(19)
(11)EP 2 950 708 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
16.01.2019 Bulletin 2019/03

(21)Application number: 14707449.6

(22)Date of filing:  22.01.2014
(51)International Patent Classification (IPC): 
A61B 5/00(2006.01)
G06T 11/00(2006.01)
G01J 3/28(2006.01)
(86)International application number:
PCT/IB2014/058455
(87)International publication number:
WO 2014/118674 (07.08.2014 Gazette  2014/32)

(54)

IMAGING SYSTEM WITH HYPERSPECTRAL CAMERA GUIDED PROBE

BILDGEBUNGSSYSTEM MIT HYPERSPEKTRALER KAMERAGEFÜHRTER SONDE

SYSTÈME D'IMAGERIE AVEC SONDE GUIDÉE PAR CAMÉRA HYPERSPECTRALE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 30.01.2013 US 201361758312 P

(43)Date of publication of application:
09.12.2015 Bulletin 2015/50

(73)Proprietor: Koninklijke Philips N.V.
5656 AE Eindhoven (NL)

(72)Inventors:
  • HENDRIKS, Bernardus Hendrikus Wilhelmus
    NL-5656 AE Eindhoven (NL)
  • SCHLEIPEN, Johannes Joseph Hubertina Barbara
    NL-5656 AE Eindhoven (NL)
  • REICH, Christian
    NL-5656 AE Eindhoven (NL)
  • LUCASSEN, Gerhardus Wilhelmus
    NL-5656 AE Eindhoven (NL)
  • MÜLLER, Manfred
    NL-5656 AE Eindhoven (NL)
  • PULLY, Vishnu Vardhan
    NL-5656 AE Eindhoven (NL)

(74)Representative: Cohen, Julius Simon 
Philips Intellectual Property & Standards High Tech Campus 5
5656 AE Eindhoven
5656 AE Eindhoven (NL)


(56)References cited: : 
US-A1- 2007 024 946
US-A1- 2013 003 064
US-A1- 2009 326 383
  
      
    Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


    Description

    FIELD OF THE INVENTION



    [0001] The present invention relates to the field of medical devices or equipment. More specifically, the invention relates to a medical imaging system and a method for identifying a target structure in biological tissue, such as for effective tumor margin assessment.

    BACKGROUND OF THE INVENTION



    [0002] During tumor resection surgery the evaluations of the boundaries of the resected tissue whether they are tumor free is of vital importance. When the tumor is not complete resected the treatment will not be curative with all devastating consequences for the patients. Although histopathology is the gold standard for evaluating whether the resection was complete, this can only be done after the surgery has been completed because of the time required for the evaluation of the resected specimen. Although surface probes are under development that can investigate the outer layer of the resected tissue to a certain depth for positive margins, these methods are still rather time consuming because the exterior surface has to be investigated with sufficient accuracy and with sufficient resolution. Furthermore, since these measurements are point like (or small area measurements) they are prone to fluctuations because of heterogeneous tissue structure that may render these measurements less stable.

    [0003] The problem is thus how to inspect the surface of a resected tumor tissue for positive margins in a time efficient and a robust way.

    [0004] US 2009/326383 A1 describes a hyperspectral image system for identifying a suspect region of a subject. An optional contact probe module is used to collect a signal of the suspect region for medical diagnosis.

    [0005] US 2007/024946 A1 describes a hyperspectral imaging system for providing information physiologic measurements that support early detection of shock and outcomes therefrom.

    SUMMARY OF THE INVENTION



    [0006] It would be advantageous to provide a system and method allowing a stable and at the same time efficient identification of a target structure, e.g. tumor tissue, within a biological tissue.

    [0007] Such a system of the invention is defined in claim 1 and a method of the invention is defined in claim 15. In a first aspect, the invention provides a medical imaging system arranged to identify a target structure in an associated biological tissue, the system comprising
    • a hyperspectral camera system arranged to provide a hyperspectral image covering a first surface area of the biological tissue, wherein pixels of the hyperspectral image contains information at different light wavelength bands,
    • a probe arranged to provide a probe measurement of a second surface area of the biological tissue with information indicative of the target structure, and to generate probe measurement data accordingly, wherein the second surface area is smaller than the first surface area, and wherein the probe is selected to provide a higher specificity with respect to identification of the target structure than the hyperspectral camera system, and
    • a processor functionally connected to receive the hyperspectral image and the probe measurement data, wherein the processor is arranged to analyze the hyperspectral image according to a hyperspectral processing algorithm in order to identify a part of the hyperspectral image which possibly contains the target structure, wherein the processor is further arranged to analyze the probe measurement data and to calibrate the hyperspectral processing algorithm in response thereto, so as to provide a calibrated hyperspectral processing algorithm with an enhanced sensitivity to identify the target structure in the hyperspectral image.


    [0008] The invention solves the mentioned problem in that the hyperspectral image can be used to provide an image of a sufficiently large area very quickly, but with a rather limited spectral resolution, and thus a rather low specificity regarding identification of the target structure within the biological tissue. However, still the hyperspectral image can be used to indentify suspicious areas, i.e. areas where it can be expected that there is a certain possibility that the target structure is present. The hyperspectral image can then be used by the operator, e.g. by means of a visual indication on a display, as a guide towards selecting which location(s) within the first surface area covered by the hyperspectral image to be measured by a probe selected such that it provides a higher specificity with respect to identification of the target structure. Thus, the probe provides better data for identifying the target structure, e.g. by providing a higher spectral resolution than the hyperspectral image, but at the price of a limited measurement area, e.g. in the form of point or point-like measurements. Using the probe measurement data to modify or calibrate the hyperspectral image processing algorithm, an improved image with enhanced contrast and/or sensitivity with respect to identify the target structure, e.g. tumor tissue, can be obtained. After one or a limited number of probe measurements have been performed, and the hyperspectral image processing algorithm has been calibrated accordingly, the real-time hyperspectral images processed with the calibrated algorithm can be used to provide very quick inspection of further target structures.

    [0009] Thus, the high speed of the hyperspectral camera is used to point out suspicious area to be examined with the more time consuming probe. However, it turns out that even one single point-like probe measurement can be used to significantly improve the hyperspectral image towards increasing sensitivity to identify the target structure. In this way, the medical imaging system according to the invention can provide a robust and time efficient identification of positive margins after resection. This allows the system to be used during surgery and thus increases the chance of removing all tumor tissue during oncological surgery without the need to redo the surgery, if it turns out at a later examination that not all tumor tissue had been successfully removed. It is to be understood, that the type of probe and subsequent processing of the probe measurement data should preferably be selected towards providing a high sensitivity towards indentifying the specific target structure in question.

    [0010] In the case of identifying a tumor, the hyperspectral camera system can be used to make visible images and hyperspectral images of the surface of the tissue specimen, a contact probe or surface probe capable of measuring locally inside the tissue near the surface of the specimen. The hyperspectral images are analyzed for suspicious areas based on look-up-table data and capable of analyzing the probe for suspicious signatures. The probe is capable of producing a signal indicative whether the probed surface contains a positive margin, and the area to be inspected with the surface probe is indicated in the visible image based on the hyperspectral image analysis. E.g. signals from the surface probe, taken at a few locations, are used to calibrate the 2D-image processing algorithms used for analyzing the hyperspectral images. The probe data may especially be acquired with a larger spectral range, with higher spectral resolution, and may also cover part of the tissue below the surface, hence allowing an accurate assessment of these tissue points. The idea is that these surface probe measurements needs to be done only a few times during an intervention (e.g. when the 2D-spectral image processing indicates a large reconstruction uncertainty), and that once this calibration with the surface probe data has been performed and the system is re-calibrated, the operator can use the real-time 2D results obtained from the hyperspectral images according to the invention. Especially, the signals of the surface probe measurements are annotated in the visible image and or hyperspectral image based on image analysis of the visible image to register the probe tip location with the specimen at the time of measurement.

    [0011] Apart from identifying tumor tissue, the imaging system can also be used to identify other target structure, e.g. the presence of blood vessels, nerve bundles and the like below the surface of biological tissue, e.g. during surgery. Hereby it is possible to avoid an accidental section of such blood vessels and nerve bundles within the need for a time consuming and/or interventional probes.

