BACKGROUND: Apparent diffusion coefficient (ADC) histogram analyses have been used to differentiate tumor grades and predict therapeutic responses in various anatomic sites with moderate success. PURPOSE: To determine the ability of diffusion-weighted imaging (DWI) with a whole-tumor ADC histogram analysis to differentiate benign peripheral neurogenic tumors (BPNTs) from soft tissue sarcomas (STSs). STUDY TYPE: Retrospective study, single institution. SUBJECTS: In all, 25 BPNTs and 31 STSs. FIELD STRENGTH/SEQUENCE: Two-b value DWI (b-values = 0, 1000s/mm2 ) was at 3.0T. ASSESSMENT: The histogram parameters of whole-tumor for ADC were calculated by two radiologists and compared between BPNTs and STSs. STATISTICAL TESTS: Nonparametric tests were performed for comparisons between BPNTs and STSs. P < 0.05 was considered statistically significant. The ability of each parameter to differentiate STSs from BPNTs was evaluated using area under the curve (AUC) values derived from a receiver operating characteristic curve analysis. RESULTS: The mean ADC and all percentile parameters were significantly lower in STSs than in BPNTs (P < 0.001-0.009), with AUCs of 0.703-0.773. However, the coefficient of variation (P = 0.020 and AUC = 0.682) and skewness (P = 0.012 and AUC = 0.697) were significantly higher in STSs than in BPNTs. Kurtosis (P = 0.295) and entropy (P = 0.604) did not differ significantly between BPNTs and STSs. DATA CONCLUSION: Whole-tumor ADC histogram parameters except kurtosis and entropy differed significantly between BPNTs and STSs. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.
BACKGROUND: Apparent diffusion coefficient (ADC) histogram analyses have been used to differentiate tumor grades and predict therapeutic responses in various anatomic sites with moderate success. PURPOSE: To determine the ability of diffusion-weighted imaging (DWI) with a whole-tumor ADC histogram analysis to differentiate benign peripheral neurogenic tumors (BPNTs) from soft tissue sarcomas (STSs). STUDY TYPE: Retrospective study, single institution. SUBJECTS: In all, 25 BPNTs and 31 STSs. FIELD STRENGTH/SEQUENCE: Two-b value DWI (b-values = 0, 1000s/mm2 ) was at 3.0T. ASSESSMENT: The histogram parameters of whole-tumor for ADC were calculated by two radiologists and compared between BPNTs and STSs. STATISTICAL TESTS: Nonparametric tests were performed for comparisons between BPNTs and STSs. P < 0.05 was considered statistically significant. The ability of each parameter to differentiate STSs from BPNTs was evaluated using area under the curve (AUC) values derived from a receiver operating characteristic curve analysis. RESULTS: The mean ADC and all percentile parameters were significantly lower in STSs than in BPNTs (P < 0.001-0.009), with AUCs of 0.703-0.773. However, the coefficient of variation (P = 0.020 and AUC = 0.682) and skewness (P = 0.012 and AUC = 0.697) were significantly higher in STSs than in BPNTs. Kurtosis (P = 0.295) and entropy (P = 0.604) did not differ significantly between BPNTs and STSs. DATA CONCLUSION: Whole-tumor ADC histogram parameters except kurtosis and entropy differed significantly between BPNTs and STSs. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.