D S Gunawardena1, S Burrows2, D B Taylor3. 1. Department of Radiology, Royal Perth Hospital, Wellington Street, Perth, Western Australia, 6004, Australia. Electronic address: dsguna@yahoo.com. 2. School of Medicine, University of Western Australia, GPO Box X2213, Perth, Western Australia 6847, Australia. 3. Department of Radiology, Royal Perth Hospital, Wellington Street, Perth, Western Australia, 6004, Australia; Faculty of Health and Medical Sciences, Medical Schools, University of Western Australia, Western Australia, Australia.
Abstract
AIM: To review the ultrasound (US) patterns of pure ductal carcinoma in situ (DCIS) using a non-mass-like (NML) versus mass-like (ML) classification and to investigate histopathological associations. MATERIALS AND METHODS: The present study was a retrospective analysis of sonographically evident pure DCIS lesions detected in a mammographic (MG) screening programme over a 7-year period from 2008. All lesions had undergone US-guided 14 G core biopsies with no upgrades to invasive disease on surgical histopathology. Lesions that were three-dimensional with convex margins were classified as ML and all others as NML. ML lesions were subdivided into solid, cystic, or mixed, and NML lesions into ductal and non-ductal. Imaging and pathological characteristics of NML versus ML lesions were investigated using logistic regression. RESULTS: There were 78 lesions in 75 participants. NML lesions accounted for 45 (58%) lesions, comprising 27 (60%) ductal and 18 (40%) non-ductal subtypes. There were 33 (42%) ML lesions; the largest subgroup being solid (n=21, 64%). Significant associations between lesion type and lesion size on US (<15 versus ≥15 mm), presence of US and mammographic calcification and posterior shadowing on sonography were identified. NML lesions had fivefold higher odds (OR=5.41 95% confidence interval [CI]: 2.03, 14.39, p=0.001) to be high grade and sevenfold higher odds (OR=7 95% CI: 1.75, 27.99, p=0.006) to have comedo necrosis on histopathology. CONCLUSION: DCIS lesions can be successfully classified using ML and NML lesion descriptors and NML morphology on US is associated with histological features of "high-risk" DCIS. Crown
AIM: To review the ultrasound (US) patterns of pure ductal carcinoma in situ (DCIS) using a non-mass-like (NML) versus mass-like (ML) classification and to investigate histopathological associations. MATERIALS AND METHODS: The present study was a retrospective analysis of sonographically evident pure DCIS lesions detected in a mammographic (MG) screening programme over a 7-year period from 2008. All lesions had undergone US-guided 14 G core biopsies with no upgrades to invasive disease on surgical histopathology. Lesions that were three-dimensional with convex margins were classified as ML and all others as NML. ML lesions were subdivided into solid, cystic, or mixed, and NML lesions into ductal and non-ductal. Imaging and pathological characteristics of NML versus ML lesions were investigated using logistic regression. RESULTS: There were 78 lesions in 75 participants. NML lesions accounted for 45 (58%) lesions, comprising 27 (60%) ductal and 18 (40%) non-ductal subtypes. There were 33 (42%) ML lesions; the largest subgroup being solid (n=21, 64%). Significant associations between lesion type and lesion size on US (<15 versus ≥15 mm), presence of US and mammographic calcification and posterior shadowing on sonography were identified. NML lesions had fivefold higher odds (OR=5.41 95% confidence interval [CI]: 2.03, 14.39, p=0.001) to be high grade and sevenfold higher odds (OR=7 95% CI: 1.75, 27.99, p=0.006) to have comedo necrosis on histopathology. CONCLUSION: DCIS lesions can be successfully classified using ML and NML lesion descriptors and NML morphology on US is associated with histological features of "high-risk" DCIS. Crown