    [0012] In a preferred embodiment, the probe measurement is acquired with a higher spectral resolution than the hyperspectral image. Especially, the hyperspectral image may be acquired with overlapping spectral bands of a rather low spectral resolution, or the spectral bands can be non-overlapping and thus have spectral bands without any information. Thus, in both cases, the probe may provide measurement data with a higher spectral resolution which thus allows a higher specificity with respect to identify the target structure within the biological tissue.

    [0013] It is preferred that the second surface area is selected to be located within the first surface area, and thus that the probe measurement is performed within the area covered by the hyperspectral image. In case the probe is an interventional device, it is to be understood that by "within the first surface area" is understood that the probe measurement is performed in a point or small area in the biological tissue below the surface of the first surface area.

    [0014] The system may be arranged to visually display the hyperspectral image after being processed with the hyperspectral processing algorithm together with a visual indication of the part which is. E.g. special shading, coloring, highlighting, or encircling can be used to graphically indicate the part or parts which are identified as possibly containing the target structure.

    [0015] The system is preferably arranged to visually display the hyperspectral image after being processed with the calibrated hyperspectral processing algorithm. Hereby, the operator, e.g. a physician, can inspect the resulting hyperspectral image, which after processing by the calibrated algorithm is now enhanced with respect to better visualize the target structure.

    [0016] The processor may be arranged to analyze the probe measurement data according to a probe processing algorithm comprising comparing the probe measurement data with a look-up-table. Alternatively or additionally, the processor may be arranged to analyze the probe measurement data according to a probe processing algorithm comprising calculating a plurality of measurement parameters and translating the plurality of measurement parameters into physiological parameters indicative of the target structure.

    [0017] The processor may be arranged to calibrate the hyperspectral processing algorithm by refining a grouping of pixels in the hyperspectral image which are identified in the probe measurement data as belonging to the target structure. Such improved grouping of pixels in the hyperspectral image will provide an enhanced visualization of the target structure. E.g. the grouping of pixels assigned with the same color or brightness may be changed or tuned based on parameters extracted from the probe measurement data.

    [0018] The probe may comprise a surface probe, and/or an interventional device. The probe may comprise an optical probe, and/or an ultrasound probe. The selection of which probe to use depends on the type of biological tissue and the target structure to be identified therein. Especially the optical probe may comprise one of: a diffuse reflectance spectroscopy probe, a diffuse optical tomography probe, a differential path length spectroscopy, a fluorescence probe, a Raman spectroscopy probe, a confocal reflectance probe, a confocal fluorescence probe, a two-photon fluorescence probe, an optical coherence tomography probe, and a scanning fiber microscope arranged to perform confocal imaging.

    [0019] Preferably, the hyperspectral camera system is arranged to provide two-dimensional real-time images, thus allowing real-time navigation of the probe tip if the hyperspectral images are display on a monitor to the probe operator. The hyperspectral camera system may in general be based on one of: (1) a dedicated hyperspectral camera which, based on light from a broadband light source, has wavelength separation inside the camera, (2) a light source with tunable wavelength or tunable filter set and a normal camera, and (3) a broadband light source and a tunable filter in front of normal camera.

    [0020] The hyperspectral cameras, i.e. type (1), are typically expensive, while type (2) and (3) are based on a normal camera. However, they suffer from other advantages. E.g. the solution with illumination with different wavelength lights which requires the environments to be dark to produce high quality images. See "An introduction to hyperspectral imaging and its application for security, surveillance and target acquisition", P. W.T. Yuen, M. Richardson, The Imaging Science Journal, Vol. 58, 2010, p. 241-253 regarding hyperspectral camera systems. Thus, in one embodiment, the hyperspectral camera system comprises one of:
    • a hyperspectral camera arranged to provide the hyperspectral image,
    • a normal camera and a plurality of light sources with different narrow band wavelengths and a controller arranged to provide images with the biological tissue sequentially illuminated by the plurality of light sources, and
    • a normal camera and a tunable optical filter arranged between the normal camera and the biological tissue, so as to provide sequential images with different settings of the tunable optical filter.


    [0021] The processor may be arranged to process the hyperspectral image according to an image processing algorithm in order to identify a tip of the probe in the hyperspectral image, so as to identify to which part of the hyperspectral image the probe measurement data corresponds. In this embodiment, it is possible to automatically identify to which part of the hyperspectral image a given probe measurement applies, which is preferred in order to perform the calibration of the hyperspectral processing algorithm in a fast and reliable manner compared to other types of location means for determining the location of the probe during measurement.

    [0022] As already mentioned, the imaging system can be used for identification of different target structures, such as: tumor tissue, a nerve bundle, a blood vessel (artery or vein), lymph nodes, and oxygenation level of tissue.

    [0023] It is to be understood that the processor is preferably controlled by a control algorithm which can be implemented as a software program arranged for execution on a suitable processor system or computer. The processor of the system may be implemented in a dedicated device, e.g. including a display to display the hyperspectral images, or the processor of the system may be implemented as a software program arranged to execution on a server or a general purpose computer.

    [0024] The size and shape of the interventional device and/or surface probe, such as thickness and number of the optical fibers in case of an optical probe, is understood to be selected for a given application.

    [0025] In a second aspect, the invention provides a method for detecting a target structure in an associated biological tissue, the method comprising
    • providing a hyperspectral image covering a first surface area of the associated biological tissue, wherein pixels of the hyperspectral image contains information regarding different light wavelength bands,
    • providing probe measurement data in accordance with a probe measurement of a second surface area of the biological tissue with information indicative of the target structure, wherein the second surface area is smaller than the first surface area and located within the first surface area, and wherein the probe measurement is selected to provide a higher specificity with respect to identification of the target structure than the hyperspectral camera system,
    • analyzing the hyperspectral image according to a hyperspectral processing algorithm in order to identify a part of the hyperspectral image which may contain the target structure,
    • analyzing the probe measurement data, and
    • calibrating the hyperspectral processing algorithm in response to a result of the probe measurement data analysis, so as to provide a calibrated hyperspectral processing algorithm with an enhanced sensitivity to identify the target tissue structure in the hyperspectral image.


    [0026] It is understood that "providing" hyperspectral image and "providing" probe measurement data covers a version of the method where "providing" means merely receiving data, e.g. in digital format, e.g. stored or transmitted data. However, in another version of the method, "providing" covers actually taking the hyperspectral image of the biological tissue, and also performing the probe measurement on the biological tissue.

    [0027] It is appreciated that the same advantages and embodiments of the first aspect apply as well for the second aspect. In general the first and second aspects may be combined and coupled in any way possible within the scope of the invention. These and other aspects, features and/or advantages of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

    [0028] It is to be understood that the system according to the first aspect and the method according to the second aspect can be used in a number of applications, such as: identification of a positive margin after oncological surgery, identification of blood vessels (artery or vein) below a tissue surface, identification of a nerve bundle below a tissue surface, lymph nodes, and oxygenation level of tissue.

    BRIEF DESCRIPTION OF THE DRAWINGS



    [0029] Embodiments of the invention will be described, by way of example only, with reference to the drawings, in which

    FIG. 1 illustrates a simple block diagram of an imaging system embodiment,

    FIG.2 illustrates a sketch of an optical probe tip suitable for the invention,

    FIG. 3 illustrates a hyperspectral camera system embodiment based on a normal camera and with different narrow band wavelength light sources.

    FIG. 4a illustrates a visualization of a hyperspectral image with indications of suspicious areas where tumor tissue may be present,

    FIG. 4b illustrates the image of FIG. 4a after the hyperspectral processing algorithm has been calibrated, thus revealing clearly tumor as a white shading,

    FIGs. 5a-5c illustrate different strategies for selecting location of point measurement with a probe,

    FIG. 6 illustrates a block diagram of a diagnostic embodiment where a tumor indication is shown on a display together with the hyperspectral image after calibration,

    FIG. 7 illustrates a graph showing absorption spectra for blood, water and fat,

    FIG. 8 illustrates a graph showing intrinsic fluorescence curves for A: collagen, B: elastin, C: NADH, and D: FAD, and

    FIG. 9 illustrates a block diagram of a method embodiment.


