The usefulness of diffusion-weighted magnetic resonance (MR) imaging (DWI) for differentiating central lung cancer from postobstructive lobar collapse (POC) was investigated. Thirty-three cases suspected of lung cancer and POC on chest bolus computed tomography (CT) underwent thoracic MR imaging examinations. MR examinations were performed using a 1.5-T clinical scanner. Scanning sequences were T1-weighted imaging, T2-weighted imaging (T2WI) and DWI with b=0, 500 s/mm(2), four excitations and segmented breath-holding. The densities and signals of cancer and postobstructive collapsed lung were compared on bolus-enhanced CT, T2W and DW images. Statistical analyses were performed with chi-square test, paired t-test, non-parameter test and kappa statistics. Differentiation between cancer and POC was possible on bolus CT, T2W and DW images in 14, 21 and 26 patients, respectively. Eight cases that were impossible to differentiate on T2W images were distinguishable on DWI, demonstrating that DWI is complementary to T2WI. Using a combination of T2W and DW images, 88% (29/33) of cases were differentiated on MR imaging. Thus, a combination of T2W and DW imaging is superior to bolus-CT or T2WI alone. The contrast-to-noise ratio of DWI was significantly higher than that of T2WI. Agreement between two independent observers on the differential ability of lung cancer and POC was higher for DWI (kappa=0.474) than for T2WI (kappa=0.339). The degree of consolidation around the cancer was negatively correlated with the degree of artifact and degree of deformation. It is feasible to use DWI to differentiate lung cancer from POC. DWI played a role in confirming and providing complementary information to that obtained from T2WI. Our data indicate that using a combination of the two scanned sequences was the best means of distinguishing between lung cancer and POC.