Apparent diffusion coefficient histogram parameters of the whole tumor based on whole-body magnetic resonance imaging can better characterize the tumor burden of multiple myeloma and predict survival prognosis
TL;DR
Whole tumor histogram parameters can better characterize the tumor burden of multiple myeloma (MM) patients than those of the focal lesion groups and the Cox regression risk model based on the whole tumor ADC histogram parameters and clinical parameters can predict the survival status.
Abstract
Background Newly diagnosed multiple myeloma (NDMM) is characterized by a spatially heterogeneous tumor burden. Whole-body magnetic resonance imaging (WB-MRI)-derived apparent diffusion coefficient (ADC) histogram analysis may enable comprehensive characterization of tumor burden. This study aimed to investigate the feasibility of using whole-tumor ADC histogram parameters based on WB-MRI images to assess tumor burden and evaluate their prognostic value in NDMM patients. Methods There were 95 NDMM patients categorized into focal lesion and whole-tumor groups in this retrospective study. Correlation analysis of ADC histogram parameters with clinical parameters was performed in the two groups respectively. In the survival analysis section, 64 patients were included, Cox regression analysis was used to conduct univariate and multivariate analysis of potential risk factors for overall survival (OS). X-tile plots were used to determine the optimal cutoffs based on the highest χ2 value defined by Kaplan-Meier survival analysis and log-rank test. Results Apart from ADCmax, there was statistically significant difference in the remaining ADC histogram parameters between two groups (all P<0.05). In the whole-tumor group, ADCmean (r=0.434, P<0.01), ADC5th (r=0.334, P<0.01), ADC10th (r=0.371, P<0.01), ADC25th (r=0.416, P<0.01), ADC50th (r=0.431, P<0.01), ADC75th (r=0.424, P<0.01), ADC90th (r=0.398, P<0.01), and ADC95th (r=0.402, P<0.01) were positively correlated with bone marrow plasma cell (BMPC). Multivariate Cox regression showed that BMPC (P=0.022), β2-MG (P=0.016), ADC25th (P=0.023), ADC95th (P<0.001) and entropy (P=0.001) were all independent prognostic factors for poor OS. X-tile plots and the Kaplan-Meier survival curves showed that BMPC (>70%), β2-MG (>10.68 mg/L), ADC25th (>0.37 mm2/s), ADC95th (>1.03 mm2/s) and low entropy (≤3.56) achieved worse prognosis. Conclusions Whole tumor histogram parameters can better characterize the tumor burden of multiple myeloma (MM) patients than those of the focal lesion groups. The Cox regression risk model based on the whole tumor ADC (wtADC) histogram parameters and clinical parameters can predict the survival status.