    DESCRIPTION OF EMBODIMENTS



    [0030] FIG. 1 illustrates basic parts of a system embodiment with a hyperspectral camera HSC real-time 2D hyperspectral images HSI which are applied to a processor P and processed in a hyperspectral processing algorithm HSA. The hyperspectral camera HSC provides a hyperspectral image of a first area A1 of a biological tissue BT to be tested, e.g. a part of a human body to be tested with respect to the presence of a target structure in the form or tumor tissue TS. The hyperspectral image after processing I1 is then displayed in color or grey shadings on a display D, e.g. with a visual indicators graphically highlighting suspicious parts of the image which are identified as possible for containing tumor tissue TS.

    [0031] An optical surface probe PR provides probe measurement data PMD of a single measurement covering a second area A2 located within the first area A1 covered by the camera HSC. The probe measurement data PMD are applied to the processor and processed by a probe measurement algorithm PMA. This algorithm PMA may include extracting one or more physiological parameters PP based on the probe measurement data PMD. The probe is selected such that it provides measurement data PMD with a higher specificity with respect to identifying the tumor tissue TS, e.g. by providing a higher spectral resolution than the hyperspectral camera HSC. Thus, even though the probe PR covers a limited area A2, e.g. a single point, it provides probe measurement data PMD including information that allows identification of the tumor tissue TS with a higher certainty than possible based on the hyperspectral image, e.g. due to a measurement technique allowing information further below the tissue surface than it is possible with a hyperspectral image.

    [0032] The physiological parameters PP extracted from the probe measurement data PMD are used to calibrate the hyperspectral processing algorithm HSA, e.g. by refining the grouping of image pixels in the hyperspectral image in order to better visually enhance physiological structures of the tumor tissue in the hyperspectral image. This results in a calibrated hyperspectral processing algorithm HSA, and the subsequently processed hyperspectral images 12 are then displayed on the display D, thus resulting in an image where tumor tissue TS relevant structure are visually enhanced, thus allowing the operator to more clearly identify tumor tissue TS. However, the hyperspectral image allows identification of parts of the area A1 covered where there is a possibility that tumor tissue is present, and this can be used as a guide for an operator towards selecting the point or area A2 to perform probe measurements, thus saving time consuming probe measurements covering the whole area A1. Especially, probes PR providing point measurements can be used which allow a very high specificity with respect to identification of tumor tissue TS, thus providing a highly reliable identification of tumor tissue TS within the first area A1 at a high speed.

    [0033] The probe can be a surface probe or an interventional device. In case the probe is an optical probe, optical fibers of the probe are preferably connected to an optical console. In a specific embodiment, the optical detector can resolve light with a wavelength substantially in the visible and infrared regions of the wavelength spectrum, such as from 400 nm to 1700 nm and it may allow for diffuse reflectance measurements in both the visible and the near infrared wavelength range. For a detailed discussion on diffuse reflectance spectrum (DRS) measurements see R. Nachabe, B.H.W. Hendriks, M.V.D. Voort, A. E, and H.J.C.M. Sterenborg, "Estimation of biological chromophores using diffuse optical spectroscopy : benefit of extending the UV-VIS wavelength range to include 1000 to 1600 nm," Optics Express, vol. 18, 2010, pp. 879-888, and R. Nachabe, B.H.W. Hendriks, A.E. Desjardins, M. van der Voort, M.B. van der Mark, and H.J.C.M. Sterenborg, "Estimation of lipid and water concentrations in scattering media with diffuse optical spectroscopy from 900 to 1600 nm," Journal of Biomedical Optics, vol. 15, May. 2010, pp. 037015-10.

    [0034] The information regarding the progress and whether the area under A1 under test contains tumor tissue can be provided in several ways by the processor P to the operator, e.g. a physician. For instance a light indicator or an indication on the display D can be used that when showing a red light still tumor tissue is detected and with green light no tumor and/or optionally with yellow light suspects tumor. Another way is using an audible sound signal. Yet another way of informing the operator of the result is to calculate a probability of tumor detection by means of a graphical indication on the display D, such as a bar and/or with numbers and/or colors.

    [0035] Since the hyperspectral camera HSC provides real-time images, it is possible to visually indicate the presence of the probe in the displayed image, thus allowing the operator to place the probe in the suspicious area and perform the probe measurements therein. The processor may then include an image processing algorithm allowing a detection of the probe tip, so as to automatically assign the probe measurement data PMD to the correct location within the hyperspectral image.

    [0036] In hyperspectral imaging a 2D-image of a distant object is taken, whereby for each pixel of the 2D-image the spectral content of the registered photons is being analyzed. This effectively results in a 3D-data set where the image intensity is stored as function of (X,Y)-coordinate and wavelength. Although originally developed for space borne and airborne applications (mineralogy, agriculture) requiring highly sensitive and expensive imaging systems, hyperspectral imaging cameras are now commercially available for the professional and consumer market, and are also being applied in e.g. healthcare and surveillance. Depending on the required field-of-view, frame-rate and the number and width of the spectral windows, different technologies are being used based on point-scanning or line-scanning devices and diffractive, refractive or interferometric spectral filtering. A typical example of a hyperspectral imaging camera is shown in Figure 1 illustrating two hyperspectral cameras from manufacturers Headwall and Horiba. A more detailed overview of hyperspectral imaging can be found in Shaw G.A.; Burke H.K., "Spectral Imaging for Remote Sensing", Lincoln Laboratory Journal 14 (1), 2003 , pp. 3-28.

    [0037] For healthcare applications, Medical Hyperspectral Imaging (MHSI) is typically used to provide two-dimensional tissue oxygenation maps. During imaging, reflected light can be collected at multiple wavelengths and processed to separate the oxy- and deoxyhemoglobin signals, and dedicated algorithms can be used to display data in an oxygenation map of tissue viability HyperMed Inc. "OxyVu Hyperspectral Imaging Technology." from www.hypermed-inc.com. Recently, hyperspectral imaging systems with dedicated algorithms for identifying areas with indications for the presence of tumor tissue have been introduced [Panasyuk, S. V., S. Yang, et al. (2007). "Medical hyperspectral imaging to facilitate residual tumor identification during surgery." Cancer Biology & Therapy 6(3): 439-446], and [Hamed Akbari, Kuniaki Uto, Yukio Kosugi, Kazuyuki Kojima and Naofumi Tanaka (2011), Cancer detection using infrared hyperspectral imaging, Cancer Science 102(4) 852-857]. These systems image surface areas either in the visible wavelength range from 400nm up to 800nm [Panasyuk, S. V., S. Yang, et al. (2007). "Medical hyperspectral imaging to facilitate residual tumor identification during surgery." Cancer Biology & Therapy 6(3): 439-446] or in the near-infrared range (1000-2500nm) [Hamed Akbari, Kuniaki Uto, Yukio Kosugi, Kazuyuki Kojima and Naofumi Tanaka (2011), Cancer detection using infrared hyperspectral imaging, Cancer Science 102(4) 852-857] and color the image according to a look-up table to identify suspicious spots.

    [0038] However, the sensitivity is often hampered by the presence of extravagating blood on the tissue surface, which commonly occurs in open surgery. The limited wavelength range and spectral resolution also limit the specificity, since the major distinguishing parameters are blood content and blood oxygenation, and tumor discriminators which are less prominent in the visible wavelength range (such as tissue scattering and absorption by bile, water, fat, and others) may not be properly considered.

    [0039] In some embodiments, the hyperspectral camera system can be arranged to obtain high spectral images using line scanning images, see e.g. [K. Hamada, K. Fujita, N.I. Smith, M. Kobayashi, and S. Kawata, "Raman microscopy for dynamic molecular imaging", J Biomed Optics. 13(4), 044027 (1-4) (2008)]. Here, the sample is illuminated using a white light source and the scattered/reflected photons after absorption is detected from the slit in front of the detector/camera. One direction the detector is used to capture the spectral information and the other direction is used to capture the lateral information along the line. By scanning the line along the sample, the spectral information along the second lateral direction is obtained. This provides with a huge hyperspectral data set illustrating a three dimensional spectral overview of the region of interest. By selecting a wavelength, respective images signifying different molecules/compounds can be obtained. This can further be used for performing ratio imaging of two different wavelengths. This modality enables faster imaging compared to point measurement imaging as described above.

