Glioblastoma are brain tumors currently incurable, however, optimized treatment gives better prognosis and quality of life. In case of surgical treatment, there is still need to help surgeons to determine whether a tissue is tumorous or not. Within the framework of the design of a new autofluorescence probe for this issue, optically calibrated gel phantoms have been developed using "tumorous" inclusions in a "healthy" environment. Depending on "tumor" shape, size and localization, the sensitivity of the probe is evaluated. The probe sensitivity for fluorescence spectroscopy will be presented. The probe configuration is also taken into account and compared to simulated results.

The survival outcome of patients suffering from gliomas is directly linked to the complete surgical resection of the tumour. To help the surgeons to delineate precisely the boundaries of the tumour, we developed an intraoperative positron probe with background noise rejection capability. The probe was designed to be directly coupled to the excision tool such that detection and removal of the radiolabelled tumours could be simultaneous. The device consists of two exchangeable detection heads composed of clear and plastic scintillating fibres. Each head is coupled to an optic fibre bundle that exports the scintillating light to a photodetection and processing electronic module placed outside the operative wound. The background rejection method is based on a real-time subtraction technique. The measured probe sensitivity for (18)F was 1.1 cps kBq(-1) ml(-1) for the small head and 3.4 cps kBq(-1) ml(-1) for the large head. The mean spatial resolution was 1.6 mm FWHM on the detector surface. The gamma-ray rejection efficiency measured by realistic brain phantom modelling of the surgical cavity was 99.4%. This phantom also demonstrated the ability of the probe to detect tumour discs as small as 5 mm in diameter (20 mg) for tumour-to-background ratios higher than 3:1 and with an acquisition time around 4 s at each scanning step. These results indicate that our detector could be a useful complement to existing techniques for the accurate excision of brain tumour tissue and more generally to improve the efficiency of radio-guided cancer surgery.

Autofluorescence spectroscopy from brain tissue may help to discriminate cancerous from healthy tissue. The characteristics of our probe are studied on phantoms and confronted to Monte Carlo simulations. Geometrical origins of fluorescence light are evaluated.

Using the POCI camera, we recently demonstrated the clinical impact of per-operative imaging techniques thanks to a successful clinical trial. Taking advantage of both the POCI experience and the availability of new pixelated detectors, we are developing a new hand held gamma camera TReCam (tumor resection camera). The first prototype offers a 49 × 49 mm2 field of view. It combines a 15 mm thick parallel hole collimator (with an efficiency of 5.10-4 and a spatial resolution of 7.5 mm at 50 mm), a 5 mm thick LaBr3:Ce crystal optically coupled to a Hamamatsu H9500 flat panel multianode photomultiplier tube (MAPMT). The read-out of the MAPMT is ensured thanks to a new specific integrated circuit called HARDROC2 (hadronic Rpc detector readout chip). This chip was initially designed for the digital hadronic calorimeter (DHCAL) of the linear collider. The TReCam presents an intrinsic spatial resolution of 0.65 mm (FWHM) and an energetic resolution of 17.4% at 122 keV. The new read-out by the HARDROC2 chip is currently being implemented and using this new device energetic resolution is expected to be better.

Surgery is considered as the primary therapeutic procedure for gliomas and several recent clinical studies have shown that total tumor resection is directly associated with longer survival when compared to subtotal resection. In order to refine the resection in the boundaries of gliomas, we are developing an intraoperative probe specifically dedicated to the localization of residual tumor labeled with positron emitters. The probe is designed to be compact and electrically safe in order to be directly coupled to the excision tool leading to simultaneous detection and removal of tumor tissues. It is built with clear and plastic scintillating fibers held in a closed packed annular arrangement ensheathing the excision tool. The annihilation gamma ray background is eliminated by a real-time subtraction method. Validation of the technical choice and optimization of the probe geometry were performed by preliminary measurements and Monte Carlo simulations based on the MCNP-4C code and an anthropomorphic brain phantom. The theoretical probe sensitivity was found to be 82 cps/muCi/ml with a gamma ray rejection efficiency of 99.6%. The expected minimum radiotracer detectable concentration for tumors labeled with 18 F-FET was 0.10 muCi/ml. When compared to the 0.29 muCi/ml average concentration in the bulk of the tumor, this result demonstrate the potential ability of the probe to define more accurately the extent of brain tumor resection

