Objective:To evaluate the application value of fat-suppressed T2WI sequence in Modic changes at a 1.5T magnetic Resonance (MR) system. Methods:Total 276 cases from November 2013 to March 2014 with lumbocrural pain had been performed MR scan with T1WI,T2WI and T2WI-FS. A retrospective analysis of Modic classification was applied on these data. Two musculoskeletal radiologists used two methods to classify Modic changes with the double-blind method:one according to T1WI and T2WI,and the another according to T1WI,T2WI and T2WI-FS. Results:Doctor A and B observed the Modic changes using method 1 and 2,respectively. The coincidence rate between two observers were 98.66%,Kappa=0.925 (method 1,P < 0.001);99.49% and Kappa=0.975(method 2,P < 0.001). Modic changes were divided into type Ⅰ,Ⅱ,Ⅲ and mixed. There were statistically significant differences between the two methods about the detection rate. Meanwhile,mixed type was further divided into 7 subtypes,namely typeⅠ/Ⅱ,typeⅡ/Ⅰ,type Ⅲ/Ⅱ,type Ⅱ/Ⅲ,typeⅠ/Ⅲ,type Ⅲ/Ⅰand typeⅠ/Ⅱ/Ⅲ. Among them,typeⅠ/Ⅱ and type Ⅱ/Ⅰconstituted the majority. Conclusion:Fat-suppressed T2WI sequence combined with T1WI and T2WI could increase the positive detection rate of Modic I type and Modic mixed type. It is superior to traditional T1WI and T2WI sequence in the classification of Modic changes. The mixed type can also be further classified into different subtypes according to the signs of MRI.