    [0040] The processor P may be implemented in various ways. The processor P may be integrated in a dedicated device including front-ends for the hyperspectral camera HSC and the probe PR. However, the processor P may be distributed to one or more separate computer (s) or processor system(s) remotely located from the camera HSC and probe PR.

    [0041] FIG. 2 illustrates a tip of a possible optical contact probe for point-based probe measurements. A source fiber end SF emits light, and a detector fiber end DF detects the reflected light. E.g. the tip could be the tip of the PR shown in FIG. 1. The two fibers end SF, DF are arranged with a centre-to-centre distance of FD. This can be used e.g. for DRS, Fluorescence or Raman spectroscopy. DRS measurements with a contact probe can be executed very fast over a continuous wavelength range, and the probed volume can extend up to several millimeters into the tissue depending on the effective source-to-detector fiber(s) distance FD (e.g. with a probing depth ≈ ½ fiber distance). Thus, single point DRS provides an enhanced sensitivity and specificity compared to hyperspectral imaging by measuring over a wavelength range extending into the NIR (e.g. from 400nm to 1600nm), which can include major contributions from additional tumor discriminators. See also [R. Nachabe, D. Evers, BHW Hendriks, GW Lucassen, M. Van der Voort, EJ Rutgers, M-J Vrancken Peeters, JA van der Hage, HS Olderburg, J. Wesseling and TJM Ruers, "Diagnosis of breast cancer using optical spectroscopy from 500 to 1600 nm: a comparison of classification methods, J. Biomed. Opt. 16 (2011) p087010.].

    [0042] FIG. 3 illustrates a hyperspectral camera system based on a normal digital camera CM which provides 2D images of the biological tissue BT in a sequential manner controlled by a controller, where light sources L1, L2, and L3 sequentially applies their different narrow band light to the surface of the biological tissue BT. Often it is possible to achieve a good discrimination between tissue types by measuring reflectivity at only a few carefully selected wavelengths, and using some basic techniques to create contrast like dividing reflectivity at two different wavelengths. With such system it is possible to replace the expensive hyperspectral camera with a normal camera CM and use various narrow-band light sources instead. Preferably, biological tissue BT is located in a dark place during the sequential illumination by light sources L1-L3 to avoid any interference from other light sources.

    [0043] FIGs. 4a and 4b show a constructed illustration of the same hyperspectral image before (FIG. 4a) and after (FIG 4b) calibration of the hyperspectral processing algorithm. In the hyperspectral image FIG. 4a, suspicious areas which cannot be unambiguously classified by the hyperspectral image processing can be visually indicated to an operator, here illustrated as two areas indicated by dashed white lines. This is a guide to the operator to perform probe measurements within these suspicious areas, e.g. using a point measurement probe. The increased sensitivity and specificity of such point measurements for (residual) tumor discrimination can be used to recalibrate the hyperspectral processing image, e.g. a look-up table used therein, and thereby essentially enhancing the sensitivity and specificity of each pixel in the hyperspectral image. The result of such calibration is clearly seen in FIG. 4b, where the upper area in FIG 4a is seen to actually hide a tumor (white shading) which is clearly indicated in FIG. 4b, whereas the lower are indicated as suspicious in FIG 4a is seen in the enhanced image of FIG. 4b not to include tumor tissue.

    [0044] The point probe can also be used with fluorescence spectroscopy, which allows for enhanced tumor discrimination, e.g. by providing further information on the local metabolism, such as NAD+/NADH ratio, or the presence of autofluorescent tumor-specific markers in the tissue, such as cytostatic drugs like doxorubicin or paclitaxel. Using Raman spectroscopy, chemical constituents associated with the presence of tumor cells can be identified with a very high accuracy and can give complement information for the recalibration of the hyperspectral processing algorithm.

    [0045] FIGs. 5a-5c show examples of different strategies for probe measurements with point or at least small area measurements serving to cover a larger area. FIG. 5a shows point measurements, while FIGs. 5b and 5c show clusters of points covering a region of interest. E.g. the strategy shown in FIGs. 5b and 5c can be obtained using a combination of a single point probe and a scanning fiber microscope. Thus, in some embodiments, the probe may comprise both a single point probe and a scanning fiber microscope [BHW Hendriks, WCJ Bierhoff, JJL Horikx, AE Desjardins, CA Hezemans, GW 't Hooft, GW Lucassen, and N Mihajlovic, "High-resolution resonant and non resonant fiber-scanning confocal microscope", J. Biomed. Opt. 16 (2011) p026007]. Such scanning fiber microscope enables confocal imaging of the biological tissue under examination with a significant spatial resolution, field of view (FOV), and contrast. The combination in addition to the hyperspectral image provides a significant advancement over a single point measurement, e.g. as FIG. 5a. This modality enables multi-spectral imaging, full wavelength regions, over a spatial region of interest (∼200µm) with a submicron resolution (∼0.6µm). The modality can be used to obtain full spectral information either from a single region of interest, such as shown in FIG. 5b, or in combination as shown in FIG. 5c to cover many individual regions of interested, thus allowing information of a significantly larger region.

    [0046] The combination probe system could further employ different optical imaging modalities to be readily implemented such as confocal reflectance [K. Carlson, M. Chidley, K-B. Sung, M. Descour, A. Gillenwater, M. Follen, and R. Richards-Kortum, "In vivo fiber-optic confocal reflectance microscope with an injection-molded plastic miniature objective lens," Appl. Opt. 44(10), 1792-1797 (2005)], confocal fluorescence [J. C. Jung, A. D. Mehta, E. Aksay, R. Stepnoski, and M. J. Schnitzer, "In vivo mammalian brain imaging using one- and two-photon fluorescence microendoscopy," J. Neurophysiol. 92(5), 3121-3133 (2004)], two-photon fluorescence [J. C. Jung and M. J. Schnitzer, "Multiphoton endoscopy," Opt. Lett. 28(11), 902-904 (2003)] and [M. T. Myaing, D. J. MacDonald, and X. Li, "Fiber-optic scanning two-photon fluorescence endosocope," Opt. Lett. 31(8), 1076-1078 (2006)], and optical coherence tomography [G. J. Tearney, M. E. Brezinski, B. E. Bouma, S. A. Boppart, C. Pitris, J. F. Southern, and J. G. Fujimoto, "In vivo endoscopic optical biopsy with optical coherence tomography," Science 276(5321), 2037-2039 (1997)], could enable better characterization of the sample under examination.

    [0047] FIG. 6. illustrates a diagnostic embodiment where, referring to FIG. 1, the hyperspectral image 12 after calibration of the hyperspectral image algorithm HSA is applied to the display D. A diagnosis algorithm DA processes the same image 12 and detects therein the area where the target structure is probable to be, e.g. a tumor. A target indicator TI is then applied to the display, e.g. a highlighting in the form of a color, a shading, a line around the area or the like.

    [0048] In the following, examples details regarding the extraction of features of parameters from the one or more probe measurement data will be described. An example of extracting the physiological parameter is by fitting the acquired spectra using a custom made Matlab 7.9.0 (Mathworks, Natick, MA) algorithm. In this algorithm, a widely accepted analytical model was implemented, namely the model introduced by the reference [T.J. Farrel, M.S. Patterson and B.C. Wilson, "A diffusion theory model of spatially resolved, steady-state diffuse reflectance for the non-invasive determination of tissue optical properties," Med. Phys. 19 (1992) p. 879-888], which is hereby incorporated by reference in entirety. The input arguments for the model of the reference [T.J. Farrel, M.S. Patterson and B.C. Wilson, "A diffusion theory model of spatially resolved, steady-state diffuse reflectance for the non-invasive determination of tissue optical properties," Med. Phys. 19 (1992) p. 879-888] are the absorption coefficient µa(λ), the reduced scattering coefficient µ'a(λ) and the center-to-center distance between the emitting and collecting fibers at the tip of the probe. For the camera image there is no fiber distance and the model of Farrel is no longer applicable. Instead it is possible to use algorithms based on the well-known Kubelka-Munk model. An example of how reflectivity is derived from scattering and absorption properties is given e.g. in [M. Müller, G. F. Zhou " Modeling electrophoretic paper-like displays", Proc. IDW 1-3, 1287-1290 (2007)]) For a complete description of the diffusion theory model, we refer to [T.J. Farrel, M.S. Patterson and B.C. Wilson, "A diffusion theory model of spatially resolved, steady-state diffuse reflectance for the non-invasive determination of tissue optical properties," Med. Phys. 19 (1992) p. 879-888].

