Diffusion Tensor Imaging (DTI)
Structure of the historical evidence of invention
The 5,560,360 Patent (click here to download; PDF format) starts with a list of “Foreign Application Priority Data” and this lists the provisional filings.
Provisional filings have the status of fully competent verified data of what was done. The exact text is saved by the UK government and the exact date and time of filing is stamped on the document. These filings provide unavailable facts as to historical events of invention. An inventor has one year after the initial provisional is filed at which time the definitive patent application is filed, but it is entitled to the date of priority of any part of the invention that appears in the provisionals as of the time that the particular idea or proof at issue is filed.
A key filing is the May 21, 1992 provisional (GB9210810; click here to download; PDF format) that states that numerous directions of diffusion gradients will be used to generate numerous diffusion images. The data for each voxel will then be treated to a mathematical analysis to determine a direction and magnitude for each voxel. This is a very fundamental idea of the invention:
“If the major axis of diffusional anisotropy is not known, a plurality of images are acquired with different orientations of the diffusion sensitizing gradients. The image data can then be processed to give a parameter associated with each pixel (or voxel in a 3D data set) which is a measure of the diffusional anisotropy at that point reflecting both magnitude and direction.”
GB9210810 (May 21, 1992) page 17, paragraph 1. This is key because this was the status in the UK when only Filler and Howe were inventors – the day before Filler returned to Seattle to resume work with Todd Richards which had commenced in 1988.
A similar statement appears in the March 13, 1992 provisional GB9205541 at page 13 (click here to download; PDF format). This shows that the basic invention was fully articulated within one week of the very initial filing on March 9. In the filing on Monday, March 9th, 1992 (GB9205058; click here to download; PDF format) – Filler and Howe described using vectors for three orientation of gradients. In the filing five days later on Friday March 13 filing this had been clarified and progressed to explain the idea of using a large number of directions of gradients (GB9205541; click here to download; PDF format).
This is then made more explicit in the following July 21, 1992 priority document (GB9210810) which states:
“The use of vector analysis algorithms of this sort, or involving the treatment or coordinate transformation of MR diffusional anisotropy data with TENSORS of various rank can improve the generality and tlexibility of neurographic imaging. The example described above demonstrates that by the application of TENSOR and/or vector analysis methods such as algorithms similar to those developed for the evaluation of e.g. magnetic, thermal, or structural anisotropy data, it is possible to greatly improve the flexibility and generality of image techniques for neurological diagnosis.”
GB9216383 (July 21, 1992; click here to download; PDF format) page 21, paragraph 4. This is important because the text references the use of tensor analysis in magnetic and structural data. This refers to our work at Harvard using data obtained from multiple directions for geological samples to determine the orientation of magnetic fields in paleontological work and for the use of this with strain gauges. There is no need to describe in a patent methods that are already in use in hundreds of publications because these are considered to be well known to a “PHOSITA” a Person Having Ordinary (non-inventive) Skill In The Art.
We then go on to explain how the vectors will be used for tractography:
“Such techniques can also he used to follow continuous serial changes in the direction of maximum anisotropy of a nerve or neural tract as it travels along its natural course.”
The first tractograms ever created appear as figures 12 and 13 in this provisional. They have been on file in the UK patent office for 25 years with absolutely crystal clear context. Nothing Peter Basser (who sometimes claims to have made this invention later) or anyone else can say will make these figures go away. Basser – who filed the pork loin rotation patent, states himself that in 1994 (two years later) he thought the idea of doing tractography was still “science fiction” (his own words).
The first public presentation of any of this material was in August of 1992 in Berlin at the ISMRM meeting.
Therefore, the month before Basser’s abstract and before Todd Richard’s abstract were shown to the public, we had filed with the patent office the July 1992 statement of obtaining the diffusion date from multiple different directions, using tensors and doing tractography and had actually filed images of the first such tractograms – of monkeys who were either normal or who an experimental allergic encephalomyelitis (EAE) – to show that tractography based on anisotropy could detect the encephalomyelitis that did not show on standard MRI. In Basser’s filing they were just manually rotating a pork loin.
In the patent itself, we clarify the “plurality” term:
“Alternatively, gradient coils oriented in three planes can be simultaneously activated in various combinations to achieve the effect of an infinite variety of differently oriented gradients.”
US 5,560,360 Column 20, Line 5 to 8 (March 8, 1993)
Peter Basser’s patent was originally filed on August 6, 1993. That means that the earliest provable priority document could not have been earlier than August 6, 1992 (the day before the Berlin meeting). Yet we had filed for multiple angles, tensors and tractography more than two week earlier with a priority document on July 21, 1992, although we had filed the concept in May, and in fact first described the concept in the March 13, 1992 document – five months ahead of Basser’s filing.
While it is easy to understand Basser’s level of frustration, that counts for nothing in answering the question of who invented. He may have filed a limited focused patent provisional on pork loin data the day before the Berlin meeting (which he had not moved beyond by the time of the definitive patent filing on August 6, 1993 – a year later), while we had filed a complete tractographic image of a live monkey’s brain obtained by multi–directional diffusion imaging – on July 21 (17 days before the Berlin meeting) that does not make our claim erasable or irrelevant as he seems to argue in various communications. It just means that Basser had too little too late to compete on inventing and should not try to take credit for that. His most important work was in promoting the technology later.
Academic Interactions on DTI Near the Time of Initial Disclosure
In this correspondence (click to download; PDF format), we can see that in a letter of June 16, 1992, Filler was writing to Professor John Griffiths (University of London/St. George’s) about discussions with GE and specifically describe using tensor algorithms to tract trace. However, this correspondence doesn’t have the official proof of date and content provided by the GB patent provisional filings.
