Elisabetta De Bernardi

Optimization of rapid acquisition with relaxation enhancement (RARE) pulse sequence parameters for 19F-MRI studies

To optimize signal‐to‐noise ratio (SNR) in fast spin echo (rapid acquisition with relaxation enhancement [RARE]) sequences and to improve sensitivity in 19F magnetic resonance imaging (MRI) on a 7T preclinical MRI system, based on a previous experimental evaluation of T1 and T2 actual relaxation times.

PET-CT scanner characterization for PET raw data use in biomedical research

The purpose of this paper is to describe the experiments and methods that led to the geometrical interpretation of new-generation commercial PET-CT scanners, finalized to off-line PET-based treatment verification in ion beam therapy. Typically, the geometrical correspondence between the image domain (i.e., the dicom PET) and the sinogram domain (i.e., the PET raw data) is not explicitly described by scanner vendors. Hence, the proposed characterization can be applied to commercial PET-CT scanners used in biomedical research, for the development of technologies and methods requiring the use of PET raw data, without having access to confidential information from the vendors.

Optimized Bayes variational regularization prior for 3D PET images.

A new prior for variational Maximum a Posteriori regularization is proposed to be used in a 3D One-Step-Late (OSL) reconstruction algorithm accounting also for the Point Spread Function (PSF) of the PET system. The new regularization prior strongly smoothes background regions, while preserving transitions. A detectability index is proposed to optimize the prior. The new algorithm has been compared with different reconstruction algorithms such as 3D-OSEM+PSF, 3D-OSEM+PSF+post-filtering and 3D-OSL with a Gauss-Total Variation (GTV) prior. The proposed regularization allows controlling noise, while maintaining good signal recovery; compared to the other algorithms it demonstrates a very good compromise between an improved quantitation and good image quality.

Background based Gaussian mixture model lesion segmentation in PET

PURPOSE Quantitative (18)F-fluorodeoxyglucose positron emission tomography is limited by the uncertainty in lesion delineation due to poor SNR, low resolution, and partial volume effects, subsequently impacting oncological assessment, treatment planning, and follow-up. The present work develops and validates a segmentation algorithm based on statistical clustering. The introduction of constraints based on background features and contiguity priors is expected to improve robustness vs clinical image characteristics such as lesion dimension, noise, and contrast level. METHODS An eight-class Gaussian mixture model (GMM) clustering algorithm was modified by constraining the mean and variance parameters of four background classes according to the previous analysis of a lesion-free background volume of interest (background modeling). Hence, expectation maximization operated only on the four classes dedicated to lesion detection. To favor the segmentation of connected objects, a further variant was introduced by inserting priors relevant to the classification of neighbors. The algorithm was applied to simulated datasets and acquired phantom data. Feasibility and robustness toward initialization were assessed on a clinical dataset manually contoured by two expert clinicians. Comparisons were performed with respect to a standard eight-class GMM algorithm and to four different state-of-the-art methods in terms of volume error (VE), Dice index, classification error (CE), and Hausdorff distance (HD). RESULTS The proposed GMM segmentation with background modeling outperformed standard GMM and all the other tested methods. Medians of accuracy indexes were VE <3%, Dice >0.88, CE <0.25, and HD <1.2 in simulations; VE <23%, Dice >0.74, CE <0.43, and HD <1.77 in phantom data. Robustness toward image statistic changes (±15%) was shown by the low index changes: <26% for VE, <17% for Dice, and <15% for CE. Finally, robustness toward the user-dependent volume initialization was demonstrated. The inclusion of the spatial prior improved segmentation accuracy only for lesions surrounded by heterogeneous background: in the relevant simulation subset, the median VE significantly decreased from 13% to 7%. Results on clinical data were found in accordance with simulations, with absolute VE <7%, Dice >0.85, CE <0.30, and HD <0.81. CONCLUSIONS The sole introduction of constraints based on background modeling outperformed standard GMM and the other tested algorithms. Insertion of a spatial prior improved the accuracy for realistic cases of objects in heterogeneous backgrounds. Moreover, robustness against initialization supports the applicability in a clinical setting. In conclusion, application-driven constraints can generally improve the capabilities of GMM and statistical clustering algorithms.

Regional MLEM reconstruction strategy for PET-based treatment verification in ion beam radiotherapy.

