Independent predictors of breast malignancy in screen-detected microcalcifications: biopsy results in 2545 cases

Abstract

Background:Mammographic microcalcifications are associated with many benign lesions, ductal carcinoma in situ (DCIS) and invasive cancer. Careful assessment criteria are required to minimise benign biopsies while optimising cancer diagnosis. We wished to evaluate the assessment outcomes of microcalcifications biopsied in the setting of population-based breast cancer screening.Methods:Between January 1992 and December 2007, cases biopsied in which microcalcifications were the only imaging abnormality were included. Patient demographics, imaging features and final histology were subjected to statistical analysis to determine independent predictors of malignancy.Results:In all, 2545 lesions, with a mean diameter of 21.8 mm (s.d. 23.8 mm) and observed in patients with a mean age of 57.7 years (s.d. 8.4 years), were included. Using the grading system adopted by the RANZCR, the grade was 3 in 47.7% 4 in 28.3% and 5 in 24.0%. After assessment, 1220 lesions (47.9%) were malignant (809 DCIS only, 411 DCIS with invasive cancer) and 1325 (52.1%) were non-malignant, including 122 (4.8%) premalignant lesions (lobular carcinoma in situ, atypical lobular hyperplasia and atypical ductal hyperplasia). Only 30.9% of the DCIS was of low grade.Mammographic extent of microcalcifications >15 mm, imaging grade, their pattern of distribution, presence of a palpable mass and detection after the first screening episode showed significant univariate associations with malignancy. On multivariate modeling imaging grade, mammographic extent of microcalcifications >15 mm, palpable mass and screening episode were retained as independent predictors of malignancy. Radiological grade had the largest effect with lesions of grade 4 and 5 being 2.2 and 3.3 times more likely to be malignant, respectively, than grade 3 lesions.Conclusion:The radiological grading scheme used throughout Australia and parts of Europe is validated as a useful system of stratifying microcalcifications into groups with significantly different risks of malignancy. Biopsy assessment of appropriately selected microcalcifications is an effective method of detecting invasive breast cancer and DCIS, particularly of non-low-grade subtypes.

Microcalcifications Detected as an Abnormality on Screening Mammography: Outcomes and Followup over a Five-Year Period

“…The incidence of malignancy in women with screened detected calcifications who underwent pathological diagnosis at assessment was 26.0% ( n = 38) which is comparable to previous studies (20%–30%) [4, 7] although not as high as others [3, 10] who record values of up to 47.9% [3]. Our work differs from previous studies by including a five-year followup for subsequent development of malignancy of all women who were initially recalled for evaluation of microcalcification, including those who were given a benign diagnosis based on the imaging and clinical examination alone without the need for biopsy.…”

“…Of all calcifications categorised as indeterminate, suspicious, or malignant, the incidence of associated malignancy has been reported by previous studies by approximately 10%–48% [3–12]. While most studies report a malignancy rate of 20%–30%, a recent study by Farshid et al [3] gives a figure of 47.9%. …”

Improved Diagnostics by Assessing the Micromorphology of Breast Calcifications via X-Ray Dark-Field Radiography

“…Position, size, shape and appearance of microcalcifications within the breast indicate their possible origin as either benign or cancerous cell processes345. Morphology and distribution descriptors (BIRADS) help to categorize microcalcifications and stratify the malignancy risk of the corresponding breast tissue, hence determining the need for invasive assessment and follow-up procedures678.…”

Towards better veracity for breast cancer detection using Gabor analysis and statistical learning

“…speculated lesions, circumscribed masses and calcification clusters [6]- [8] as shown in figure 4. The motivation of the proposed method is to detect these types with greater accuracy in comparison to the contemporary systems.…”