Trainee Project
Title:
Multimodal spectro-imaging and machine learning methods for bio-optical characterization of kidney stones in
the UV-Visible-NIR-FIR-THz ranges
Dates:
2024/03/04 - 2024/09/04
Supervisor(s):
Other supervisor(s):
Alexandre Locquet (alocquet@georgiatech-metz.fr)
Description:
Medical and scientific context
Kidney stones are hard crystals developing in kidneys and urinary tracts (lithiasis disease). To destroy these
stones, during the endoscopic examination, the clinician uses a laser (low frequency, few Watts) which will
fragment or disintegrate them into dust. Once extracted, the fragments are analyzed with microscope and
infrared spectrophotometer (SPIR) in order to determine their biochemical composition and consequently
establish their metabolic origins i.e. the appropriate treatments for the patient (diet, drug, etc.). However, with
technical progress, the destruction of the stones is increasingly based on dusting: in this case, stones are
vaporized into such small pieces that usual morpho-functional analysis is no longer possible.
Thus, an in situ biophotonics characterization of the kidney stones, that is during endoscopy and before
dusting, would be a major alternative and innovation for the identification of stones but also for time saving
(real-time knowledge of the stone composition) allowing the clinician to propose appropriate treatments
immediately after the surgery, without waiting SPIR results for several weeks.
The biophotonics characterization of kidney stones using new in vivo spectro-imaging methods is
therefore a prior and major issue to address in the development of new solutions compatible with clinical
translation. In this frame, the originality of our research project is to study the potential of two complementary
methods, hyperspectral imaging (HIS) in the visible spectral range [400-900]nm and terahertz time-domain
spectroscopy and pulsed imaging featuring a [0.1-5]THz spectral bandwidth, for the biophotonics
characterization and classification of the different types of stones. This project gathers the competence and
expertise of (i) the urology department at CHRU Nancy in ureteroscopy, laser treatment and biological analysis
of kidney stones, (ii) the Photonics and Terahertz group at GeorgiaTech-CNRS (GT-CNRS) in THz
spectroscopy/imaging instrumentation and data analysis and (iii) the CRAN team in UV-Vis-NIR spectro-
imaging, light-tissue interactions and data processing/analysis (spectra, images, feature extraction/selection,
classification).
The main objectives of the internship are:
- Conduct a detailed analysis of the scientific literature on existing studies related to kidney stone HIS and
THz imaging
- Taking charge of the experimental systems Hyperspectral imaging and Double Integrating spheres (CRAN)
and THz imaging (GT-CNRS) including calibration procedures
- Based on the feasibility test protocol validated on a preliminary (limited) set of stones, define then
conduct an experimental protocol of measurements to build an extended database
- Propose and implement algorithmic solutions (on Matlab) to process HS images and THz signals acquired
in order to extract discriminant features allowing for an efficient automatic classification (machine learning) of
the various types of kidney stones.
Kidney stones are hard crystals developing in kidneys and urinary tracts (lithiasis disease). To destroy these
stones, during the endoscopic examination, the clinician uses a laser (low frequency, few Watts) which will
fragment or disintegrate them into dust. Once extracted, the fragments are analyzed with microscope and
infrared spectrophotometer (SPIR) in order to determine their biochemical composition and consequently
establish their metabolic origins i.e. the appropriate treatments for the patient (diet, drug, etc.). However, with
technical progress, the destruction of the stones is increasingly based on dusting: in this case, stones are
vaporized into such small pieces that usual morpho-functional analysis is no longer possible.
Thus, an in situ biophotonics characterization of the kidney stones, that is during endoscopy and before
dusting, would be a major alternative and innovation for the identification of stones but also for time saving
(real-time knowledge of the stone composition) allowing the clinician to propose appropriate treatments
immediately after the surgery, without waiting SPIR results for several weeks.
The biophotonics characterization of kidney stones using new in vivo spectro-imaging methods is
therefore a prior and major issue to address in the development of new solutions compatible with clinical
translation. In this frame, the originality of our research project is to study the potential of two complementary
methods, hyperspectral imaging (HIS) in the visible spectral range [400-900]nm and terahertz time-domain
spectroscopy and pulsed imaging featuring a [0.1-5]THz spectral bandwidth, for the biophotonics
characterization and classification of the different types of stones. This project gathers the competence and
expertise of (i) the urology department at CHRU Nancy in ureteroscopy, laser treatment and biological analysis
of kidney stones, (ii) the Photonics and Terahertz group at GeorgiaTech-CNRS (GT-CNRS) in THz
spectroscopy/imaging instrumentation and data analysis and (iii) the CRAN team in UV-Vis-NIR spectro-
imaging, light-tissue interactions and data processing/analysis (spectra, images, feature extraction/selection,
classification).
