To evaluate the role of pretreatment intravoxel incoherent motion (IVIM) parameters in predicting treatment response in patients with locally advanced rectal cancer who underwent chemoradiation therapy (CRT), with postoperative pathological results as the reference standard. Forty-six patients (36 men, 10 women; mean age, 59.9 years; age range, 27-79 years) with rectal cancers underwent pelvis magnetic resonance imaging (MRI) examination before and after CRT. All pelvis MRI examinations were performed in 3.0T MR unit (Philips 3.0T Ingenia, Philips Medical System, The Netherlands) including high spatial resolution T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) sequences. Totally, eight b values (0, 25, 50, 75, 150, 400, 800 and 1000s/mm2) were used in DWI sequence, and IVIM parameters (D, pure diffusion; f, perfusion fraction; D*, pseudo-diffusion coefficient) were measured independently by two radiologists using bi-exponential model analysis. Patients were categorized as responders to CRT (patients with cancers characterized as T3 or T4 by the first MR imaging converted to lower stage at the second MR imaging or at pathological examination, patients with tumor volume reduction) or non-responders (patients with stable or progressive trend)[5]. The IVIM parameters between CRT responders and non-responders were compared by using independent samples t test or Mann-Whitney U test. Receiver operating characteristic (ROC) analysis was used to evaluate the diagnostic performance of IVIM parameters in predicting the response to CRT. The area under the ROC curve (AUC) and the optimal cut-off values were calculated, meanwhile accuracy rate, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) was determined. Interobserver agreement of IVIM parameter values were evaluated using Kappa values. P＜0.05 was considered to indicate a statistically significant difference.Interobserver agreement for IVIM parameters measurement was good with kappa values range from 0.683 to 0.856. There were 25 CRT responders and 21 non-responders. D values were significantly lower in CRT responders (D =[1.03±0.31]×10^-3mm²/s) than in non-responders (D =[1.25±0.24]×10^-3mm²/s), while D* and f values showed different trend (CRT responders: D*=[34.51±24.12]×10^-3mm²/s, [18.34±6.90]%; non-responders: D*=[14.35±13.19]×10^-3mm²/s, [16.81±11.33]%) ( Figure 1-2). According to ROC curve, D and D* values showed diagnostic significance with the AUC values of 0.773, 0.773, respectively. The cutoff values for D and D* were 1.052×10^-3mm²/s ( D values of CRT non-responders was greater than this value; accuracy rate 73.91%, sensitivity 80.95%, specificity 68.00%, PPV 68.00%, NPV 80.95% ), 23.516×10-3mm2/s (D* values of CRT responders was greater than this value; accuracy rate 73.91%, sensitivity 68.00%, specificity 80.95%, PPV 80.95%, NPV 68.00% ), respectively.Pretreatment IVIM parameters (D, D*) measurements were reproducible and may be used as a non-invasive index to evaluate response to CRT in locally advanced rectal cancer.