Patient population: we retrospectively identified all patients who underwent breast biopsies classified as B1 in the period between June 2012 and June 2015 in our Institute,
searching for “B1” and “inadequate” in the RIS of our Hospital.
Patients who received a diagnosis of B1-lesion without a follow-up of at least 2 years and without a second histologic assessment (percutaneous or surgical) were excluded [1,
2].
All the other patients who received a diagnosis of B1-lesion were included.
Data acquisition: all patients underwent digital mammography and ultrasound.
Digital mammography was performed with a Giotto SD or Giotto 3DL (IMS Italy,
Bologna) by trained radiographers,
performing the two standard mammographic projections (Cranio-Caudal and Medio-Lateral-Oblique) of each breast.
Ultrasound examinations were performed by six radiologists with at least 10 years of experience in breast imaging with three different equipments: MyLab Class C (Esaote),
Logiq E9 (GE-Healthcare) or Iu22 (Philips),
all of them equipped with 5-17 MHz high frequency linear probes.
Ultrasound-guided biopsies were performed either with automatic devices (Bard,
Magnum Biopsy System,
Tempe,
AZ,
USA) with 14 or 18 Gauge needle and with 15 / 23 mm window width,
either with semi-automatic devices (Precisa HS Hospital Service,
Rome,
Italy) with 14 or 18 Gauge needle and with 10 / 20 mm window width.
The standard number of cores obtained in each procedure was from 3 to 5.
Biopsies under stereotactic guidance were performed using an automatic biopsy system (Bard,
Encor,
Tempe,
AZ,
USA) with 10 or 12 Gauge needle.
The standard number of cores obtained in each procedure was 12.
None of the biopsies performed under MR guidance was classified as B1.
Data analysis: for each patient we assessed age,
personal/family history of breast cancer,
presence or absence of symptoms,
sampling guidance (mammography/ultrasound/MR),
number of cores obtained during the procedure,
needle’s Gauge,
lesion size (≤10mm,
≥11mm),
BIRADS category (BI-RADS; ≤R4a,
≥R4b) and presence or absence of cellular atypia in the specimen.
The degree of suspicion of biopsied lesions was retrospectively assessed by two radiologists (>10 years of experience in breast imaging) in consensus,
according to BI-RADS classification of the American College of Radiology [10].
All the histological diagnosis were obtained from pathologist’s reports.
All the other data were obtained from radiological reports.
Statistical analysis: according to the final diagnosis,
obtained with re-biopsy,
lumpectomy or follow-up ≥2 years as standard of reference,
patients were classified into 2 groups: Breast Cancer (BC) and Non Breast Cancer (NBC).
The two groups were compared with regards to the above-mentioned features (age,
personal/family history of breast cancer,
presence or absence of symptoms,
sampling guidance,
number of cores obtained during the procedure,
needle’s Gauge,
lesion size,
BIRADS category and presence or absence of cellular atypia in the specimen).
T-test and chi-square test were used.
Values of p <0.05 were considered statistically significant.
A linear regression test was used to identify a possible correlation between the presence of symptoms and the age of patients in predicting the presence of a tumor.
Statistical analysis was performed using MedCalc statistical analysis software (Medcalc,
Mariakerke,
Belgium).