Promising molecular techniques may allow for testing of novel and complex hypotheses such as defining gene expression profiles in specific cells, tumors, or their microenvironments. For most large cancer epidemiologic and population-based studies, however, application of such promising techniques may not be possible owing to constraints of specimen preservation from paraffin-embedded tissues. Alternative methods would ideally preserve tissue morphology and not degrade DNA or RNA. We conducted a comparison of snap-freezing (freezing with liquid nitrogen), ethanol-fixation with low melt polyester wax embedding, and RNAlater-preservation techniques to determine which method was optimal for subsequent assessment of gene expression changes in cervical cancer. From each of 15 women with cancer and 30 without, we procured 3 pieces of cervical tissue and compared snap-freezing, ethanol-fixation, and RNAlater-preservation techniques. Despite slight loss in morphologic quality from snap-frozen cervix tissues, RNA quality was equivalent to or better than RNAlater-preserved tissues and significantly exceeded that from ethanol-fixed/polyester wax embedded tissue. In conclusion, despite the moderate logistical constraints in set-up that required either liquid nitrogen or dry ice on-site for snap-freezing tissue, the ease of downstream processing and consistent high quality RNA made it preferable to the other 2 methods.