U.S. patent application number 14/549057 was filed with the patent office on 2015-05-21 for systems and methods for hyperspectral analysis of cardiac tissue.
The applicant listed for this patent is The George Washington University. Invention is credited to Narine Sarvazyan.
Application Number | 20150141847 14/549057 |
Document ID | / |
Family ID | 53173995 |
Filed Date | 2015-05-21 |
United States Patent
Application |
20150141847 |
Kind Code |
A1 |
Sarvazyan; Narine |
May 21, 2015 |
SYSTEMS AND METHODS FOR HYPERSPECTRAL ANALYSIS OF CARDIAC
TISSUE
Abstract
Systems and methods for hyperspectral analysis of cardiac tissue
are provided. In some embodiments, a method for visualizing
ablation lesions includes illuminating at one or more illumination
wavelengths a surface of tissue having an ablation lesion;
collecting a spectral data set comprising spectral images of the
illuminated tissue acquired at multiple spectral bands each at one
or more acquisition wavelengths; distinguishing between the
ablation lesion and an unablated tissue based on one or more
spectral differences between the ablation lesion and unablated
tissue; and creating a composite image of the tissue showing the
ablation lesion and the unablated tissue.
Inventors: |
Sarvazyan; Narine; (Potomac,
MD) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
The George Washington University |
Washington |
DC |
US |
|
|
Family ID: |
53173995 |
Appl. No.: |
14/549057 |
Filed: |
November 20, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61906769 |
Nov 20, 2013 |
|
|
|
Current U.S.
Class: |
600/478 ;
600/476 |
Current CPC
Class: |
A61B 2018/00357
20130101; A61B 5/0036 20180801; A61B 18/1492 20130101; A61B
2018/0212 20130101; A61B 2017/00061 20130101; A61B 5/0071 20130101;
A61B 2090/373 20160201; A61B 5/4836 20130101; A61B 2017/00057
20130101; A61B 2018/00351 20130101; A61B 90/37 20160201; A61B
2090/3735 20160201; A61B 5/0084 20130101; A61B 5/0075 20130101 |
Class at
Publication: |
600/478 ;
600/476 |
International
Class: |
A61B 5/00 20060101
A61B005/00 |
Claims
1. A method for visualizing ablation lesions, the method
comprising: illuminating at one or more illumination wavelengths a
surface of tissue having an ablation lesion; collecting a spectral
data set comprising spectral images of the illuminated tissue
acquired at multiple spectral bands each at one or more acquisition
wavelengths; distinguishing between the ablation lesion and an
unablated tissue based on one or more spectral differences between
the ablation lesion and unablated tissue; and creating a composite
image of the tissue showing the ablation lesion and the unablated
tissue.
2. The method of claim 1 wherein the surface of tissue is an
endocardial surface of atrial tissue.
3. The method of claim 1 wherein the one or more illumination
wavelengths are between about 350 nm and about 400 nm.
4. The method of claim 1 wherein the one or more illumination
wavelengths are between about 400 nm and about 700 nm.
5. The method of claim 1 wherein the one or more illumination
wavelengths are between about 700 nm and about 900 nm.
6. The method of claim 1 wherein the one or more illumination
wavelengths are below a range of acquisition wavelengths by between
about 10 nm and about 50 nm.
7. The method of claim 1 further comprising filtering light
returning from the illuminated heart tissue using a set of bandpass
filters.
8. The method of claim 1 further comprising filtering light
returning from the illuminated heart tissue using a tunable
filter.
9. The method of claim 1 further comprising detecting a light
returning from the illuminated heart tissue, the detected light
including fluorescence, reflectance and scattering components,
wherein the fluorescence component being detected at the
acquisition wavelength between about 400 nm to about 500 nm, the
reflectance component being detected at the acquisition wavelength
between about 450 nm and about 700 nm, and the scattering component
detected across the entire visible spectra.
10. The method of claim 1 further comprising classifying each pixel
of a digital image of the illuminated tissue as either ablated or
unablated tissue.
11. The method of claim 10 further comprising constructing a depth
map of the ablation lesion from the image based on a bulk density
of the pixels of the digital image classified as ablated
tissue.
12. The method of claim 1 further comprising identifying the
ablated lesion as a lesion created by radiofrequency; and setting
the illumination wavelength to between about 400 nm and about 700
nm and the acquisition wavelength to about 400 nm and about 700
nm.
13. The method of claim 1 further comprising identifying the
ablation lesion as a lesion created by cryoablation or
radiofrequency; and setting the illumination wavelength to between
about 350 nm to 400 nm and the range of acquisition wavelengths to
about 380 nm to 500 nm.
14. The method of claim 1 further comprising illuminating at one or
more illumination wavelengths a specific wavelength or using wide
band illumination with a known spectral distribution.
15. The method of claim 1 further comprising distinguishing between
the ablation lesion and an unablated tissue based on one or more
spectral differences of a pre-selected set of multiple spectral
differences.
16. A method for visualizing atrial ablation lesion, the method
comprising: illuminating one or more discrete illumination
wavelengths a surface of heart tissue having an ablation lesion;
collecting a spectral data from the illuminated heart tissue;
distinguishing between the ablation lesion and an unablated tissue
based on one or more spectral differences between the ablation
lesion and unablated tissue; and creating an image of the heart
tissue illustrating ablated tissue and unablated tissue.
17. The method of claim 16 wherein the collecting of the spectral
data includes one of a detector-based Hyperspectral Imaging (HSI)
system, a detector-based HSI system using static devices with
tunable filters, an alternative source-based HSI including a
changeable wavelength of illuminating light or some combination
thereof.
18. The method of claim 16 wherein the collected spectra data is
one of matched to existing spectral libraries, subjected to a
principal component analysis, subjected to related principal
component analysis algorithms or some combination thereof.
19. A system for imaging tissue comprising: a catheter having a
distal region and a proximal region; a light source; an optical
fiber extending from the light source to the distal region of the
catheter to illuminate a tissue having a lesion site in proximity
to the distal end of the catheter; an image bundle for collecting
light reflected from the illuminated tissue; a camera connected to
the image bundle, the camera being configured to gather
hyperspectral data comprising spectral images of the illuminated
tissue acquired at multiple spectral bands or at each illumination
wavelength; an image processing unit in communication with the
camera, the unit being configured to distinguish between the
ablation lesion and an unablated tissue based on one or more
spectral differences between the ablation lesion and unablated
tissue and creating an image of the heart tissue illustrating the
ablated tissue and the unablated tissue.
20. The system of claim 19 wherein the light source includes one of
one or more bands, a set of switchable light sources or a light
source with at least one tunable filter.
21. A system for imaging heart tissue comprising: an illumination
device configured to illuminate a tissue having a lesion site; an
imaging device configured to gather hyperspectral data; an image
processing unit in communication with the imaging device, the image
processing unit configured to processing gathered hyperspectral
data to generate an image that reveals the lesion site, wherein the
generated image enabling to distinguish between an ablation lesion
and an unablated tissue based on one or more spectral differences
between the ablation lesion and unablated tissue and creating an
resulting image of the heart tissue illustrating the ablated tissue
and the unablated tissue.
22. The system of claim 21 wherein the imaging device is configured
to gather hyperspectral data dependent by either spectrally
selective illumination or dependent spectrally selective filtering
prior to image acquisition.
Description
RELATED APPLICATIONS
[0001] This application claims the benefit of and priority to U.S.
Provisional Application Ser. No. 61/906,769, filed on Nov. 20,
2013, which is incorporated herein by reference in its
entirety.
FIELD
[0002] The present disclosure generally relates to optical imaging
to reveal structures of effected biological tissue within
biological specimens for biomedical purposes. In particular, the
present disclosure relates to devices, systems and methods of
hyperspectral or multispectral modality for the identification and
visualization of cardiac ablation lesions.
BACKGROUND
[0003] Atrial fibrillation (AF) is the most common sustained
arrhythmia. In the United States alone, AF is projected to affect
over 10 million people by the year 2050. AF accounts for one third
of all hospital admissions for cardiac rhythm disturbances. AF is
associated with increased mortality, morbidity and an impaired
quality of life. Its incidence sharply increases with age. It is an
independent risk factor for stroke, as it increases stroke
probability by fivefold. AF prevalence increases significantly with
age. Annual costs related to the management of AF in the US alone
are approximately $7 billion. These costs consistently rank AF as
leading public health expenditure.
[0004] Radiofrequency ablation (RFA), laser ablation and
cryoablation are the most common technologies of catheter-based
mapping and ablation systems used by physicians to treat atrial
fibrillation. Physician uses a catheter to direct energy to either
destroy focal triggers or to form electrical isolation lines
isolating the triggers from the heart's remaining conduction
system. The latter technique is commonly used in what is called
pulmonary vein isolation (PVI). However, the success rate of the AF
ablation procedure has remained relatively stagnant with estimates
of recurrence to be as high as 50% one-year post procedure. The
most common reason for recurrence after catheter ablation is one or
more gaps in the PVI lines. The gaps are usually the result of
ineffective or incomplete lesions that may temporarily block
electrical signals during the procedure but heal over time and
facilitate the recurrence of atrial fibrillation.