    [0049] In the following part, the model will be explained briefly. The used formulas are mainly based on work of Nachabé et al., and reference is thus made to [R. Nachabe, B.H.W. Hendriks, M.V.D. Voort, A. E, and H.J.C.M. Sterenborg, "Estimation of biological chromophores using diffuse optical spectroscopy : benefit of extending the UV-VIS wavelength range to include 1000 to 1600 nm," Optics Express, vol. 18, 2010, pp. 879-888] which is hereby incorporated by reference in entirety, and furthermore reference is made to [R. Nachabe, B.H.W. Hendriks, A.E. Desjardins, M. van der Voort, M.B. van der Mark, and H.J.C.M. Sterenborg, "Estimation of lipid and water concentrations in scattering media with diffuse optical spectroscopy from 900 to 1600 nm," Journal of Biomedical Optics, vol. 15, May. 2010, pp. 037015-10] which is hereby incorporated by reference in entirety.

    [0050] A double power law function can be used to describe the wavelength dependence of the reduced scattering, where the wavelength λ is expressed in nm and is normalized to a wavelength value of λ0 =800 nm. The parameter α corresponds to the reduced scattering amplitude at this specific wavelength.



    [0051] In this equation the reduced scattering coefficient is expressed as the sum of Mie and Rayleigh scattering where ρMR is the Mie-to-total reduced scattering fraction. The reduced scattering slope of the Mie scattering is denoted b and is related to the particle size.

    [0052] FIG. 7 shows logarithmic plots of absorption spectra for blot Hb, Hb02, water W, and fat F. For a homogeneous distribution of absorbers, the total light absorption coefficient µa(λ) can be computed as products of the extinction coefficients and volume fraction of the absorbers shown in FIG. 7:



    [0053] Instead of modeling the absorption coefficient µa(λ) as the sum of absorption coefficients weighted by the respective concentrations of the four chromophores of interest, it was decided to express the tissue absorption coefficient as:

    , where

    corresponds to the absorption by blood and

    corresponds to absorption by water and lipid together in the probed volume. The volume fraction of water and lipid is vWL = [Lipid]+[H2O], whereas vBlood represents the blood volume fraction for a concentration of hemoglobin in whole blood of 150 mg/ml.

    [0054] The factor C is a wavelength dependent correction factor that accounts for the effect of pigment packaging and alters for the shape of the absorption spectrum. This effect can be explained by the fact that blood in tissue is confined to a very small fraction of the overall volume, namely blood vessels. Red blood cells near the center of the vessel therefore absorb less light than those at the periphery. Effectively, when distributed homogeneously within the tissue, fewer red blood cells would produce the same absorption as the actual number of red blood cells distributed in discrete vessels. The correction factor can be described as:

    , where R denotes the average vessel radius expressed in cm. The absorption coefficient related to blood is given by:

    , where

    and

    represent the basic extinction coefficient spectra of oxygenated hemoglobin HbO2 and deoxygenated hemoglobin Hb, respectively. The oxygenated hemoglobin fraction in the total amount of hemoglobin is noted αBL = [HbO2]/([HbO2]+[Hb]) and is commonly known as the blood oxygen saturation.

    [0055] The absorption due to the presence of water and lipid in the measured tissue is defined as:



    [0056] In this case the concentration of lipid related to the total concentration of lipid and water together can be written as αWF = [Lipid]/([Lipid]+[H2O]), where [Lipid] and [H2O], correspond to the concentration of lipid (density of 0.86g/ml) and water, respectively.

    [0057] This way of relating the water and lipid parameters in the expression of the absorption coefficient defined in Eq.6, rather than estimating separately the water and lipid volume fraction corresponds to a minimization of the covariance of the basic functions for fitting resulting in a more stable fit cf. the reference [R. Nachabe, B.H.W. Hendriks, M.V.D. Voort, A. E, and H.J.C.M. Sterenborg, "Estimation of biological chromophores using diffuse optical spectroscopy : benefit of extending the UV-VIS wavelength range to include 1000 to 1600 nm," Optics Express, vol. 18, 2010, pp. 879-888]. For further explanation and validation of this theorem reference is made to [R. Nachabe, B.H.W. Hendriks, A.E. Desjardins, M. van der Voort, M.B. van der Mark, and H.J.C.M. Sterenborg, "Estimation of lipid and water concentrations in scattering media with diffuse optical spectroscopy from 900 to 1600 nm," Journal of Biomedical Optics, vol. 15, May. 2010, pp. 037015-10], which is hereby included by reference in entirety.

    [0058] Another way to discriminate differences in spectra is by making use of a principal components analysis. This method allows classification of differences in spectra and thus allows discrimination between tissues. It is also possible to extract features from the spectra.

    [0059] FIG. 8 shows intrinsic fluorescence curves for A: Collagen, B: Elastin, C: NADH, and D: FAD. Such parameters could also be measured with a probe. Especially, the ratio NADH/FAD, which is called the optical redox parameter, could be calculated, since this ratio is an indicator for the metabolic state of the tissue, which is assumed to change upon effective treatment of cancer cells, see e.g. [Q. Zhang, M.G. Mueller, J. Wu and M.S. Feld, "Turbidity-free fluorescence spectroscopy of biological tissue" Opt. Lett. 25 (2000) p1451].

    [0060] FIG. 9 illustrates a method embodiment according to the invention for identifying possible tumor tissue within a human or animal tissue. The method comprises the step R_HSI of receiving a hyperspectral image of a first area of the tissue under test, optionally also the step of providing the hyperspectral image with a hyperspectral camera. The hyperspectral image is then analyzed L_TT with an algorithm for identifying and/or localizing a part or parts of the hyperspectral image which may contain tumor tissue. In case such suspicious areas are localized in the hyperspectral image, the next step D_SA is displaying the hyperspectral image with indication of the suspicious areas, e.g. using a line or other graphical symbol overlaid the hyperspectral image. In this step D_SA, the hyperspectral image is processed with an initial setting or calibration of a hyperspectral image algorithm, e.g. a setting selected based on the target structure, in this case tumor tissue, to be identified. The operator then has a visual indication of which parts of the tissue under test to perform probe measurements. The next step R PMD is receiving probe measurement data, optionally also the step of providing the probe measurements resulting in the probe measurement data. E.g. the probe measurement data are obtained with an optical surface probe providing measurement data in a second area being essentially in one point within the first area, preferably within the suspicious area indentified and displayed.

    [0061] The probe measurement data is then analyzed A PMD with an algorithm arranged to identify one or more physiological parameters, e.g. based on a measured optical absorption coefficient, the reduced scattering coefficient and the center-to-center distance between light emitting and collecting fibers at the tip of the probe. The result is then used for calibrating C_HSA the processing algorithm used to analyze the hyperspectral image, so as to adapt the processing algorithm to provide an image with an enhanced contrast between target structure, in this example tumor tissue, and the tissue under test. Subsequently, the hyperspectral image is analyzed A HSI using the calibrated hyperspectral image processing algorithm. Finally, the re-analyzed hyperspectral image is displayed D HSI, e.g. on a color monitor.

