WO2017005838A1

WO2017005838A1 - Non-invasive biological tissue examination based on full field high definition optical coherence tomography imaging

Info

Publication number: WO2017005838A1
Application number: PCT/EP2016/066088
Authority: WO
Inventors: Marc Boone; Frans Dhaenens
Original assignee: Agfa Healthcare
Priority date: 2015-07-09
Filing date: 2016-07-07
Publication date: 2017-01-12

Abstract

A depth profile of a region of interest in a 3D image of a biological tissue obtained through full field time domain high definition optical coherence tomography imaging is subjected to curve fitting. Parameters linked to optical anisotropy characteristics of this tissue are deduced from fitted curves and are compared to set values of these parameters, the set values being indicative of normal, precancerous or cancerous tissue characteristics.

Description

Non- invasive biological tissue examination based on full field high definition optical coherence tomography imaging.

[DESCRIPTION]

FIELD OF THE INVENTION

The present invention relates to non-invasive skin or biological tissue examination, more specifically to such examination by means of full field time domain high definition optical coherence

tomography (FF-TD-HD-OCT) .

BACKGROUND OF THE INVENTION

Optical coherence tomography (OCT) is a non-invasive imaging technique based on a measurement of optical reflections of internal microstructures inside a tissue specimen in vivo and in situ.

Currently the most important areas of application of OCT in human and veterinary medicine are dermatology, ophthalmology, cardiology, surgery, histology, diagnosis of cancer as well as follow-up of treatment .

Due to its light scattering properties biological tissue (in vivo in situ or ex vivo) is particularly suitable for diagnostic examination by means of OCT .

Since relatively low light intensities are sufficient for OCT examination and because the used wavelengths are situated in the near infrared range (750 nm to 1350 nm) , unlike ionizing radiation the method does not contaminate biological tissue with radiation. It is therefore particularly significant for medical applications.

However the skin is due its protective nature against all kinds of radiation ranging from UV, visible light and NIR, very reflective and scattering towards spectral radiation and allows only

penetration in certain "windows" of the infrared. Specially the horny layer (stratum corneum) of skin is highly in-penetrable for all kinds of light, preventing outward transmission of information from deeper layers of the epidermis and the dermis.

The OCT technique is comparable to the ultrasound medical imaging technique. However, in the OCT technique broadband light with short coherence length is used instead of sound.

A sample is irradiated with light with short coherence length.

Then, the time of flight of the light reflected on different reflective boundary layers and backscattering sides in a sample are recorded with the aid of an interferometer.

This time of flight information in its turn yields spatial

information about the specimen's microstructure .

With OCT typical resolutions higher than two orders of magnitude are achieved than with ultrasound, however, the achieved measuring depth is considerably smaller. Due to optical scattering the penetration depth is rather limited, image can be obtained that reach into the tissue up to a depth of 2 millimeters.

A large number of patent applications regarding aspects of HD-OCT have been filed by Agfa: EP 1 962 049, EP 1 962 050, EP 1 962 051, EP 1 962 052, EP 1 962 079, EP 1 962 080, EP 1 962 081, EP 1 962 082, EP 2 199 734, EP 2 498 048, EP 2 508 842, EP 2 508 843, EP 2 702 935, WO 2014 / 048 573.

High resolution optical coherence tomography is obtained by applying dynamic focus tracking as described in EP 1 962 050. The described method provides high lateral resolution, i.e. parallel to the skin surface .

In this patent a system and a corresponding method for high

definition optical coherence tomography are described having an interferometer for emitting light with which a full field is irradiated, the interferometer comprising a beam splitter and at least one reflector the optical distance of which from the beam splitter is changeable. Further a specimen objective is provided by means of which light emitted by the interferometer is focused in a focus plane lying within the specimen, and a detector is provided for collecting light which is reflected by the specimen.

For simpler and quicker recording of the sharpest possible images of the specimen provision is made such that during a change of the optical distance between the reflector and the beam splitter the light respectively reflected at a number of different depths of the specimen is collected by the detector, and during the collection of the light respectively reflected at the different depths of the specimen the imaging properties of the specimen objective are changed such that the focus comes within the range of the respective depth of the specimen.

EP 2 498 048 describes the use of a light source with a Gaussian filter. The use of such a light source provides enhanced resolution in axial direction (i.e. perpendicular to the surface of the skin that is examined) .

In OCT lateral and axial resolution are decoupled. High lateral resolution is obtained by the resolution and magnification power of the optics that are used.

High lateral resolution may lead to low depths of focus (depth resolution being defined by the bandwidth of the light source) .

This disadvantage can be solved by moving the focus while scanning the depth as performed in the focus tracking system described in EP 2 498 048.

By applying the adaptive focus concept a resolution up to 3μ can be obtained in all directions and at all depths, creating a homogenous 3D image with cellular resolution, comparable to microscopy and to Reflection Confocal Microscopy. Using a low coherent light source with a Gaussian filter has the following effects:

- broad bandwidth resulting in high axial resolution

- Gaussian spectrum resulting eliminating ghost images

- Laterally low coherent light resulting in the absence of image blur (by coherent cross-talk) .