The main objectives of the internship are:
- Conduct a detailed analysis of the scientific literature on existing studies related to kidney stone HIS and
THz imaging
- Taking charge of the experimental systems Hyperspectral imaging and Double Integrating spheres (CRAN)
and THz imaging (GT-CNRS) including calibration procedures
- Based on the feasibility test protocol validated on a preliminary (limited) set of stones, define then
conduct an experimental protocol of measurements to build an extended database
- Propose and implement algorithmic solutions (on Matlab) to process HS images and THz signals acquired
in order to extract discriminant features allowing for an efficient automatic classification (machine learning) of
the various types of kidney stones.
Keywords:
Photodiagnostics, kidney stones, tissue optics, hyperspectral imaging, terahertz time-domain spectro
Conditions:
Starting date: February, March or April-2024
Duration: 5-6 months
Funding: Laboratory CRAN (~660 ¬ / month)
Laboratories:
1. CRAN, UMR 7039 CNRS-UL, Département BioSIS
2. Photonics and Terahertz group, IRL 2958 GeorgiaTech-CNRS, Georgia Tech Europe
Locations:
1. CRAN, Site Brabois-Santé, Vandoeuvre-lès-Nancy
2. GeorgiaTech-CNRS, 2 rue Marconi, Metz
Supervision:
- Prof. Walter Blondel, walter.blondel@univ-lorraine.fr
- Dr. Alexandre Locquet, alocquet@georgiatech-metz.fr
- Clarice Perrin-Mozet, clarice.perrin-mozet@univ-lorraine.fr
- Dr. Marine Amouroux, marine.amouroux@univ-lorraine.fr
- Prof. Christian Daul, christian.daul@univ-lorraine.fr
Cooperations:
- Urology department at CHRU in Nancy : Prof. Jacques Hubert, Dr. Jonathan Elbeze, Dr. Maatem Caillerez
- Ecole Nationale Supérieure de Géologie de Nancy (ENSG) : Arnaud Marotel
Knowledge and skills required for application:
- Final year MSc student in signal/image processing or/and biomedical engineering
- Expected technical and scientific skills:
o Numerical signal/data processing and analysis
o Strong experience in MATLAB® programming
o Knowledge of the biomedical and/or medical imaging fields
- Very good level in English (speaking and writing)
- Autonomous, go-ahead and proactive
- Motivated by the development of innovative solutions for the field of health engineering
- Able to synthesize and exploit data from different partners associated in the project
Duration: 5-6 months
Funding: Laboratory CRAN (~660 ¬ / month)
Laboratories:
1. CRAN, UMR 7039 CNRS-UL, Département BioSIS
2. Photonics and Terahertz group, IRL 2958 GeorgiaTech-CNRS, Georgia Tech Europe
Locations:
1. CRAN, Site Brabois-Santé, Vandoeuvre-lès-Nancy
2. GeorgiaTech-CNRS, 2 rue Marconi, Metz
Supervision:
- Prof. Walter Blondel, walter.blondel@univ-lorraine.fr
- Dr. Alexandre Locquet, alocquet@georgiatech-metz.fr
- Clarice Perrin-Mozet, clarice.perrin-mozet@univ-lorraine.fr
- Dr. Marine Amouroux, marine.amouroux@univ-lorraine.fr
- Prof. Christian Daul, christian.daul@univ-lorraine.fr
Cooperations:
- Urology department at CHRU in Nancy : Prof. Jacques Hubert, Dr. Jonathan Elbeze, Dr. Maatem Caillerez
- Ecole Nationale Supérieure de Géologie de Nancy (ENSG) : Arnaud Marotel
Knowledge and skills required for application:
- Final year MSc student in signal/image processing or/and biomedical engineering
- Expected technical and scientific skills:
o Numerical signal/data processing and analysis
o Strong experience in MATLAB® programming
o Knowledge of the biomedical and/or medical imaging fields
- Very good level in English (speaking and writing)
- Autonomous, go-ahead and proactive
- Motivated by the development of innovative solutions for the field of health engineering
- Able to synthesize and exploit data from different partners associated in the project
Department(s):
| Biology, Signals and Systems in Cancer and Neuroscience |
Funds:
LUE Programme interdisciplinaire
Publications: