Conventional, unsharp masking, and slit-mask (combining slit radiography and unsharp masking) techniques were compared in a clinical nodule detection study in cancer patients who were at risk for metastatic lung disease. Unsharp masking improved detection rates for nodules located in poorly penetrated areas of the chest (25% vs. 52%), with no difference in detection rates for other areas of the image and no change in false-positive detection rates. Results with the slit-mask technique did not differ significantly from those obtained with conventional techniques. In general, the differences in nodule detection rates noted between different techniques in the clinical study were less than had been observed in earlier phantom-nodule detection studies. Possible reasons for these findings are discussed.