Moseley called Filler after seeing the posters on neurography and on tractography at the August 8-14, 1992 Berlin meeting of the SMRM (Society for Magnetic Resonance in Medicine) to ask permission to use the images at his plenary session in March of 1993 at the San Francisco meeting of the SMRI (Society for Magnetic Resonance Imaging). On page 24-25 of the abstract book, we see that Denis LeBihan was the session moderator. The Moseley abstract uses our terminology “MR Neurograms” which we used in the patent for all of the techniques in relation to diffusion and white matter. Basser and Le Bihan filed their patent nearly five months after LeBihan was on the podium where our work was presented.
In his following letter of April 8, 1993 (click to download; PDF format), Moseley acknowledges the use of the slides and states he will be suggesting including this in a book that LeBihan is working on.
Note that others, such as John Mazziotta (now vice chancellor of UCLA and CEO of UCLA) went around the world showing our images (Letter of August 26, 1997; click to download; PDF format).
Mosely Society for Magnetic Resonance Imaging 1993 Eleventh Annual Meeting, Plenary Symposia (click to download; PDF format).
It may sound plausible to some if Basser says it never happened, but at the time, Moseley and Mazziotta were praising it, and the filed patent documents make the specifics 100% provable.
PTAB ruling (Patent Trial and Appeals Board) and Failure of All Attacks on Validity of the Patent
The PTAB is described by former Chief Judge of the Court of Appeals of Federal Circuit Randall Rader as a “death squad for patents” (Address at the American Intellectual Property Law Association Annual Meeting (Oct. 25, 2013) and a place where “good patents go to die” (http://whoswholegal.com/news/features/article/33156/how-united-states-patent-office-became-place-patents-go-die/). Any patent that survives reexamination by the PTAB is therefore “platinum plated” as to validity.
After two attempts at reexams and full litigation, the PTAB knocked out six of our 64 claims and left every single Diffusion Anisotropy claim intact (see Decision on rehearing of February 18, 2015; click to download; PDF format. This was confirmed upon review by the Court of Appeals of the Federal Circuit. Nine years of full scale attack on the 5,560,360 patent in Federal Courts by Siemens, GE, Philips, Toshiba, Hitachi, Medtronic, BrainLAB, The State of California and the the United States Department of Justice (in Neurografix v. United States in the Court of Federal Claims) all failed to dent the validity of the patent. Despite an army of lawyers with unlimited budgets from seven large multi-national corporations using the largest and most feared patent firms in the world resulted in the patent being still valid. So when Peter Basser says he thinks the patent is invalid (click to download; PDF format)– it is a statement in utter contradiction with the facts on the ground and clearly recorded in the public record.
Markman hearing as to Tensors and Vectors
Most importantly, the Markman Claim Construction hearing in the Multi-District Litigation finally and definitively determines that the claims cover both tensor and vector methods (see Order on Claim Construction of August 19, 2016 (page 24, paragraph 2; click to download; PDF format). These were clearly described in the text – there was a dispute over whether the claims covered the invention adequately, but that too has been decided by the Federal Courts in favor of the Filler patent.
Also attached is the interview in which Basser states that although he used Mathematica to make what he thought tracts would look like – he states that in 1994 he thought that actually making tractograms from MRI images was “science fiction” see – page 10 (click to download; PDF format).
So – Peter Basser admits that two years after the Filler patent documents showed the first tractograms – he thought it was impossible to do what had already been done. How could he be ahead of the Filler group on this? Basser needs to accepts the incontrovertible facts on the record.
Publications that explain the history of the invention
The 2010 article recounts the various contributions (click to download; PDF format).
The 2009 article has some further mathematical work by Filler (click to download; PDF format).
Filler’s undergraduate thesis used multi-demsional math, and factor analysis on a BMDP system to analyze brain measurements (1977), working under Charles Oxnard at University of Chicago.
Filler taught about the use of tensor analysis to evaluate strain gauge data in the Bio21 course at Harvard in 1980 to 1984, during which time Basser – who was an undergraduate – would have heard these course lectures. Filler also taught a course in Anthro on biomechanics that dealt with the use of vector and tensor methods for analysis of anatomical data. Publications from the Pilbeam group while Filler was working his PhD prove that the exact mathematics used in DTI was used by the Harvard Anthro group Filler was working in the early 1980s (see Tauxe 1990; click to download; PDF format) in relation to Professor David Pilbeam’s (form Dean of Harvard College and Chairman of the Department of Anthropology at Harvard) paleontological work in the Siwalik mountains studying fossils as evidence of human evolution (Lisa Tauxe and Catherine Badgley).
Basser’s work in use of tensors in analyzing strain gauge data was his field before he got interested in LeBihan’s work. Basser then applied the strain gauge tensor math to LeBihan’s brain diffusion data. However, as shown from the Basser and LeBhian patent (US), the Filler group was two or three years ahead of them. The Filler project started in 1988 at University of Washington.
Filler’s Pioneer in Medicine Award
Filler was awarded the Pioneer in Medicine Crystal Award in 2016 by the Society of Brain Mapping & Therapeutics.
“The SBMT Board is well aware of your dedication to scientific excellence and significant contributions to scientific advancement in the area of brain mapping and therapeutics. Your unique contributions to the field of brain imaging (US Patent 5,560,360) have been translated to clinically important advances in the diagnosis and potential treatment of patients with brain diseases worldwide.”