Characterizing Cardiac Tissue Using Catheterized Light Scattering Spectroscopy

Through multi-arm arrangement of fiber optic probes and machine learning technology, the problem of difficult-to-identify lesions such as fibrosis in the deep heart tissue is solved, deep characterization and lesion detection of heart tissue are achieved, and the diagnosis and treatment effect of heart disease is improved.

CN114981695BActive Publication Date: 2024-11-12UNIV OF UTAH RES FOUND
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202080088229.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-12-17
Filing Date
2020-12-17
Publication Date
2024-11-12
Estimated Expiration
2040-12-17

AI Technical Summary

Technical Problem

Existing cardiac tissue characterization methods cannot effectively characterize the relevant tissue depth and microstructure scale, resulting in challenges in the diagnosis and treatment of heart diseases such as myocardial infarction and atrial fibrillation, especially the depth distribution of fibrotic tissue is difficult to accurately. Identify.

Method used

A tissue characterization probe was designed, using multi-arm elongated members to realize the depth characterization of cardiac tissue through multiple illumination and detection fibers, which can effectively characterize anisotropic tissues within depths greater than 100μm or even up to 30mm. , combined with machine learning technology to analyze spectral data, provide detailed components and distribution information of heart tissue.

Benefits of technology

Accurate detection and monitoring of deep fibrotic tissue, allograft rejection, myocarditis and other lesions in the heart tissue, improve the accuracy and efficiency of diagnosis and treatment of heart disease, and reduce the costly equipment and invasive surgery. dependency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114981695B_ABST
    Figure CN114981695B_ABST
Patent Text Reader

Abstract

Disclosed are devices, systems, and methods for characterizing tissue using light scattering spectroscopy. A tissue characterization probe comprises: an elongated member having a proximal end and a plurality of distal probe tips disposed at a distal end. A plurality of illumination fibers extend through the elongated member to the distal probe tips, such that each distal probe tip includes at least one illumination fiber. A plurality of detection fibers also extend through the elongated member, such that each probe tip includes at least one detection fiber. The disclosed devices and systems are advantageously capable of characterizing tissue within a depth greater than 100 μm. The disclosed devices and systems are also capable of effectively characterizing anisotropic tissues, such as myocardium.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Cross-references to related applications

[0002] This application claims priority and benefit to U.S. Provisional Patent Application Serial No. 62 / 949,290, filed December 17, 2019, entitled “Cardiac Tissue Characterization Using Catheterized Light Scattering Spectroscopy”, the entire disclosure of which is incorporated herein by reference.

[0003] Statement on Federally Funded Research

[0004] This invention was completed with government funding under license numbers HL128813 and HL135077 granted by the National Institutes of Health. The government holds certain rights to this invention. Background Technology

[0005] Heart disease imposes a significant disease burden on society. Abnormal cardiac tissue microstructure is frequently associated with heart disease. Heart diseases associated with microstructural abnormalities include allogeneic transplant rejection, myocarditis, amyloidosis, hypertrophy, and other cardiomyopathies.

[0006] One specific type of this remodeling is fibrosis, which occurs as an maladaptive response to metabolic, hemodynamic, and ischemic stress. Fibrosis is defined as the excessive formation of connective tissue, specifically including the extracellular matrix, fibroblasts, and myofibroblasts. During the development of fibrosis, extracellular matrix proteins, including collagen-1 and fibronectin-1, are overproduced and released into the atrial tissue. Fibrosis significantly alters the mechanical properties of cardiac tissue. One effect is that fibrosis reduces myocardial mechanical function, such as myocardial mechanical function quantified, for example, by radial strain and ejection fraction.

[0007] Important examples of heart diseases associated with fibrosis are myocardial infarction and atrial fibrillation (AF). During a myocardial infarction, muscle tissue is replaced by fibrotic tissue, which can cause arrhythmias. Fibrosis is also thought to maintain arrhythmias such as AF. Typical AF treatment involves using medications (such as beta-blockers) to control heart rate and using anticoagulants to prevent thromboembolism. However, for patients who still have symptoms, rhythm control is often achieved using antiarrhythmic drugs and / or catheter ablation. Catheter ablation involves selectively destroying areas of tissue, a process typically achieved by applying radiofrequency (“RF”) energy to heat the tissue. Several trials have shown that catheter ablation can result in long-term sinus rhythm maintenance in patients with AF. However, the recurrence rate of AF after ablation is as high as 50%. Furthermore, 20–40% of AF patients will undergo multiple ablation procedures.

[0008] Other important examples of heart disease are characterized by changes in the density of cells and their nuclei due to infiltration and proliferation. These examples include myocarditis and allogeneic transplant rejection. Similarly, cardiac hypertrophy is characterized by changes in the density of cardiomyocytes.

[0009] Improving cardiac tissue characterization and mapping could potentially enhance the diagnosis and / or treatment of heart diseases associated with microstructural abnormalities, such as atrial fibrillation (AF), allogeneic transplant rejection, myocarditis, hypertrophy, and amyloidosis. Unfortunately, conventional methods for identifying and diagnosing cardiac tissue do not provide effective characterization.

[0010] Macroscopic areas of cardiac tissue can be visualized using magnetic resonance imaging (MRI) (e.g., late-stage gadolinium-enhanced MRI). However, not all care centers have access to the necessary and relatively expensive MRI equipment, and such procedures are associated with high costs. Furthermore, while MRI imaging can detect certain tissue abnormalities at a macroscopic scale, it has limited resolution and does not provide in-depth understanding of the microscopic distribution and composition of microstructural abnormalities.

[0011] Fiber confocal microscopy (FCM) can be used as an optical method for imaging cardiac tissue. However, suitable FCM systems require expensive hardware. Furthermore, FCM has a limited depth of penetration and therefore cannot provide information about tissues of interest at depths exceeding approximately 100 μm.

[0012] The established clinical tool for assessing cardiac tissue microstructure is endomyocardial biopsy (EMB), which requires an invasive procedure for tissue extraction. Furthermore, due to its high complication rate, this procedure is rarely performed in the atria.

[0013] Therefore, there has long been a need for devices and methods capable of characterizing cardiac tissue at relevant tissue depths and microstructural scales. Such advances would be beneficial in improving outcomes and reducing the disease burden. Summary of the Invention

[0014] In one embodiment, the tissue characterization probe includes an elongated member having a proximal end and a plurality of distal probe tips disposed at or near the distal end of the elongated member to form a multi-arm arrangement. A plurality of illumination fibers extend at least partially through the elongated member, each illumination fiber extending to a corresponding probe tip of the multi-arm arrangement such that each probe tip includes at least one illumination fiber. A plurality of detection fibers also extend at least partially through the elongated member such that each probe tip of the multi-arm arrangement includes at least one detection fiber, and optionally includes a plurality of detection fibers.

[0015] In one embodiment, the tissue characterization probe includes an elongated member having a proximal end and a distal probe tip. An illumination fiber extends through the elongated member to the distal probe tip and is configured to deliver light to and beyond the probe tip into the target tissue. A plurality of detection fibers also extend through the elongated member to the probe tip and are configured to receive light scattered from the target tissue.

[0016] The detection fiber is arranged relative to the illumination fiber to advantageously enable tissue characterization at depths greater than 100 μm, such as up to about 4 mm, or up to about 8 mm, or up to about 12 mm, or up to about 16 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm. The detection fiber is also arranged to effectively characterize anisotropic tissues, such as myocardium.