Objective. This study investigated a new technique for automatic model-based segmentation of broadband ultrasound attenuation (BUA) images of the calcaneus. We determined whether this technique was able to improve osteoporotic fracture discrimination. Methods. The segmentation process included 2 major steps: a model-building stage and the automatic segmentation of new image data sets via an elastic deformation of contour models. Broadband ultrasound attenuation was then averaged within the final contour (BUA(whole)). The results of the segmentation were validated on a database of 256 patients by comparison of the clinical results obtained with the automatic circular region of interest (BUA(circ)) currently implemented on a commercially available ultrasonography unit. All patients were selected by the same physician, who assessed that the fractures were caused by bone fragility on the basis of the circumstances under which fractures occurred and radiologic data. Results. Short-term reproducibility assessed in 49 women was 3.5% and 3.98% for BUA(circ) and BUA(whole), respectively. Both BUA(circ) (age-adjusted T score, -3.78; P < .0005, age-adjusted odds ratio, 1.92; 95% confidence interval, 1.34-2.75; area under the receiver operating characteristic curve, 0.70) and BUA(whole) (age-adjusted T score, -2.73; P < .01, age-adjusted odds ratio, 1.57; 95% confidence interval, 1.12-2.21; area under the curve, 0.67) performed equally well in discriminating healthy postmenopausal patients (n = 150) from those with fractures (n = 60). Conclusions. Fully automatic segmentation by parametrically deformable elastic models for contour using Fourier descriptors can be achieved with reasonable reproducibility and fracture risk prediction. The method is similar to existing methods (automatic circular region of interest), however, the new contour-based region of interest allows more flexible region of interest geometries and placement and potential adaptation to individual anatomy. The method could also possibly be extended to quantitative ultrasonographic imaging at different skeletal sites.

La précision diagnostique de l’échoendoscopie dans le bilan d’extension ganglionnaire des cancers digestifs pourrait être améliorée par l’analyse paramétrique de la texture des images. Nous avons cherché à évaluer les résultats de cette méthode dans une étude prospective.

This study investigates a new technique for automatic model-based segmentation of quantitative ultrasound images of the calcaneus. The segmentation process included two major steps : a model building stage derived from interactive segmentation of a sample set of 52 training images and the automatic segmentation of new image data sets via an elastic deformation of contour models. The clinical validation on a database of 256 patients showed that fully automatic segmentation by parametrically deformable elastic models for contour using Fourier descriptors can be achieved with reasonable reproducibility and fracture risk prediction.

For absorptiometry measurements, soft tissue may have an impact on quantitative ultrasound (QUS) measurements. ill the present study, we focused primarily on the quantification of measurement error on speed of sound (SOS) caused by surrounding soft tissue. To meet our goal, SOS measurements were taken at the right heel using a QUS imaging device in 21 healthy subjects. Site-matched measurements of soft tissue thickness (STT) and bone width were performed using magnetic resonance imaging (MRI) of the heel. Several bone velocities were calculated either by accounting for bone width (SOSBW) only or by taking into account the exact path lengths of all major components traversed by ultrasound (V-b). Given that soft tissue composition is difficult to determine in vivo, we chose to estimate lower and upper error bounds on bone velocity (V-b (lower) and V-b (upper)) by spanning the full range of available values in the literature. Accounting for true BW only resulted in no significant difference between SOS (1533 +/- 37) and SOSBW (1531 +/- 33). By contrast, accounting for both true BW and surrounding soft tissue resulted in an increase in the calculated bone velocity and statistically significant differences between SOS and V-b (upper) (1568 +/- 36) and V-b (lower) (1542 +/- 34). Root mean square errors between SOS and the calculated velocities were 0.34, 2.32, and 0.70% for SOSBW, V-b (upper) and V-b (lower) respectively.

The ability of computerized parameters to discriminate benign from malignant breast nodules from digitized ultrasonic acquisitions has been assessed. The images of 75 lesions, including 19 lesions proved to be malignant at histology and 56 found to be benign, were digitized and characterized by morphometric and texture parameters. The texture parameters, derived from first-order statistics, run-length matrices and co-occurrence matrices, were computed in the entire lesion and in a ring-shape region surrounding the contour of the lesion. The strongest features were found to be issued from the second region. Further investigations confirmed that the discriminant information was contained in the external part of the lesion and, to a lesser extent, in the neighboring tissue. A linear discriminant analysis using three features yielded a sensitivity of 94.7% for a specificity of 80.4% and the "leave-one-out" technique confirmed the results. Comparison with the classifications given by radiologists let us assume that information revealed by texture features is able to help the physician in reducing the number of unnecessary biopsies.

