview of analytical chemistry. Early NMR instruments could fill a large room, and scientists would have to fit their samples into 5 mm tubes, which sat in the two-inch space between magnets that could weigh more than a ton. For all intents and purposes NMR was a technique for looking closely at isolated chemicals in solution.

Throughout the late 1950s the application of NMR began to tiptoe into biochemistry as researchers began studying organic compounds like human eye tissue and cervical mucus. In 1959 a researcher at UC Berkeley named Jay Singer took it a step further when he used NMR to track blood flow in live mice. This was the first indication that NMR could be safely used to study a living creature, and it opened up a world of possibilities. But the impacts of high-strength magnetic fields on a living body were still largely unknown, and the potential clinical applications of NMR (beyond monitoring blood flow) hadn’t been explored. It would take a great leap into the unknown to dislodge NMR from its place in the lab and bring it into the world of medicine.

Eleven years after Singer’s study was published Damadian conducted his landmark experiment. He hypothesized that the organization of water differed in cancerous and non-cancerous cells and used an NMR spectrometer to examine samples from the tumors of six tumor-infested rats and samples from healthy rats. He recorded a significant difference in their T₁ relaxation times (the average time it takes for the atomic nuclei being studied to return to their equilibrium after being disturbed by a pulse of radio waves) and published his findings in Science. Damadian’s conclusion that relaxation times could be definitively used to identify cancerous tissue was later disputed by other researchers when his results failed to replicate,³ but it gave him an idea for a machine that could be used to “scan” the body for any malignancies, for which he quickly filed a patent. He had just begun the process of setting up an NMR lab of his own when he had the unfortunate (for him) luck of meeting Paul Lauterbur, a chemist on the faculty of SUNY at Stony Brook.

The Best Burger of Your Life

An unusual set of circumstances brought Lauterbur and Damadian together in the summer of 1971. Lauterbur didn’t have a summer salary that year, and hadn’t secured grant funding for the summer. It was under these conditions that he’d agreed to take over as the acting president and chairman of the board of NMR Specialties. (This back-and-forth between academia and industry is something that would be quite uncommon today.) Lauterbur had lent a hand in setting up the business in the late 1950s while running an NMR lab at Carnegie Mellon, and in the spring of 1971 the company was close to bankruptcy.

As it happened, Damadian had visited NMR Specialties in the previous year when they agreed to let them use their NMR machine for his rat tumor experiment. While trying to save the sinking business, Lauterbur fielded an order for a superconducting magnet that Damadian had put in for an NMR project at his lab. Lauterbur didn’t think they could build it and refused the order. Damadian has since remarked that this was the first of many slights and obstacles he believes Lauterbur subjected him to throughout the decades they spent independently developing MRI.⁴

While Lauterbur’s feelings about Damadian in those days is unclear, we do know that he was aware of Damadian’s initial insights into applying NMR in medical diagnostics. But while Lauterbur was intrigued by the idea of extending NMR into medicine, he wondered whether NMR was limited to producing data about the makeup of a particular tissue, which required taking biopsies and scanning them while the patient waited (as Damadian had done for his paper in Science). If one could somehow use NMR to get spatial data, he reasoned, one could create an image, allowing a doctor to see tissues inside the body without resorting to cutting “little chunks” out of the patient. (This, of course, is the fundamental insight behind CT, but EMI wouldn’t release the first CT scanner for another year.)

The idea struck him one night that summer while he was out having a burger: you could get spatial information if you used magnetic field gradients! He reasoned that varying the strength of the magnetic field across the sample, in every direction, would result in relaxation times (in the form of RF signals) that varied in direct proportion to the strength of the magnetic field. Over the next few days he worked out a technique for translating the signals from individual atoms along a magnetic field gradient into spatial information, and eventually into an image, using an iterative algebraic reconstruction algorithm.

Lauterbur’s mathematical solution for reconstructing a two-dimensional image from a series of one-dimensional projections happened to be the same technique that Hounsfield used in his prototype CT scanner. Of course, Lauterbur couldn’t have known that; in fact, he was under the impression that he had invented a new field of applied mathematics. He wrote the ideas down in a notebook and had them witnessed, hoping to eventually patent his breakthrough.