[0017] In one embodiment, a first set of detection optical fibers is arranged along a first detection line, which is orthogonal to the illumination axis. A second set of detection optical fibers is arranged along a second detection line. The second detection line is transverse to the first detection line, and preferably orthogonal to the first detection line. Preferably, both the first and second sets of detection optical fibers have at least two detection optical fibers.

[0018] In one embodiment, a method for characterizing tissue includes the following steps: (i) providing a tissue characterization system; (ii) guiding the distal probe tip of the tissue characterization system to a target anatomical location; (iii) at the target anatomical location, manipulating the tissue characterization probe to obtain spectral data at depths greater than about 100 μm (such as up to about 1 mm, or up to about 1.5 mm, or up to about 2 mm, or up to about 2.5 mm, or up to about 3 mm, or up to about 3.5 mm, or up to about 4 mm, or up to about 5 mm, or up to about 7.5 mm, or up to about 10 mm, or up to about 15 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm); and resolving the spectral data to characterize the target tissue.

[0019] The target tissue may be cardiac tissue. The methods described herein are particularly suitable for characterizing cardiac tissue within a blood-filled, beating heart. Characterizing the target tissue may include detecting, measuring, or monitoring one or more of the following: fibrotic tissue, allogeneic transplantation acceptance or rejection, myocarditis, amyloidosis, other cardiomyopathy, or one or more tissue parameters, such as nuclear density. One method may include determining the volume fraction of the target tissue components and / or the spatial distribution of the target tissue within the heart.

[0020] In the methods described herein, the step of parsing spectral data to characterize the target tissue may include the use of one or more machine learning techniques. Machine learning techniques may include supervised and / or unsupervised techniques. Attached Figure Description

[0021] Various objects, features, characteristics, and advantages of the present invention will become apparent and more readily understood from the following description of embodiments, taken in conjunction with the accompanying drawings and appended claims, all of which form part of this specification. In the drawings, similar reference numerals may be used to designate corresponding or similar portions in the various figures, and the various elements depicted are not necessarily drawn to scale, wherein:

[0022] Figure 1 A cross-section of the atrial wall is shown as an example of a cardiac chamber, and the depth below the endocardial surface is made visible, where the tissue of interest is typically located, showing that the tissue of interest is typically outside the optical imaging range;

[0023] Figure 2 An exemplary embodiment of a tissue characterization probe is shown, the tissue characterization probe having a multi-arm arrangement of the distal probe tip;

[0024] Figures 3A to 3C A frontal view of the distal end of an exemplary tissue characterization probe is shown, which is configured to characterize cardiac tissue at clinically relevant depths using light scattering spectroscopy (LSS).

[0025] Figure 4 An extended diagram of an exemplary tissue probe assembly is shown;

[0026] Figure 5 An exemplary tissue characterization system that can be used for in situ or in vivo characterization of cardiac tissue is shown;

[0027] Figure 6 The count of cell nuclei in cardiac tissue is shown, with a scatter plot of the measured ND relative to age and a regression fit.

[0028] Figure 7A and Figure 7B The diagram shows spectral clustering measured from cardiac tissue, where Figure 7A The clusters identified in the principal component plot are shown. Figure 7B The diagram shows the differences in the nominative index (ND) of several identified clusters compared to other clusters (the bracketed markers indicate that the clusters are statistically distinct from each other at the 5% confidence level); and

[0029] Figure 8 The confusion matrix of aggregated CNN predictions from the ND of the spectrum is shown, which demonstrates that the classification of the spectrum is effective. Detailed Implementation

[0030] introduction

[0031] This disclosure relates to apparatus, systems, and methods for characterizing tissues, particularly cardiac tissues, using LSS. The embodiments described herein, including probes, systems, and methods, can be used in combination and / or in conjunction with the apparatus, systems, and methods described in patent document PCT / US2018 / 016314 (disclosed as WO2018144648A1), the entire contents of which are incorporated herein by reference.

[0032] For example, the tissue characterization probe assembly described herein can be added to any endovascular device described in patent document PCT / US2018 / 016314, thereby adding tissue characterization capabilities to the imaging, localization, therapeutic (e.g., ablation), and / or electrographic functions of the endovascular device of patent document PCT / US2018 / 016314. Similarly, the tissue characterization method described herein can be added to any method of generating and / or mapping tissue atlases described in patent document PCT / US2018 / 016314, thereby adding to or enhancing the effective tissue characterization of the atlas.

[0033] For example, as described in more detail below, the tissue characterization using LSS in conjunction with the optimized embodiments described herein advantageously enables the characterization of tissue at greater depths than is possible using conventional methods. Therefore, this additional and / or more accurate tissue characterization information can enhance the tissue atlases generated using the embodiments described in patent document PCT / US2018 / 016314 (published as WO2018144648A1).

[0034] It should also be understood that while many of the examples detailed below relate to the detection of fibrosis in cardiac tissue, the same principles and characteristics can be readily applied to other applications that ensure the detection, diagnosis, and / or treatment of abnormal tissue microstructure. Therefore, the embodiments can be used to monitor the risk of allogeneic transplant rejection, myocarditis, amyloidosis, and other cardiomyopathies. Cardiac nuclear density (ND) is a parameter that can be measured and effectively characterized using the described embodiments to provide enhanced insights into cardiac microstructure for disease monitoring, diagnosis, and / or treatment purposes.

[0035] Figure 1 A cross-section of cardiac tissue (e.g., from the atrial wall) is shown. As illustrated, the tissue of interest is typically located deeper than the depth of focus in FCM. In particular, fibrotic extracellular tissue may be located at a depth of approximately 1 mm or greater, while conventional optical imaging such as FCM may only have a maximum imaging depth of approximately 100 μm. The structures and tissues of the cardiac conduction system may also be located in relatively deep tissue layers.

[0036] Furthermore, cardiac tissue is composed of muscle fibers with directional and anisotropic structures. This anisotropy makes tissue imaging and characterization difficult. For example, even when using LSS, the scattered light detection signal is affected by the anisotropic arrangement of the target cardiac tissue, making it difficult to accurately characterize the tissue (e.g., as a contrast between fibrosis and normal tissue).

[0037] Therefore, while optical imaging may be sufficient to characterize surface-level microstructures such as epithelial cells, it cannot provide information about underlying tissues. This is a particular drawback in cardiac tissue applications, where the tissues of interest typically extend beyond the immediate surface level. Furthermore, although conventional LSS can theoretically provide information about deeper tissue layers, the anisotropic nature of cardiac tissue makes effective characterization difficult.

[0038] Tissue characterization probes and systems

[0039] Figure 2 An exemplary embodiment of a tissue characterization probe 400 is shown, which includes a plurality of probe tips 403 disposed in a multi-arm arrangement. The probe 400 includes an elongated member 401 through which illumination and detection optical fibers can be arranged.

[0040] Each probe tip 403 in the multi-arm arrangement can be independently constructed according to any other probe tip configuration described herein, such as those described in more detail below. Figures 3A to 4 The ones shown are not included. However, in some embodiments, one or more probe tips 403 may be constructed differently. For example, one or more probe tips 403 may include only a single detection fiber and a single illumination fiber. That is, while the entire probe 400 includes multiple detection fibers (e.g., at least one detection fiber in each individual probe tip 403), each particular probe tip 403 need not include multiple detection fibers.

