Is Mammography Superior to Breast Ultrasound? Some remarks to: Moon HJ, Jung I, Park J, et al. Ultraschall in Med 2015; 36: 255 – 263

With great interest we have read the study of Moon et al. in which the authors compare the performance of mammography and breast ultrasound in a population with average risk for breast cancer.

In the study, 3895 of the total of 4394 women received a negative mammography result according to BIRADS 1 and 2. Among the latter group, 2005 women received a supplemental ultrasound.

In the mammography group of 4394 women, 17 women were diagnosed with breast cancer. In the group of negative mammograms of 2005 women, 4 additional carcinomas were found by ultrasound.

Moon et al. compared mammography and breast ultrasound with respect to PPV, sensitivity, specificity and accuracy for breast cancer.

The conclusion of the study should be that ultrasound reveals additional carcinomas in mammographically negative women with an average risk of about 2.0 per thousand. It is interesting and not well known that the additional cancers found by ultrasound were discovered in fatty and dense breasts.

In order to draw general conclusions regarding the performance of the two methods, the same study population should be examined by both methods and ideally blinded to each other. However, the results of ultrasound in positive mammograms were not mentioned, although they probably were performed during the course of the study. In the case in which only the subgroup with mammography and ultrasound examination is analyzed, the sensitivity of ultrasound would be 100 % and that of mammography would be 0 %. This is, however, not a reasonable conclusion regarding the performance of the two methods in general.

Therefore, if we assume that ultrasound would diagnose all positive mammographic cases with a similar accuracy, which may not be the case, the sensitivity in the whole collective would be 76.47 % for mammography and 100 % for ultrasound. This emphasizes the fact that it is necessary to consider the whole population in order to obtain an appropriate comparison of the two methods.

Another point of discussion is the accounting of BIRADS 3 findings as false positives. By definition, BIRADS 3 is a benign finding with a probability of 98 %. Therefore, it is close to BIRADS 2 and not deemed to be suspect. Consequently, if a lesion is categorized as BIRADS 3 and is assumed to be benign, it seems logical to count only cases that proved to be malignant as false-negative, but to count cases which proved to be benign as true-negative. This seems to be a critical point in the analysis of the performance of the methods.

Is the capacity to correctly diagnose a fibroadenoma by ultrasound a disadvantage and the failed representation by mammography an advantage? We do not think so.

Furthermore, Moon et al. investigated BIRADS 3 cases by mammography with further diagnostic management which leads to categorization in other groups. For that reason, a comparison of BIRADS 3 findings by ultrasound without any further work-up is not possible. The percentage of 26.6 % of BIRADS 3 cases in an average risk population by ultrasound (more than every 4^th case!) by Moon et al. is unusual and seems questionable.

To make a profound comparison between mammography and ultrasound in a screening setting, we have to know the total cancer yield of each method. For mammography, the result of the study is 0.3. However, for ultrasound the result of only one subgroup is 0.2, but it could be as high as 0.5 in the total group. 

For this reason, we do not support the conclusion drawn from the communicated data that mammography is superior to ultrasound. For the benefit of participants in a breast cancer screening program, we would ultimately prefer a setting in which potential carcinoma will not be overlooked.

• Reference

• 1 Berg WA, Blume JD, Cormack JB et al. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA 2008; 299: 2151-2163
• 2 D’Orsi CJ, Sickles EA, Mendelson EB et al. ACR BI-RADS® Atlas, Breast Imaging Reporting and Data System. Reston, VA: American College of Radiology; 2013
• 3 Barr RG, Zhang Z, Cormack JB et al. Probably benign lesions at screening breast US in a population with elevated risk: prevalence and rate of malignancy in the ACRIN 6666 trial. Radiology 2013; 269: 701-712
• 4 Chang JM, Koo HR, Moon WK. Radiologist-performed hand-held ultrasound screening at average risk of breast cancer: results from a single health screening center. Acta Radiol 2015; 56: 652-658
• 5 Hooley RJ, Greenberg KL, Stackhouse RM et al. Screening US in patients with mammographically dense breasts: initial experience with Connecticut Public Act 09-41. Radiology 2012; 265: 59-69