A complex series of events (including a costly fight with the NMR Specialties’ lawyers and a conservative approach from the university research office at Stony Brook) prevented Lauturber from securing a patent for his projection method. Undeterred, he set out to publicize his discovery, and in 1973 he published an article about his new NMR imaging method in Nature. That article included the first published MRI image, a fuzzy black-and-white picture of two test tubes of water, one with a solution of D2O the other with a solution of MnSO4. Lauterbur didn’t cite Damadian’s previous study specifically, though the paper does mention the possibilities of using NMR imaging for the "in vivo study of malignant tumors, which have been shown to give proton nuclear magnetic resonance signals with much longer water spin-lattice relaxation times than those in the corresponding normal tissues." The article established Lauterbur’s standing as a founder of the field, and initiated a decade-long, international race to develop MRI.

The Race Begins

As it so often happens in the history of science, at least one of Lauterbur and Damadian’s contemporaries had made the same breakthrough at roughly the same time. In 1973 Peter Mansfield, a physicist at the University of Nottingham, published his research using NMR to create one-dimensional projections of proton density, applying an approach similar to Lauterbur but for solids, rather than liquids. He was presenting his work at a conference in Poland when someone in the audience directed him to Lauterbur’s newly-published article in Nature. Lauturber’s images came as an unfortunate surprise to Mansfield, who, siloed in the world of physics, had missed the hubbub around Lauterbur’s discovery. When Mansfield got back to the UK he soon began building on Lauterbur’s work, pivoting to focus on NMR imaging in liquids.

Though it may appear that Mansfield’s insights, coming on the heels of Lauterbur’s publication, would put him in direct competition with Lauterbur, he was more concerned with the other teams jumping into the field in the UK than with the two lone-wolf scientists toughing it out on their own in the states. Another two physicists at the University of Nottingham, Raymond Andrew and Waldo Hinshaw, heard Lauterbur speak about his research and by 1974 they published their discovery of a new imaging technique that improved on Lauterbur’s method, significantly reducing the amount of computing power required to produce an image. Meanwhile, a doctor at the University of Aberdeen in Scotland named John Mallard was assembling a team to begin building an MRI machine of their own, and it would only be a few years until Ian Young, a recording physicist in the central research lab at EMI in London, would start work on the first MRI machine developed in the private sector.

Early on, Lauterbur’s pursuit of support for the burgeoning field of MRI found him pitching the potential of MRI to representatives from the medical industry. But the lab tours and meetings with executives made little difference – industry was almost universally uninterested. With the notable exception of EMI, MRI would remain in academia until an intense period of investment and private competition began in the late 1970s. There were many unknowns in the early stages of MRI, and the CT market was booming; Lauterbur was pitching a machine that appeared to do the same thing as CT, except that it was slower, produced images of poorer quality, and was 20 times more expensive. Add to that the complete lack of information about the possible long-term health effects of high-strength magnetic fields, and Lauterbur’s failure to interest industry starts to make sense.

Meanwhile, as the American researchers were being snubbed by industry, the UK teams had significant government support from the Medical Research Council (the British answer to the NIH) and the Department of Health and Social Security (DHSS). The MRC helped Mallard purchase the magnet needed to build the first full-body MRI scanner, and a DHSS program designed to support the development of technology that might be used in hospitals partnered with EMI to fund Young and his team (with support and guidance from Mansfield) to build another early full-body MRI machine. In addition to government support, Hinshaw and Andrew’s team also received grant funding from the Wolfson Foundation, a private philanthropic organization in England, which they used to build their own full-body scanner.

Soon after the Nature article was published Lauterbur began receiving extramural grants; $100,00 a year from the National Cancer Institute (NCI), and later a $200,000 per year grant from the National Heart, Lung and Blood Institute (NHLBI). From then on he received consistent support from the NIH, in the form of grants from NCI, NHLBI, and other institutes. Damadian’s early work was supported with small grants from the NCI, the American Cancer Society, the New York City Health Research Council, and donations from private philanthropists. Overall, early MRI researchers in the US did not get significant government support to develop MRI technology, due to the NIH’s primary interest in supporting basic research rather than new technology.

Despite the lack of funding, Damadian was determined to stay at the front of the pack. In the time since he’d submitted that first patent application CT had taken over the field of medical imaging, and his idea for a machine that could “detect” cancerous tissue had transformed into something more like a CT scanner using NMR technology - an MRI. In 1976 he and his team built a homemade superconducting magnet large enough for a person to fit inside, and the next year published images of the torso of a member of Damadian’s research team, painstakingly captured over 5 hours. Damadian’s lead didn’t last long; the next year Mansfield’s team at Nottingham published images of Mansfield’s own abdomen, taken in 40 minutes. By the late 1970s image quality had improved significantly, and the imaging process was speeding up. At the same time, the CT market was going through a shakeout, and the companies that had emerged as winners had the resources to invest in something new.