[0041] Furthermore, although the specific structural relationship between the illumination fiber and the detection fiber is about Figures 3A to 3C The probe is described, but one or more probe tips 403 may have different configurations, such as by setting the detection fiber radially around the illumination fiber, or setting the detection fiber along a grid, or setting the detection fiber in a random orientation relative to the illumination fiber, etc.

[0042] It should be understood that although the illustrated embodiment includes a specific number of probe tips 403 (i.e., arms), other embodiments may include more or fewer arms. Generally, a greater number of arms is preferred, provided they can be included within given space and / or cost constraints.

[0043] As shown, any probe tip 403 may include one or more electrodes 407 configured to provide navigation, mapping, and / or positioning capabilities. For example, electrodes 407 may be used to determine the position of the probe tip(s) within a three-dimensional anatomical workspace, such that measurements can be correlated with their corresponding positions within a target anatomical location. The correlation between position and measurement data can be used to generate a three-dimensional map of the target anatomical location (e.g., the tissue microstructure of the target anatomical location). As shown, electrodes 407 may be formed as loops. In a preferred embodiment, multiple loops are positioned at different longitudinal locations along the length of the distal portion of the probe tip 403. Other embodiments may additionally or alternatively use other types of electrodes known in the art.

[0044] Figure 2 The multi-arm arrangement shown offers several benefits. In particular, it allows for more rapid characterization and mapping of target anatomical locations. This can be especially important for invasive and / or expensive procedures, such as those involving cardiac catheterization and cardiac tissue characterization. Furthermore, the multi-arm arrangement increases the speed of characterization and / or mapping not merely by the number of tips / arms included. For example, a probe tip / arm can be positioned in a given location for reading, and the probe can then be rotated to radially reposition the tip / arm for additional readings. Conversely, a single-arm design cannot provide any additional information simply by rotating the probe, and the probe tip must be moved to a new position for each reading.

[0045] Figures 3A to 3C A frontal view of the distal end of an exemplary tissue characterization probe is shown, which is configured to provide effective tissue characterization at clinically relevant depths and when characterizing anisotropic tissues, such as cardiac tissue. Figure 3A A front view of a specific distal probe 100 is shown. The distal probe tip may be referred to herein as the "distal tip" or "probe tip" synonymously.

[0046] The probe 100 includes an elongated member 102 forming the external structure of the device, which is configured to penetrate the patient's vascular system to reach the heart. An illumination fiber 104 extends through the elongated member 102 and is configured to carry source light and propagate it distally into the target tissue. As shown, a plurality of detection fibers 106 are also disposed within the elongated member 102. The detection fibers 106 are configured to receive scattered light and return it towards the proximal end of the elongated member 102.

[0047] The illumination fiber 104 defines the illumination axis of the probe (from... Figure 3A (The viewing angle extends through the paper). The lighting axis can be substantially centered within the slender member 102, such as... Figure 3A As shown in the embodiments, although other embodiments may position the illumination fiber off-center from the center of the elongated member (e.g.) Figure 3B (As shown in the embodiment). The "first detection line" is defined as a line extending orthogonally from the illumination axis, as shown in line 108. The "second detection line" is also defined as a line extending orthogonally from the illumination axis, as shown in line 110.

[0048] As shown in the figure, the first detection line 108 and the second detection line 110 are transverse to each other (i.e., not parallel to each other), and preferably orthogonal to each other (i.e., perpendicular to each other). Figure 3A In the cross-sectional view along the illumination axis, the first detection line 108 and the second detection line 110 may intersect each other at the illumination axis to form a lateral angle, such as about 30° to about 150°, or about 60° to about 120°, or preferably about 90°.

[0049] The first set of detection fibers 112 is arranged substantially along the first detection line 108, and the second set of detection fibers 114 is arranged substantially along the second detection line 110. It has been found that arranging the detection fibers 106 in this manner provides effective functionality, and in particular, it has been found to be effective for characterizing cardiac tissue (including anisotropic tissue) at clinically relevant depths.

[0050] As used herein, if the detection fibers are radially offset from the detection line by no more than about 30 degrees, or no more than about 25 degrees, or no more than about 20 degrees, or no more than about 15 degrees, or no more than about 10 degrees, or no more than about 5 degrees, they are considered to be "substantially arranged," "substantially aligned," and / or "substantially positioned" along the corresponding first or second detection line. This refers to... Figure 3C As best illustrated. If detection line 108 is defined as starting from illumination fiber 104 and extending across one of the detection fibers (in this case, 106a), then any other detection fiber in that group should be close to detection line 108, but not necessarily perfectly aligned with it. For example, detection fiber 106b is not perfectly aligned with detection line 108, but is radially offset from detection line 108 by an angle “A”, where illumination fiber 104 defines the apex.

[0051] Refer again Figure 3A The first set of detection fibers 112 and the second set of detection fibers 114 preferably each comprise at least two detection fibers. Providing at least two detection fibers in one set allows for depth sensitivity. It has been found that providing a second set of detection fibers laterally offset from the first set (i.e., the two sets form a non-parallel, preferably perpendicular angle with respect to each other, with the illumination fiber 104 serving as the vertex) advantageously reduces the directional sensitivity of the spectrum in anisotropic tissues.

[0052] For example, it has been found that arranging the first set of detection fibers 112 and the second set of detection fibers 114 in a lateral and, in particular, orthogonal manner, provides an overall averaging effect when probing anisotropic tissues, allowing for effective tissue characterization despite the high level of anisotropy in the target tissue. Therefore, depth sensitivity and anisotropic sensitivity are achieved through the first set of detection fibers 112 having at least two detection fibers and the second set of detection fibers 114 having at least two detection fibers radially corresponding to the first set of fibers 112.

[0053] It should be noted that although four detection fibers 106 are shown in this embodiment (two along the first detection line 108 and two along the second detection line 110), other embodiments may include other numbers of detection fibers. As described above, the tissue characterization probe preferably includes at least four detection fibers (two sets of detection fibers, each set containing two detection fibers, each set of detection fibers arranged along the transverse detection line) to provide effective depth sensitivity and anisotropic sensitivity. Additional detection fibers may be arranged along the transverse detection line and / or at other locations to further improve resolution and / or sensitivity. However, in some cases, space constraints may favor a minimum number of detection fibers.

[0054] The spacing of the detection fiber 106 along the corresponding detection lines 108 and 110 can be varied. By changing the spacing and arrangement of the fibers, the characterization system can be constructed to meet specific application needs. For example, the detection fiber 106 can be spaced apart from the illumination fiber 104 and / or spaced apart from each other at a distance relevant to specific application needs. In one example, it was found that for some applications, the combination of the adjacent (illumination fiber 104) detection fiber and the distal detection fiber produces the most accurate results.

[0055] Therefore, although the spacing of the detection fibers 106 can vary depending on the specific application requirements, some embodiments minimize the spacing so that the detection fibers are substantially adjacent to other detection fibers in a group (e.g., within about 135 μm) and each group of detection fibers is substantially adjacent to the illumination fiber 104 (e.g., within about 135 μm).