In ion beam radiotherapy, PET-based treatment verification provides a consistency check of the delivered treatment with respect to a simulation based on the treatment planning. In this work the region-based MLEM reconstruction algorithm is proposed as a new evaluation strategy in PET-based treatment verification. The comparative evaluation is based on reconstructed PET images in selected regions, which are automatically identified on the expected PET images according to homogeneity in activity values. The strategy was tested on numerical and physical phantoms, simulating mismatches between the planned and measured β+ activity distributions. The region-based MLEM reconstruction was demonstrated to be robust against noise and the sensitivity of the strategy results were comparable to three voxel units, corresponding to 6 mm in numerical phantoms. The robustness of the region-based MLEM evaluation outperformed the voxel-based strategies. The potential of the proposed strategy was also retrospectively assessed on patient data and further clinical validation is envisioned.

Proposal of a 4D ML reconstruction strategy for PET-based treatment verification in ion beam radiotherapy

The aim of this work is to propose an adaptation of a 4D Maximum Likelihood (ML) reconstruction strategy as a tool to improve the sensitivity of PET-based treatment verification in ion beam radiotherapy. PET images acquired during/shortly after the treatment (Measured PET) and an estimate of the same PET images derived from the treatment plan (Estimated PET) are considered as two frames of a 4D dataset. The algorithm iteratively estimates the annihilation events distribution in a reference frame and the deformation motion fields that map it in the Expected and Measured PET frames. Expected PET images can be then mapped into the Measured PET frame to verify the treatment. The details of the algorithm are presented and the strategy is preliminarily tested on an analytically simulated dataset. Convergence at different count statistics and ability to detect mismatches are assessed.

Clinical investigations of a 4D ML reconstruction strategy for PET-based treatment verification in ion beam therapy

In ion beam therapy, Positron Emission Tomography (PET) imaging can be applied to provide a consistency check of the delivered treatment with respect to the treatment plan. Treatment verification relies on the comparison between a PET distribution measured during or after treatment delivery (“measured PET”) and a PET simulation calculated on the basis of the treatment plan (“expected PET”). The method is challenged by the poor image quality of the measured PET, which may result in a low sensitivity to detectable mismatches between planned and applied treatment. In this work we propose the application of a motion-aware 4D Maximum Likelihood (ML) reconstruction strategy for PET-based treatment verification in ion beam therapy. This approach is specifically tested on simulated clinical-like scenarios, as well as on a real clinical dataset. The measured PET and the expected PET are interpreted as two different motion states of a 4D PET dataset. The idea is to estimate the deformation field mapping the expected PET onto the measured PET as a measure of mismatches. An enhanced measured PET is therefore obtained by warping the expected PET according to the estimated motion field. Clinical-like scenarios were reproduced by means of Monte Carlo simulations corresponding to a hypo-fractionated carbon ion treatment of liver tumour. The measured PET was attributed to different simulations introducing positioning mismatches by means of rigid translations, to be compared to the reference expected PET simulation without mismatch. Real clinical dataset include PET data acquired shortly after treatment at a dedicated PET-CT scanner installed at HIT (Heidelberg Ion beam Therapy Center, Germany), in comparison to the corresponding expected PET coming from Monte Carlo simulations. The accuracy of mismatch estimation and the robustness to noise are shown.

Reconstruction of uptake patterns in PET: The influence of regularizing prior

Purpose The effects of regularizing priors on the maximum likelihood (ML) reconstruction of activity patterns in Positron Emission Tomography (PET) were assessed. Methods Two edge‐preserving priors (one originally proposed by Nuyts et al. and nowadays implemented and commercialized by General Electric Medical Systems as Q.Clear software, and a second one originally proposed by Rapisarda et al. and our group) were assessed and compared to a standard Ordered Subset (OS)‐ML reconstruction, assumed as reference. The main difference between the two priors is that Nuyts prior (NY‐p) penalizes relative voxel differences while Rapisarda prior (RP‐p) absolute ones. Prior parameters were selected by imposing a reference noise texture inside uniform regions with activity comparable to that measured in 18F‐FluoroDeoxyGlucose (FDG) patient livers overall the field of view. Comparisons were then made: (a) on phantom data in terms of sphere recovery coefficients, ability to correctly reconstruct uniform irregularly shaped objects and heterogeneous patterns in patient backgrounds; (b) on patient data in terms of lesion detectability and image quality. Results On phantoms, both priors succeeded in improving all the assessed features with respect to standard OS‐ML reconstruction, mainly thanks to the better signal convergence and to the noise breakup control. On 10 mm spheres, an average recovery coefficient augment of 9% (NY‐p) and 34% (RP‐p) was obtained; homogeneity of uniform activity objects augmented of 4% (NY‐p) and 11% (RP‐p); accuracy in reconstructing heterogeneous lesions improved on average of 5% (NY‐p) and 15% (RP‐p). On patients, lesion detectability resulted improved (on 27 of 30 lesions), regardless of lesion anatomical districts and position in the scanner field of view. NY‐p provides a spatial resolution and a noise texture more uniform in the field of view and an image quality similar to standard OS‐ML. RP‐p has instead a behavior more dependent on the local counting statistics that imposes a trade‐off between spatial resolution uniformity and noise texture homogeneity. Conclusions The assessed regularizing priors improve PET uptake pattern reconstruction accuracy. Therefore, they should be considered both for oncological lesion detection and uptake spatial distribution assessment. Pitfalls and open challenges are also discussed.