We’ll Take it from Here

It was around this time that the head of the radiology department at the University of California San Francisco (UCSF), Alex Margulis, had dreams of getting UCSF in on MRI early. He leased a space in South San Francisco for a brand new UCSF MRI lab and put physicist Leon Kaufman in charge. But they faced the same problems with funding that Damadian and Lauterbur were facing: "When you go in for an NIH contract," Kaufman explained, “the first thing they want is what's your hypothesis. Well, my hypothesis is I can make it work. When you're inventing a machine, you don't know how it's going to work. . . . Instruments don't have hypotheses." They found a solution to this problem through a partnership with Pfizer; later, when Pfizer had moved out, Diasonics took over, and was eventually replaced by the Japanese company Toshiba.

This partnership is one of many encounters between industry and academia that marked MRI’s shift, beginning in the late 1970s and extending into the 1980s, from academia into the private sphere. Sometimes it took the form of partnership as when Andrew’s team at Nottingham, who were seeking support to begin clinical trials, partnered with US company Picker (which had acquired EMI in 1978).

A few companies interested in getting involved in MRI simply poached researchers directly from academia, as the American company Technicare tried with Mallard’s team at Aberdeen. They’d supported Mallard’s research with gifts of instruments, and attempted to move the whole team to the US. The bid was unsuccessful, but they did end up luring away Hinshaw from Andrew’s team at Nottingham. GE appropriated several other members of the UK research teams using similar tactics.

Companies who could afford to built their own machines from scratch, and relied on previous medical industry experience and brand loyalty to help them catch up. Notably, two researchers formed their own companies. In 1978 Damadian rounded up support from friends and family to found his own MRI company, FONAR, and produced the first commercial MRI scanner (“The Indomitable”) in 1980. Mallard, meanwhile, went to Japan in search of industry backing, and started his own company, M&D Technology Ltd., with the support of Japanese beer company Asahi.

By the Grace of the FDA

In 1976 the Medical Device Regulation Act established a new protocol by which the Federal Drug Administration (FDA) would approve new medical technology. The new protocol differentiated between devices that were extensions of existing technologies which were available on the market by 1976 or had been approved by the FDA, and new devices, which had to undergo FDA-approved clinical trials to gain “Pre-Market Approval” (PMA). MRI was the first technology categorized as a novel technology requiring PMA before it could be marketed. Despite a joint appeal by over a dozen companies claiming that MRI was simply an extension of NMR the FDA maintained its decision, requiring all companies interested in bringing an MRI machine to market to run FDA-approved clinical trials.

Securing placement for clinical trials was tricky. FDA rules didn’t allow companies to make a profit on a device before it was approved but even at-cost the machines were exorbitantly expensive (upwards of $700,000), and the final price of the machine also included the installation of a shield to protect the rest of the hospital from the strong magnetic field. In addition, due to the small number of companies producing magnets of the scale needed in an MRI machine, commissioning a machine could take more than a year. Companies needed to find hospitals that were willing to spend the money on installing the machine without knowing whether it would pass clinical trials.

Luckily there were some advantages for hospitals for taking part. Some companies offered financial incentives, like letting hospitals defer the full cost of the machine until after it had been transferred from research to clinical use. Hospitals could get government or company-funded research grants to help them offset the costs of the machine, or, in some cases, share costs with an affiliated university. Hospitals could also benefit from taking part in clinical trials when applying for a “Certificate of Need'' (CON). The idea of requiring hospitals to apply for a CON was developed as a policy instrument to respond to increases in the problem of technology-fueled cost-inflation in healthcare. Each state’s CON board operated differently, but in general their goal was to limit unnecessary expenditures on new technologies in hospitals by regulating purchases which either exceeded some threshold in cost, or which represented a change in the hospital’s service offerings. Many states had a CON policy in place that required hospitals to demonstrate that the desired equipment wasn’t already available at other local hospitals, thereby creating surplus capacity. Hospitals that adopted the technology first reduced the chances that a state CON board would deny their application to purchase the expensive new machines.

The first FDA approvals began rolling out in 1984 and by 1988 eleven companies had conducted clinical trials and received PMAs.⁵ This convinced the FDA that MRI technology was no longer “novel,” leading to reclassification. MRI’s new FDA status allowed anyone to bring new MRI models or accessories to market without a PMA. The ensuing influx of products came mostly from two sources: established, multinational medical imaging companies who had won the majority of the market, and Japanese companies like Toshiba, Hitachi and Shimadzu, who had been selling machines in Japan since 1982. By 1990 there were over 2200 MRI machines installed in the US, and over 3000 machines installed worldwide.

