Abstract
There is a paucity of definitive guidelines on how to approach an architectural distortion detected on mammography. The purpose of this study was to identify any imaging features that can assist in devising a pattern-based approach to image-detected architectural distortion. This prospective study included females who underwent screening or surveillance mammography in the department. Patients with architectural distortion (AD) not attributed to treatment or primary breast cancers were included. Recorded imaging features included lesion size, presence or absence of central lucency, and convergence of radiating spicules with or without thickened periphery. Biopsy of ADs was performed using image guidance with core needle (CNB) or vacuum-assisted biopsy (VAB) techniques. Fisher's exact test was used for statistical evaluation. Out of a total of 1307 ADs, a cohort of 25 patients was eventually included for evaluation, amongst which 14 underwent tissue sampling. The final histopathology confirmed five malignant and nine benign lesions. ADs associated with malignancy demonstrated statistically significant differences in lesion size (p value, 0.02) and spicules converging to a thickened periphery (p value, 0.03) as compared to the ADs with no evidence of malignancy. No significant difference was observed for presence of a correlate on conventional mammography and ultrasound. We concluded that the potential imaging features to be considered when evaluating ADs were lesion size, lucent center, and spicules converging to a thickened periphery. Whenever considering biopsy, VAB allows better yield as compared to CNB for histopathological assessment of these findings.