Performances of Principal Component Analysis for the extraction of respiratory signal from Time-of-Flight PET coincidences stream

Recently Principal Component Analysis (PCA) was suggested as a potential way to extract motion signals (e.g: cardiac beat and respiratory signals) from the coincidences stream of the PET scan. Proofs of principle ensued. Aim: To assess minimum requirements of the signal for PCA to successfully recover it, in terms of temporal resolution, of total counts needed and of strength of the motion signal in the data. The use of Time-Of-Flight technology, to increase signal-to-background ratio was investigated to see whether it would give improved results. Materials and methods: A General Electrics Medical Systems Discovery-690 PET/CT was used. A phantom with an uniform background and hot spherical sources placed on a moving platform was used. Motion period was set to about 4 seconds and motion range at 10 mm and 20 mm. The motion was recorded with an RPM device. PCA was applied to obtain motion signals with 80, 160 and 320 ms sampling. It was applied to data relative to 327, 163, 81 and 41 seconds of total acquisition time. The correlation to the RPM signal, which for this phantom is virtually noiseless, was used to measure the power of PCA signal to be an effective indicator of the motion signal. PCA was applied to TOF, traditional non-TOF data and non-TOF data with a rejection for all events with a time signature indicating events outside a 40 cm diameter. Data were also analyzed for a cardiac scan of a patient, with samplings of 20, 40 and 80 ms, to try to recover both cardiac and respiratory signal. Results: Correlation coefficients of 0.80 or greater were found in all cases for the phantom with 20 mm motion. For the 10 mm motion markedly lower correlations are found. At 80 ms temporal resolution the correlations are too low or absent to allow motion signal extraction. High enough values (r>0.7) are found only at 320 ms sampling for 81 s (or longer) acquisitions or at 160 ms sampling for 327 s acquisitions. The use of TOF data did not improve results for this, relatively small, phantom. Nonetheless exploiting TOF to improve rebinning resolution and to reject random coincidences improved a bit the results.In the example patient analyzed both the cardiac and the respiratory signal could be extracted at 20 ms sampling with 2 minutes of total duration Conclusion: Principal Component Analysis proved to be effective as a tool to extract motion signal from PET coincidences stream. Sampling durations as fast as 160 ms for respiratory signals with few moving objects in the Field of View or 40 ms sampling for dual motion extraction with large motion signals in the FOV are feasible.

Radiomics of the primary tumour as a tool to improve 18 F-FDG-PET sensitivity in detecting nodal metastases in endometrial cancer

BackgroundA radiomic approach was applied in 18F-FDG PET endometrial cancer, to investigate if imaging features computed on the primary tumour could improve sensitivity in nodal metastases detection. One hundred fifteen women with histologically proven endometrial cancer who underwent preoperative 18F-FDG PET/CT were retrospectively considered. SUV, MTV, TLG, geometrical shape, histograms and texture features were computed inside tumour contours. On a first group of 86 patients (DB1), univariate association with LN metastases was computed by Mann-Whitney test and a neural network multivariate model was developed. Univariate and multivariate models were assessed with leave one out on 20 training sessions and on a second group of 29 patients (DB2). A unified framework combining LN metastases visual detection results and radiomic analysis was also assessed.ResultsSensitivity and specificity of LN visual detection were 50% and 99% on DB1 and 33% and 95% on DB2, respectively. A unique heterogeneity feature computed on the primary tumour (the zone percentage of the grey level size zone matrix, GLSZM ZP) was able to predict LN metastases better than any other feature or multivariate model (sensitivity and specificity of 75% and 81% on DB1 and of 89% and 80% on DB2). Tumours with LN metastases are in fact generally characterized by a lower GLSZM ZP value, i.e. by the co-presence of high-uptake and low-uptake areas. The combination of visual detection and GLSZM ZP values in a unified framework obtained sensitivity and specificity of 94% and 67% on DB1 and of 89% and 75% on DB2, respectively.ConclusionsThe computation of imaging features on the primary tumour increases nodal staging detection sensitivity in 18F-FDG PET and can be considered for a better patient stratification for treatment selection. Results need a confirmation on larger cohort studies.