This appears to be relatively fast adoption, but an analysis by Hillman and Schwartz in 1985 comparing the diffusion of CT and MRI in the US suggests the adoption of MRI by the medical community was actually slower than CT. They state that over 400 CT units were installed in the first 4 years of clinical availability, and they think that that statistic is actually conservative relative to how many hospitals/clinics put in orders because manufacturing couldn’t keep pace with demand. E.g. in 1975 there were twice as many orders for CT scanners as could be filled.

Investment from the medical device industry was a clear inflection point in the clinical implementation of MRI, but if it were not for the half-baked experiments of an ambitious doctor trying to cure cancer, the lack of summer grant money that pushed an academic chemist onto the board of a failing NMR business, competition that spurred academic research teams in San Francisco, Zurich, Aberdeen and London, a well-timed shakeout in the CT market, and the advances in computing and mathematics that went into the development of image reconstruction algorithms for CT, MRI may never have made it to the world.

Development Timeline

• 1971
  • Raymond Damadian publishes the results of his rat tumor experiments, which seem to indicate that NMR relaxation times could be used to identify cancerous tissues
  • Paul Lauterbur invents a technique for translating RF signals from NMR scans into spatial information to generate an image
• 1973
  • Paul Lauterbur publishes his breakthrough paper, which includes the first published MRI image
• 1975
  • Mansfield and his team proposed a technique which, in 1977, led to the first image of in vivo human anatomy, a cross section through a finger
• late 1970s
  • The CT market goes through a shakeout, and the companies that emerge as winners have resources to invest in new technologies, including MRI
  • Partnership between industry and academia marked MRI’s shift into the private sphere
• 1984
  • The FDA approves the first three companies to bring MRI to the US market
• 1985
  • Medicare begins reimbursing for MRI scans
• 1988
  • The FDA changes the status of MRI, allowing MRI companies to skip the PMA process
• 1990
  • Annual units installed and scans taken in the US increased tenfold throughout the 80s after FDA approval, with 2,200 units installed and 5.5 million scans taken

Appendix: Some choice quotes from Bettyann Kevles’s Naked To The Bone on EMI’s decision not to license CT technology

“But salespeople at EMI had no doubts about the importance of images as they made marketing decisions. Broadway advised granting patent licenses to American companies. He reasoned that EMI did not have the resources, or more important, the familiarity with the medical market, to be able to compete with established medical firms. EMI had already been contacted by every big medical equipment manufacturer, with the exception of Siemens, about cooperative ventures. EMI, however, flush with its recording bonanza (although company officials denied that the Beatles' billions had anything to do with their decision), opted to go it alone.
(Kelves 162)

“In the rush to upgrade CT scanners, the machines were fine-tuned, but the prices did not drop. EMI faced increasing competition. It had brought out its own second- and third-generation models and by 1975 had enlarged the computer display from 80 x 80 pixels, to 160 x 160. EMI, still a major manufacturer with machines all over the United States and Europe, fought to protect its lead by introducing a maze of lawsuits to obstruct the competition. Alan Cormack, who was asked to testify as an expert witness, describes EMI's legal manipulations as skilled brinkmanship in which they initiated many cases but stopped short each time before reaching open court, sealing the record as part of the out-of-court settlement.

EMI successfully held off competitors by changing the rules in an old game. A representative from GE, testifying at a congressional committee, explained:

Before CT, few X-ray companies bothered with patents, since all the X-ray companies recognized that no one company had a monopoly on the patents, each would have to license from the other to stay in business the result being "why bother with patents?" EMI, which was new to the X-ray business, patented their CT designs, and by the end of the decade was requiring substantial royalties from all the CT suppliers. As a defensive measure, the X-ray companies substantially changed their patent policies; for example, we went from having a part-time use of one patent attorney to having three full-time patent attorneys. This, I believe, was totally the result of EMI's changing the practice of an industry.

If EMI had been committed to medical instrumentation, it would have been in its interest to keep costs down and cooperate with its competitors. But EMI both underestimated and misunderstood the medical market. After Hounsfield's Chicago success, it sold the remaining prototype machines for $310,000 and in the next few years improved the machine and raised the price to about $500,000. EMI never tried to diversify into other medical instruments but looked on CT as a single product in the X-ray market. It had a large lead and competed successfully against the ACTA and Delta scanners. But, with the entry of GE and Siemens into the market, EMI found that the medical imaging business demanded frequent redesigns and a team of field representatives to attend to broken or damaged units. Pasadena's Huntington Memorial Hospital was not the only one that, fed up with relying on this or that new company for maintaining its very expensive machine, returned for their next purchases to GE and Siemens, whose representatives they knew and trusted.