[0056] Figure 3B Another exemplary embodiment of a distal probe 200 having features similar to those of the distal probe 100 is shown, unless otherwise stated. Figure 3A Similar to the previous embodiment, the illustrated embodiment includes an elongated member 202, an illumination fiber 204 extending through the elongated member 202, and a plurality of detection fibers 206 arranged in coordination with the illumination fiber 204 to achieve LSS using a distal probe 200.

[0057] In the illustrated embodiment, the illumination fiber 204 is offset from the longitudinal axis of the elongated member 202. Similar to the distal probe 100, detection lines 208 and 210 extend from the illumination fiber 204, and detection fibers 206 are substantially aligned on the detection lines, wherein a first set of detection fibers 212 is substantially aligned on detection line 208, and a second set of detection fibers 214 is substantially aligned on detection line 210. It should be noted that in this embodiment, the detection fibers 206 are spaced apart from each other and from the illumination fiber 204.

[0058] As shown, the detection fibers 206 can be spaced substantially equidistantly on each corresponding detection line. For example, along the first detection line 208, the space between each fiber (including the illumination fiber 204 and the detection fiber 206) is substantially equal. Preferably, the spacing is repeated in a similar manner on the second detection line 210, such that each detection line in the detection line similarly spaces the corresponding detection fibers, although other embodiments may include different spacings.

[0059] As shown, by moving the illumination fiber 204 away from the center of the elongated member 202, the internal space of the elongated member 202 is utilized more effectively, allowing for a smaller overall diameter of the elongated member 202 and / or the use of additional components. For example, the illustrated embodiment includes a support wire 216 that extends at least partially through the elongated member 202 and is configured to increase the bending stiffness of the distal tip 200 and / or provide a structure capable of forming a bent / shaped tip.

[0060] The increased stiffness provided by the support filament 216 helps to hold the distal tip 200 in place during surgery. For example, when taking measurements within a blood-filled, beating heart, it can be difficult to keep the device positioned on the target tissue without losing contact or slipping out of place. The increased structure and stiffness make it easier for the user to maintain the correct position throughout multiple heartbeats without causing tissue damage or complicating vascular navigation. The support filament 216 also allows the user to “shape” the tip with desired bending and / or orientation. Bending the tip is beneficial for navigating specific vascular system channels and / or providing other desired structural arrangements, such as… Figure 2 The probe tip is arranged circumferentially as shown.

[0061] The support wire 216 may have a quadrilateral cross-sectional shape. A quadrilateral cross-sectional shape can advantageously provide a specific bending plane. That is, the bending stiffness of the support wire 216 along a direction aligned with the edge of the cross-sectional shape will be less than the bending stiffness along other directions (e.g., the diagonal directions of the shape). In some embodiments, as shown, the support wire 216 has a rectangular shape. A rectangular cross-sectional shape may be desirable in some cases because it can provide bending planes with different bending stiffnesses. For example, the bending stiffness along a direction aligned with the long axis of the rectangular cross-section will be greater than the bending stiffness along a direction aligned with the short axis of the rectangular cross-section.

[0062] The support wire 216 can be positioned anywhere within the distal tip 200. Preferably, the support wire 216 is configured such that its cross-section forms an acute angle between the detection lines 208 and 210, as in the illustrated embodiment. This position effectively utilizes the space within the distal tip 200 and thus provides greater design flexibility, dimensional control, etc.

[0063] Therefore, tissue characterization probe 200 is structurally and functionally similar to tissue characterization probe 100. However, tissue characterization probe 200 shows that the illumination fiber 204 (and therefore the illumination axis) does not necessarily need to be aligned with the central axis of the elongated member 202. As shown, by moving the illumination fiber 204 out of the center, the internal space of the elongated member 202 is utilized more efficiently, allowing for a smaller overall diameter of the elongated member 202.

[0064] Figure 4 An extended view of tissue probe assembly 501 is shown, which can be combined with any tissue catheter insertion probe assembly described herein, including one or more. Figure 3A and Figure 3B The distal tip 100 and / or 200. Furthermore, although only a single distal tip 500 is shown, it should be understood that other tips such as... Figure 2 The multi-arm arrangement shown.

[0065] The probe assembly 501 shown includes an elongated member 502 extending between a proximal and a distal end. A distal tip 500 is located at the distal end. A combiner 505 receives multiple component portions at its proximal end and positions them at the distal tip 500 in a manner necessary for proper function. For example, as shown, the combiner 505 may receive one or more illumination fibers 504 and one or more sets of detection fibers 512, 514. These can then be connected to a connector (not shown) capable of connecting to corresponding components (e.g., light sources and spectrometers) of a tissue characterization / LSS system.

[0066] The probe assembly 501 is configured to be introduced (primarily distally) through the lumen or working channel of a catheter, sheath, endoscope, or other such intravascular delivery device. The delivery device may be steerable, for example, in the form of a guideable catheter or a steerable sheath.

[0067] Figure 5 An example of a tissue characterization system 301 is shown. System 301 includes a tissue characterization probe 300, which may be configured as tissue characterization probes 100 and 200 as described herein. An illumination fiber 304 extends from probe 300 and is operatively connected to a suitable light source 316 (e.g., a tungsten halide source or other suitable LSS source). A detection fiber 306 extends from probe 300 and is operatively connected to a suitable spectrometer 318 (or multiple spectrometers). System 301 can be used to characterize target tissue 30.

[0068] Methods for characterizing tissues

[0069] In one embodiment, a method for characterizing tissue includes the following steps: (i) providing a tissue characterization system; (ii) guiding the distal probe tip of the tissue characterization system to a target anatomical location; (iii) at the target anatomical location, manipulating the tissue characterization probe to obtain spectral data at a depth greater than about 100 μm (such as up to about 1 mm, or up to about 1.5 mm, or up to about 2 mm, or up to about 2.5 mm, or up to about 3 mm, or up to about 3.5 mm, or up to about 4 mm, or up to about 5 mm, or up to about 7.5 mm, or up to about 10 mm, or up to about 15 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm); and resolving the spectral data to characterize the target tissue.

[0070] The target tissue may be cardiac tissue. The methods described herein are particularly suitable for characterizing cardiac tissue within a blood-filled, beating heart. Characterizing the target tissue may include detecting, measuring, or monitoring one or more of the following: fibrotic tissue, allogeneic transplantation acceptance or rejection, myocarditis, amyloidosis, other cardiomyopathy, or one or more tissue parameters such as nuclear density. One method may include determining the volume fraction of the target tissue component and / or the spatial distribution of the target tissue within the heart.

[0071] Some embodiments may further include characterizing tissue at multiple target locations to obtain one or more characterization data points; at each location, manipulating a tissue characterization probe to determine the position of the probe tip within a three-dimensional anatomical workspace; associating each characterization data point with its corresponding determined position within the anatomical workspace; and generating a three-dimensional map of the anatomical workspace based on the characterization data points and their corresponding positions. Additional details regarding these methods are described in more detail in patent document PCT / US2018 / 016314 (disclosed as WO2018144648A1), the entire disclosure of which is incorporated herein by reference.

[0072] LSS spectroscopy using machine learning

[0073] In the methods described herein, the step of parsing spectral data to characterize target tissues may include the use of one or more machine learning techniques. These machine learning techniques may include supervised or unsupervised techniques.

