The aim of the present study was to evaluate the thymuses of non-thymomatous myasthenia gravis (MG) patients by computed tomography (CT) for differentiating lymphoid follicular hyperplasia (LFH) thymus from normal/involuted thymus in order to assist surgeons in determining whether a non-thymomatous MG patient requires an operation. In the present retrospective review over 10 years, 80 patients who received CT scan and thymectomy at the Affiliated General Hospital of Tianjin Medial University (Tianjin, China) were included. According to the pathological records, 54 of the cases initially detected on CT were confirmed as LFH thymus. Thymic measurements, including anteroposterior and transverse dimensions, width (the longest axis of the lobe on a transverse scan) and thickness (the largest dimension perpendicular to the long axis of the lobe) and CT attenuation of the thymus region, adipose tissue and chest wall musculature in each CT slice were included to assess differences between the LFH group and the normal/involuted thymus group. Although a negative association between patient age and the CT attenuation of the thymus region was identified (r=-0.779, P<0.05, Pearson's correlation test), the LFH thymus group featured nodular changes on CT, while no such changes were observed in the normal/involuted thymus group. The mean age of disease onset in the LFH thymus group was significantly lower than that in the normal thymus group (40.2±17.3 vs. 59.2±9.3 years). Furthermore, significant differences in CT attenuation were identified between the LFH group and the normal/involuted thymus group [-41.21±54.42 vs.-108.23±8.72 Hounsfield units (HU) on unenhanced CT;-25.57±58.65 vs.-117.40±6.22 HU on contrast-enhanced CT]. In the LFH group, the difference in mean CT attenuation between the thymus region and adipose tissue was significant, while no significant difference was observed in the normal/involuted thymus group. In conclusion, CT may be used to distinguish LFH thymus from normal/involuted thymus in MG patients.