In most applications a small amount of optical gel is applied to the skin or to the examining probe in order to ensure good contact and to increase the penetration depth by adapting the optical reflection index of the skin by filling pores and skin asperities.

Whereas with the above-described techniques for examination of skin or biological tissue satisfactory results can already be obtained, there is a constant need for improvement. More specifically there is a need for an improved method to discriminate normal tissue from precancerous or cancer tissue (dysplastic, neoplastic) .

SUMMARY OF THE INVENTION

The above-mentioned aspects are realised by a method having the specific steps set out in claim 1.

Specific features for preferred embodiments of the invention are set out in the dependent claims.

The invention provides an aid to perform a more sensitive and more specific discrimination in the diagnosis of (malignant) lesions in the skin (in vivo, in situ in the skin) and other tissues compared with classical anatomo-pathological imaging of normal skin and tissues, based on mathematical analyses of the optical anisotropic scattering properties of a depth-profile of the skin and tissues in specified anatomical areas. The described method may assist the dermatologist when making a diagnostic separation between normal, dysplastic (pre-cancerous) and neoplast / cancerous tissue in a range of skin cancers like melanoma and non-melanoma skin and epithelial cancers as well as other categories of skin diseases. The results are significantly better than the results obtained with classic OCT and HD-OCT anatomo- pathological cellular imaging in 3D.

Further advantages and embodiments of the present invention will become apparent from the following description.

DETAILED DESCRIPTION OF THE INVENTION

The present invention describes a discriminative mathematical analysis method based on the anisotropic scattering characteristics of normal and diseased skin or epithelial tissue, which can be applied for in vivo in situ as well as non- invasive ex vivo

diagnostic examination of tissue.

According to the present invention a specimen, more particularly a biological tissue (in vivo or ex vivo) which can be normal, precancerous or melanoma skin or epithelial cancer, is subjected to full field time domain high definition optical coherence tomography imaging (HD-OCT imaging) , a technique as described in the

introduction. The imaging is performed by means of a full field time domain high definition optical coherence tomography apparatus denominated as Skintell^® of Agfa Healthcare, Mortsel, Belgium.

The full field high definition time domain optical coherence tomography imaging method uses adaptive optics to create a

homogeneous 3D image block with cellular resolution at all depths.

Preferably the full field time domain OCT technique is applied with light having a wavelength of about 1300 nm. Also preferably the emitted light has a Gaussian bandwidth of 200 nm.

In one embodiment such a produced 3D image block (in the described specific embodiment the example measuring 1.5 x 1.8 x0.570 mm³) contains 200 en-face slices of about 3 μπι thickness (exact value : 2.9 μπι) parallel to the surface of the tissue.

These en-face slice images can be represented as 3D DICO images.

In a next step of the method of the present invention, these 3D DICOM images are fed into a standard, commercially available image processing software package. An example of such a suitable software package is Image J, being a public domain Java-based image

processing program developed at National Institutes of Health. This software package can display, edit, analyze, process, save, and print color and grayscale images.

The following consecutive processing steps are applied to the voxels pixels of the 3D images:

In a specific embodiment a removal is applied of the high reflective entrance offset image between a gel which is in most occasions applied on the tissue and the lens of the optical HD OCT imaging probe to improve contact at imaging. This is performed by means of a normalization operation.

By means of this normalization operation differences due to patient differences, differences between OCT devices, age, skin type can also be removed.

Next the method comprises the step of determining a region of interest (ROI) . In a specific embodiment this ROI consists of a region of 300μτη x300 μπι x570 μπι depth through a lesion of interest. When defining the ROI artifacts which could have an influence on the signal are avoided. Such artifacts may e.g. be a hair root, pores etc .

The sequence of the actions may be different. E.g. first the RI is determined, then a normalization operation is performed and next a depth profile is generated. The next step of the method of the present invention consists of the computation of an (averaged) depth profile of this ROI . An average value is calculated for the ROI in every slice in the depth profile. This (averaged) depth profile is preferably displayed as a graph.

This graph of the depth profile represents the degree of an- isotropical scattering properties of the imaged tissue and its cellular and extracellular constituents and the scattering is different for normal, precancerous and cancerous tissue.

This graph of the normalized and averaged depth profile is split into at least 2 parts or sections depending on the specific pattern of the successive histological tissue layers along the depth

(division into sections is performed based on the peak value pattern of the depth profile) .

In a following step, a mathematical analysis of the above-mentioned parts of the depth profile (signal as f (depth)) is performed. This mathematical analysis comprises at least one curve fitting process for each section.

However, several curve fitting methods may be applied such as negative exponential function, power series, and piece-wise local linear fittings (Levenberg-Marquard) .

Then certain parameters are deduced from the curve resulting from the mathematical fitting. Among such parameters are amplitude

(expressing reflected signal values) , slope (reflecting degree of anisotropy) and half layer value (HLV, reflecting thickness of an abnormal layer in the tissue) . These are parameters that directly represent optical characteristics of the imaged tissue.