Background: The sensitivity of visual interpretation of endoscopic ultrasonography (EUS) images in pancreatic diseases does not exceed 70-80%, especially for discriminating cancer and pancreatitis. Aims : to assess the ability of ultrasound image texture analysis to improve diagnostic results and to provide objective diagnostic criterion. Methods : 73 patients were included. The final diagnosis was obtained from patient's history and follow-up, and histology when available. There were 30 cases of normal pancreas, 22 cases of chronic pancreatitis and 21 cases of pancreatic cancer. Images were captured during EUS and computerized on a PC, then transferred on workstation for analysis. 127 Regions of Interest (ROI) were selected. 39 parameters were studied for each ROI (2 parameters of the 1st order, 30 from from cooccurrence matrices, and 7 from rangelength matrices). Parameters were analyzed by logistic regression, then by multivariate analysis with the 4 most significant parameters, for discriminating patients in 2 groups : normal pancreas vs pathological pancreas and cancer vs non-cancer. Results : Analysis of variance showed that the 39 parameters were able to discriminate one group from the 2 others. 7 parameters were able to discriminate groups by pairs. For discriminating the normal pancreas from the pathological pancreas (pancreatitis + cancer), sensitivity was 93%, specificity was 80%, and diagnostic accuracy was 87.7%. For discriminating pancreatic cancer from non-cancer pancreas, se was 90.5%, sp was 84.6% and diagnostic accuracy was 86.3%. Conclusion : These results show that texture measurements obtained in patients with a normal pancreas, a cancer of the pancreas or a pancreatitis are significantly different, and that texture analysis is likely to improve the diagnostic capability of EUS.

This paper presents a computerized method for automated detection of the boundary of the os calcis on in vivo ultrasound parametric images, using an active dynamic contour model. The initial contour, defined without user interaction, is an iso-contour extracted from the textural feature space. The contour is deformed through the action of internal and external forces, until stability is reached. The external forces, which characterize image features, are a combination of gray-level information and second-order textural features arising from local cooccurrence matrices. The broadband ultrasound attenuation (BUA) value is then averaged within the contour obtained. The method was applied to 381 clinical images. The contour was correctly detected in the great majority of the cases. For the short-term reproducibility study, the mean coefficient of variation was equal to 1.81% for BUA values and 4.95% for areas in the detected region. Women with osteoporosis had a lower BUA than age-matched controls (p = 0.0005). In healthy women, the age-related decline was -0.45 dB/MHz/yr. In the group of healthy post-menopausal women, years since menopause, weight and age were significant predictors of BUA. These results are comparable to those obtained when averaging BUA values in a small region of interest.

Three-dimensional (3D) high-resolution ultrasonography has proved to be useful for in vitro assessment of cartilage remodeling due to osteoarthritis. The diagnosis is performed by computation of the mean thickness of the cartilage, which reveals hypertrophy or thinning, and by 3D reconstruction of the data, which provides essential information about the size, extent, and localization of the lesion. In both cases, preliminary segmention of the cartilage is necessary. This article proposes an algorithm for automatic segmentation of the cartilage from 3D ultrasonic acquisitions of the rat patella, which includes the detection of the cartilage surface and the cartilage/bone interface. The method was designed on the assumption of regularity and smoothness of the interfaces. The use of a global threshold was sufficient to separate the patella area from the background. The cartilage/bone interface was detected by selection of regions of interest (ROIs) encompassing the interface, followed by the detection of the interface within these ROIs using the graph theory. The method was applied to 162 samples. The detection accuracy was judged to be very good or good in 99% of the cases for the cartilage surface and in 86% of the cases for the cartilage/bone interface. The mean cartilage thickness value in the central part of the patella obtained from the automatic detection method was compared to that obtained manually. The coefficient of correlation between the two measurements was 0.92. These results show that our method is reliable. Thus, fast processing of a large number of acquisitions and a more complete analysis of the cartilage surface become possible.