Fitting intravoxel incoherent motion model to diffusion MR signals of the human breast tissue using particle swarm optimization

Abstract

Intravoxel incoherent motion (IVIM) modeling offers the parameters f, D and D *from diffusion MR signals as biomarkers for different lesion types and cancerstages. Challenges in fitting the model to the signals using the availableoptimization algorithms motivate new studies for improved parameter estimations.In this study, one thousand value sets of f, D, D * for human breast tissue areassembled and used to generate five thousand diffusion MR signals consideringnoise-free and noisy situations exhibiting signal-to-noise ratios (SNR) of 20, 40,80 and 160. The estimates of f, D, D * are obtained using Levenberg-Marquardt(LM), trust-region (TR) and particle swarm (PS) algorithms. On average, thealgorithms provide the highest fitting performance for the noise-free signals(R 2adj =1.000) and great fitting performances for the noisy signals with SNR>20(R 2adj >0.988). TR algorithm performs slightly better for SNR=20 (R 2adj =0.947).TR and PS algorithms achieve the highest parameter estimation performance forall the parameters, while LM algorithm reveals the highest performance for f andD only on the noise-free signals (r=1.00). For the noisy signals, performancesincrease with an increase in SNR. All algorithms accomplish poor performancesfor D * (r=0.01-0.20) while TR and PS algorithms perform same for f (r=0.48-0.97)and D (r=0.85-0.99) but remarkably better than LM algorithm for f (r=0.08-0.97)and D (r=0.53-0.99). Overall, TR and PS algorithms demonstrate better butindistinguishable performances. Without requiring any user-given initial value, PSalgorithm may facilitate improved estimation of IVIM parameters of the humanbreast tissue. Further studies are needed to determine its benefit in clinical practice.