Z3J» 12-
`
`DEVELOPMENT OF NOVADAO SPY™ CABDIACIMAGING INVENTION
`
`I was interested in what happened to the heart when blood supply was stopped (e.g. heart attack
`due to vessel blockage, or stopping flow for surgery) and then restarted again (e.g angioplasty or
`restoration of flow at surgery). I was mainly using magnetic resonance imaging and spectroscopy
`techniques to investigate what was happening to the heart muscle during this ischemic period and
`at reperfusion. I was also interested in ways to look at what was happening to die blood vessels
`during this insult It was well known that during ischemia-reperfuslon (I-R) there can be
`extensive edema. I wanted to know if this was because the blood vessel walls were becoming
`more leaky i.e. 1 wanted a method to asses vascular permeability during I-R. In the course of.
`discussions .with other scientists at IBD it was mentioned that indocyanine green bound
`extensively to plasma proteins and fluoresced with the excitation and emission maxima being in
`tire near infrared part of die spectrum. We thought that this might be a good marker for vascular
`permeability as under normal conditions the ICG would stay within the blood vessels and, as It
`was bound to proteins, only leak out if there was significant damage to the blood vessel walL We
`decided to investigate this in our ischemia-reperfuston model in isolated rat hearts. This seemed
`like an ideal project for a graduate student so we put the project on hold until a suitable candidate
`was identified.
`
`In 1997 Rick'Manga t started as a graduate student in pharmacology at the University of Manitoba.
`Before starting on their thesis work in this program graduate students must rotate through a
`couple of labs. Rick came to work in mine (I had an adjunct professorship in the pharmacology
`department) as part of a rotation and at the end'Of this rotation decided to undertake the ICG
`imaging project for his PhD thesis. We spent a considerable amount of time trying to obtain
`fluorescence images of ICG. During this time we' were collaborating with a group of scientists
`and engineers from the Spectroscopy group at IBD — these people were experts in optical
`technologies including spectroscopy and fluorescence. From the literature (mainly Bob Flower’s
`publications) we learned that maximal ICG fluorescence was observed at concentrations in die
`range 10-30 pg/ml. We therefore reasoned that if we wanted to see ICG fluorescence in the blood
`vessels with a gradual shift out of the vessels as damage occurs during I-R we should infose ICG
`at this concentration over a prolonged period and observe the accumulation in the extravascular
`compartment Rather than use animals in all of these early experiments we used some phantoms
`(e.g grapes) in addition to isolated rat hearts.
`
`Our experimental set up included a very sensitive camera with quite long integration times (a few
`frame? per second??) equipped wife a liquid crystal tunable filter. The camera cost irithe region
`of $40K, Le. this was a state of the art set up as might be expected in a leading
`optics/fluorescence research lab. This was used in conjunction wife a 50 mWatt laser. We were
`unable to observe any fluorescence. Rick contacted Bob Flower by phone on several occasions
`seeking advice. Bob even sent us one of his 2 Watt lasers and power driver but still no success.
`We approached fee group leader of spectroscopy (Dr. Henry Mhntsch) for funds to bring Bob to
`Winnipeg, This request was refused and the project was put on the backbumer while we turned
`our attention to a different approach - the use of NIR spectroscopy to look at other aspects of I-R
`injury.
`
`This project was a bit disappointing in that in spite of the advances that had been made in
`technology the data that we could acquire was no more informative than fee information that Dr.
`Britton Chance had gleaned 20 years previously. We did not want to pursue this. We turned our
`attention hade to the fluorescence'imaging project, but still with no success even when we tried
`things like pre binding ICG to albumin, purifying this product by molecular sieve
`chromatography and infusing this into rat hearts. In desperation we approached the Btosytems
`
`Page I of5
`
` VISIONSENSE - 1011
` VISIONSENSE v.
` NOVADAQ TECHNOLOGIES
` Page 1 of 5
`
`

`

`group leader. Dr. Roxanne Deslauriers (John Docherty’s direct supervisor) and requested funding
`to bring Bob Flower to Winnipeg stating that if he could not come we would close dawn the'
`project
`
`Bob Flower visited Winnipeg in December 1998 for a weekend of lab experiments. We acquired
`very promising images of coronary arteries in the rat heart The two key changes that allowed this
`to happen were:
`
`1. The delivery of ICG as a bolus of higher concentration that arrived at the field of ■
`view at the appropriate concentration by mixing with bulk perfusate.
`2. Replacing our $4QK camera that acquired with long integration times with an $800
`camera that acquired images at video rate, i.e. 30 frames per second
`
`The imaging system at this point basically consisted of a 2 watt laser passing through a lens and
`positioned far enough back from the heart to illuminate the entire surface and a Hitachi camera
`with a 330 nm endwStena1 bandpass filter in front of the lens. We worked on tills for a short time
`optimizing the image quality (changing ICG dose, slight changes to filter etc) but soon realized
`that the ICG does not leak out of the vasculature within the time frame that we are looking at plus
`the background flush caused by filing of the micro vasculature precluded any way of looking at
`vascular leakage within the spatial resolution that we were looking aL So at this point we realized
`that what we could acquire were angiograms -how could we use this tool in our research. At this
`point I was starting to collaborate with molecular biologists who were working on a series of
`transgenic animal models of human disease - could we look at the intact animal. At this point we
`started doing some studies on mice (mot transgenic models were mice) and started with the
`femoral artery vascular bed with the idea o Hooking at changes in diameter in response to various
`drugs in normal mice and transgenics.
`
`As these studies were moving along a colleague who was investigating the delivery of
`cardioplegia in isolated blood perfused pig hearts asked if we could use our system on his model.
`He was using MR perfusion imaging techniques but the resolution was not quite good enough to
`answer his questions. We imaged some isolated pig hearts and acquired some quite outstanding
`angiograms, We then thought -“could this be of clinical benefit, especially to surgeons”. We lad
`Roxanne send a fax to Dr. Wilbur Keon (with whom she had collaborated on a number of
`scientific projects) explaining what we were doing and if he thought that it would be of benefit.
`Dr. Keon was one of the thought leaders in cardiac surgery in Canada (and a Senator in the upper
`house of the Canadian Parliament). This very busy person responded within a day to say “yes,
`this is very important - I will be visiting Winnipeg within .the next few weeks, show me the
`images and we can discuss”. Dr. Keon saw the images, explained that it was his practice to inject
`a very high dose of ICG down bypass grafts and lode at tire surface of the heart to see where it
`turned green. What we had just done was a quantum leap ahead of what he was doing (his words).
`He arranged for us to contact one of his junior surgeons with a view to testing this in humans
`when an appropriate device was ready for human use. Attention then turned to developing and
`validating such a device.
`
`We (largely Rick) had a prototype device EHDN0Tt built by a company in Toronto. It was very
`basic,' bringing the laser and camera together in one box to make it a simple point and shoot
`device rather than messing about for 20 minutes independently positioning foe 2 components.
`This device was used to acquire some data from pig hearts for Rick’s thesis (the data never did
`make it into his thesis). A second prototype was built by the same company - basically making it
`more ergonomically friendly. At this time I received grant funding from a local agency to perform
`a proof of concept study in pigs after they had bypass graft surgery (a collaboration with a local
`,Page2ofS
`
` VISIONSENSE - 1011
` VISIONSENSE v.
` NOVADAQ TECHNOLOGIES
` Page 2 of 5
`
`

`

`cardiac surgeon). The results of this study proved that we could successfully image bypass grafts
`*
`* ENDnOTE
`in pigs.
`
`The company CWNOTENe'4 was formed (April 2000) and an agreement entered into with Colorado
`MedTec to design and build a device that could be used in humans (Pall 2000). A two day kick
`’ off session in Boulder, CO was used to define the design requirements for the device. A Medtec
`team was put in place - optical engineer, electrical engineer, mechanical engineer, industrial
`designer) and first prototype ENDN0TK N*s was ready for human use in January 2001, The device
`was again primarily bringing together the laser and camera in one unit with a mobile cart and
`articulated arm. The first images were terrible. We had used the same filters (830 nm bandpass)
`endnotb «■>•< ^ had been used for the pig experiment Of course the pigs were relative babies
`(even though they weighed 70 kg) and the arteries were sitting right on the surface of the heart
`with no fat present The patients were not so young, had arteries that did not Sit on the surface of
`the heart and there were very large deposits of fat- we were not getting enough signal.KNDN0T8tN^
`7 The next few months were spent trying out a series of filters (extensive input from Bob Flotoer,
`including trying out some samples of his filters) and we ended up with an 815 nm cut filter that
`very efficiently transmits light >815 nm so that we could now capture pretty much all of the
`emitted fluorescence. endno1®n“'s&* This early device made use of a VCR to capture the images
`(even though we started with PC and frame-grabber in the lab) but a year later we switched over
`to PC to increase image quality and ease of image distribution etc.
`
`ENDNOTES;
`
`endnote No. x: -pjjjj js a typographical error. The filter comprises a barrier bandpass filter of 830
`nm, commensurate with the peak fluorescence wavelength of ICG.
`
`prototype device employed a Class Mb Lasiris Magnum 810-3000-20°
`endnote No. i!
`semiconductor laser having a nominal optical power output of 2.7 Watt at 808 rim. The laser was
`tuned to a wavelength of 806 nm. The delivered optical power at the illuminated field (7.5 cm x
`7J cm; 30 cm working distance) was 2J25 Watt, for a power density of 40 mW/cm2. The camera
`was a Hitachi ECPM2 RN ’A” CCD Monochrome Video Camera with a bandpass filter (CVA
`Laser Coip, FIQ—830.0-4-2.00 (830 ± 2 ran bandpass, 10 nm full width half maximum
`(FWHM)) placed between the camera and the illuminated tissue to suppress excitation light A
`CosmicariPentax 16 ram f 1.4 lens was used with an f-stop of F12, The laser was housed in the
`imaging head.
`
`endnote N*.Ji 7^5 study involved ICG imaging of bypass grafts for pig hearts in which ICG
`images of excellent quality were acquired. We confirmed that this method was also applicable
`for CABG in humans by the time of this study, at the latest. Since the invention was completed,
`we filed a provisional application to the United States Patent and Trademark Office on September
`24, 1999, AU the practical adjustments stated in this memo after this Endnote No. 3, including'
`the selection of the filters, were conducted based on a device, etc., identical to the invention.
`
`endnote jbe company referred to is Novadaq Technologies Inc., headquarted at Toronto,
`CANADA.
`
`Page 3 of5
`
` VISIONSENSE - 1011
` VISIONSENSE v.
` NOVADAQ TECHNOLOGIES
` Page 3 of 5
`
`

