Percutaneous needle core biopsy is the standard of care in the assessment of suspicious breast lesions. The diagnostic term “Atypia” is used in breast biopsy reporting when histologic appearances are suspicious but not diagnostic of malignancy. Multiple histopathological appearances are encompassed by the umbrella term “atypia”, including atypical ductal hyperplasia (ADH), columnar cell change with atypia (CCCWA), which is also known as flat epithelial atypia (FEA), and a miscellaneous group of diagnoses, known as atypia NOS.

A pathologic diagnosis of “atypia” in breast core biopsies usually leads to a recommendation to surgically excise the lesion. Many studies have correlated the diagnosis of “atypia” in core biopsies with the subsequent finding of carcinoma in the surgical excisions, and the percentage of carcinoma found represents the positive predictive value (PPV) of the diagnosis. To date, there is no agreed target PPV for the diagnosis of breast atypia on biopsy, but the most studies have demonstrated a PPV of 20- 40%. Individual “atypia” diagnoses such as CCCWA have an even lower PPV of 10-15%.

One method of performance review is an audit of the average breast-atypia PPV within individual pathology departments, which then can be monitored and studied over time, to detect trends and “diagnostic drift” at an early stage. In addition, assessment of the individual PPV of each breast pathologist allows for analysis of the consistency of the diagnostic practice of each individual with their colleagues. Surprisingly, there have been no major studies assessing the intradepartmental range of PPVs for breast atypia diagnoses to date. In contrast, the American College of Radiology has designed the BIRADS classification system in order to audit and monitor the PPV of breast imaging in diagnosing malignancy.

We undertook to measure the departmental PPV for malignancy following a biopsy diagnosis of breast atypia, and performed an anonymized subanalysis in order to establish the range of PPVs of atypia diagnoses between the sub-specialized breast pathologists within the department.

This study established that the baseline PPV in our department is comparable to previously reported studies at 24%, while the range of PPV for an atypia diagnosis between pathologists is between 22.8 and 25% for 5 of 6 pathologists, with one pathologist demonstrating a higher PPV of 36.8%. ADH was the most common diagnosis of the atypia subtypes, and the PPV for ADH alone was 29.9%. 15% of ADH diagnoses were described as “ADH bordering on low grade DCIS”; within this subgroup the PPV was 48.5%. The PPV for a diagnosis of CCCWA alone was 10%.

This study demonstrates that the PPV for breast atypia in a major tertiary cancer center is approximately 24%. We have demonstrated very reproducible use of this diagnostic term within the department. We plan to use the findings of this study to identify subgroups of patients with a sufficiently low PPV to justify a decision not to proceed to surgical intervention. We aim to develop an algorithm for use in the clinical setting in order to direct further patient management. The ultimate aim of this research is to reduce the number of patients undergoing unnecessary surgical interventions.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-02-09.