OBJECTIVES: The study compares contrast-detail and microcalcification detectability of a full-field digital mammography (FFDM) to a state-of-the-art conventional screen-film mammography (SFM) by using different doses in the digital system. MATERIALS AND METHODS: The investigations were performed with an FFDM (Senographe 2000 D, GEMS) and an SFM system (Senographe DMR, GEMS) using a contrast-detail mammography phantom (CDMAM) and an anthropomorphic breast phantom with superimposed microcalcifications. The digital detector was exposed with standard dose of SFM and with a dose reduction of up to 75%. Contrast-detail curves and correct observation ratio (COR) were performed for the CDMAM phantom. ROC analysis with a confidence level ranging from 1 to 5 was done with the results of the anthropomorphic phantom. RESULTS: Digital mammography with the same dose revealed at least an equivalent or even higher detectability rate than conventional mammography, COR could be increased at about 10-25%. The ROC analysis yielded better results for the FFDM system. The same lesion detectability in digital mammography as in the conventional method was reached at a dose reduction of about 25%, concerning spot views even at higher reduction. Dose reduction in the anthropomorphic phantom resulted in a linear loss of detectability. The same detectability as in conventional mammography was reached, however, by a dose reduction of about 50%. CONCLUSION: The results suggest that FFDM is at least equivalent to or--as far as spot views are concerned--superior to conventional SFM concerning the detectability of simulated lesions. Thus, a potential of dose reduction is suggested.
OBJECTIVES: The study compares contrast-detail and microcalcification detectability of a full-field digital mammography (FFDM) to a state-of-the-art conventional screen-film mammography (SFM) by using different doses in the digital system. MATERIALS AND METHODS: The investigations were performed with an FFDM (Senographe 2000 D, GEMS) and an SFM system (Senographe DMR, GEMS) using a contrast-detail mammography phantom (CDMAM) and an anthropomorphic breast phantom with superimposed microcalcifications. The digital detector was exposed with standard dose of SFM and with a dose reduction of up to 75%. Contrast-detail curves and correct observation ratio (COR) were performed for the CDMAM phantom. ROC analysis with a confidence level ranging from 1 to 5 was done with the results of the anthropomorphic phantom. RESULTS: Digital mammography with the same dose revealed at least an equivalent or even higher detectability rate than conventional mammography, COR could be increased at about 10-25%. The ROC analysis yielded better results for the FFDM system. The same lesion detectability in digital mammography as in the conventional method was reached at a dose reduction of about 25%, concerning spot views even at higher reduction. Dose reduction in the anthropomorphic phantom resulted in a linear loss of detectability. The same detectability as in conventional mammography was reached, however, by a dose reduction of about 50%. CONCLUSION: The results suggest that FFDM is at least equivalent to or--as far as spot views are concerned--superior to conventional SFM concerning the detectability of simulated lesions. Thus, a potential of dose reduction is suggested.