5/44 patients were excluded (2 due to technical problems and 3 for lack of surgical information).
Breast level
Before NCT,
both DWI and DCE-MRI identified 24 unifocal tumors and 15 multiple disease.
The mean size was 49 mm (range 14-87mm) at DWI and 37,5 (range 30–50mm) at DCE-MRI.
After the treatment,
at both DWI and DCE-MRI concentric shrinkage was observed in 21 cases while fragmentation occurred in the remaining 18.
At DWI the mean size was 34 mm (range 0-82 mm) while at DCE-MRI it was 34 (range 0-58mm).
The assessment of response for DWI and DCE-MRI is summarized in Table 2.
18 patients underwent to conservative surgery and 21 patients to mastectomy.
DWI correctly suggested surgical treatment in 33/39 cases,
planning BCS in 16/18 cases: 13 mass lesions with concentric shrinkage and 3 non-mass lesions.
Mastectomy was correctly planned by DWI in 17/21 patients: 4 mass lesions with concentric shrinkage and 13 non-mass lesions.
Figure 1 and 2.
False positive cases were 2 when DWI overrated the lesion extent due to fragmentation after NCT; false negative cases were 4 when DWI underrated the lesion extension after fragmentation.
All the discordances were observed in non mass lesions.
Sensitivity,
specificity and accuracy of DWI in planning breast surgery were 81% (17/21),
89% (16/18) and 85% (33/39),
respectively.
Table 3.
Axillary level
Concerning lymph node,
before NCT DWI identified 14 patients as negative (short axis <10mm) and 25 patients as positive (mean short axis 16mm; range 10-30); after the treatment 15/25 patients resulted as negative (mean short axis 5mm; range 0-9) and 10/25 as positive (mean short axis 11mm; range 11-20).
Pathological nodal status was classified as positive in 21 patients and negative in 18 patients.
Concordance between DWI data and the reference standard was observed in 20/39.
False positive cases were 4,
while false negative cases were 15.
Figure 3.
Sensibility,
specificity and accuracy of DWI in the evaluation of axillary response to NCT were 30% (6/21),
78% (14/18) and 51% (20/39),
respectively.
Table 4.