    [0062] In this way one or a few point measurements with a high spectral resolution can be utilized to add information about the tissue in suspicious areas of the standard-processed hyperspectral image so as to modify the hyperspectral processing algorithm. This is then used to process the hyperspectral image in order to provide an image with enhanced contrast or sensitivity towards distinguishing between the tissue under test and the target structure, e.g. tumor tissue. Hereby, the method assists the operator, e.g. a medical doctor or other medical staff, in evaluating whether tumor tissue is present within an area under test, e.g. in the search for positive margins. Additionally, the method may include indicating suspicious areas also in the re-analyzed hyperspectral image, and/or indicating the probability of tumor tissue being present.

    [0063] It is to be understood that some of the steps may be performed in a different order than shown in FIG. 9 and than in the description of the above method embodiment.

    [0064] To sum up, the invention provides a medical imaging system for identifying a target structure, e.g. a tumor, in a biological tissue. A hyperspectral camera system is used for imaging a surface area of the tissue, e.g. with a limited spectral resolution, but enough to allow identification of suspicious areas where the target structure may be, e.g. such areas can be visually indicated on a display to the operator. A probe, e.g. an optical surface probe, is used to provide probe measurement of a smaller surface area of the tissue, but with more information indicative of the target structure. The probe is selected to provide a higher specificity with respect to identification of the target structure than the hyperspectral camera. The hyperspectral processing algorithm is then calibrated based on probe measurement data performed within the suspicious areas, thus providing a calibrated hyperspectral processing algorithm resulting in images with an enhanced sensitivity to identify the target structure. Only few probe measurements are required to significantly improve the resulting image, thereby providing a reliable and fast target structure identification.

    [0065] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage. A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.


    Claims

    1. A medical imaging system arranged to identify a target structure in an associated biological tissue (BT), the system comprising

    - a hyperspectral camera system (HSC) arranged to provide a hyperspectral image (HIS) and configured to cover a first surface area (A1) of the biological tissue (BT), wherein pixels of the hyperspectral image (HIS) contains information at different light wavelength bands,

    - a probe (PR) arranged to provide a probe measurement of a second surface area (A2) of the biological tissue (BT) with information indicative of the target structure (TS), and to generate probe measurement data (PMD) accordingly, wherein the probe (OR) is configured so that the second surface area (A2) covered is smaller than the first surface area (A1) that is covered by the hyperspectral camera system (HSC), and wherein the probe (PR) is selected to provide a higher specificity with respect to identification of the target structure (TS) than the hyperspectral camera system (HSC), and

    - a processor (P) functionally connected to receive the hyperspectral image (HSI) and the probe measurement data (PMD), wherein the processor (P) is arranged to analyze the hyperspectral image (HIS) according to a hyperspectral processing algorithm (HSA) in order to identify a part of the hyperspectral image (HSI) which possibly contains the target structure (TS), characterised in that the processor (P) is further arranged to analyze the probe measurement data (PMD) and to calibrate the hyperspectral processing algorithm (HSA) in response thereto, so as to provide a calibrated hyperspectral processing algorithm (HSA) with an enhanced sensitivity to identify the target structure (TS) in the hyperspectral image (HSI).


     
    2. Medical imaging system according to claim 1, wherein the probe measurement is acquired with a higher spectral resolution than the hyperspectral image (HSI).
     
    3. Medical imaging system according to claim 1, wherein the second surface area (A2) is selected to be located within the first surface area (A1).
     
    4. Medical imaging system according to claim 1, wherein the system is arranged to visually display the hyperspectral image (II) after being processed with the hyperspectral processing algorithm (HSA) together with a visual indication of the part which is identified as possibly containing the target structure (TS).
     
    5. Medical imaging system according to claim 1, wherein the system is arranged to visually display the hyperspectral image (12) after being processed with the calibrated hyperspectral processing algorithm (HSA).
     
    6. Medical imaging system according to claim 1, wherein the processor (P) is arranged to analyze the probe measurement data (PMD) according to a probe processing algorithm (PMA) comprising comparing the probe measurement data (PMD) with a look-up-table.
     
    7. Medical imaging system according to claim 1, wherein the processor (P) is arranged to analyze the probe measurement data (PMD) according to a probe processing algorithm (PMA) comprising calculating a plurality of measurement parameters and translating the plurality of measurement parameters into physiological parameters (PP) indicative of the target structure (TS).
     
    8. Medical imaging system according to claim 1, wherein the processor (P) is arranged to calibrate the hyperspectral processing algorithm (HSA) by refining a grouping of pixels in the hyperspectral image (HSI) which are identified in the probe measurement data (PMD) as belonging to the target structure (TS).
     
    9. Medical imaging system according to claim 1, wherein the probe (PR) comprises one of: a surface probe, and an interventional device.
     
    10. Medical imaging system according to claim 1, wherein the probe (PR) comprises at least one of: an optical probe, and an ultrasound probe.
     
    11. Medical imaging system according to claim 1, wherein the probe (PR) comprises one of: a diffuse reflectance spectroscopy probe, a diffuse optical tomography probe, a differential path length spectroscopy, a fluorescence probe, a Raman spectroscopy probe, a confocal reflectance probe, a confocal fluorescence probe, a two-photon fluorescence probe, an optical coherence tomography probe, and a scanning fiber microscope arranged to perform confocal imaging.
     
    12. Medical imaging system according to claim 1, wherein the hyperspectral camera system comprises one of:

    - a hyperspectral camera (HSC) arranged to provide the hyperspectral image (HSI),

    - a normal camera (CM) and a plurality of light sources (L1, L2, L3) with different narrow band wavelengths and a controller (C) arranged to provide images with the biological tissue (BT) sequentially illuminated by the plurality of light sources (L1, L2, L3), and

    - a normal camera and a tunable optical filter arranged between the normal camera and the biological tissue, so as to provide sequential images with different settings of the tunable optical filter.


     
    13. Medical imaging system according to claim 1, wherein processor (P) is arranged to process the hyperspectral image (HSI) according to an image processing algorithm in order to identify a tip of the probe (PR) in the hyperspectral image (HSI), so as to identify to which part of the hyperspectral image (A2) the probe measurement data (PMD) corresponds.
     
    14. Medical imaging system according to claim 1, wherein the target structure (TS) is one of: tumor tissue, a nerve bundle, a blood vessel, lymph nodes, and oxygenation level of tissue.
     
    15. A method for detecting a target structure in an associated biological tissue, the method comprising

    - providing (R_HSI) a hyperspectral image to cover a first surface area of the associated biological tissue, wherein pixels of the hyperspectral image contains information regarding different light wavelength bands,

    - providing (R_PMD) probe measurement data in accordance with a probe measurement configured to cover a second surface area of the biological tissue with information indicative of the target structure, wherein the second surface area is smaller than the first surface area and located within the first surface area, and wherein the probe measurement is selected to provide a higher specificity with respect to identification of the target structure than the hyperspectral camera system,

    - analyzing (A_HSI) the hyperspectral image according to a hyperspectral processing algorithm in order to identify a part of the hyperspectral image which may contain the target structure,

    - analyzing (A_PMD) the probe measurement data, and characterised by,

    - calibrating (C_HSA) the hyperspectral processing algorithm in response to a result of the probe measurement data analysis, so as to provide a calibrated hyperspectral processing algorithm with an enhanced sensitivity to identify the target tissue structure in the hyperspectral image.