[0074] In one embodiment, unsupervised machine learning techniques for parsing spectral data include cluster analysis. Cluster analysis can apply dimensionality reduction of the spectrum through principal component analysis (PCA). Euclidean distance can be used to generate the similarity between nodes of the first and second principal components of the spectrum, and the eigenvalues ​​of the spectral clusters can indicate the number of groups a sample data point falls into. Additional details related to using spectral clustering to parse spectral data for tissue characterization are included in the Examples section.

[0075] In another embodiment, supervised machine learning techniques for parsing spectral data include the use of convolutional neural networks (CNNs). A CNN can be trained to classify spectra into a selected number of groups / categories. Various methods known in the art can be used to train a CNN. A particularly useful training method is the Nl method. For example, multiple CNNs (each with the same topology) are trained on a dataset that includes all spectra in the dataset except for one sample. A different sample is omitted from each dataset, such that the number of training datasets equals the number of samples, and each of the multiple CNNs is trained on a different dataset depending on which sample was omitted.

[0076] A CNN can be trained for a set number of epochs and / or until a loss threshold is reached. The CNN can then be tested on spectra omitted from their respective training sets. One or more optimal CNNs can then be selected for characterization and / or for further training. CNNs can also be trained on different subgroups of spectral wavelengths to determine if the CNN has a dependency on a specific wavelength range.

[0077] Preprocessing of spectral data prior to characterization and / or training may include various filtering and / or normalization (e.g., intensity normalization) steps. Other spectral data processing techniques known in the art may be used additionally or alternatively.

[0078] Example

[0079] Tissue representation function

[0080] Experimental tissue stacking was used to test tissue characterization probes. Thin tissue slices (200 μm thick) of myocardial or aortic tissue were stacked on top of each other. The height of the stack was limited to up to eight cross-sectional layers (1.6 mm). Based on the fundamental principle that the heart is primarily myocardium and the ventricles are mainly free of fibrotic infiltration, the matrix layer was prepared using myocardial tissue from the free wall of the ventricles of a formalin-fixed adult canine heart. This also allowed for the study of tissue anisotropy. Based on the fundamental principle that the aorta is composed of approximately 50% elastin and collagen by weight and is easily accessible and identifiable in vivo, the target tissue was prepared from tissue from the ascending aorta and aortic arch of a fixed adult canine heart.

[0081] To test the depth sensitivity of the tissue characterization probe, two slices of the target tissue were sequentially lowered into the tissue stack to determine the depth of the structure the system could detect. The tissue detection probe has been able to detect target tissue at depths of approximately 1.5 mm to 2 mm with good accuracy, and is expected to detect target tissue up to approximately 4 mm in at least some cases. In this experimental setup, the overall accuracy of depth detection was approximately 95.45 ± 3.99%.

[0082] Tests were also conducted to determine the sensitivity of the tissue characterization probe to detecting the volume fraction of fibrotic tissue, an important indicator of tissue functional capability. Different volume fractions of fibrotic tissue within the entire stack were tested by varying the relative percentage of fibrotic tissue in the structure by 12.5% ​​(i.e., a 200 μm slice) until the entire volume fraction of the structure was fibrotic tissue. Although some "masking" effect was initially expected due to the upper layer of target tissue concealing the lower layer, the overall mean accuracy of the tissue characterization probe in determining the volume fraction of fibrotic tissue was very high, at approximately 80.56 ± 2.12%.

[0083] Tests were also conducted to characterize the mixed / replacement portion, in which the total volume fraction of the aortic portion was 50%. Although it was initially expected that such “dissection” and / or mixed / embedded tissue would be difficult for the system to characterize accurately, the tissue characterization probe was able to characterize such mixed / embedded tissue with a high accuracy of approximately 84.25 ± 1.88%.

[0084] Exploratory measures were also taken throughout the canine heart (n=2) using spectra acquired from several different locations, including the right atrium, atrioventricular node, diaphragmatic RA along the Todaro tendon, septal joint of the tricuspid valve, aorta, ventricular wall, and sinoatrial node. Tissue characterization probes provided accuracy up to 95%. LSS spectral analysis was performed using machine learning.

[0085] Spectroscopy of cardiac tissue:

[0086] The LSS system was used to acquire spectra from sheep ventricular tissue. We obtained transmural tissue samples from the free walls of the right ventricle (RV) and left ventricle (LV) and the interventricular septum of formalin-fixed hearts of 18 animals ranging in gestational age from 4.3 to 56 months. The samples were examined using the LSS to identify changes in ND (density suture). The LSS probe was positioned at multiple locations on the epicardial surface to collect 20 spectra from all samples. Spectral acquisition occurred at 5 Hz. Spectra from the tissue samples were recorded at a resolution of 0.6 nm within the wavelength range of 500–1100 nm.

[0087] The raw spectra show a high degree of variability in intensity both within and between samples. After normalization of their corresponding averages, the spectra from a single sample show a high correlation (R0) with other spectra from the same sample. 2 ≥0.999). The differences in the spectra between different samples are subtle (R0.999). 2 =0.996±0.002).

[0088] Histology:

[0089] Tissue samples were sectioned within a sampling area at the center of each sample. 100 μm thick sections were cut from a transmural plane perpendicular to the epicardial surface and parallel to the heart wall. After sectioning, extracellular matrix glycoconjugates and cell membrane glycoproteins were labeled using fluorophore-bound wheat germ lectin (WGA). Cell nuclei were stained with 1 mM 6-diamidinyl-2-phenylindole (DAPI). ND was calculated by dividing the number of detected nuclei by the total tissue area. The average ND per heart provided the basis for machine learning. Figure 6 The study summarized the decrease in ND with age.

[0090] Cluster analysis:

[0091] Cluster analysis was applied to reduce the dimensionality of the spectra using Principal Component Analysis (PCA). We used the MATLAB function `spectralcluster` to identify spectral clusters from different ND preparations. This function allows selection of the data, distance metric, and desired number of groups. Euclidean distance was used to generate a similarity graph between the nodes of the first and second principal components of the spectra. Analysis of the eigenvalues ​​from the spectral clustering function showed that the spectra fell into 5 distinct groups. Therefore, the desired number of groups varied between 3 and 5.

[0092] Approximately 95% of the spectrum's variance is described in the first two principal components. Analysis of the eigenvalues ​​reveals five distinct groups. A scatter plot of the first two principal components is shown. Figure 7A These clusters are shown. Cluster groups are labeled from 1 to 5 as the number of possible values ​​(ND) increases. Analysis of variance revealed differences in ND among many clusters. Figure 7B The brackets connect clusters that are statistically distinct from each other at a 5% confidence level.

[0093] Convolutional Neural Networks:

[0094] We designed a CNN to classify the spectrum into three groups of ND counts: up to 2000, 2000-3800, and above 3800 kernels / mm. 2 The loss value of a CNN is calculated by applying a softmax activation to the cross-entropy loss compared to the base case. CNNs are systematically trained and validated using the Nl method. This method requires training 22 different CNNs with the same topology, each with a different training set that includes all spectra from the entire dataset, rather than collecting spectra from a single sample. Each of the 22 unique training datasets excludes spectra from a different heart. CNN training is terminated after 2000 epochs, or when the training loss no longer decreases after more than 600 epochs. The CNN weights associated with the lowest loss value are saved and restored to the CNN for testing. These CNNs are then tested on spectra excluded from their respective training sets.

