Abstract： Objective To investigate the ability of single photon emission computed tomography (SPECT) and MRI in detecting skull-base invasion in nasopharyngeal carcinoma. Methods Sixty-one patients with nasopharyngeal carcinoma received whole body and skull-base tomography SPECT, and nasopharynx and skull-base MRI before radiotherapy. The results were double-blind compared and evaluated. Results The overall positive rates of skull-base invasion detected by SPECT and MRI were 51% and 46% (P=0.508). In paitents with headache, cranial nerve palsy or both, the rates were 83% and 86% (P=1.000) ,80% and 80% (P=1.000), 88% and 94% (P=1.000), respectively. In patients with T1+T2 and T3+T4lesions,the rates were 22% and 0(P=0.031) ,74% and 82% (P=0.250) ,repectively. In patients with N0+N1and N2+N3lesions,they were 50% and 48% (P=1.000) ,53% and 40% (P=0.500) ,respectively. The conformation rate between SPECT and MRI was 85%. Binary Logistic regression analysis showed that T stage was a risk factor for positive SPECT(χ2=4.23,P=0.040, OR=3.04). Headache tended to be a risk factor for both positive SPECT and positive MRI (χ2=3.13, P=0.077, OR=4.54;χ2=3.64,P=0.056,OR=12.00). Conclusions The detection sensitivity of SPECT in skull-base invasion in nasopharyngeal carcinoma is equivalent to that of MRI. The consistency between SPECT and MRI is good. Moreover, there is a good correlation between SPECT and symptoms, signs and stage. SPECT of skullbase tomography is necessary for patients with severe headache, negative CT and those who can not receive MRI. When SPECT result is positive,skull-base should be considered to be invaded and should be defined as gross tumor volume in radiotherapy planning.