CRAN - Campus Sciences
BP 70239 - 54506 VANDOEUVRE Cedex
Tél : +33 (0)3 72 74 52 90
cran-secretariat@univ-lorraine.fr
 
 
Ph. D. Project : Multi-scale quantitative spectroscopy for In Vivo diagnosis
Dates : 2017/10/01 - 2020/09/30
Manager(s) CRAN: Walter BLONDEL , Marine AMOUROUX
Other Manager(s): Dr. Anne PLANAT-CHRETIEN (anne.planat-chretien@cea.fr)
Full reference: Wide-field multispectral imaging (MSI) has been recently studied in the framework of medical diagnosis because it may provide spectral information on a wide region of interest. Thanks to the acquisition of surface diffuse reflectance at specific wavelengths, such a technique allows biological tissues’ optical properties (i.e. absorption and scattering) evaluation; the estimated absorption coefficient gives access to biological tissue’s composition in chromophores that is modified along with specific diseases. Such a technique faces with 2 important clinical challenges: non-invasive and quick measurements. However current MSI devices provide a relative (and not absolute) estimation of absorption since the absorption estimation is based on a preliminary assumption on scattering. It results on hardly reproducible absorption values.


To achieve a wide-field quantitative measurement, CEA-LETI developed a « Dual-Step Technique » that couples MSI to a local measurement of Spatially-Resolved – Diffuse Reflectance Spectroscopy (SR – DRS). Diffuse Reflectance measured at several distances allows to separate and estimate mean global optical properties (OP) for the tissue that is probed punctually. However, this technique does not provide quick and easy 2D mapping of OP on a large area (several cm²). Since the scattering coefficient correlates mainly to tissue structures, the ‘Dual-Step approach consists in using the SR-DRS measurement to estimate the scattering property of common zoned of interest; followed by a MSI measurement to estimate the absorption. This provides fast and non-invasive quantitative 2D mapping of both types of coefficients.

The current PhD project aims at developing a new diagnostic device for clinics that allows to target a depth-resolved region of interest on a quantitative spectral image. The device will provide a representative measurement of local metabolism, medium structure and disease’s depth extension with a much better axial resolution compared with current SR-DRS devices. Our strategy is to develop a DRS approach based on a matrix sensor coupling MSI and DRS that gives access to a tremendous potential of information of depth-resolved spectral information. The matrix sensor’s resolution will allow diffuse reflectance measurements at a large number of source-to-detector - distances. Therefore, it will allow to probe the biological tissue at very high axial and lateral resolutions. The sequential surface-bottom reconstruction of OP will allow chromophores’ local concentrations’ quantitative estimation of an heterogeneous medium instead of considering such a biological medium as an homogeneous one like it was the case before. Since the sensor that will be used for the heterogeneous medium’s quantitative measurements of local optical parameters is a matrix, it will be also used for the wide-field MSI measurements. This type of coupled setup allows to consider it as an integrated device into several modalities.


Clinical applications of such a quantitative multi-scale spectroscopy device are numerous. Information from deep layers may be helpful when it is necessary to get rid of the effect of a first layer’s - such as epidermis - impact that may hide an identical underlying perfusion/hydration phenomenon if it is not taken into account. Depth sensitivity might be crucial in order to help physicians specify lateral and axial surgical resection limits when resecting skin cancers (carcinomas or melanomas). Finally, depth may be crucial for evaluating a drug’s therapeutic effect and help localize the therapeutic action’s site. Expected innovations during this PhD project are technological (development of non-contact DRS based on a matrix sensor) and methodological (evaluation of matrix sensor’s potential for a depth-resolved tissue exploration and optical parameters’ robust estimation). Moreover, when considering technical solution based on matrix DRS may allow the development of non-contact, non-invasive and simple technical solution. This PhD first year of work will be dedicated to showing that depth information is accessible through non-contact matrix DRS. An experimental setup will have to be developed on the basis of existing devices found in the CEA laboratory; current known limitations of such devices will have to be taken into account: illumination profile, parasitic reflections, integration time, the influence of the sample’s bending radius. In order to correctly size the experimental setup, the PhD student will be able to get help from Monte Carlo simulations adapted to the new acquisition geometry in a multi-layered medium. Therefore, the first year of work will be dedicated to conception and development of the matrix DRS acquisition chain. Measurement’s sensitivity to depth information will have to be quantified as a function of the medium’s optical parameters. During the second year of the PhD work, the local depth information will be used to solve the inverse problem, in other words estimate optical parameters’ values in a multi-layered model of the explored medium. Depth scanning may be gradual in order to evaluate optical parameters in the sequential layers thanks to the camera and lens couple’s pixel resolution used for probing medium’s depth. Experimental validation will be made (i) on synthetic phantoms mimicking diffusion and absorption properties of biological tissues of interest, then according to the progress report (ii) on hybrid phantoms i.e. made of synthetic materials and ex vivo tissues and/or on ex vivo and/or in vivo skin samples whose histological status will be assessed based on conventional histology analysis. The problematic of a multi-layered model quickly becomes highly complex; work is currently under progress in both managing teams that aims at solving the problem through sources’ localization approach and/or meta-modeling, in order to find a compromise between simplicity and accuracy. Third year will explore coupling depth estimation obtained on a local basis to a wide field multi-spectral image. It will be about demonstrating the diagnosis interest of the proposed device for instance on a deep quantitative oximetry measurement. Experimental validations may be realized in comparison to gold standard devices currently used in clinics (to measure transcutaneous oxygen pressure).
Keywords: Diffuse reflectance, multiscale spectroscopy, multispectral imaging, depth resolution
Conditions: - MSc graduation in physics/optics consolidated by skills in instrumentation/optoelectronics
- Good knowledge in programming : Matlab,
- Autonomous and go-ahead in bibliographic analysis, scientific communications (oral presentation, article writing…)
- Motivated by R&D of optical systems for medical and biotechnologies
- Able to develop skills in experimental work and data processing
Specific and complementary expertise will be brought by the supervision team and the partners.
Department(s):
Health - Biology - Signal
Financial aspects: NC (Region Grand Est or CONACYT or other)
Publications: hal-01246724v1, hal-01416842v1, hal-01246723v1, hal-01316456v1, hal-00757157v1    + CRAN - Publications