[0095] To optimize the CNN parameters, we systematically varied the batch size, learning rate, and convolutional filter kernel size. The CNN configurations used for these parameter variations are described in Table 1. Seven batch sizes, ranging from 50 to 300, were tested. Ten different learning rates, ranging from 0.0001 to 0.015, were tested. Fifteen different convolutional filter kernel sizes, ranging from 5 to 40, were tested. The full-resolution spectrum was then divided into three wavelength ranges: 500–700 nm, 700–900 nm, and 900–1100 nm. These partial spectra were used to train the CNN to determine the classification's dependence on specific wavelengths.

[0096]

[0097]

[0098] Table 1: CNN Configuration

[0099] Parameter optimization yielded a kernel size of 20, a batch size of 100, and a learning rate of 0.003. Training the CNN using the N-1 method achieved an accuracy of 95.23 ± 12.20%. Figure 8 The confusion matrix of the CNN prediction is shown in the figure. Training the CNN on different spectral wavelength ranges yielded accuracies of 90.23±21.56%, 58.86±36.02%, and 81.59±34.52% for the wavelength ranges of 500-700nm, 700-900nm, and 900-1100nm, respectively. These results indicate that ND prediction depends more on the low (500-700nm) and high (900-1100nm) wavelength ranges than on the medium (700-900nm) range.

[0100] Additional Examples

[0101] The following embodiments are presented as examples. It should be understood that embodiments may include properties and features (e.g., components, parts, elements, components, and / or portions) described in other embodiments. Therefore, while certain features are described in these embodiments, various features of a given embodiment may be combined with and / or incorporated into other embodiments of this disclosure. Therefore, the disclosure of certain features of specific embodiments of this disclosure should not be construed as limiting the application or inclusion of said features to the specific embodiments. Rather, it should be understood that other embodiments may also include these features.

[0102] Example 1: A tissue characterization probe, comprising: an elongated member having a proximal end and a distal end; a plurality of distal probe tips disposed at or near the distal end of the elongated member to form a multi-arm arrangement; a plurality of illumination optical fibers extending at least partially through the elongated member, each illumination optical fiber extending to a corresponding probe tip of the multi-arm arrangement such that each probe tip includes at least one illumination optical fiber; and a plurality of detection optical fibers extending at least partially through the elongated member, each detection optical fiber extending to a corresponding probe tip of the multi-arm arrangement such that each probe tip includes at least one detection optical fiber.

[0103] Example 2: The tissue characterization probe according to Example 1, wherein the probe is configured to be introduced through the lumen or working channel of a catheter, a guideable catheter, a steerable sheath, or an endoscope, and extends beyond the tip of the catheter, sheath, or working channel.

[0104] Example 3: A tissue characterization system, comprising: a tissue characterization probe as described in Example 1 or Example 2; a light source operatively coupled to an illumination fiber; and one or more spectrometers operatively coupled to a detection fiber.

[0105] Example 4: The tissue characterization system according to Example 3, wherein the system is configured to characterize tissues at depths greater than about 100 μm, such as up to about 1 mm, 1.5 mm, 2 mm, 2.5 mm, 3 mm, 3.5 mm, 4 mm, 5 mm, 7.5 mm, 10 mm, 15 mm, 20 mm, 25 mm or 30 mm.

[0106] Example 5: A tissue characterization system according to Example 3 or Example 4, wherein the system is configured to characterize structurally anisotropic tissues, such as cardiac tissue, in a manner that reduces the influence of probe tip rotation on the measured spectrum.

[0107] Example 6: A tissue characterization probe, comprising: an elongated member having a proximal end and a distal end; a distal probe tip disposed at the distal end of the elongated member; an illumination fiber extending at least partially through the elongated member to the probe tip and configured to define an illumination axis; and a plurality of detection fibers extending at least partially through the elongated member to the probe tip and configured to receive light scattered from a target tissue, wherein a first set of detection fibers is disposed substantially along a first detection line orthogonal to the illumination axis, and wherein a second set of detection fibers is disposed substantially along a second detection line orthogonal to the illumination axis and transverse to the first detection line.

[0108] Example 7: The tissue characterization probe according to Example 6, wherein the probe is configured to be introduced through the lumen or working channel of a catheter, a guideable catheter, a steerable sheath, or an endoscope, and extends beyond the tip of the catheter, sheath, or working channel.

[0109] Example 8: The tissue characterization probe according to Example 6 or Example 7, wherein each detection fiber is radially offset from its corresponding detection line by no more than about 30 degrees, or no more than about 25 degrees, or no more than about 20 degrees, or no more than about 15 degrees, or no more than about 10 degrees, or no more than about 5 degrees.

[0110] Example 9: A tissue characterization probe according to any one of Examples 6-8, wherein the first detection line and the second detection line intersect each other at the illumination axis to form a lateral angle of about 30 to about 150, or about 45 to about 135, or about 60 to about 120, or about 75 to about 105.

[0111] Example 10: A tissue characterization probe according to any one of Examples 6-9, wherein the second detection line is orthogonal to the first detection line.

[0112] Example 11: The tissue characterization probe according to any one of Examples 6-10, wherein the first set of detection optical fibers includes at least two detection optical fibers.

[0113] Example 12: The tissue characterization probe according to any one of Examples 6-11, wherein the second set of detection optical fibers includes at least two detection optical fibers.

[0114] Example 13: A tissue characterization probe according to any one of Examples 6-12, wherein the first set of detection optical fibers, the second set of detection optical fibers, or both are substantially adjacent to the illumination optical fibers.

[0115] Example 14: The tissue characterization probe according to any one of Examples 6-12, wherein the detection optical fiber and the illumination optical fiber are spaced apart.

[0116] Example 15: The tissue characterization probe according to Example 14, wherein the detection optical fibers are spaced apart from each other.

[0117] Example 16: The tissue characterization probe according to any one of Examples 6-15 further includes a support wire that extends at least partially through an elongated member to the probe tip and is configured to increase the bending stiffness of the probe tip.

[0118] Example 17: The tissue characterization probe according to any one of Examples 6-15 further includes a support wire that extends at least partially through an elongated member to the probe tip and is configured to form a bend at the distal end of the probe tip.

[0119] Example 18: A tissue characterization probe according to Example 16 or Example 17, wherein the support wire has a quadrilateral cross-sectional shape.

[0120] Example 19: The tissue characterization probe according to Example 18, wherein the support wire has a rectangular cross-sectional shape.

[0121] Example 20: A tissue characterization probe according to any one of Examples 6-19, wherein the tissue characterization probe or a plurality of such tissue characterization probes are incorporated into a multi-arm tissue characterization probe as described in Example 1 or Example 2.

[0122] Example 21: A tissue characterization probe according to any one of Examples 1-20, wherein the probe further comprises one or more of the following imaging components configured to provide microstructural imaging of the target tissue; a positioning component configured to provide positional information of the distal tip in a three-dimensional anatomical workspace; and / or a treatment component having one or more treatment parts disposed at the distal tip for treating the target tissue.

[0123] Example 22: The tissue characterization probe according to Example 21, wherein the positioning component includes one or more electrodes, magnetic, optical or other positioning components to provide means for positioning the distal tip.

