Surgical specimen can be replaced by core samples in assessment of ER, PR and HER-2 for invasive breast cancer.

OBJECTIVE

To compare 14-gauge SCNB (stereotactic core needle biopsy) with surgery and to investigate tissue-heterogeneity of estrogen receptors (ER), progesterone receptors (PR) and human epidermal growth-factor receptor (HER-2) for nonpalpable breast cancers. To determine the number of cores needed for assessment of these factors.

MATERIALS AND METHODS

Cores of 41 invasive cancers were collected in three containers: the 1st into A, the 2nd and 3rd into B and subsequent cores into C. ER, PR and HER-2 were scored by immunohistochemistry and if 2+ or 3+, by chromogenic-in-situ-hybridisation (CISH) for containers and for surgical specimen.

RESULTS

Between SCNB and surgical specimen concordance was 83% (kappa = 0.39) for ER, 88% (kappa = 0.69) for PR and HER-2 and 93% (kappa = 0.63) for HER-2 after CISH. For the most discordant cases, status was positive in cores but negative in surgery: 5/7 for ER (p = 0.459), 5/5 for PR (p = 0.063), and 4/5 for HER-2 (p = 0.375), after CISH 3/3 (p = 0.250), but the difference was not statistically significant. Concordances between containers of cores was 100% (kappa = 1), 85% (kappa = 0.66) and 85% (kappa = 0.66), respectively. With more than three cores, sensitivities of 95%, 100% and 100% were reached.

CONCLUSIONS

SCNB is at least as sensitive as surgery in assessment of ER, PR and HER-2. Three cores are needed for reliable assessment of HER-2 after adding CISH and more than three cores for PR, possibly due to tissue heterogeneity. For ER sensitivity remained lower, 95%, even in multiple cores, therefore ER-negative cases should be further investigated from surgical specimens.