     


    Ansprüche

    1. Medizinisches Bildgebungssystem, das dafür eingerichtet ist, eine Zielstruktur in einem zugehörigen biologischen Gewebe (BT) zu identifizieren, wobei das System Folgendes umfasst:

    - ein hyperspektrales Kamerasystem (HSC), das dafür eingerichtet ist, ein hyperspektrales Bild (HIS) bereitzustellen, und dafür konfiguriert ist, einen ersten Oberflächenbereich (A1) des biologischen Gewebes (BT) abzudecken, wobei Pixel des hyperspektralen Bilds (HIS) Informationen bei verschiedenen Lichtwellenlängenbändern enthalten,

    - eine Sonde (PR), die dafür eingerichtet ist, eine Sondenmessung eines zweiten Oberflächenbereichs (A2) des biologischen Gewebes (BT) mit Informationen, die eine Zielstruktur (TS) angeben, bereitzustellen, und entsprechend Sondenmessdaten (PMD) zu erzeugen, wobei die Sonde (PR) derartig konfiguriert ist, dass der abgedeckte zweite Oberflächenbereich (A2) kleiner ist als der erste Oberflächenbereich (A1), der durch das hyperspektrale Kamerasystem (HSC) abgedeckt wird, und wobei die Sonde (PR) ausgewählt ist, um eine höhere Spezifizität in Bezug auf die Identifizierung der Zielstruktur (TS) bereitzustellen als das hyperspektrale Kamerasystem (HSC), und

    - einen Prozessor (P), der funktional verbunden ist, um das hyperspektrale Bild (HSI) und die Sondenmessdaten (PMD) zu empfangen, wobei der Prozessor (P) dafür eingerichtet ist, das hyperspektrale Bild (HIS) gemäß einem hyperspektralen Verarbeitungsalgorithmus (HSA) zu analysieren, um einen Teil des hyperspektralen Bilds (HSI) zu identifizieren, der möglicherweise die Zielstruktur (TS) enthält,

    dadurch gekennzeichnet, dass der Prozessor (P) ferner dafür eingerichtet ist, die Sondenmessdaten (PMD) zu analysieren und den hyperspektralen Verarbeitungsalgorithmus (HSA) als Reaktion darauf zu kalibrieren, um so einen kalibrierten hyperspektralen Verarbeitungsalgorithmus (HSA) mit einer verbesserten Empfindlichkeit zum Identifizieren der Zielstruktur (TS) in dem hyperspektralen Bild (HSI) bereitzustellen.
     
    2. Medizinisches Bildgebungssystem nach Anspruch 1, wobei die Sondenmessung mit einer höheren spektralen Auflösung als das hyperspektrale Bild (HSI) erfasst wird.
     
    3. Medizinisches Bildgebungssystem nach Anspruch 1, wobei der zweite Oberflächenbereich (A2) so ausgewählt ist, dass er sich innerhalb des ersten Oberflächenbereichs (A1) befindet.
     
    4. Medizinisches Bildgebungssystem nach Anspruch 1, wobei das System dafür eingerichtet ist, das hyperspektrale Bild (I1) nach der Verarbeitung mit dem hyperspektralen Verarbeitungsalgorithmus (HSA) zusammen mit einer visuellen Angabe des Teils, der als möglicherweise die Zielstruktur (TS) enthaltend identifiziert wurde, visuell anzuzeigen.
     
    5. Medizinisches Bildgebungssystem nach Anspruch 1, wobei das System dafür eingerichtet ist, das hyperspektrale Bild (I2) nach der Verarbeitung mit dem kalibrierten hyperspektralen Verarbeitungsalgorithmus (HSA) visuell anzuzeigen.
     
    6. Medizinisches Bildgebungssystem nach Anspruch 1, wobei der Prozessor (P) dafür eingerichtet ist, die Sondenmessdaten (PMD) gemäß einem Sondenverarbeitungsalgorithmus (PMA) zu analysieren, der das Vergleichen der Sondenmessdaten (PMD) mit einer Nachschlagetabelle umfasst.
     
    7. Medizinisches Bildgebungssystem nach Anspruch 1, wobei der Prozessor (P) dafür eingerichtet ist, die Sondenmessdaten (PMD) gemäß einem Sondenverarbeitungsalgorithmus (PMA) zu analysieren, der das Berechnen einer Vielzahl von Messparametern und das Übersetzen der Vielzahl von Messparametern in physiologische Parameter (PP) umfasst, welche die Zielstruktur (TS) angeben.
     
    8. Medizinisches Bildgebungssystem nach Anspruch 1, wobei der Prozessor (P) dafür eingerichtet ist, den hyperspektralen Verarbeitungsalgorithmus (HSA) zu kalibrieren, indem eine Gruppierung von Pixeln in dem hyperspektralen Bild (HSI) verfeinert wird, die in den Sondenmessdaten (PMD) als zu der Zielstruktur (TS) gehörend identifiziert wurden.
     
    9. Medizinisches Bildgebungssystem nach Anspruch 1, wobei die Sonde (PR) eines von Folgendem umfasst: eine Oberflächensonde oder eine interventionelle Vorrichtung.
     
    10. Medizinisches Bildgebungssystem nach Anspruch 1, wobei die Sonde (PR) eines von Folgendem umfasst: eine optische Sonde oder eine Ultraschallsonde.
     
    11. Medizinisches Bildgebungssystem nach Anspruch 1, wobei die Sonde (PR) eines von Folgendem umfasst: eine diffuse Reflexionsspektroskopiesonde, eine diffuse optische Tomographiesonde, eine differentielle Pfadlängenspektroskopie, eine Fluoreszenzsonde, eine Raman-Spektroskopiesonde, eine konfokale Reflexionssonde, eine konfokale Fluoreszenzsonde, eine Zwei-Photonen-Fluoreszenzsonde, eine optische Kohärenztomographiesonde oder ein Rasterfasermikroskop, dafür vorgesehen, konfokale Bildgebung durchzuführen.
     
    12. Medizinisches Bildgebungssystem nach Anspruch 1, wobei das hyperspektrale Kamerasystem eines von Folgendem umfasst:

    - eine hyperspektrale Kamera (HSC), die dafür eingerichtet ist, das hyperspektrale Bild (HSI) bereitzustellen,

    - eine normale Kamera (CM) und eine Vielzahl von Lichtquellen (L1, L2, L3) mit verschiedenen Schmalbandwellenlängen und eine Steuereinheit (C), die dafür eingerichtet ist, Bilder mit dem biologischen Gewebe (BT) bereitzustellen, die sequentiell durch die Vielzahl von Lichtquellen (L1, L2, L3) beleuchtet werden, und

    - eine normale Kamera und einen abstimmbaren optischen Filter, der zwischen der normalen Kamera und dem biologischen Gewebe angeordnet ist, um sequentielle Bilder mit unterschiedlichen Einstellungen des abstimmbaren optischen Filters bereitzustellen.


     
    13. Medizinisches Bildgebungssystem nach Anspruch 1, wobei der Prozessor (P) dafür eingerichtet ist, das hyperspektrale Bild (HSI) gemäß einem Bildverarbeitungsalgorithmus zu verarbeiten, um eine Spitze der Sonde (PR) in dem hyperspektralen Bild (HSI) zu identifizieren, um so zu identifizieren, welchem Teil des hyperspektralen Bilds (A2) die Sondenmessdaten (PMD) entsprechen.
     
    14. Medizinisches Bildgebungssystem nach Anspruch 1, wobei die Zielstruktur (TS) eines von Folgendem ist: Tumorgewebe, ein Nervenbündel, ein Blutgefäß, Lymphknoten oder Sauerstoffanreicherungsgrad von Gewebe.
     
    15. Verfahren zum Detektieren einer Zielstruktur in einem zugehörigen biologischen Gewebe, wobei das Verfahren Folgendes umfasst:

    - Bereitstellen (R HSI) eines hyperspektralen Bilds, um einen ersten Oberflächenbereich des zugehörigen biologischen Gewebes abzudecken, wobei Pixel des hyperspektralen Bilds Informationen in Bezug auf verschiedene Lichtwellenlängenbänder enthalten,

    - Bereitstellen (R_PMD) von Sondenmessdaten gemäß einer Sondenmessung, die konfiguriert ist, um einen zweiten Oberflächenbereich des biologischen Gewebes mit Informationen, die eine Zielstruktur angeben, abzudecken, wobei der zweite Oberflächenbereich kleiner ist als der erste Oberflächenbereich und sich innerhalb des ersten Oberflächenbereichs befindet, und wobei die Sondenmessung ausgewählt wird, um eine höhere Spezifizität in Bezug auf die Identifizierung der Zielstruktur bereitzustellen als das hyperspektrale Kamerasystem,

    - Analysieren (A HSI) des hyperspektralen Bilds gemäß einem hyperspektralen Verarbeitungsalgorithmus, um einen Teil des hyperspektralen Bilds zu identifizieren, der möglicherweise die Zielstruktur enthält,

    - Analysieren (A_PMD) der Sondenmessdaten, und
    gekennzeichnet durch

    - Kalibrieren (C_HSA) des hyperspektralen Verarbeitungsalgorithmus als Reaktion auf ein Ergebnis der Analyse der Sondenmessdaten, um so einen kalibrierten hyperspektralen Verarbeitungsalgorithmus mit einer verbesserten Empfindlichkeit zum Identifizieren der Zielstruktur in dem hyperspektralen Bild (HSI) bereitzustellen.