[0124] Example 23: A tissue characterization system comprising: a tissue characterization probe according to any one of Examples 6-22; a light source operatively coupled to an illumination fiber; and one or more spectrometers operatively coupled to a detection fiber.

[0125] Example 24: The tissue characterization system according to Example 23, wherein the probe is configured to characterize tissue at depths greater than about 100 μm, such as up to about 1 mm, 1.5 mm, 2 mm, 2.5 mm, 3 mm, 3.5 mm, 4 mm, 5 mm, 7.5 mm, 10 mm, 15 mm, 20 mm, 25 mm or 30 mm.

[0126] Example 25: A method for characterizing tissue, comprising: providing a tissue characterization system; guiding a distal probe tip to a target anatomical location; at the target anatomical location, manipulating the tissue characterization probe to obtain a spectrum at a depth greater than about 100 μm, such as up to about 1 mm, 1.5 mm, 2 mm, 2.5 mm, 3 mm, 3.5 mm, 4 mm, 5 mm, 7.5 mm, 10 mm, 15 mm, 20 mm, 25 mm, or 30 mm; and resolving the spectrum to characterize the target tissue.

[0127] Example 26: The method according to Example 25, wherein the target tissue is heart tissue.

[0128] Example 27: The method described in Example 26, wherein the target anatomical location is a blood-filled, beating heart.

[0129] Example 28: The method according to any one of Examples 25-27, wherein the tissue characterization system is the system as described in any one of Examples 3-5 or 23-24.

[0130] Example 29: The method according to any one of Examples 25-28, wherein characterizing the target tissue includes detecting, measuring or monitoring one or more of fibrosis, allogeneic transplantation acceptance or rejection, myocarditis, amyloidosis, hypertrophy or nuclear density.

[0131] Example 30: The method according to any one of Examples 25-29, wherein characterizing the target tissue includes determining the volume fraction of the components of the target tissue within the heart and / or the spatial distribution of the components of the target tissue.

[0132] Example 31: The method according to any one of Examples 25-30 further includes: characterizing tissue at multiple target locations and obtaining one or more data points characterizing the tissue; determining the position of the probe tip in a three-dimensional anatomical workspace at each data acquisition location; associating each data point with its corresponding determined position in the anatomical workspace; and generating a three-dimensional map of the tissue microstructure based on the data points and their corresponding positions.

[0133] Example 32: The method according to Example 31, wherein the three-dimensional map is a fiber map.

[0134] Example 33: The method according to Example 31 or Example 32, wherein the step of characterizing tissues at multiple target locations includes simultaneously characterizing tissues at multiple target locations.

[0135] Example 34: The method according to any one of Examples 25-33, wherein the step of resolving the spectrum to characterize the target tissue includes utilizing unsupervised machine learning techniques.

[0136] Example 35: The method according to Example 34, wherein the unsupervised machine learning technique includes cluster analysis.

[0137] Example 36: The method according to Example 35, wherein the cluster analysis includes dimensionality reduction of the spectrum by principal component analysis (PCA).

[0138] Example 37: The method according to Example 36, wherein the cluster analysis includes measuring the similarity of at least a first principal component and a second principal component from the spectrum.

[0139] Example 38: The method according to Example 37, wherein the similarity of at least the first principal component and the second principal component from the spectrum is based on Euclidean distance.

[0140] Example 39: The method according to any one of Examples 25-33, wherein the step of resolving the spectrum to characterize the target tissue includes utilizing supervised machine learning techniques.

[0141] Example 40: The method according to Example 39, wherein the supervised machine learning technique includes a convolutional neural network (CNN).

[0142] Example 41: The method according to Example 40, wherein a set of previous measurements from scattering in the tissue is used to train and test the CNN.

[0143] Example 42: The method according to Example 40 or Example 41, wherein the batch size for training the CNN is varied from 50 to 300.

[0144] Example 43: The method according to any one of Examples 40-42, wherein the learning rate of the CNN varies from 0.0001 to 0.015.

[0145] Example 44: The method according to any one of Examples 40-43, wherein the kernel size of the convolutional filter of the CNN varies from 5 to 40.

[0146] Example 45: The method according to any one of Examples 40-44, wherein the CNN is trained from a spectrum within a specific wavelength range and / or trained using reduced sampling.

[0147] Summarize

[0148] While certain embodiments of this disclosure have been described in detail with reference to specific constructions, parameters, components, elements, etc., these descriptions are illustrative and should not be construed as limiting the scope of the claimed invention. Furthermore, it should be understood that for any given element or component in the described embodiments, any possible alternatives listed for that element or component may generally be used alone or in combination with each other, unless implied or expressly stated otherwise.

[0149] Furthermore, unless otherwise stated, all figures used in the specification and claims to represent quantities, components, distances, or other measurements are to be understood as potentially modified by the term "about" or its synonyms. When the terms "about," "approximately," "substantially," etc., are used with a specified quantity, value, or condition, they are to be understood as deviations from the specified quantity, value, or condition less than 20%, less than 10%, less than 5%, or less than 1%. Without at least attempting to limit the application of the doctrine of equivalence to the claims, each numerical parameter shall be interpreted according to the reported number of significant digits and by applying general rounding techniques.

[0150] Any headings and subheadings used herein are for organizational purposes only and are not intended to limit the scope of the description or claims. It should also be noted that, as used in this specification and the appended claims, the singular forms “a,” “an,” and “the” do not exclude plural indicators unless the context otherwise indicates. Thus, for example, a reference to a singular indicator (e.g., “part”) may also include two or more such indicators.

Claims

1. A tissue characterization probe, comprising: an elongated member having a proximal end and a distal end; a plurality of distal probe tips disposed at or near the distal end of the elongated member to form a multi-arm arrangement; a plurality of illumination optical fibers extending at least partially through the elongated member, each illumination optical fiber extending to a respective probe tip of the multi-arm arrangement such that each probe tip includes at least one illumination optical fiber; as well as A plurality of detection optical fibers extend at least partially through the elongated member, each detection optical fiber extending to a respective probe tip of the multi-arm arrangement such that each probe tip includes at least one detection optical fiber.

2. The tissue characterization probe according to claim 1, wherein: The probe is configured to be introduced through a navigable catheter, a steerable sheath, a lumen or working channel of a catheter, or a working channel of an endoscope and to extend beyond the tip of the catheter, sheath, or working channel.

3. A tissue characterization system, comprising: The tissue characterization probe according to claim 1 or claim 2; a light source operatively coupled to the illumination fiber; as well as One or more spectrometers are operatively coupled to the detection fiber.

4. The tissue characterization system according to claim 3, wherein: The system is configured to characterize tissue within a depth of greater than about 100 μm, such as up to about 1 mm, or up to about 1.5 mm, or up to about 2 mm, or up to about 2.5 mm, or up to about 3 mm, or up to about 3.5 mm, or up to about 4 mm, or up to about 5 mm, or up to about 7.5 mm, or up to about 10 mm, or up to about 15 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm.

5. The tissue characterization system according to claim 3, wherein: The system is configured to characterize structurally anisotropic tissue, such as cardiac tissue, optionally in a manner that reduces the effect of rotation of the probe tip on the measured spectrum.

