Nuclear magnetic resonance (NMR) was discovered in the late 1940s by groups at Harvard and Stanford, led by Edward Purcell and Felix Bloch, and turned into a rudimentary imaging modality by Paul Lauterbur and Peter Mansfield. All of these men won the Nobel Prize.
GRC retiree William (Bill) Edelstein played a key role in translating these preliminary ideas into high quality magnetic resonance imaging (MRI) and a robust industry, one that continues to be a flagship modality for GE Healthcare today. “For his pioneering developments leading to the commercialization of MRI for medical applications,” Bill was awarded the American Institute of Physics prize for the Industrial Applications of Physics in 2005. He passed away in February while still actively working to make MRI machines better.
Never Give Up
Bill’s early scientific career was not auspicious. Unable to find a job in the US after earning an experimental nuclear physics PhD, he landed a postdoc in Glasgow, Scotland. Instead of detecting gravitational waves per plan, he wooed and wed Fiona Jones, a Scottish lass.
The Edelsteins moved north to another postdoc at Aberdeen. The small team there was working on a new application of nuclear magnetic resonance, making anatomical images. At that time, a few groups had demonstrated tomographic reconstructions of small objects, like test tubes and mice, whereas the Aberdeen team’s goal was a whole body scanner.
In stereotypical Scottish form, the project’s budget only supported a weak and poor quality whole-body electro-magnet, so the team spent most of its time dealing with mundane engineering tasks like avoiding spurious signals from AM radio stations. But the poor quality magnet had a benefit – it forced the team to find an alternative to the popular projection reconstruction tomography scheme. They needed an imaging method that was less sensitive to their magnet’s deficiencies. The Aberdeen team recounts how while trying to understand another researcher’s paper, Bill recognized about five problems with that paper’s method, and more importantly ways to change the approach to overcome these limitations. They dubbed their approach “spin warp” and it remains central to every commercial MRI scanner to this day. Early in 1980 they had a whole body image with clearly recognizable features (Fig1.a). In later years Bill described getting the job in Aberdeen as winning the scientific lottery, and frequently reminded co-workers about the importance of luck.
Bill arrived at the then-GE Corporate Research and Development (CRD) in the late summer of 1980 joining Rowland “Red” Redington’s team. Red, the legendary manager who already had led the very successful CT projects at CRD, had slowly assembled a team to build an MRI machine using a superconducting magnet at 1.5 Tesla(T), more than 35 times stronger than the Aberdeen field and more than 3 times stronger than any other whole-body scanner. At 1.5T they planned not only to view anatomy but also to perform chemical spectroscopy of functional metabolites.
MRI is a tradeoff among imaging time, contrast, and field strength; the performance should get better as the magnetic field strength increases. At that time there was a commonly held belief that the RF fields needed at 1.5T might just get into the head, but would certainly not penetrate the torso. Whole body imaging wouldn’t work. More than one CRD manager referred to the literature and listened to external experts like the author of “the book” on nuclear magnetic resonance, who also happened to be the president of the National Academy of Sciences, and periodically reminded Red’s team of their folly.
The Niskayuna team systematically worked through many engineering difficulties associated with getting good images at this dramatically higher field. Bill was often leading the charge with a keen sense of the critical roadblocks, and developing or inventing (he authored more than 50 patents) technology along the way to solve them. You may view Bill’s own description of this work to some current students here. He touched a wide range of crucial barriers; novel, robust high frequency antennas for head and body imaging shielded gradient coils to reduce noise and interactions with the magnet, pulse sequences to optimize contrast to noise, surface coils and their associated electronics, fundamental studies to understand electronic signal to noise, and phased arrays of surface coils for large-area, high resolution images. In Lauterbur’s original experiments and the low magnetic fields used in Aberdeen, MRI is a curiosity, whereas at 1.5T it is a formidable clinical tool and a viable commercial industry. (Figure 1.b)
Science with Style
Bill’s technical work had style, a unique fingerprint, as it were. When Bill worked on a project it always included a technical memo describing the problem’s key features, the underlying physics, and the experimental data describing the result and comparing it to the theory. Not only did Bill do this himself, he insisted everybody else on the team carefully document their progress in technical memos that he annually collated and distributed to the team and grateful business engineers. Although I only worked closely with Bill for about five years in these early NMR days, I’ve continued to find bookshelf space for these valuable tomes. (Figure 2).
Bill understood customer focus and FastWorks long before the terms even existed. During a visit some customers whined about patient claustrophobia in a head coil where wide copper strips blocked the view. Grumbling, Bill disappeared into the lab, returning later to demonstrate excellent performance from a new coil on a clear Lexan tube where the wide copper strips had been replaced by thin wires that provided minimal obstruction. The customers went away delighted.
Bill did everything with passion and optimism, and loved to share his learning with others. He vehemently preferred the technical term NMR to the abbreviation MRI because he felt the public needed to understand that the atomic nucleus was at the center of all atoms, and most of these nuclei were perfectly safe. It was only the anomalous radioactive ones where caution was needed. Not only did he promote NMR in the technical literature, he also publicly advocated for the scientific term on the license plate of his family car.
When introduced to a computer algebra program he used for NMR calculations, he was compelled to show it to everybody and encouraged us to use it for all engineering analyses. To advocate for stronger use of online library services, he campaigned for a seat on the Whitney Library committee. In response to a question from his young son, he calculated the radiation experienced by astronauts during warp-speed travel. Finding lethal exposures, Bill reveled in the extensive press coverage generated by his conference talk despite disbelieving reactions from Star Trek fans. Through it all Bill defended scientific integrity with a strong commitment to logical and technical accuracy, along with fierce advocacy for peer-reviewed publication.
To boldly go …
For launching GE into the magnetic resonance business, Bill shared the 1983 Dushman Award. Later he would receive a Coolidge Fellowship, the Gold Medal of the Society of Magnetic Resonance in Medicine, and become a fellow of the American Physical Society, the Institute of Physics (UK), and the International Society of Magnetic Resonance in Medicine. In 2013 he received the Alumni Achievement Award from the University of Illinois, his undergraduate alma mater. As the highest award presented to an alumnus, his portrait hangs in honor in the university’s Illini Union.
After retiring from GE in 2001, Bill served as a senior research associate at Case Western Reserve University and visiting professor at Rensselaer Polytechnic Institute before moving to Baltimore as a Visiting Distinguished Professor at Johns Hopkins. He remained dedicated to improving the MRI patient experience, by reducing acoustic noise and examination time, up until his death.
In addition to Fiona, his wife of thirty six years, children Arthur, Jean, and Elspeth, and two grandsons, he leaves behind unforgettable memories amongst a large group of grateful mentees, collaborators and colleagues, and copious fingerprints on MRI imagers that will continue to benefit people for years to come.
* W A Edelstein et al 1980 Phys. Med. Biol. 25 (4) 751 doi: 10.1088/0031-9155/25/4/017. (Copyright Institute of Physics and Engineering in Medicine. Published on behalf of IPEM by IOP Publishing Ltd. All rights reserved).