[0005] To perform radiofrequency (RF) or cryo ablation procedures,
a catheter is threaded into the heart and the tip is guided into
the atria. A transseptal puncture is then performed to crossover
from the right atrium into the left atrium where the crux of the
ablation is performed. The most common treatment of AF consists of
placing ablation lesions in a circular fashion around the ostium of
pulmonary veins to isolate ectopic sources from the rest of the
atria. Cryoablation involves freezing target tissue with the same
ultimate goal to destroy abnormal sources of activity.
[0006] It believed surgical ablation when compared to
pharmacological treatment provides the patient for a longer term of
survival. Yet, the ablation surgery is often needed to be performed
multiple times due to the lack of complete isolation of abnormal
source from the tissue. Recurrence rate of AF after ablation
surgery reaches as high as 50%, of which 90% of culprits can be
linked to gaps between ablation lesions. These viable gaps can
occur largely due to an inability of the surgeon to directly
visualize tissue damage while performing percutaneous AF ablations.
Whether the surgeon's goal is to complete pulmonary vein isolation
with no gaps or site-targeted ablation, it is critical to know the
degree of tissue damage at the site of the ablation. This is
because the extent of tissue damage beneath the catheter is not a
simple function of applied energy, for example, it depends on many
factors including: contact between the catheter tip and the tissue,
the thickness of the myocardium, the degree of blood flow nearby,
the presence of fatty tissue and collagen and other factors.
[0007] Identification of gaps in prior circumferential ablation is
possible with MRI technology; however an MRI cannot be done in real
time in current EP labs. Thus, there is a need in the art for a
device that provides for in vivo, real time analysis of the area
that is being ablated. There is a need for a device that provides
for high resolution visualization of lesion boundary, quantitative
determinations of the gaps between the lesions and the lesion
depth. There is a further need for a device that allows
determination of the presence of scarred tissue at previously
ablated sites in order to avoid re-ablating the same area
[0008] As of today, ablations are performed in essentially `blind`
fashion, with electrical isolation of focal sources being the main
indicator of ablation efficiency. There are at least two
limitations of this approach. The first one is that the extent of
the lesions cannot be measured during the procedure. The second is
that the specific cause of electrical isolation cannot be
determined. It may result from tissue necrosis, functional changes
in reversibly injured cells, as well as by temporary edema. In the
case of edema, it will subside after a few weeks, potentially
restoring electrical conduction between the pulmonary veins and the
left atrium and the return of AF. Indeed, despite an initial return
to sinus rhythm after ablation therapy, AF has a high degree of
recurrence.
[0009] Therefore, there is need, among many needs, in reducing the
number of times for ablation surgery due to the lack of complete
isolation of abnormal source from the tissue by providing better
ways in forming and verifying proper lesions in real time and
overall. Further, there is at least one need in improving ablation
procedures by having real-time in-surgery visualization of lesions
and gaps between them.
SUMMARY
[0010] Systems and methods for hyperspectral analysis of cardiac
tissue are provided.
[0011] According to embodiments, devices, systems and methods of
the present disclosure hyperspectral imaging can be used as a tool
to distinguish between ablated and unablated atrial tissue based on
spectral differences between the two. At least one method of the
present disclosure employs spectral unmixing of a hyperspectral
hypercube dataset to reveal the sites of thermal ablations and gaps
between the ablated and unablated atrial tissue.
[0012] In some aspects, there is provided a method for visualizing
ablation lesions that includes illuminating at one or more
illumination wavelengths a surface of tissue having an ablation
lesion; collecting a spectral data set comprising spectral images
of the illuminated tissue acquired at multiple spectral bands each
at one or more acquisition wavelengths; distinguishing between the
ablation lesion and an unablated tissue based on one or more
spectral differences between the ablation lesion and unablated
tissue; and creating a composite image of the tissue showing the
ablation lesion and the unablated tissue.
[0013] In some aspects, there is provided a method for visualizing
atrial ablation lesion that includes illuminating one or more
discrete illumination wavelengths a surface of heart tissue having
an ablation lesion; collecting a spectral data from the illuminated
heart tissue; distinguishing between the ablation lesion and an
unablated tissue based on one or more spectral differences between
the ablation lesion and unablated tissue; and creating an image of
the heart tissue illustrating ablated tissue and unablated
tissue.
[0014] In some aspects, there is provided a system for imaging
tissue that includes a catheter having a distal region and a
proximal region; a light source; an optical fiber extending from
the light source to the distal region of the catheter to illuminate
a tissue having a lesion site in proximity to the distal end of the
catheter; an image bundle for collecting light reflected from the
illuminated tissue; a camera connected to the image bundle, the
camera being configured to gather hyperspectral data comprising
spectral images of the illuminated tissue acquired at multiple
spectral bands or at each illumination wavelength; an image
processing unit in communication with the camera, the unit being
configured to distinguish between the ablation lesion and an
unablated tissue based on one or more spectral differences between
the ablation lesion and unablated tissue and creating an image of
the heart tissue illustrating the ablated tissue and the unablated
tissue.
[0015] In some aspects, there is provided a system for imaging
heart tissue that includes an illumination device configured to
illuminate a tissue having a lesion site; an imaging device
configured to gather hyperspectral data; an image processing unit
in communication with the imaging device, the image processing unit
configured to processing gathered hyperspectral data to generate an
image that reveals the lesion site, wherein the generated image
enabling to distinguish between an ablation lesion and an unablated
tissue based on one or more spectral differences between the
ablation lesion and unablated tissue and creating an resulting
image of the heart tissue illustrating the ablated tissue and the
unablated tissue.
BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The presently disclosed embodiments will be further
explained with reference to the attached drawings, wherein like
structures are referred to by like numerals throughout the several
views. The drawings shown are not necessarily to scale, with
emphasis instead generally being placed upon illustrating the
principles of the presently disclosed embodiments.
[0017] FIG. 1A illustrates a hyperspectral imaging (HSI) system
with an acousto-optical tunable filter (AOTF) along with software
(i.e. principal component analysis software (PCA)).
[0018] FIG. 1B illustrates a sequence of x-y images acquired at
different wavelengths to create hyperspectral data cube.
[0019] FIG. 2A illustrates an embodiment of a system of the present
disclosure with a pushbroom (spatial scanning) system.
[0020] FIG. 2B illustrates an embodiment of a system of the present
disclosure with a turnable filter.
[0021] FIG. 2C illustrates an embodiment of a system of the present
disclosure a turnable filter at the illumination side.
[0022] FIG. 3A illustrates a diagram showing at least one design of
clinical catheter and hyperspectral imaging components according to
aspects of the present disclosure.
[0023] FIG. 3B illustrates a system architecture diagram of an
embodiment system of the present disclosure.
[0024] FIG. 3C illustrates a block diagram of an embodiment system
of the present disclosure.
[0025] FIG. 3D illustrates a diagram showing an exemplary computer
system suitable for use with the methods and systems of the present
disclosure.
[0026] FIG. 3E illustrates a view of a specialty catheter in
accordance with an embodiment of the present disclosure.
[0027] FIG. 3F illustrates a close-up photo of an inflated catheter
balloon and tip in accordance with an aspect of the present
disclosure.
[0028] FIG. 4A, FIG. 4B, FIG. 4C and 4D show flow diagrams of a
method in accordance with the present disclosure.
[0029] FIG. 5A illustrates an example of excitation-emission
matrices (EEM) that provide information related to the physical
phenomena behind spectral changes caused by ablation such as
reflectance, scattering, absorption and fluorescence.
[0030] FIG. 5B illustrates an example of excitation-emission
matrices (EEM) of a heart muscle.
[0031] FIG. 6 illustrates some of the major differences in a way
light interacts with radiofrequency ablation (RFA) lesions and cryo
lesion in atrial tissue vs ventricular tissue.
[0032] FIG. 7A and FIG. 7B illustrate the difference in the left
and the right atrial tissue appearance of an opened excised porcine
atria, wherein the left atria image illustrates significantly
higher layers of endocardial collagen as compared to the right
atria image of an animal or human.
[0033] FIG. 7C and FIG. 7D illustrate an endocardial surface of an
excised fresh human atria. FIG. 7C shows the endocardial surface of
the excised fresh human atria with cuts to show muscle tissue
beneath. FIG. 7D shows the human atrial tissue stained with
triphenyltetrazolium chloride (TTC), wherein cross sections show
darkly stained muscle layers and white collagen layer on the
endocardial side.
[0034] FIG. 8A, FIG. 8B, FIG. 8C and FIG. 8D illustrate imaging
lesion boundaries with a HSI system, wherein FIG. 8A and FIG. 8B
show visual appearances of the lesions, FIG. 8C and FIG. 8D
illustrate lesions identified by Hyperspectral Imaging (HSI)
approach.
