The aim of this study was to evaluate the therapeutic effect more accurately and predict the prognosis of treated non-small cell lung cancer by using contrast-enhanced magnetic resonance imaging (CE-MRI). Contrast-enhanced computed tomography (CE-CT) and CE-MRI were examined 90 non-small cell lung cancer patients treated with conservative therapies. Enhancement patterns of CE-MRI were classified into three types: peripheral; mottled; and homogeneous. Reduction ratio of tumor size (RRT) based on the World Health Organization response criteria and a new response rate; reduction ratio of viable tumor size (RRVT) which evaluates not only the reduction of tumor size but also changes in necrosis and/or cavity size, were evaluated. Changes of enhancement pattern were compared and correlated with pathological diagnosis. The RRTs, RRVTs, and interobserver agreements evaluated by all modalities were compared. The RRTs and RRVTs in each subgroup were correlated and compared with prognoses. Change of enhancement pattern depended on therapy had no tendency (p = 0.06). Enhancement pattern had significant correlation with pathological diagnosis (p < 0.0001). Partial response (PR) case of RRVT had significant difference between imaging techniques (p = 0.04). The RRVT of other cases and RRT had no significant difference. Interobserver agreements of RRT and RRVT were almost perfect (n > or = 0.93). Prognosis had better correlation with RRVT than with RRT. Differences of relapse-free survival and survival between patients considered as having no change (NC) by RRT and PR by RRVT (NC-PR) and patients considered as having NC by RRT and RRVT were significant (p = 0.03, p = 0.01). There were no significant differences of relapse-free survival and survival between NC-PR patients and patients considered as having PR by RRT and RRVT. The CE-MRI technique could accurately evaluate the therapeutic effect and predict the prognosis of treated non-small cell lung cancer.