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.

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

PURPOSE: 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. MATERIALS AND METHODS: Relative SNR changes were theoretically calculated at given relaxation times (T1, T2) and mapped in RARE parameter space (TR, number of echoes, flip back pulse), at fixed acquisition times. T1 and T2 of KPF6 phantom samples (solution, agar mixtures, ex vivo perfused brain) were measured and experimental SNR values were compared with simulations, at optimal and suboptimal RARE parameter values. RESULTS: The optimized setting largely depended on T1, T2 times and the use of flip back pulse improved SNR up to 30% in case of low T1/T2 ratios. Relaxation times in different conditions showed negligible changes in T1 (below 14%) and more evident changes in T2 (-95% from water solution to ex vivo brain). Experimental data confirmed theoretical forecasts, within an error margin always below 4.1% at SNR losses of ~20% and below 8.8% at SNR losses of ~40%. The optimized settings permitted a detection threshold at a concentration of 0.5 mM, corresponding to 6.22 × 10¹6 fluorine atoms per voxel. CONCLUSION: Optimal settings according to measured relaxation times can significantly improve the sensitivity threshold in ¹9F MRI studies. They were provided in a wide range of (T1, T2) values and experimentally validated showing good agreement.