[0035] FIG. 9A and FIG. 9B illustrate an excised porcine left atria
with two RF lesions imaged using Perkin-Elmer Nuance FX
multispectral imaging system equipped with AOTF filter in front of
the camera. FIG. 9A shows a visual appearance of the tissue,
wherein FIG. 9B shows the individual components revealed by HSI
cube principal component analysis (lesion sites 900, collagen 910,
and muscle 920). Tissue was illuminated with incandescent white
light; the reflected light was acquired within 450-950 nm range
using 20 nm steps.
[0036] FIG. 10A, FIG. 10B, FIG. 10C and FIG. 10D illustrate an
excised human left atria with four RF lesions. FIG. 10A illustrates
a visual appearance of the tissue with regions of interest used to
decompose a HSI cube into individual components. FIGS. 10B-10C show
images of a muscle, collagen and lesions, respectively. FIG. 10E
shows a graph of the raw and normalized reflectance spectra, where
the tissue was illuminated with incandescent white light, the
reflected light was acquired within 450 nm to 750 nm range with 20
nm steps. FIG. 10F illustrates an overlay of two individual HSI
components, collagen 910 and the lesions 900.
[0037] FIG. 11A, FIG. 11B, FIG. 11C and FIG. 11D illustrate an
excised porcine left atria with three RF lesions. FIG. 11A shows a
visual appearance of tissue under UV illumination, FIG. 11B shows a
HSI composite image, and FIG. 11C and FIG. 11D show individual
components revealed by principal component analysis. FIG. 11C shows
the individual components revealed by principal component analysis
including lesion component. FIG. 11D shows the individual
components revealed by principal component analysis including
non-ablated component.
[0038] FIG. 11E shows tissue that was crosscut through line 1130 in
FIG. 11C shown above and folded in half in order to expose ablated
muscle tissue (see arrows) beneath the collagen layer, showing
muscle 1110, ablated muscle 1100 and collagen 1120.
[0039] FIG. 11F shows the corresponding target spectra of FIG. 11E,
wherein the tissue was illuminated with 365 nm LED UV light, and
the reflected light was acquired from 420 nm to 600 nm range using
10 nm steps.
[0040] FIG. 12A, FIG. 12B, FIG. 12C and FIG. 12D show an image of
excised human left atria with four RF lesions of different
strengths.
[0041] FIG. 13A and FIG. 13B, illustrate a porcine left atrial
tissue with RF ablation lesions on the endocardial surface. FIG.
13A upper shows a single deep lesion and corresponds to the 3D
surface plot below it, and FIG. 13B upper illustrates two lesions
of different depths with unablated tissue and the corresponding 3D
plot below them, the unablated tissue is what appears as an
elevation between the two lesion depths.
[0042] While the above-identified drawings set forth presently
disclosed embodiments, other embodiments are also contemplated, as
noted in the discussion. This disclosure presents illustrative
embodiments by way of representation and not limitation. Numerous
other modifications and embodiments can be devised by those skilled
in the art which fall within the scope and spirit of the principles
of the presently disclosed embodiments.
DETAILED DESCRIPTION
[0043] The present disclosure generally relates to using
hyperspectral and/or multispectral modality for identification and
visualization of cardiac ablation lesions. In particular, using
hyperspectral imaging as a tool to distinguish between ablated and
unablated atrial tissue based on spectral differences between the
two.
[0044] Due to the hearts different types of tissues and tissue
structures, the aspect of revealing sites of thermal ablations and
gaps between the ablated and unablated atrial tissue can be very
complex and challenging to overcome.
[0045] According to embodiments of the present disclosure, a method
for visualizing ablation lesions, the method includes illuminating
at one or more illumination wavelengths a surface of tissue having
an ablation lesion. Collecting a spectral data set comprising
spectral images of the illuminated tissue acquired at multiple
spectral bands each at different acquisition wavelengths.
Distinguishing between the ablation lesion and an unablated tissue
based on one or more spectral differences between the ablation
lesion and unablated tissue. Finally, and creating a composite
image of the tissue showing the ablated tissue and the unablated
tissue.
[0046] In some embodiments, the present systems and methods may be
employed to visualize ablated lesions in heart tissue (endocardial,
epicardial, atrial and ventricular tissue). However, the presently
disclosed methods and systems may also be applicable for analyzing
lesions in other tissue types. The lesions to be analyzed may be
created by ablation during ablation procedure. In some embodiments,
existing lesions, created by ablation or by other means, may also
be analyzed using methods and systems disclosed herein.
[0047] FIG. 1A and FIG. 1B illustrate Hyperspectral Imaging (HSI)
that is based on collecting and storing individual tissue images
across multiple spectral bands. For example, FIG. 1A illustrates
the components of HSI 100 which include a camera 135A, an
acousto-optical tunable filter (AOTF) 141 along with a computer 140
having software, i.e. principal component analysis software (PCA)).
Typically, when there are more than 10 spectral bands it is called
hyperspectral, and when there are less than 10 spectral bands it is
called multispectral. The HSI device used includes the math that
decides what images are to be displayed and in what order is based
on an Independent Component Analysis (PCA), but with a few
proprietary modifications specific to the Nuance software (and
Maestro, which is a whole animal imaging system based on the same
technology). Please note, that the PCA algorithm can be used or
another commercially available similar type of program could be
used. However, the user chooses which images to be used to
calculate spectra. Each image is thresholded to form a mask of
positive/negative pixels. Those masks are compared and only unique
regions are kept (if two images have the same pixels in their mask
those pixels are discarded). After that the spectra within the
remaining regions are averaged and a typical "compute pure
spectrum" calculation is done, using the spectrum and the mixed
spectrum.
[0048] Still referring to FIG. 1A and FIG. 1B, Hyperspectral
Imaging (HSI) involves acquisition of a three-dimensional dataset
called hypercube, with two spatial dimensions and one spectral
dimension and it can be accomplished using different hardware
configurations. FIG. 1B illustrates a sequence of x-y images
acquired at different wavelengths (i.e. UV, visible and
near-infrared wavelengths), that creates the hyperspectral data
cube, wherein the spectral information from each pixel is then used
to classify the pixels into different subtypes. HSI can visualize
ablation lesions in both atrial and ventricular tissue, as well as
provide visualization of healed lesions which results in tissue
scarring. The methods and systems of the present disclosure are
suitable for identifying lessions created by various types of
ablation, including, but not limited to, laser, microwave, focused
ultrasound induced lesion, acute cryo and radiofrequency.
Delineation of lesion boundary and estimates of lesion depth are
based on revealing changes in spectral properties of ablated
regions. This is accomplished by acquiring images of ablated areas
with specific ranges of UV, visible and near-infrared light and
analysis of spectrum of a sample at each point in the imaging
plane. Spectra from each pixel are then matched to existing
spectral libraries, or are subjected to principal component
analysis or related algorithms. Aspects include real-time in vivo
analysis of the area that is being ablated, including high
resolution visualization of lesion boundary, quantitative
determination of the gaps between the lesions and an estimate of
lesions depth. It will also provide determination of the presence
of scarred tissue at previously ablated sites to avoid re-ablation
of the same area. Some improvement include shortening the time and
improving the efficiency of thermal ablation for treatment of AF,
and minimize unnecessary tissue injury, which can lead to
post-ablation complications such as pulmonary vein stenosis and
esophageal injury, including erythema, ulcers and, in worst case
scenario, left atrial-esophageal fistulas, and decrease
post-ablation recurrence of AF and the need for multiple hospital
readmissions for repeated ablations.
[0049] FIG. 2A, FIG. 2B and FIG. 2C show different optical
configurations that can be used to collect hyperspectral
information, wherein one or more may be used. For example, FIG. 2A
illustrates an optical configuration that is enabled to collect
hyperspectral information, a detector-based Hyperspectral Imaging
(HSI) system that can be built using a movement of the object or
the acquisition device such as a pushbroom system. FIG. 2B
illustrates an optical configuration that is enabled to collect
hyperspectral information, a detector-based HSI system that can be
built using static devices with tunable filters or a filter wheel.
In these configurations, the tissue 200 may be illuminated from a
light source 210 with a light having an illumination wavelength.
The camera 220 may be equipped with a light modifier 230, such as a
filter or a prism, to collect from the illuminated tissue 200 light
having an acquisition wavelength at different spectral bands. FIG.
2C is illustrates an alternative source-based HSI approach, where
it is a wavelength of illuminating light that is being changed,
while a camera records whatever light is being transmitted to it.
In such embodiments, the tissue 200 is illuminated with lights 210
having various spectral bands or wavelength (due to the filter or
preselected light source), and the camera 220 collects all light
from the illuminated tissue.
[0050] Referring to FIG. 2A, FIG. 2B and FIG. 2C, the spatially
resolved spectral imaging obtained by HSI can provide diagnostic
information about the tissue physiology, morphology, and
composition. The spectra from each pixel can be classified into
different subsets using principal component analysis or other
mathematical algorithms referred hereafter as spectral unmixing.
Thus, according to aspects of the present disclosure spectral
un-mixing of hyperspectral hypercube dataset can reveal the sites
of thermal ablations and gaps between them. Alternatively, they can
be matched to pre-existing spectral libraries. Further, pixels with
spectra that match the target spectrum to a specified level of
confidence are then marked as potential targets. It is possible to
use the one or more different optical configurations that collect
hyperspectral information in combination as well as with
pre-existing spectral libraries.
