Nacim Betrouni
  • E-mail :[email]
  • Phone : 33 3 20 44 67 22
  • Location : Lille, France
Last update 2011-03-24 22:15:11.602

Nacim Betrouni Prostate cancer: Computer Aided Diagnosis and focal image guided therapies

Course and current status

Research Scientist

French National Institute of Health and Medical Research (INSERM)

Image Guided Therapies - U703

Scientific summary

In the recent years, tumors treatment techniques have known an important development especially in targeting the lesions while sparing the healthy tissues becoming increasingly conservative. However, such treatments require an accurate diagnosis and a precise planning. The purpose is to develop techniques for the detection, characterization and quantification for prostate cancer and associated tools to assist focused therapies by laser interstitial thermotherapy (LITT) and photodynamic therapy (PDT).

 Prostate cancer diagnosis by multi-parametric MR images

 Recently, prostate cancer diagnosis has experienced an important expansion through the combination of Prostatic Specific Antigen (PSA) and Digital Rectal Examination (DRE) guided by ultrasound. However, when dealing with small tumors, transrectal ultrasound-guided (TRUS-G) biopsies may sometimes fail to identify cancer. New Magnetic Resonance Imaging (MRI) techniques such as dynamic contrast-enhanced (DCE-MRI), spectroscopic (MRSI), and diffusion weighted imaging (DWI), improve pre-operative detection and local staging of prostate cancer.

They are usually combined because they provide complementary information. One of the important challenges in this field is the development of computer aided diagnosis (CAD) software based on these images for the optimal detection and staging of the cancer.

These developments need pre-treatments for the registration of the MR images and the zonal segmentation of the prostate (apex, base, left, right, peripheral zone and transition zone).

Firstly we were interested by perfusion MR images, we developed a software allowing 2D and multislice 2D contouring of suspicious areas based on a seeded growing region algorithm, and area labeling based on zonal anatomy. Tumor volume assessment, and semiquantitative analysis of DCE-MRI sequences can be performed. Based on the analysis of median wash-in and wash-out values of 121 malignant (n=74) or benign (n=47) prostatic lesions showing a suspicious high intensity and early enhancement at DCE-MRI, we designed a standardized 5-levels cancer suspicion score (ranging from "probably benign" to "Highly suspicious"). This comprehensive score provides a scaled likelihood of malignancy in the region of interest taking account of its location, according to prostate zonal anatomy. We compared its accuracy with the visual evaluation of time-intensity curves performed by specialized and nonspecialized radiologists.

In another work, we proposed a method based on T2 weighted MR images. We combined fractal and multifractal features to perform texture analysis of the images. Fractal dimension was computed using the Variance method and the multifractal spectrum estimated by an adaptation of a multifractional Brownian motion model. Voxels were labeled into tumor/non-tumor via non-linear supervised classification. Two classification algorithms were tested: Support Vector Machine (SVM) and AdaBoost.

Prostate cancer: treatment by laser therapies : LITT and PDT guided by ultrasound imaging


Laser interstitial thermotherapy (LITT) and photodynamic therapy (PDT) are two approaches used for laser treatment in oncology. The LITT heats and destroys deep tumors while minimizing the impact on healthy adjacent structures. As to PDT, it acts on the tumor neovascularization through the production of free intracellular radicals. Both methods have proved effective and are potentially part of the therapeutic arsenal. However, the conformation of the treatment to the tumor area remains a difficult task especially when, for large targets, several fibers must be used. Medical imaging plays an important role to assist the planning, fibers placement, per-operative control of the thermal evolution and post-operative monitoring.  Ultrasound imaging is one of the sole medical imaging modalities allowing all these features with low costs.

PDT and LITT are well designed for prostate because cancer is usually confined and techniques to perform interstitial treatment (as in brachytherapy) already exist and could be adapted. These 2 therapies may provide a second chance to heal after a local recurrence brachytherapy or radiotherapy. However, few teams are using them in clinical routine. This limitation is due to the technical difficulties underlying.


The aim of this axis is to develop these tools by using ultrasound imaging as a guidance tool. Developments include tracking tools, target deformations quantification methods and dosimetric planning. This part of the project constitutes the continuation of the previous team works in conformal radiotherapy of prostate cancer (Betrouni et al. CMIG 2005, Pasquier, Betrouni et al. IJROBP 2007, Pasquier et al. Can. Rad. 2008).

 Works in this field concern treatment planning and treatment guidance.

Treatment planning

This part is deduced to the development of software tools to optimize fibres positioning to ensure best coverage of the target.

Treatment guidance

We work on the integration of non-rigid registration algorithms to the planification platform to register MR planning images and ultrasound images. This registration will allow the urologist to set the fibres into the planned positions.

Prostate cancer: Contribution of the metabolic imaging

At present, Positron Emission Tomography (PET) imaging is an important modality in many cancers, but it is less useful in prostate cancer because glucose metabolism is less increased in this cancer than in most others. The sole radiopharmaceutical easily available is Fluoro-DeoxyGlucose (FDG) labeled by fluorine 18. It is a glucose analogue and its uptake is related to glucose metabolism.

Among some potentially useful tracers to explore prostate tumours, choline is one of the most promising. Choline was in a first step labeled by carbon 11 but the period of carbon 11 is short (20 minutes), so it can only be used by centre with cyclotron. In a second step, choline was labelled by fluorine 18, with a longer period (110 minutes), so its distribution is possible, with the same logistics than for 18F-FDG. Fluorocholine (or fluoro-methyl-choline) is now used in neighbouring countries, with publications of where more than 100 patients [Cimi2006, Husa 2008]. In France, TenonÕs team has an experience [Huch2007] and initiates a national PHRC (ICHOROPRO) in the evaluation of the usefulness of fluoro-choline in the detection of occult recurrences of prostate cancers.

The real place of fluoro-choline PET in prostate cancer is not yet clearly established. In the initial staging, in spite of increase of the resolution of PET scanners, its value seems limited and it could not avoid lymph node dissection for the research of micrometastases. In the identification of occult recurrences, it had proved its usefulness and can detect unknown local recurrence or/and distant metastases. The result of fluoro-choline PET has so a direct influence in the therapeutic strategy, according to the number and the localisation of detected lesions. If the sensitivity of fluoro-choline PET is function of PSA level, time-to-progression of PSA, initial GleasonÕs score and time to recurrence, a negative fluoro-choline PET seems be a factor of good prognosis and in CimitanÕs study [Cimi2006], any patient with a negative fluoro-choline PET has a recurrence 6 months later and the PSA level was stable.

PET is frequently used to evaluate therapy response and, at least with FDG, can detect an earlier response than with conventional imaging.

The aim is to study the potential usefulness of fluoro-choline PET as evaluation tool of the metabolic response to local treatments of prostate tumours with Laser interstitial thermotherapy (LITT) or Photodynamic therapy (PDT).


Image d’exemple