Examples of such optical characteristics are the attenuation coefficient, absorption coefficient, scattering coefficient, degree of anisotropy, mean free path and the scattering regime (either Rayleigh or Mie (cellular) for the used 1300 nm wavelength. In one embodiment the following curve fitting technique is used and the following parameter is evaluated.

The mathematical curve fitting to each of said sections to obtain fitted curves is performed with a mathematical curve expressed as R

= p.e⁽-"-^af*z) ;

R expressing reflection signal,

p being equal to the (starting) reflection value in a section, μτα/being a relative attenuation factor and expressing the level of anisotropic scattering in tissue and being proportional to a reduced scatter coefficient μ's = ps(\-g) wherein μ expresses attenuation of an isotropic tissue and g has a value between 0 and 1 and expresses the amount of anisotropy (g is 0 for isotrope scattering (healthy) and is close to 1 for anisotrope (cancerous) scattering, and z represents depth in said tissue.

This formula is based on the formula disclosed by Jacques S,

Samatham R, Choudhury N, Fu Y, Levitz D (2008) in

"Measuring tissue optical properties in vivo using reflectancemode confocal microscopy and optical coherence tomography."

In: Wax A, Backman W (eds) Biomedical applications of light

scattering II, Proc . SPIE 6864, 686410. doi : 10.1117/12.761803

Next, the value of the mathematical parameter juraf of said fitted curve is computed, μα being related to anisotropic scattering of said tissue.

Normal, precancerous and cancerous tissue have different optical properties at certain wavelengths (in this example a wavelength about 1300 nm is applied) due to the size and composition of their constituent cells. The cells in the tissue mainly cause Mie scattering with the wavelength of 1300 nm, with a different

intensity depending on the degree of anisotropy which is directly related to the stage of the cancer (normal, precancerous,

cancerous) .

According to this invention, tables are used in which parameter values obtained from the mathematical analysis of the depth profiles through the tissue are set out for 3 discriminative cases of tissue behaviour, being BN (normal), DN (precancerous) and MM cancerous) . Such values can be retrieved from lists coupling the parameter values with the tissue statuses.

The result of the mathematical analysis of the depth profile of the tissue image obtained by HD OCT imaging as described above is then to be compared with the values in a table. This may aid the medical doctor to draw a conclusion as to the normal, precancerous or cancer characteristic of the examined tissue.

The mathematical analysis of the depth profile of the optical characteristics of the imaged tissue with the HD-OCT apparatus thus allows a more accurate diagnosis (normal, pre-cancerous , cancerous) of the tissue than the classical morphological imaging methods such as dermoscopy, confocal microscopy and OCT.

This mathematical analysis can be automated and guided based upon a database comprising collected data of existing cases.

Having described in detail preferred embodiments of the current invention, it will now be apparent to those skilled in the art that numerous modifications can be made therein without departing from the scope of the invention as defined in the appending claims.

Claims

1. Method of discriminating biological tissue conditions comprising the steps of

Generating a homogeneous 3D image of said biological tissue by applying a full field, time domain high definition optical coherence tomography (HD OCT) imaging technique to said tissue,

Determining a region of interest (ROI) in said 3D image, Generating a depth profile of said ROI, said depth profile expressing slice-averaged and normalized signal values against depth inside said tissue ,

Dividing said depth profile into at least two sections based on the peak value pattern of said depth profile,

(corresponding to visual appearance of said 3D image of tissue)

Applying mathematical curve fitting to each of said sections to obtain fitted curves, wherein said mathematical curve is expressed as R = p.e ;

R expressing reflection signal,

p being equal to the maximum value in a section,

μτα being a relative attenuation factor and expressing the level of anisotropic scattering in tissue and being

proportional to a reduced scatter coefficient μ'ς = ^(l-g) wherein ps expresses attenuation of an isotropic tissue and g has a value between 0 and 1 and expresses the amount of anisotropy, and

z represents depth in said tissue,

Computing said mathematical parameter //ra/of said fitted curve, juraf being related to anisotropic scattering of said tissue ,

Comparing the value of juraf in said section of the depth profile with a set values for said parameter, said set values being associated with a tissue condition, Selecting a tissue condition corresponding with the set value that is closest to said computed mathematical

parameter in said part of said depth profile, as tissue condition of said biological tissue.

2. A method according to claim 1 wherein said region of interest is an artifact free region of interest.

3. A method according to claim 1 wherein said full field time domain high definition OCT technique is applied with light having a wavelength of about 1300 nm.

4. A method according to claim 3 wherein said emitted light has a Gaussian bandwidth of about 200 nm.

5. A method according to claim 1 wherein said time domain high definition optical coherence tomography (HD OCT) imaging technique applies adaptive optics to create said 3D image with cellular resolution at all depths in said 3D image.

6. A method according to claim 1 wherein said tissue is biological tissue in vivo.

7. A method according to claim 1 wherein said tissue is biological tissue ex vivo.

8. A method according to claim 1 applied to said tissue in situ .

9. A method according to claim 1 wherein said tissue is skin tissue .

10. A method according to claim 1 wherein said tissue is epithelial tissue.