[0051] Cardiac tissue has a thick layer of collagen on both endo
and epicardial sides that covers layers of muscle. This is
particularly true for human left atria, which is the most
clinically relevant site that is usually ablated to stop
progression of atrial fibrillation. At least one aspect of any
surgical ablation is to stop arrhythmias and inflict damage to
muscle layers lying beneath the collagen layers. Yet collagen,
white and highly fluorescent, masks most of the optical changes
that accompany thermal ablations. Therefore, in cases when layer of
collagen exceeds 100 microns, ablation-induced damage to the muscle
is not readily visible to a naked eye. However, ablation alters
spectral signature of the ablated tissue. Some spectral changes may
include, but are not limited to, loss of the NADH fluorescence best
reveals itself at about 350 nm to about 370 nm illumination and
about 450 nm to 480 nm emission ranges, but is not limited to these
ranges; a decrease in collagen fluorescence best reveals itself at
about 330 nm to 360 nm illumination and about 430 nm to 460 nm
emission ranges, but is not limited to these ranges; an increase in
tissue scattering can lead to a larger amount of photons returning
to the imaging detector that have lower energy than that of
illuminating light. It is noted that this effect can be largely
wavelength-independent and can occur across UV, visible and IR
ranges. The result of it can be an elevated shoulder of a
reflectance spectrum. An increase in optical tissue density can be
due to heat-induced tissue drying. It is noted that this effect can
be largely wavelength-independent and can occur across UV, visible
and IR ranges. Change in absorption can be due to e.g., myoglobin
to methmyoglobin and other intercellular chromophore transitions.
Accordingly, the presently disclosed systems and methods can take
advantage of these changes to distinguish between ablated tissue
and unablated tissue.
[0052] FIG. 3A illustrates a diagram showing at least one design of
clinical catheter 305A and hyperspectral imaging components
according to aspects of the present disclosure. The diagram of FIG.
3A shows a visualization catheter 305A for live visualization of RF
ablation lesions and gaps during percutaneous ablation procedure.
The catheter 305A can include a computer 340, a light source (not
shown), a camera 335A with a turnable filter 335AA, and a
fiberoptic cable 345. In some embodiments, an inflatable balloon
355 may be included to displace blood between the fiberoptic cable
345 and the tissue surface. The fiberoptic cable 345 is then
connected to a hyperspectral camera 335A. An increasing number of
commercial hyperspectral cameras and related imaging processing
software packages are now available. Most recent ones allows 30
different spectral bands or greater to be captured in parallel at
video rate speed and to be analyzed in real-time with principle
component analysis discrimination algorithms. The above described
hyperspectral visualization catheter 305A can be also combined with
RF or cryo ablator making it a single catheter 305A.
[0053] According to aspects of the present disclosure, the systems
and methods may be extended from the above-described thermal
lesions, i.e. RF and cryoinjury, to visualization of ablation
lesions made by other means, such as laser-based, microwave or
focused ultrasound based tissue destruction.
[0054] FIG. 3B and FIG. 3C illustrate diagrams showing an ablation
visualization system (AVS) that incorporates hyperspectral imaging
components as noted in FIG. 3A, according to some aspects of the
present disclosure. FIG. 3B shows at least one embodiment of an
ablation visualization system (AVS) 306 that incorporates
hyperspectral imaging components as noted in FIG. 3A. FIG. 3C shows
at least another embodiment of an ablation visualization system
(AVS) 307 that incorporates hyperspectral imaging components as
noted in FIG. 3A. FIG. 3B and FIG. 3C show a light source 330A that
is external to the body of a patient and a light delivery fiber
330B for delivering light from the light source 330A to within the
body of the patient, a camera 335A with appropriate filtering, if
necessary, and an image bundle 335B connected to the camera, and a
computer system 340 having one or more displays 340A (for a
technician) and 340B (for a physician) with image processing
software on its processor or controller. Aspects of the camera will
be further discussed.
[0055] FIG. 3D shows, by way of example, a diagram of a typical
processing architecture 308, which may be used in connection with
the methods and systems of the present disclosure. A computer
processing device 340 can be coupled to display 340AA for graphical
output. Processor 342 can be a computer processor 342 capable of
executing software. Typical examples can be computer processors
(such as Intel.RTM. or AMD.RTM. processors), ASICs,
microprocessors, and the like. Processor 342 can be coupled to
memory 346, which can be typically a volatile RAM memory for
storing instructions and data while processor 342 executes.
Processor 342 may also be coupled to storage device 348, which can
be a non-volatile storage medium, such as a hard drive, FLASH
drive, tape drive, DVDROM, or similar device. Although not shown,
computer processing device 340 typically includes various forms of
input and output. The I/O may include network adapters, USB
adapters, Bluetooth radios, mice, keyboards, touchpads, displays,
touch screens, LEDs, vibration devices, speakers, microphones,
sensors, or any other input or output device for use with computer
processing device 340. Processor 342 may also be coupled to other
type of computer-readable media, including, but are not limited to,
an electronic, optical, magnetic, or other storage or transmission
device capable of providing a processor, such as the processor 342,
with computer-readable instructions. Various other forms of
computer-readable media can transmit or carry instructions to a
computer, including a router, private or public network, or other
transmission device or channel, both wired and wireless. The
instructions may comprise code from any computer-programming
language, including, for example, C, C++, C#, Visual Basic, Java,
Python, Perl, and JavaScript.
[0056] Program 349 can be a computer program or computer readable
code containing instructions and/or data, and can be stored on
storage device 348. The instructions may comprise code from any
computer-programming language, including, for example, C, C++, C#,
Visual Basic, Java, Python, Perl, and JavaScript. In a typical
scenario, processor 204 may load some or all of the instructions
and/or data of program 349 into memory 346 for execution. Program
349 can be any computer program or process including, but not
limited to web browser, browser application, address registration
process, application, or any other computer application or process.
Program 349 may include various instructions and subroutines,
which, when loaded into memory 346 and executed by processor 342
cause processor 342 to perform various operations, some or all of
which may effectuate the methods for managing medical care
disclosed herein. Program 349 may be stored on any type of
non-transitory computer readable medium, such as, without
limitation, hard drive, removable drive, CD, DVD or any other type
of computer-readable media.
[0057] It is possible the light source 330A may include a cart 332.
In some embodiments, the system may further include a specialty
catheter 305A comprising an inflatable balloon 355. In some
embodiments, the image bundle 335B and the light delivery fiber may
extend from the outside of the catheter 305A to a distal region of
the catheter 305A inside the balloon 355. It is contemplated that
there could be multiple components of each component added to the
above disclosed system. The system may further include a guidewire
for the catheter 305C, a EP Fluoroscopy System 360, a sterable
sheath 365A, a guidewire for steerable sheath 365B, an introducer
sheath kit 365C, an indeflator 370 and a trasseptal kit 380.
[0058] FIG. 3C is a block diagram of an exemplary system in
accordance with the present disclosure. The AVS system includes
external equipment 325 having a light source 330A, a camera 335A
with appropriate filtering, if necessary, and a computer system
(not shown) having one or more displays 340A with image processing
software. The AVS system includes internal equipment including an
ablation device 338, an illumination device 330B and an imaging
device 335B, wherein the internal components are within an internal
balloon 355 associated with a catheter 305A. It is noted that the
internal equipment including the catheter 305A with an inflatable
balloon catheter 355, 305A is coupled to external equipment 325. In
some embodiments, the illumination device 330B and an imaging
device 335B may utilize a fiber-optic waveguide to pass the light
to and from the treated tissue.
[0059] Still referring to FIG. 3B and FIG. 3C, the light source
330A may be selected to illuminate the tissues, such as an
endocardial surface of the heart, at various wavelengths.
[0060] According to some aspects of FIG. 3B and FIG. 3C, a laser
generated light may provide much more power for illumination and
its wavelength can be pure at whatever number of nanometers that
may be required. There are sources of commercial lasers that can
emit in a desired illumination band and they are available in many
power settings near 50 to 200 mW and higher. The instant system, in
some embodiments, uses a laser with adjustable power up to 150
mW.
[0061] Still referring to FIG. 3B and FIG. 3C, the catheter 305A
can be employed to perform many functions including, without
limitations, vascular navigation, blood displacement, propagation
of light from the light source 330A to the myocardium, and image
gathering of the fluorescence light. One example of a suitable
catheter 305A is disclosed in jointly-owned U.S. application Ser.
No. 13/624,902, which is incorporated herein in its entirety. In
some embodiments, the ablation technology is housed with or
incorporated within the system and catheter 305A embodiment.
[0062] In reference to FIG. 3E and FIG. 3F, the catheter 305A may
include a balloon 355 at or near the distal end of the catheter
305A. Since blood absorbs the illumination and fluorescence
wavelengths, the balloon 355 may displace blood from the myocardial
surface. To do so, the balloon 355 may be expandable and compliant
to seat well within the anatomy--especially the pulmonary veins.