`

`endnote,no. s; j^e la$er wa$ a 2.6 Watt Class IDb Coherent Semiconductor F-81-2600C-200-B
`laser having a nominal optical power output of 2,6 Watt at 808 run. The laser was tuned to a
`wavelength of 806 nm and emitted 2.0 Watt, for a power density of 36 mW/cm1 at the illuminated
`held. This system utilized the same bandpass filter (830 ± 2 nm bandpass, 10 run FWHM) that
`• was used for the pig studies, hi the clinical device, the laser light is delivered to the imaging head
`by means of fiber optics. The only components of the illumination system now housed in the
`• imaging head are the fiber guide and the lens for expanding the laser output from the fiber.
`
`endnote No.*, The example given, in the patent relates to imaging of the small animal femoral
`• bed and the camera employs a 50 mm lens to provide sufficient magnification. The 50 mm lens
`has a different diameter compared to the 16 nun lens used in the pig heart studies, so that lens
`filter used in the pig studies was not compatible in size for this study. The 830 nm bandpass filter
`used has lOnm FWHM, whereas the 845DF25 bandpass filter disclosed in the patent has 25 nm
`FWHM. However, there is no significant difference between the two filters with respect to the
`transmitted fluorescence signal intensity in spite of their different FWHM, because the bandpass
`of the 830 nm filter is closer to the peak fluorescence of ICG than the bandpass of the 845 run
`filter. Therefore, both filters should perform equally well in either application.
`
`endnote n«. 7j |n add-on to the noted anatomic differences between the hearts of the young
`pigs and the hearts of the elderly human patients, the imaging camera was enclosed in a head
`inside a sterile drape, which precluded changing the camera’s f-stop in the operating room during
`imaging. Notwithstanding the above, in feet, we could .obtain some signal which was sufficient
`for the purpose of the invention. However, i f we compared these to the images we obtained from
`■the pigs, they were not as good as expected, which is why I stated, “27ie first images were
`terrible”. That is, while the images were in feet acceptable, they were simply not as impressive
`as those obtained with the pigs. Thus, some practical adjustments were made in order to obtain
`better images.
`
`to quickly improve the image quality without making extensive
`endnote Nn.si
`adjustments to the setup, we elected to use an 815 nm cut filter, instead of the 830 nm bandpass
`filter previously used. The 815 nm cut filter efficiently transmits light at wavelengths greater
`than 815 nm, while blocking substantially all excitation light. The 815 nm cut filter has about a
`five-fold higher transmission of the total fluorescence signal emitted by the ICG dye over the 830
`nm bandpass filter, but can also transmit a certain amount of unwanted ER. outside the ICG
`fluorescence spectra.
`
`endnote no.9, jt j$ important to note that the 830 ran bandpass filter ultimately worked
`successfully on human grafts in clinical trials. A study was conducted at two (2) Canadian
`institutions, namely: the University of Ottawa Heart Institute in Ottawa, Ontario, CANADA;
`and the Sunnybrook and Women’s Health Sciences Centre in Toronto, Ontario, CANADA. The
`imaging device (Le., SPY™ cardiac imaging device, provided by Novadaq Technologies Inc,, of
`Toronto, Ontario, CANADA) comprised a laser diode and a driver that produced light at a
`wavelength of 806 nm and at a maximum output of 2.7 W.
`
`The laser output was de-collimated to provide illumination, spread uniformly over
`a 7.5 x 7.5 cm field of view at a working distance of 30 era, which was identical to tire
`dimensions used for the earlier pig heart studies. Images were acquired at a rate of 30 frames /
`second, using a charge-coupled device (CCD) camera sensitive to near-infrared light and
`equipped with an optical filter for the selective transmission of light at 830 nm. The laser, optics,
`
`Page 4 of 5
`
` VISIONSENSE - 1011
` VISIONSENSE v.
` NOVADAQ TECHNOLOGIES
` Page 4 of 5
`
`

`

`camera, and filter were integrated into an Imaging head supported by .a mobile arm and connected
`to a wheeled cart, which allowed for the system to be moved close to the surgical table at a
`correct focal distance above the area of Interest
`
`Page 5 of 5
`
` VISIONSENSE - 1011
` VISIONSENSE v.
` NOVADAQ TECHNOLOGIES
` Page 5 of 5
`
`