Background: Molecular breast imaging (MBI) is a novel breast imaging technique that uses Cadmium Zinc Telluride (CZT) gamma cameras to detect the uptake of Tc-99m sestamibi in breast tumors. Current techniques employ an administered dose of 20-30 mCi Tc-99m, delivering an effective dose of 6.5-10 mSv to the body. This is ∼5-10 times that of mammography. The goal of this study was to reduce the radiation dose by a factor of 5-10, while maintaining image quality. Methods: A total of 4 dose reduction schemes were evaluated - a) optimized collimation, b) improved utilization of the energy spectrum below the photopeak, c) adaptive geometric mean algorithm developed for combination of images from opposing detectors, and d) non local means filtering (NLMF) for noise reduction and image enhancement. Validation of the various schemes was performed using a breast phantom containing a variety of tumors and containing activity matched to that observed in clinical studies. Results: Development of tungsten collimators with holes matched to the CZT pixels yielded a 2.1-2.9 gain in system sensitivity. Improved utilization of the energy spectra yielded a 1.5-2.0 gain in sensitivity. Development of a modified geometric mean algorithm yielded a 1.4 reduction in image noise, while retaining contrast. Images of the breast phantom demonstrated that a factor of 5 reduction in dose was achieved. Additional refinements to the NLMF should enable an additional factor of 2 reduction in dose. Conclusion: Significant dose reduction in MBI to levels comparable to mammography can be achieved while maintaining image quality.