The medium used to inflate the balloon 355 may also be optically
transparent and yet ideally be fluoroscopically opaque for
navigation purposes. Suitable inflation medium include, but are not
limited to, Deuterium (heavy water) and CO.sub.2, which meet both
requirements. The balloon 355 may also be constructed of a material
that is optically clear in at least the wavelengths of concern for
both illumination of the myocardium and fluorescence. The balloon
355 may be either, made of non-compliant materials but with
optimally variable sizes of best fit into pulmonary veins and other
structures, or, made of a compliant material such as silicone or
urethane. In some embodiments, the balloon 355 may be optically
transparent in the UV range of 330 nm to 370 nm.
[0063] In some embodiments, the balloon 355 is optically clear from
330 nm to 370 nm for UV illumination and from 400 nm to 500 nm for
the fluorescence wavelengths. Suitable UV-transparent materials for
the balloon 355 include, but are not limited to, silicone and
urethane.
[0064] Still referring to FIG. 3E and FIG. 3F, the catheter 305A
may also be used to efficiently deliver the illuminating light from
the external light source 330A to the balloon 355 and out of the
balloon 355 to the heart tissue. In some embodiments, a laser
delivery fiber, usually made of quartz due to its UV efficiency and
small diameter, may be used to deliver illuminating light from a UV
laser light source.
[0065] The catheter 305A of FIG. 3E and FIG. 3F may also be
employed to collect and transfer the light from the illuminated
tissue to an external camera. In some embodiments, this may be
accomplished via an imaging fiber bundle extending from the distal
region of the catheter 305A to the external camera. In some
embodiments, the image bundle may include one or more of
individual, single-mode fibers that together maintain image
integrity while transporting it along the length of the catheter
305A to a camera and a filter, as necessary. The imaging bundle,
though flexible and small in diameter, may be able to achieve a
sufficient field of view for imaging the target tissue area covered
by the balloon 355.
[0066] The camera 335A, can be connected to the computer system 340
for receiving the light from the illuminated tissue for use in
connection with the HSI method. In some embodiments, the digital
image that is produced by the camera 335A is used to do the 2D and
3D reconstruction. In some embodiments, the image bundle may be
connected to the camera 335A, the camera 335A may generate a
digital image from the light received from the illuminated tissue,
which can be displayed on the computer.
[0067] In reference to FIG. 4A, operation of the systems and
methods of the present disclosure are illustrated. It is noted that
systems and methods include the analysis of multiple wavelengths of
light reflected from acutely ablated atrial tissue and follow-up
spectral imaging analysis, wherein a hyperspectral imaging device
is used as a tool to distinguish between ablated and unablated
atrial tissue based on subtle spectral differences between the two.
At least one aspect of the systems and methods includes spectral
unmixing of hyperspectral cube datasets to reveal sites of thermal
ablations and gaps between them.
[0068] Initially, step 410 of FIG. 4A discloses inserting a
catheter into affected area. Step 415 of FIG. 4A includes
illuminating the affected area with the light source. As used
herein, "light" refers generally to electromagnetic radiation of
any wavelength, including the infrared, visible, and ultraviolet
portions of the spectrum. A particularly portion of the spectrum of
illuminating light can be the portion which produces the largest
spectral differences between reflected light coming out of the
native (non-ablated) and ablated tissue. To illuminate the surface,
the catheter is inserted into the area of heart tissue affected by
the atrial fibrillation, such as the pulmonary vein/left atrial
junction or another area of the heart. Blood is removed from the
visual filed, for example, by the balloon. A transparent balloon
surrounding the fiber optic waveguide can be used to displace the
blood at the pulmonary vein/left atrial junction. The affected area
may be illuminated by light from the light source and the optical
fiber or another illumination device.
[0069] Step 420 of FIG. 4A includes ablating the tissue in the
illuminated area. It is possible while collecting the reflected
light to the imaging device to also conduct ablation procedures.
For example, for atrial fibrillation ablation a tissue in the
illuminated area may be ablated using an ablation device, either
before or after illumination. Either point-to-point RF ablation or
cryoablation or laser or other known ablation procedures may be
employed using the systems of the present disclosure. Ablation
proceeds by threading the tip through the central lumen of the
catheter or outside the catheter. After the procedure, the ablation
tip may be retracted. In some embodiments, an ablation tip may be
incorporated into the catheters disclosed herein.
[0070] Step 425 includes taking an image of the illuminated area,
collecting and directing the reflected light to an imaging device.
The latter can be done via a fiberoptic cable leading to a camera
or directly by a small imaging chip.
[0071] Step 430 of FIG. 4A includes producing a display of the
imaged illuminated area. Step 435 of FIG. 4A includes identifying
ablated and unablated tissue in the imaged area using changes in
spectral signature between the ablated and unablated tissue, as
described above.
[0072] The application software, executing on the computer system
by the processor or computer, can provide the user with an
interface to the physician. Some of the main functions can include:
a laser control, a camera control, an image capture, an image
conditioning (brightness and contrast adjustment, etc.), a lesion
identification, a lesion depth analysis, a procedure event
recording, and a file manipulation (creation, editing, deleting,
etc.).
[0073] FIG. 4B is a flow chart of an exemplary method for
constructing a HSI image of the ablated tissue. Step 440 of FIG. 4B
includes illuminating at a specific wavelength endocardial surface
of the atrial tissue that has the ablation lesion. Step 445 of FIG.
4B includes collecting a spectral data cube comprising of images of
the illuminated endocardial surface acquired at multiple spectral
bands. Specifically collecting data to form a three dimensional
dataset (spectral data cube or hyperspectral data cube (HSDC)) that
comprises of images of illuminated endocardial surface acquired at
multiple spectral bands. Step 450 of FIG. 4B includes processing
the spectral data cube to identify pixels corresponding to ablated
tissue regions and to non-ablated areas. Specifically, the
processing of the spectral data to identify pixels corresponding to
the ablated tissue regions and the non-ablated areas. The latter is
composed of three dimensions: two spatial (X, Y) and one spectral
(X). HSDC can complied from two spatial dimensions (X, Y) acquired
simultaneously while the spectrum is built by sequentially scanning
through wavelengths (.lamda.) using a tunable optical band-pass
filter. Alternatively HSDC can be compiled from one spatial and one
spectral dimension (Z, X) acquired simultaneously, while the HSDC
is built by sequentially scanning second spatial dimension (Y).
[0074] Step 455 of FIG. 4B includes using imaging processing
algorithms to distinguish between the ablated lesion and an
unablated tissue based on one or more spectral differences between
the ablated lesion and unablated tissue. Specifically, classifying
spectra from each pixel into different subsets using principal
component analysis or related mathematical algorithms referred
thereafter as spectral unmixing to distinguish between the ablated
lesion and an unablated tissue based on one or more spectral
differences between the ablated lesion and unablated tissue.
Alternatively, spectra can be classified based on match with
pre-existing spectral libraries. The spectrum of each pixel is then
assumed to be a linear combination of pre-defined spectra and least
squares approach is taken to fit these spectra to the observed
pixel spectrum.
[0075] Step 460 of FIG. 4B includes making a composite image that
shows ablation lesions, gaps between the lesions and lesion depth.
Further, constructing an abundance map of each type of tissue, i.e.
ablated vs unablated, collagen vs muscle, to display the fractional
amount of its presence at each pixel. Thus, making a composite
image that shows ablation lesions, gaps between the lesions and the
lesion depth.
[0076] FIG. 4C illustrates another exemplary method of the present
disclosure. Step 465 of FIG. 4C discloses illuminating using
several specific wavelengths (.lamda..sub.1, .lamda..sub.2,
.lamda..sub.3, .lamda..sub.4 . . . .lamda..sub.n) the endocardial
surface of atrial tissue having an ablation lesion. Step 470 of
FIG. 4C discloses collecting light reflected from the endocardial
surface illuminated using each of the above noted illumination
wavelength. Step 475 of FIG. 4C discloses composing spectral data
cube comprising of images of the endocardial surface illuminated at
multiple spectral bands. Step 480 of FIG. 4C discloses processing
the spectral data cube to identify pixels corresponding to ablated
tissue regions and to non-ablated areas. Step 485 of FIG. 4C
discloses distinguishing between the ablated lesion and an
unablated tissue based on one or more spectral differences between
the ablated lesion and unablated tissue. Step 490 of FIG. 4C
discloses making a composite image that show ablation lesions, gaps
between the lesions and lesion depth.
[0077] FIG. 4D illustrates a flow chart of the determining the
lesion depth process. In some embodiments, a depth map of the
ablated lesions may be constructed, as shown, for example, in FIGS.
13A-13B. Step 495 of FIG. 4D discloses identifying ablated and
unablated tissue in the imaged area via application software from
computer display. Step 500 of FIG. 4D discloses identifying an
image or images of interest specific to a lesion or lesions to
begin the lesion depth analysis. Step 505 of FIG. 4D discloses
identifying an area of healthy tissue within the image of lesion of
interest. By way of a non-limiting example, due to changes in
spectral changes of ablated tissue, the lesion site may have a dark
appearance which may become more intense as lesion depths
increases, gaps or healthy tissue having lighter appearance. Once
the lesion or lesions are identified, they are selected for lesion
depth analysis. Step 510 of FIG. 4D discloses normalizing the
entire image using a ratio of intensity observed at each pixel to
that observed in the identified healthy tissue. Step 515 of FIG. 4D
discloses processing the resulting normalized image data via an
algorithm derived from a pre-established dataset correlating
normalized intensity ratio to lesion depth.
