Fig 1.
Workflow for patients with additional lesions detected on MRI.
Table 1.
Tumour characteristics.
Table 2.
Frequency of molecular subtypes in the study population.
Fig 2.
Additional lesions detected on MRI.
NME = Non-mass enhancement, LNs = Lymph nodes, UDH = Usual ductal hyperplasia, DCIS = Ductal carcinoma in-situ, Ix = Investigation.
Fig 3.
Flowchart outcome for conventional imaging only and conventional imaging plus MRI cohorts.
(BCS = Breast-Conserving Surgery, IORT = Intraoperative Radiotherapy, EBRT = External Beam Radiotherapy, LVI = Lymphovascular invasion, Ax LN = Axillary lymph node, NAC = Neoadjuvant chemotherapy).
Fig 4.
A 49-year-old woman with invasive carcinoma who subsequently underwent mastectomy due to imaging discordance.
Left MMG in (A) MLO and (B) CC views showing a spiculated lesion (thin arrow) in the mid outer region with foci of microcalcifications on a background of fatty breast parenchymal pattern (BIRADS density A). Corresponding US in (C) transverse view showing an ill-defined hypoechoic mass in the left 3 o’clock position, measuring 9 x 9mm with the presence of internal vascularity (D).
Fig 5.
A 49-year-old woman with invasive carcinoma who subsequently underwent mastectomy due to imaging discordance.
Breast MRI (A) Maximal intensity projection (MIP) shows the index tumour (thin arrow) and multifocal segmental clumped NME (HPE was DCIS) (arrowhead). (B, C) Axial subtracted post-contrast arterial phase 1 image shows index tumour (thin arrow) and multifocal lesion (arrowhead). (D) Colour coded intensity map of index tumour (red coloured area signifying the most intense enhancement) with type II kinetic curve (subset). (E) Colour coded intensity map of the multifocal lesion with Type III kinetic curves (subset).
Fig 6.
49-year-old woman with left breast invasive carcinoma who was deemed eligible for IORT based on conventional imaging but the tumour was larger than 30mm on MRI (done 4 days apart).
Left MMG in (A) MLO and (B) CC view showing an indistinct high-density mass (thin arrow) in the left upper central region in a background of dense breast parenchyma BIRADS. (C) Corresponding ultrasound of the left breast in (C) transverse and (D) longitudinal views show irregular hypoechoic lesion with angular margins. The largest dimension on conventional imaging was 22mm.
Fig 7.
49-year-old woman with left breast invasive carcinoma who was deemed eligible for IORT based on conventional imaging but the tumour was larger than 30 mm on MRI (done 4 days apart).
Breast MRI Images maximal intensity projection (MIP) in (A) sagittal and (B) axial projections showing the index tumour (thin arrow). (C, D) Axial DCE phase 1 showing index tumour (thin arrow) with a maximal diameter (AP) of 32mm on MRI.
Fig 8.
Bland-Altman plots of imaging (US, MMG, and MRI) and pathological size differences versus mean size differences.
Table 3.
True positive (TP), true negative (TN), false positive (FP) and false negative (FN) in US and MRI.
Table 4.
Concordances of tumour size by modality versus pathology.
Table 5.
Comparison of tumour size by modality versus pathology.