6. A tissue characterization probe, comprising: an elongated member having a proximal end and a distal end; a distal probe tip disposed at the distal end of the elongated member; an illumination fiber extending at least partially through the elongated member to the probe tip and configured such that the illumination fiber defines an illumination axis, wherein the illumination fiber is offset from a longitudinal axis of the elongated member; and a plurality of detection optical fibers extending at least partially through the elongated member to the probe tip and configured to receive light scattered from the target tissue, wherein the first group of detection optical fibers is substantially arranged along a first detection line, the first detection line being orthogonal to the illumination axis, and The second group of detection optical fibers is substantially arranged along a second detection line, which is orthogonal to the illumination axis and transverse to the first detection line.

7. The tissue characterization probe according to claim 6, wherein: The probe is configured to be introduced through a navigable catheter, a steerable sheath, a lumen or working channel of a catheter, or a working channel of an endoscope and to extend beyond the tip of the catheter, sheath, or working channel.

8. The tissue characterization probe according to claim 6, wherein: Each detection fiber is radially offset from its corresponding detection line by no more than about 30 degrees, or no more than about 25 degrees, or no more than about 20 degrees, or no more than about 15 degrees, or no more than about 10 degrees, or no more than about 5 degrees.

9. The tissue characterization probe according to claim 6, wherein: The first detection line and the second detection line cross each other at the illumination axis to form a lateral angle of about 30° to about 150°, or about 45° to about 135°, or about 60° to about 120°, or about 75° to about 105°.

10. The tissue characterization probe according to claim 6, wherein: The second detection line is orthogonal to the first detection line.

11. The tissue characterization probe according to claim 6, wherein: The first group of detection optical fibers includes at least two detection optical fibers.

12. The tissue characterization probe according to claim 6, wherein: The second group of detection optical fibers includes at least two detection optical fibers.

13. The tissue characterization probe according to claim 6, wherein: The first set of detection optical fibers, the second set of detection optical fibers, or both are substantially adjacent to the illumination optical fibers.

14. The tissue characterization probe according to claim 6, wherein: The detection optical fiber is spaced apart from the illumination optical fiber.

15. The tissue characterization probe according to claim 14, wherein: The detection optical fibers are spaced apart from each other. 16 . The tissue characterization probe of claim 6 , further comprising a support wire extending at least partially through the elongated member to the probe tip and configured to increase bending stiffness of the probe tip.

17. The tissue characterization probe of claim 6, further comprising a support wire extending at least partially through the elongated member to the probe tip and configured to form a bend in a distal end of the probe tip.

18. The tissue characterization probe according to claim 16, wherein: The supporting wire has a quadrilateral cross-sectional shape.

19. The tissue characterization probe according to claim 18, wherein: The support wire has a rectangular cross-sectional shape.

20. The tissue characterization probe according to claim 6, wherein: The tissue characterization probe, or a plurality of such tissue characterization probes, is incorporated into a tissue characterization probe according to claim 1 or claim 2.

21. The tissue characterization probe according to any one of claims 1 to 2 or 6 to 20, wherein: The probe further includes one or more of the following imaging components, configured to provide microstructural imaging of target tissue; a positioning component, configured to provide position information of the distal tip within a three-dimensional anatomical workspace; and / or a treatment assembly having one or more treatment components disposed at the distal tip for treating a target tissue.

22. The tissue characterization probe according to claim 21, wherein: The positioning assembly includes one or more electrodes, magnetic, optical or other positioning components to provide a means for positioning the distal tip.

23. A tissue characterization system, comprising: The tissue characterization probe according to any one of claims 6 to 20; a light source operatively coupled to the illumination fiber; as well as One or more spectrometers are operatively coupled to the detection fiber.

24. The tissue characterization system of claim 23, wherein: The probe is configured to characterize tissue within a depth of greater than about 100 μm, such as up to about 1 mm, or up to about 1.5 mm, or up to about 2 mm, or up to about 2.5 mm, or up to about 3 mm, or up to about 3.5 mm, or up to about 4 mm, or up to about 5 mm, or up to about 7.5 mm, or up to about 10 mm, or up to about 15 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm.

25. A method of characterizing a tissue, comprising: Providing a tissue characterization system according to claim 23 or 24; guiding the distal probe tip to a target anatomical location; operating the tissue characterization probe to obtain a spectrum at a depth greater than about 100 μm, such as up to about 1 mm, or up to about 1.5 mm, or up to about 2 mm, or up to about 2.5 mm, or up to about 3 mm, or up to about 3.5 mm, or up to about 4 mm, or up to about 5 mm, or up to about 7.5 mm, or up to about 10 mm, or up to about 15 mm, or up to about 20 mm, or up to about 25 mm, or up to about 30 mm, at the target anatomical location; and The spectrum is interpreted to characterize the target tissue.

26. The method according to claim 25, wherein: The target tissue is cardiac tissue.

27. The method according to claim 26, wherein: The target anatomical location is a beating heart filled with blood.

28. A method according to any one of claims 25 to 27, wherein: The tissue characterization system is a system according to any one of claims 3 to 5 or 23 to 24.

29. The method according to claim 25, wherein: Characterizing the target tissue includes detecting, measuring or monitoring one or more of fibrosis, allograft acceptance or rejection, myocarditis, amyloidosis, hypertrophy or nuclear density.

30. The method of claim 25, wherein: Characterizing the target tissue includes determining a spatial distribution of a component of the target tissue and / or a volume fraction of a component of the target tissue within the heart.

31. The method of claim 25, further comprising: characterizing tissue at a plurality of target locations and obtaining one or more data points characterizing the tissue; At each location of data acquisition, determining the position of the probe tip within the three-dimensional anatomical workspace; associating each data point with its corresponding determined location within the anatomical workspace; as well as Based on the data points and their corresponding positions, a three-dimensional map of the tissue microstructure is generated.

32. The method according to claim 31, wherein: The three-dimensional map is a fibrosis map.

33. The method according to claim 31, wherein: The step of characterizing tissue at a plurality of target locations includes simultaneously characterizing tissue at the plurality of target locations.

34. The method of claim 25, wherein: The step of interpreting the spectra to characterize the target tissue involves utilizing unsupervised machine learning techniques.

35. The method of claim 34, wherein: The unsupervised machine learning techniques include cluster analysis.

36. The method of claim 35, wherein: The cluster analysis includes reducing the dimensionality of the spectrum by principal component analysis (PCA).

37. The method of claim 36, wherein: The cluster analysis includes measuring the similarity of at least a first principal component and a second principal component from the spectra.

38. The method of claim 37, wherein: The similarity of at least the first principal component and the second principal component from the spectra is based on Euclidean distance.

39. The method of claim 25, wherein: The step of interpreting the spectrum to characterize the target tissue includes utilizing supervised machine learning techniques.

40. The method of claim 39, wherein: The supervised machine learning technique includes convolutional neural network (CNN).

41. The method of claim 40, wherein: The CNN was trained and tested using data from a set of previous measurements of scattering in tissue.

42. The method of claim 40, wherein: The batch size used to train the CNN was varied from 50 to 300.

43. The method of claim 40, wherein: The learning rate of the CNN was varied from 0.0001 to 0.

015.

44. The method of claim 40, wherein: The convolution filter kernel size of the CNN varies from 5 to 40.

45. The method of claim 40, wherein: The CNN is trained from spectra within a specific wavelength range and / or trained using reduced sampling.