[0078] Step 520 of FIG. 4D discloses the depth analysis performed
along a single line across the lesion is completed. It is also
possible that this can be done for just one single location in the
lesions from information from a single location, a line or a
region. Step 525 of FIG. 4D discloses repeating steps previous
steps along different lines parallel to the initial line, so the
depth data of each line compiles into a 3D reconstruction model of
the lesion site(s). The depth analysis process performed along a
single line across the lesion could be repeated as many times as
needed along different lines parallel to the initial line, and the
depth data of each line could be compiled into a 3D reconstruction
model of the lesion site.
[0079] FIG. 5A illustrates an example of excitation-emission
matrices (EEM) that provides information related to the physical
phenomena behind spectral changes caused by ablation such as
reflectance, scattering, absorption and fluorescence. FIG. 5B
illustrates an example of excitation-emission matrices (EEM) of a
heart muscle.
[0080] Referring to FIG. 5A and FIG. 5B, human atria can have an
average thickness of 2 mm ranging from 1-3 mm, wherein the layers
of atrial muscle are sandwiched between layers of epicardiac and
endocardial collagen. The latter can range from 50 micron to 1 mm
in thickness.
[0081] Still referring to FIG. 5A and FIG. 5B, the biological
tissues are heterogeneous in composition with spatial variations in
optical properties. The scattering properties of tissues composed
of cells and extracellular proteins, including elastin and
collagen, are caused by the small-scale inhomogeneities and the
large-scale variations in the structures they form. In addition,
different tissue components have different fluorescence profiles.
FIG. 5A shows the individual components of excitation-emission
matrices (EEM), wherein the EEM can provide the full information
from each individual component of the tissue. The 45 degree line
across the EEM of FIG. 5A stands for reflected light. If displayed
in linear form it shows the spectrum of reflected light, with its
peaks corresponding to decreased absorption at specific
wavelengths. Wherein, the diffuse scattering is represented by the
width of the reflectance line. For example, the wider reflectance
line the more diffuse scattering is observed at that specific
wavelength (photons lose energy upon what is called non-elastic
interactions so their wavelength decreases). Further, the peaks
outside the 45 degree reflectance line stand for fluorescence of
individual fluorophores. As noted above, FIG. 5A shows all three
different components including reflected, scattered, and
fluorescent light, collected from the cardiac muscle tissue. When
intact, the cardiac muscle tissue has a significant amount of
highly fluorescent NADH at 355/460 nm excitation/emission
range.
[0082] FIG. 6 illustrates some of the major differences in a way
light interacts with radiofrequency ablation (RFA) lesions and
cryolesion in atrial tissue vs ventricular tissue. For example,
thermal ablations can affect spectral properties of both collagen
and underlying muscle tissue, including changes in their
wavelength-specific absorption and fluorescence, and wavelength
independent changes in tissue scattering. The direction and
amplitude of these changes differ between specific subsets of
conditions and types of thermal lesions. Yet, in all cases,
ablation alters the spectral signature of tissue at the lesion site
in a very distinct way. This allows for the use of a hyperspectral
imaging approach to classify the pixels accordingly and to
precisely identify the location, shape, size and depth of the
ablation lesions, among other things.
[0083] Still referring to FIG. 6, regarding how the thermal
ablations can affect the spectral properties of both collagen and
underlying muscle tissue. It is noted that during atrial
radiofrequency (RF) ablations, high temperatures denature and
dehydrate surface collagen, which makes it more opaque and give it
a slightly yellow hue. The RF energy also changes the spectral
signature of the underlying muscle and lipid layers. The latter
includes altered absorption spectrum and the loss of endogenous
fluorophores, such as nicotinamide- and
flavine-adenine-dinucleotides, lipofuscin, porphyrins and others.
RF-induced ablation also causes protein coagulation which
dramatically increases light scattering. Lastly, RF ablation dries
both collagen and muscle layers increasing their optical density.
All together, these changes lead to altered spectral signature
enabling Hyperspectral Imaging (HSI) based identification of atrial
ablation lesions throughout wide range of wavelengths, including
ultraviolet, visible and infrared range. Specific changes include a
distinct increase in diffuse reflectance across most of the visible
range with the most occurring between 520-600 nm as well as a
change in fluorescence (365 nm excitation/450 nm emission range)
which corresponds to a decrease in NADH fluorescence.
[0084] Regarding the collagen and underlying muscle tissue note
above, FIG. 7A and FIG. 7B illustrate the difference in the left
and the right atrial tissue appearance of an opened excised porcine
atria. Wherein the left atria image illustrates significantly
higher layers of endocardiac collagen as compared to the right
atria image of an animal or human. FIG. 7C shows the endocardial
surface of the excised fresh human atria with cuts to show muscle
tissue beneath. FIG. 7D shows the human atrial tissue stained with
triphenyltetrazolium chloride (TTC), wherein cross sections show
darkly stained muscle layers and white collagen layer on the
endocardial side.
[0085] However, certain wavelengths of light used for illumination
and image acquisition can be more efficient than others as far as
the ability of HSI approach to reveal ablation sites with high
signal-to-noise ratio.
[0086] Wavelength Range (450 nm to 700 nm). When muscle tissue is
RF-ablated, its color turns from red-brown to yellow-white hue.
This can be easily seen by eye or recorded using color camera. In
atrial tissue, however, muscle is covered by 0.1 mm to 0.5 mm thick
collagen layer (see FIG. 7A and FIG. 7C). The latter obscures the
ablation induced changes in the color of the muscle making RF
lesion visually undetectable. Yet, very slight changes in hue and
optical density provide sufficient spectral information to
delineate the lesions using incandescent illumination and/or white
light sources. Either push-broom based or AOTF based HSI imaging
systems provide high fidelity lesion identification, as described
in more detail below.
[0087] Offset Acquisition. According to aspects of the present
disclosure, the experiments suggest that scattering is one of the
three major optical components, wherein absorption and fluorescence
being the other two. Therefore, at least one aspect of the present
disclosure discloses that the effective way for visualization of
atrial RF lesions is to illuminate the ablated surface with
wavelengths below the spectral range in which HSI hypercube is to
be acquired. For example ablated tissue can be illuminated with 460
nm LED and light acquired from 500-600 nm range. Increased
scattering raises the entire right shoulder of the spectra of the
light returning to the detector, allowing high fidelity spectral
unmixing. By excluding illumination wavelengths from HSI
acquisition and analysis, one prevents less discriminate, yet very
intense reflected light. The latter can mask the differences
between unablated and ablated tissue sites.
[0088] Infrared range (650-900 nm). Thermal ablation increases the
light scattering across the entire optical spectrum. Thus,
acquiring HSI spectral hypercube when the sample is illuminated by
an infrared source can also serve as a basis for lesion
identification. There can be additional advantages of using longer
wavelengths because of at least two main reasons. First, among many
reasons, it increases penetration for both illuminating and
scattered light enabling visualization of deeper layers of damaged
muscle. Secondly, among many reasons, it enables one to use
therapeutic window where light absorption by hemoglobin is minimal
and therefore lesion visualization can be done without the need to
fully displace blood between the fiberoptic and the tissue.
[0089] Ultraviolet A or UVA (330-400 nm). Another illumination
range where HSI can show high efficiency to reveal thermal lesions
in left atrial tissue is UVA (see FIGS. 11A-11F for data for
porcine and FIGS. 12A-12D for human left atria. Short wavelength
UVA photons that illuminate atrial tissue do not penetrate deep
into collagen layer and therefore are not reaching the muscle layer
beneath it. Instead, by employing wavelengths where collagen can be
excited (330-400 nm range), one elicit autofluorescence of atrial
collagen that illuminates muscle layers beneath it. The emission
profile of collagen is rather broad (peaking at 390 nm and reaching
500 nm), therefore photons emitted from collagen layer have longer
wavelengths and are capable of penetrating into muscle layer and
back to the detector. The dramatically increases reflectance from
the ablated muscle that lies beneath the collagen, elevating the
right shoulder of the returning light spectrum. In addition, heat
of RF ablation dries up and condenses the tissue, yielding an
increase in collagen fluorescence (see FIGS. 12A-12D). These two
factors add up allowing HSI to discriminate between unablated and
ablated tissue when illuminated in UV range.
[0090] According to aspects of the present disclosure, the
illumination wavelength can be anywhere in the UV range of about
350 nm to about 400 nm, preferably within can be about 360 nm to
about 370 nm. It is possible that the illumination wavelength can
be anywhere in the visible range 400 nm to about 700 nm, preferably
within about 400 nm to about 500 nm range. Further, the
illumination wavelength can be anywhere in the IR range 700 nm to
about 900 nm, preferably within about 700 nm to about 750 nm
range.
