Intravoxel incoherent motion (IVIM) for response assessment in patients with osteosarcoma undergoing neoadjuvant chemotherapy

Abstract

Purpose To explore the role of quantitative Intravoxel Incoherent Motion (IVIM) parameters and their histogram analysis in characterizing changes in osteosarcoma undergoing neoadjuvant chemotherapy (NACT), and to evaluate their potential in assessing therapeutic response. Methods Forty patients (N = 40; Male:Female = 30:10; mean age = 17.7 ± 5.9 years; Metastatic:Localized = 17:23) with histologically confirmed osteosarcoma treated with three cycles of NACT were prospectively analyzed. Diffusion-weighted imaging (DWI) using 11 b-values (0–800 s/mm²) was performed on a 1.5 T MRI scanner at three time-points: pre-treatment (t0), after one NACT cycle (t1), and after three cycles (t2). Non-invasive response evaluation was performed using RECIST 1.1 criteria. Quantitative DWI and IVIM parameters—Apparent Diffusion Coefficient (ADC), true diffusion coefficient (D), pseudodiffusion coefficient (D*), and perfusion fraction (f)—as well as their relative percentage changes (Δ1: t0–t1; Δ2: t0–t2) were evaluated. Histogram analysis was performed at all three time-points and compared across RECIST 1.1 response groups. Results According to RECIST 1.1, 11 patients (27.5%) were partial responders (PR), 21 (52.5%) had stable disease (SD), and 8 (20%) had progressive disease (PD). At baseline (t0), mean tumor ADC, D, D*, and f were 1.36 ± 0.33 × 10⁻³ mm²/s, 1.3 ± 0.3 × 10⁻³ mm²/s, 28.44 ± 10.34 × 10⁻³ mm²/s, and 13.95 ± 2.83%, respectively. Significant differences among response groups were observed during treatment: at t1 and t2, D*-variance (p = 0.038, 0.003) and f-skewness (p = 0.03, 0.03); at t2, D*-entropy (p = 0.001) and f-entropy (p = 0.002) and their Δ2 changes (p = 0.001, 0.003). At t1, the combination of D*-variance and f-skewness showed diagnostic performance in classifying PR (AUC = 0.77; Sensitivity = 73%; Specificity = 70%) and SD (AUC = 0.74; Sensitivity = 74%; Specificity = 75%). Correlation analyses showed moderate associations (r = 0.5–0.6) between Δ1 and Δ2 changes in D*-mean, D*-variance, D*-entropy, and f-entropy with tumor diameter and volume changes. Conclusions Quantitative IVIM parameters, particularly D* and f, and their histogram-based metrics (variance, entropy, skewness) provide useful, non-invasive biomarkers for early assessment of response to NACT in osteosarcoma.