     


    Revendications

    1. Système d'imagerie médicale agencé pour identifier une structure cible dans un tissu biologique associé (BT), le système comprenant

    - un système d'appareil de prise de vues hyperspectral (HSC) agencé pour fournir une image hyperspectrale (HIS) et configuré pour couvrir une première superficie (A1) du tissu biologique (BT), dans lequel des pixels de l'image hyperspectrale (HIS) contiennent des informations à différentes bandes de longueur d'onde de lumière,

    - une sonde (PR) agencée pour fournir une mesure de sonde d'une seconde superficie (A2) du tissu biologique (BT) avec des informations indicatives de la structure cible (TS), et pour générer des données de mesure de sonde (PMD) en conséquence, dans lequel la sonde (PR) est configurée de sorte que la seconde superficie (A2) couverte est plus petite que la première superficie (A1) qui est couverte par le système d'appareil de prise de vues hyperspectral (HSC), et dans lequel la sonde (PR) est sélectionnée pour fournir une spécificité plus élevée en ce qui concerne l'identification de la structure cible (TS) que le système d'appareil de prise de vues hyperspectral (HSC), et

    - un processeur (P) fonctionnellement relié pour recevoir l'image hyperspectrale (HSI) et les données de mesure de sonde (PMD), dans lequel le processeur (P) est agencé pour analyser l'image hyperspectrale (HIS) selon un algorithme de traitement hyperspectral (HSA) afin d'identifier une partie de l'image hyperspectrale (HSI) qui contient probablement la structure cible (TS), caractérisé en ce que

    le processeur (P) est en outre agencé pour analyser les données de mesure de sonde (PMD) et pour étalonner l'algorithme de traitement hyperspectral (HSA) en réponse à celui-ci, afin de fournir un algorithme de traitement hyperspectral étalonné (HSA) avec une sensibilité améliorée pour identifier la structure cible (TS) dans l'image hyperspectrale (HSI).
     
    2. Système d'imagerie médicale selon la revendication 1, dans lequel la mesure de sonde est acquise avec une définition spectrale plus élevée que l'image hyperspectrale (HSI).
     
    3. Système d'imagerie médicale selon la revendication 1, dans lequel la seconde superficie (A2) est sélectionnée pour être située à l'intérieur de la première superficie (A1).
     
    4. Système d'imagerie médicale selon la revendication 1, dans lequel le système est agencé pour visuellement afficher l'image hyperspectrale (I1) après traitement avec l'algorithme de traitement hyperspectral (HSA) en même temps qu'une indication visuelle de la partie qui est identifiée comme contenant probablement la structure cible (TS).
     
    5. Système d'imagerie médicale selon la revendication 1, dans lequel le système est agencé pour afficher visuellement l'image hyperspectrale (I2) après traitement avec l'algorithme de traitement hyperspectral étalonné (HSA).
     
    6. Système d'imagerie médicale selon la revendication 1, dans lequel le processeur (P) est agencé pour analyser les données de mesure de sonde (PMD) selon un algorithme de traitement de sonde (PMA) comprenant la comparaison des données de mesure de sonde (PMD) à une table de consultation.
     
    7. Système d'imagerie médicale selon la revendication 1, dans lequel le processeur (P) est agencé pour analyser les données de mesure de sonde (PMD) selon un algorithme de traitement de sonde (PMA) comprenant le calcul d'une pluralité de paramètres de mesure et la transformation de la pluralité de paramètres de mesure en paramètres physiologiques (PP) indicatifs de la structure cible (TS).
     
    8. Système d'imagerie médicale selon la revendication 1, dans lequel le processeur (P) est agencé pour étalonner l'algorithme de traitement hyperspectral (HSA) en affinant un groupement de pixels dans l'image hyperspectrale (HSI) qui est identifiée dans les données de mesure de sonde (PMD) comme appartenant à la structure cible (TS).
     
    9. Système d'imagerie médicale selon la revendication 1, dans lequel la sonde (PR) comprend l'un : d'une sonde de surface et d'un dispositif interventionnel.
     
    10. Système d'imagerie médicale selon la revendication 1, dans lequel la sonde (PR) comprend au moins l'une : d'une sonde optique et d'une sonde à ultrasons.
     
    11. Système d'imagerie médicale selon la revendication 1, dans lequel la sonde (PR) comprend l'une : d'une sonde de spectroscopie par réflectance diffuse, d'une sonde de tomographie optique diffuse, d'une spectroscopie à longueur de chemin différentielle, d'une sonde à fluorescence, d'une sonde de spectroscopie de Raman, d'une sonde à réflectance à foyer commun, d'une sonde à fluorescence à foyer commun, d'une sonde à fluorescence à deux photons, d'une sonde de tomographie à cohérence optique et d'un microscope à balayage par fibre agencé pour effectuer une imagerie à foyer commun.
     
    12. Système d'imagerie médicale selon la revendication 1, dans lequel le système d'appareil de prise de vues hyperspectral comprend l'un de :

    - un appareil de prise de vues hyperspectral (HSC) agencé pour fournir l'image hyperspectrale (HSI),

    - un appareil de prise de vues normal (CM) et une pluralité de sources de lumière (L1, L2, L3) avec des longueurs d'onde à bande étroites différentes et une unité de commande (C) agencée pour fournir des images avec le tissu biologique (BT) éclairé de manière séquentielle par la pluralité de sources de lumière (L1, L2, L3), et

    - un appareil de prise de vues normal et un filtre optique accordable agencé entre l'appareil de prise de vues normal et le tissu biologique, afin de fournir des images séquentielles avec différents réglages du filtre optique accordable.


     
    13. Système d'imagerie médicale selon la revendication 1, dans lequel le processeur (P) est agencé pour traiter l'image hyperspectrale (HSI) selon un algorithme de traitement d'image pour identifier une pointe de la sonde (PR) dans l'image hyperspectrale (HSI), afin d'identifier à quelle partie de l'image hyperspectrale (A2) les données de mesure de sonde (PMD) correspondent.
     
    14. Système d'imagerie médicale selon la revendication 1, dans lequel la structure cible (TS) est l'un : d'un tissu tumoral, d'un faisceau de nerfs, d'un vaisseau sanguin, de noeuds lymphoïdes et d'un niveau d'oxygénation de tissu.
     
    15. Procédé pour détecter une structure cible dans un tissu biologique associé, le procédé comprenant

    - la fourniture (R _HSI) d'une image hyperspectrale pour couvrir une première superficie du tissu biologique associé, dans lequel des pixels de l'image hyperspectrale contiennent des informations concernant des bandes de longueur d'onde de lumière différentes,

    - la fourniture (R_PMD) de données de mesure de sonde selon une mesure de sonde configurée pour couvrir une seconde superficie du tissu biologique avec des informations indicatives de la structure cible, dans laquelle la seconde superficie est plus petite que la première superficie et située à l'intérieur de la première superficie, et dans laquelle la mesure de sonde est sélectionnée pour fournir une spécificité plus élevée en ce qui concerne l'identification de la structure cible que le système d'appareil de prise de vues hyperspectral,

    - l'analyse (A HSI) de l'image hyperspectrale selon un algorithme de traitement hyperspectral afin d'identifier une partie de l'image hyperspectrale qui peut contenir la structure cible,

    - l'analyse (A_PMD) des données de mesure de sonde et caractérisé par,

    - l'étalonnage (C_HSA) de l'algorithme de traitement hyperspectral en réponse à un résultat de l'analyse de données de mesure de sonde, afin de fournir un algorithme de traitement hyperspectral étalonné avec une sensibilité améliorée pour identifier la structure de tissu cible dans l'image hyperspectrale.


     




    Drawing




















    Cited references

    REFERENCES CITED IN THE DESCRIPTION



    This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

    Patent documents cited in the description




    Non-patent literature cited in the description