[0091] According to aspects of the present disclosure, the
illumination light contains continuous range wavelengths that can
be anywhere in the UV range of about 350 nm to about 400 nm,
preferably within can be about 360 nm to about 370 nm. It is
possible that the illumination wavelength can be anywhere in the
visible range 400 nm to about 700 nm, preferably within about 400
nm to about 500 nm range. Further, the illumination wavelength can
be anywhere in the IR range 700 nm to about 900 nm, preferably
within about 700 nm to about 750 nm range.
[0092] According to aspects of the present disclosure, the
illumination light can be within about 350 nm to about 380 nm range
while collected light can be between about 400 nm and 700 nm range,
preferably within about 400 nm and 500 nm range. Further, the
illumination light can be in the visible range of about 400 nm to
about 700 nm), while collected light can also be in the same range
of about 400 nm to about 700 nm). It is possible the illumination
light can be in the infrared range of about 700 nm to about 900 nm,
while collected light is in the same range of about 700 nm to about
900 nm.
[0093] According to aspects of the present disclosure, the
wavelength of illuminating light can be below the acquisition range
by between about 10 nm and about 50 nm, including but limited to
examples such as about 360 nm: about 370 nm to about 470 nm, about
370 nm: about 400 nm to about 480 nm, about 450 nm: about 500 nm to
about 600 nm.
[0094] According to aspects of the present disclosure, the
filtering can have an emission from the illuminated heart tissue
using a set of individual bandpass filters, including but not
limited to about 460/25 nm, about 500/25 nm, and about 540/25 nm,
wherein using an acousto-optical filter in front of the camera
tunable to a specified range, including but not limited to about
420 nm to about 720 nm, and about 45 nm to about 900 nm.
[0095] According to aspects of the present disclosure, the
detecting light emitted from the illuminated heart tissue can
include an emission comprising of fluorescence, reflectance and
scattering components with fluorescence being the most different
between ablated and unablated sites within about 400 nm to about
500 nm range, and with reflectance being most different between
ablated and unablated sites within about 450 nm to about 600 nm
range, and with diffuse reflectance being most different between
ablated and unablated sites across visible spectra within about 550
nm to 600 nm range.
[0096] According to aspects of the present disclosure, the
constructing a depth map of the ablated lesion from the image can
be accomplished by using a density of the pixels that have target
spectra, i.e. being ablated tissue, as an indicator of lesion
depth, wherein the intensity of grey levels of each pixel can
enable then to create a 3D map.
[0097] According to aspects of the present disclosure, the
identifying the ablated lesion as a lesion created by
radiofrequency and the setting of the illumination wavelength can
be between about 400 nm to about 700 nm and the acquisition
wavelength to about 400 nm to about 700 nm. Further, the
identifying of the ablated lesion as a lesion created by
cryoablation and setting the illumination wavelength can be between
about 350 nm to about 400 nm and the acquisition wavelength to
about 380 nm to about 500 nm. Further still, in identifying whether
the ablated lesion was created by radiofrequency or cryoblation
and, if the ablated lesion was created by radiofrequency, setting
the illumination wavelength can be between about 400 nm to about
700 nm and the acquisition wavelength can be about 400 nm to about
700 nm, or, if the ablated lesion was created by cryoblation,
setting the illumination wavelength can be between about 350 nm to
about 400 nm and the acquisition wavelength can be between about
380 nm to about 500 nm.
[0098] According to aspects of the present disclosure, the spectral
ranges and types of camera used can be varied. In some embodiments,
both UV and white light illumination with hyperspectral analysis of
reflected spectra can work to visualize RF lesions. UV illumination
combined with the hyperspectral analysis of fluorescence is a
better approach to visualize cryolesions. And because fluorescence
intensity is several orders of magnitude less compared to the
intensity of reflected light, cryolesion visualization requires
hyperspectral cameras more sensitive in the near-UV range (350-500
nm), while RF lesions can be observed using less sensitive
hyperspectral cameras that collect reflected light in 400 nm-700 nm
range. Therefore, both illumination sources as well as types of
camera might need to be different for the two types of ablation
surgeries, despite the fact that the ultimate result of either cryo
or RF ablation are lesions of necrotic myocardium. By way of a
non-limiting example, the following settings may be used.
TABLE-US-00001 Images Camera LESION Type of Illumi- Illumination
acquired sensi- TYPE signal nation wavelength at tivity Radio-
Reflectance White 400 nm- 400 nm- medium frequency light 700 nm
1000 nm Cryo- Fluorescence UV light 350 nm- 380 nm- high ablation
source 400 nm 500 nm
[0099] FIG. 8A, FIG. 8B, FIG. 8C and FIG. 8D illustrate the ability
of Hyperspectral Imaging (HSI) system to reveal lesion boundaries.
FIG. 8A shows an example of visual appearance of an RFA lesion on
the endocardial surface of canine left atria. Although lesion can
be visually seen it has poorly defined boundaries. FIG. 8B shows a
visual appearance of two radiofrequency ablation lesions on the
endocardial surface of excised porcine left atria. Again,
boundaries of the lesions cannot be clearly define.
[0100] FIG. 8C and FIG. 8D show the porcine sample shown in FIG. 8B
that was imaged by a custom push-broom HSI system comprised of a
Specimen spectrograph and an Andor iXon CCD built by Middleton
Research. Spectral profiles from unablated and ablated areas of
endocardal surface of porcine left atria differed (FIG. 8C)
enabling effective unmixing of hyperspectral cube. The resulting
HSI data cube was analyzed using principal component analysis
performed by UmBio Evince software package, yielding a pseudo color
composite HSI image with two clearly delineated lesions (FIG.
8D).
[0101] FIG. 9A and FIG. 9B illustrate an excised porcine left atria
with two RF lesions imaged using a Perkin-Elmer Nuance FX
multispectral imaging system equipped with a AOTF filter in front
of the camera. FIG. 9A shows a visual appearance of the tissue, and
FIG. 9B shows the individual components revealed by HSI cube
principal component analysis (900 is the lesion sites, blue the
collagen and red is muscle). The tissue was illuminated with an
incandescent white light, wherein the reflected light was acquired
within a 450 nm-950 nm range using a 20 nm step.
[0102] FIG. 10A, FIG. 10B, FIG. 10C and FIG. 10D illustrate an
excised human left atria with four RF lesions. FIG. 10A illustrates
a visual appearance of the tissue with regions of interest used to
decompose a HSI cube into individual components. FIG. 10E shows a
graph of the raw and normalized reflectance spectra, where the
tissue was illuminated with incandescent white light, the reflected
light was acquired within 450 nm to 750 nm range with 20 nm steps.
FIG. 10F illustrates an overlay of two individual HSI components,
the collagen 910 and the lesions 900.
[0103] FIG. 11A, FIG. 11B, FIG. 11C and FIG. 11D illustrate an
excised porcine left atria with three RF lesions. FIG. 11A shows a
visual appearance of tissue under UV illumination, FIG. 11B shows a
color coded HSI composite image, and FIG. 11C and FIG. 11D show
individual components revealed by principal component analysis.
FIG. 11C shows the individual components revealed by principal
component analysis including Lesion component. FIG. 11D shows the
individual components revealed by principal component analysis
including non-ablated component.
[0104] FIG. 11E shows tissue that was crosscut through line 1130 in
FIG. 11C shown above and folded in half in order to expose ablated
muscle tissue (see arrows) beneath the collagen layer. Wherein a
color coded HSI composite image shows the aspects of the tissue,
showing muscle 1110, ablated muscle 1100 and collagen 1120.
Further, FIG. 11F shows the corresponding target spectra of FIG.
11E, wherein the tissue was illuminated with 365 nm LED UV light,
and the reflected light was acquired from 420 nm to 600 nm range
using 10 nm step.
[0105] In regard to chronic Radio Frequency (RF) lesions, HSI can
be used for visualization of healed lesions from previously
performed RF surgeries, as well as in addition to real-time
in-surgery visualization of acute RF lesions and gaps between them.
This is because collageneous scar formed at the site of the
successful ablation has a very different spectral signature that
surrounding muscle tissue. Therefore, HSI can reveal the sites of
previous RF ablations allowing physician to target the remaining
gaps and to avoid repetitively burning previously ablated areas
(the latter can lead to excessive scarring, loss of atrial
compliance and/or pulmonary vein stenosis).
[0106] In regard to using cryo ablation procedures, cryo ablation
procedures is gaining popularity as an alternative means to destroy
tissue near the ectopic sources. It creates more `clean` lesions as
compared to the RF procedures and has less incidences of
post-surgical pulmonary stenosis. The cryolesions also are more
defined and have less scar formation, in part because, in contrast
to RF, cryoablation does not lead to thermal coagulation of
collagen and underlying muscle layers. Instead it destroys cells by
formation of ice crystals that tear membranous structures. In
contrast to RF ablation, cryoablation of muscle yields in reduced
light scattering across visible spectrum.
