Purpose: To evaluate the sensitivity for breast cancer of different diagnostic tests performed in a consecutive series of women aged under 36 years.
Materials and methods: The study analyses 155 women with breast cancer incident in the Tuscany Cancer Registry from 1985 to 2000. The sensitivity of each method was evaluated in relation to the most recent test performed during the year before diagnosis and to different variables, such as tumour size and calendar period.
Results: Sensitivity was 70.3 % for physical examination, 76.0% for mammography, 69.1% for ultrasonography and 80.6% for cytology (86.2% if inadequate samples are excluded). Sensitivity was associated to pT category for physical examination (pT1=60.6%; pT2-4=86.4%; chi squared=10.2, p= 0.001) and for ultrasonography (pT1=61.9%; pT2-4=92.0%; chi squared=5.7, p=0.01) and to breast radiological density for mammography (50-100% density=75.5%; 0-50% density=91.3%; chi squared=1.85, p<0.17). Among 112 cases examined with palpation, mammography and cytology, these methods had a sensitivity of 78.5%, 75.8% and 91.9%, respectively, and an overall sensitivity of 98.2%. Among 65 cases examined with palpation, ultrasonography and cytology, these methods had a sensitivity of 67.6%, 70.7% and 92.3%, respectively, and an overall sensitivity of 95.3%. Among 123 cases examined with palpation, imaging (mammography and/or ultrasonography) and cytology, these methods had a sensitivity of 76.4%, 78.0% and 90.2%, respectively, and an overall sensitivity of 96.7%. Cancer was suspected by one method only in 2.4% of cases with palpation, 1.6% with imaging and 7.3% with cytology. Comparison between mammography and ultrasonography was possible in 58 cases undergoing both tests: sensitivity was 79.3% for ultrasonography and 74.1% for mammography, respectively chi squared=0.1, p=0.66). If sensitivity had been determined according to the test performed immediately before diagnosis it would have been higher, that is 74.4% for palpation, 76.8% for mammography, 73.5% for ultrasonography and 90.9% for cytology (97.7% if inadequate samples were excluded), and overall sensitivity would be 96.1%.
Conclusions: Differential diagnosis of breast cancer in women aged under 36 years is difficult, sensitivity being lower as compared to older women, with the only exception of cytology. The association of multiple tests (diagnostic imaging and cytology) is crucial to achieve an acceptable false negative rate. In particular, the extensive use of cytology in the presence of an even minimal doubt at palpation and/or imaging is strongly recommendable. Sensitivity must be calculated on the basis of an independent pathology registry, and false negatives must be defined within a given period after the test. Analysis limited to tests performed immediately before diagnosis leads to a biased overestimation of sensitivity.