In Britain, after EMI's astonishing initial success, there was no long-term plan. The Medical Research Council (MRC) was at least as eager to recoup its investment as EMI. They could not tap markets in countries with state-controlled health budgets like their own until the scanners had proved their value and the kinks had been worked out in the free market of American medicine. British engineers today still complain bitterly that British ingenuity, as displayed in the development of the scanner, suffered from a lack of correlative business leadership. But university-based British scientists are not complaining. The MRC continues to support basic research.” (Kelves 165)

Appendix: Key factors affecting development

• Government Funding
  • UK scientists had lots of support
  • Was hard to get funding in the US for development of technology vs. basic research
• Industry support
  • Lacking in the first 6 years during the CT boom
  • Was combined with gov’t funding in the UK
  • Brain drain of the UK, funded by CT, happened quickly
• Concurrent technological developments that improved MRI
  • Higher field strength resistive magnets
  • Faster/cheaper computers
  • Improvements in computer displays
  • Improvements in RF transmitters and receivers
• Collaboration or lack thereof
  • In the US Lauterbur and Damadian worked independently
  • No collaboration between the UCSF team and either of the NY teams
  • In the UK teams of physicists and clinical researchers teamed up
  • In the UK many teams competed against each other, and were well funded to do so
  • Lauterbur visited/consulted with the UK teams
  • Mansfield was involved with the team at EPI

Appendix: Key factors affecting adoption

• Changes in cost of machines and speed of imaging
  • Went down over time, early deals for hospitals who helped with clinical trials
  • Significant change when FDA approval allowed Japanese companies to enter the market, and they eventually made it cheaper
  • Technological improvements impacted cost a lot - changes in magnet technology, smaller cheaper machines with lower field strength
  • Fast fourier transform algorithm reduced speed and compute requirements
  • The high cost of CT scanners proved that the market was there for such an expensive machine
• FDA review process
• Certificates of Need and the influence of state health planning boards
• Reimbursement from 3rd party payers and Medicare and changes in healthcare policy over time
• Clinician education/training on how MRI differed from CT
• The proliferation of outpatient imaging centers

References

Bhidé, A., Datar, S., & Stebbins, K. (n.d.). Magnetic Resonance Imaging. https://www.hbs.edu/ris/Publication%20Files/20-001_40243e2a-ee15-46f9-9c8b-c0c91f309d20.pdf

EMRF Foundation. (2003). A short history of Magnetic Resonance Imaging from a European point of view. EMRF Online; EMRF Foundation. [https://web.archive.org/web/20070413032705/http://www.emrf.org/FAQs%20MRI%20Histor]{.underline}

Hisashige A. The introduction and evaluation of MRI in Japan. Int J Technol Assess Health Care. 1994 Summer;10(3):392-405. doi: 10.1017/s0266462300006632. PMID: 8071002.

Kevles, B. (1997). Naked to the bone : medical imaging in the twentieth century. Rutgers University Press.

Lauterbur, P. C. (1986). NMR imaging in biomedicine. Cell Biophysics, 9(1-2), 211–214. https://doi.org/10.1007/bf02797382

MacWilliams, B. (2003). Russian claims first in magnetic imaging. Nature, 426(6965), 375–375. https://doi.org/10.1038/426375a

MRI’s inside story. (2003, December 6). The Economist. https://www.economist.com/technology-quarterly/2003/12/06/mris-inside-story

Roessner, D., Bozeman, B., Feller, I., Hill, C., & Newman, N. (1997). The Role of NSF’s Support of Engineering in Enabling Technological Innovation: First Year Final Report. SRI International. https://web.archive.org/web/20041129175829/http://www.sri.com/policy/stp/techin/mri1.html

Steinberg, E. P. (1984). Nuclear Magnetic Resonance Imaging Technology: A Clinical, Industrial, and Policy Analysis. In Office of Technology Assessment Archive. Office of Technology Assessment. https://ota.fas.org/reports/8420.pdf

Volume 2. “Making the Human Body Transparent: The Impact of N.M.R. and M.R.I.; Research in General Practice; Drugs in Psychiatric Practice; The M.R.C. Common Cold Unit.” (1998, September). Wellcome Collection. https://wellcomecollection.org/works/w3rvtqfk

Young, I. R. (2004). Significant events in the development of MRI. Journal of Magnetic Resonance Imaging, 20(2), 183–186. https://doi.org/10.1002/jmri.20123

Have feedback? Find a mistake? Please let me know!