[0107] The above described unmixing of reflectance spectra that can
resolve RF lesions, can be also applicable to visualization of
cryolesions. Yet, because changes in scattering are less dramatic
for cryolesions, HSI is likely to be less effective for
visualization of cryolesions based on just change in tissue
scattering. On the other hand, for cryolesions, the loss of muscle
NADH and diminished reflectance produce changes in the SAME
direction--i.e., both results in LOWERING the right shoulder of the
returning light spectrum. In addition, if one looks at the spectral
profile of the fluorescence, the fluorescence peaks of NADH and
collagen can be clearly distinguished. Upon cryoablation, atrial
muscle is destroyed and fluorescence of NADH diminishes, while
collagen layer stays intact. Therefore while total decrease in
returning light intensity is minimal, the spectral profile of the
fluorescence coming from the lesion site changes. Use of spectral
unmixing algorithms then allows one to identify the lesions and the
gaps between them.
[0108] FIG. 12A, FIG. 12B, FIG. 12C and FIG. 12D show an image of
excised human left atria with four RF lesions of different
strengths. FIG. 12A shows a visual appearance of the tissue under
incandescent illumination. FIG. 12B illustrates HSI identified
lesions using UV illumination (365 nm LED source). FIG. 12C shows
that under UV illumination lesions appears denser. One of the
underlying spectral components is increased fluorescence intensity
from ablated collagen due at least in part to RF caused loss of
water content, as noted above. To illustrate this effect, the
tissue of FIG. 12D was crosscut through line 1220, folded in half
and imaged using the same settings. The red arrows point to
increased intensity of collagen at the lesion sites. Additional
spectral component can be change in color and increased scattering
of underlying muscle shown in previous figure.
[0109] FIG. 13A and FIG. 13B, illustrate a porcine left atrial
tissue with RF ablation lesions on the endocardial surface. FIG.
13A shows a single deep lesion and its 3D surface plot, and FIG.
13B illustrates two lesions of different depth with unablated
tissue in between. Below is shown a 3D surface plot of the two
lesions that shows the region of unablated tissue.
[0110] Regarding ablation lesions of FIG. 13A and FIG. 13B,
ablation lesions must be of adequate depth and cause cell necrosis
in a near transmural fashion while minimizing damage to non-cardiac
structures. The mean left atrial wall thickness in humans is about
1.8 mm. HSI ability of detect lesion depth is inversely related to
frequency of illuminating wavelength, with longer wavelengths (near
infrared and infrared range) penetrating and sensing the deeper
lesions. Changes in diffuse tissue reflectance caused by RF
ablation are, in large part, result of increased transmural tissue
dryness as water evaporates from the ablation site. Increased
lesion depth therefore, results in more opaque appearance of the
tissue. FIG. 13A and FIG. 13B show a reconstruction of a lesion
depth from an HSI image. FIG. 13A displays a 3D depth profile of a
deep RF lesion as well as gap between weak and strong lesion placed
side by side as viewed in FIG. 13B.
[0111] It is contemplated that the systems and methods of the
present disclosure relating to the approach of unmixing of spectral
information from the areas that contain ablated tissue can be
applied to the entire image made from multiple pixels or to limited
number of individual pixels. It is possible that in an extreme case
that there will be single a point measurement and a spectral
analysis of the light collected using a contact-type catheter.
[0112] It is possible that the systems and methods of the present
disclosure can be applied to any parts of the heart, including
epicardial and endocardial surfaces of the right and left atria,
endocardial and epicardial surfaces of the ventricles, as well as
major vessels and valve structures.
[0113] Further, it is possible that the systems and methods of the
present disclosure can be applied to identify different tissues and
sites of ablation that are performed on various organs and parts of
the human body, including, but not to limited to uterine lining
(endometrial ablation) or cancer within several organs of the body,
including the liver, kidneys, lungs, muscle or bone.
[0114] According to aspects of the present disclosure, the surface
of tissue can include heart tissue. The surface of heart tissue can
include an endocardial surface of atrial tissue. The illumination
wavelength can be between about 350 nm and about 400 nm. The
illumination wavelength can be between about 400 nm and about 700
nm. The illumination wavelength can be between about 700 nm and
about 900 nm. The illumination wavelength can be below a range of
acquisition wavelengths by between about 10 nm and about 50 nm.
[0115] According to aspects of the present disclosure, it is
possible to include filtering light returning from the illuminated
heart tissue using a set of bandpass filters. It is contemplated to
include filtering light returning from the illuminated heart tissue
using a tunable filter. It is possible to incorporate detecting a
light returning from the illuminated heart tissue, that includes
fluorescence, reflectance and scattering components, wherein the
fluorescence component being detected at the acquisition wavelength
between about 400 nm to about 500 nm, the reflectance component
being detected at the acquisition wavelength between about 450 nm
and about 700 nm, and the scattering component detected across the
entire visible spectra.
[0116] According to aspects of the present disclosure, it is
possible to include classifying each pixel of a digital image of
the illuminated tissue as either ablated or unablated tissue.
Wherein constructing a depth map of the ablation lesion can be from
the image based on a bulk density of image pixels classified as
ablated tissue.
[0117] According to aspects of the present disclosure, it is
contemplated to further include identifying the ablated lesion as a
lesion created by radiofrequency; and setting the illumination
wavelength to between about 400 nm and about 700 nm and the
acquisition wavelength to about 400 nm and about 700 nm. It is
possible to further include identifying the ablated tissue created
by cryoablation by setting the illumination wavelength to between
350-400 nm and the range of acquisition wavelengths to 380-500 nm.
It is contemplated to further include identifying the ablated
tissue created by radiofrequency by setting the illumination
wavelength to between 350-400 nm and the range of acquisition
wavelengths to 380-500 nm. Further, it is possible to include
identifying ablation lesions created by other types of thermal
ablations including laser, microwave or focused ultrasound induced
lesion. It is contemplated to further include illuminating at one
or more illumination wavelengths a specific wavelength or using
wide band illumination with a known spectral distribution. It is
possible to further include distinguishing between the ablation
lesion and an unablated tissue based on one or more spectral
differences of a pre-selected set of multiple spectral
differences.
[0118] According to aspects of the present disclosure, it is
contemplated to further include the collecting of the spectral data
that includes one of a detector-based Hyperspectral Imaging (HSI)
system, a detector-based HSI system using static devices with
tunable filters, an alternative source-based HSI including a
changeable wavelength of illuminating light or some combination
thereof. It is possible to further include the collected spectra
data can be one of matched to existing spectral libraries,
subjected to a principal component analysis, subjected to related
principal component analysis algorithms or some combination
thereof.
[0119] In some aspects, there is provided a method for visualizing
ablation lesions that includes illuminating at one or more
illumination wavelengths a surface of tissue having an ablation
lesion; collecting a spectral data set comprising spectral images
of the illuminated tissue acquired at multiple spectral bands each
at one or more acquisition wavelengths; distinguishing between the
ablation lesion and an unablated tissue based on one or more
spectral differences between the ablation lesion and unablated
tissue; and creating a composite image of the tissue showing the
ablation lesion and the unablated tissue.
[0120] In some aspects, there is provided a method for visualizing
atrial ablation lesion that includes illuminating one or more
discrete illumination wavelengths a surface of heart tissue having
an ablation lesion; collecting a spectral data from the illuminated
heart tissue; distinguishing between the ablation lesion and an
unablated tissue based on one or more spectral differences between
the ablation lesion and unablated tissue; and creating an image of
the heart tissue illustrating ablated tissue and unablated
tissue.
[0121] In some aspects, there is provided a system for imaging
tissue that includes a catheter having a distal region and a
proximal region; a light source; an optical fiber extending from
the light source to the distal region of the catheter to illuminate
a tissue having a lesion site in proximity to the distal end of the
catheter; an image bundle for collecting light reflected from the
illuminated tissue; a camera connected to the image bundle, the
camera being configured to gather hyperspectral data comprising
spectral images of the illuminated tissue acquired at multiple
spectral bands or at each illumination wavelength; an image
processing unit in communication with the camera, the unit being
configured to distinguish between the ablation lesion and an
unablated tissue based on one or more spectral differences between
the ablation lesion and unablated tissue and creating an image of
the heart tissue illustrating the ablated tissue and the unablated
tissue.
[0122] In some aspects, there is provided a system for imaging
heart tissue that includes an illumination device configured to
illuminate a tissue having a lesion site; an imaging device
configured to gather hyperspectral data; an image processing unit
in communication with the imaging device, the image processing unit
configured to processing gathered hyperspectral data to generate an
image that reveals the lesion site, wherein the generated image
enabling to distinguish between an ablation lesion and an unablated
tissue based on one or more spectral differences between the
ablation lesion and unablated tissue and creating an resulting
image of the heart tissue illustrating the ablated tissue and the
unablated tissue.
[0123] The foregoing disclosure has been set forth merely to
illustrate various non-limiting embodiments of the present
disclosure and is not intended to be limiting. Since modifications
of the disclosed embodiments incorporating the spirit and substance
of the disclosure may occur to persons skilled in the art, the
presently disclosed embodiments should be construed to include
everything within the scope of the appended claims and equivalents
thereof.
* * * * *