B2
Improved methods of detection, and prognostic significance, of vascular invasion in breast cancer
Rabab AA Mohammed1, Ian O Ellis1, Muhammad S Gill2 , Andrew R Green1, Emma C Paish1, Stewart G Martin1
1University of Nottingham, UK, 2Queens Medical Centre, Nottingham, UK
Vascular invasion, encompassing both lymphovascular (LVI) and blood vascular invasion (BVI), influences breast cancer patient prognosis. We have previously shown, using immunohistochemistry (IHC) to distinguish between LVI and BVI in 177 consecutive breast cancer patients (52 LN+ and 125 LN-), that VI was detected in 32% of specimens and that 96% of such VI were invasion of lymphatic vessels. LVI was significantly associated with metastasis, recurrence, worse disease free interval (DFI) and overall survival (OS). This original study has been expanded by examining LVI, assessed by podoplanin staining, in 1000 LN-, early stage, breast cancers. In addition to verifying initial results it also allowed an examination of LVI in different breast cancer phenotypes.
LVI was detected in 212 (21% of specimens). Use of IHC showed that VI was missed in 13% of patients when assessed by conventional H&E methods. The 20-year OS rate was 85% in LVI- compared with 68% in LVI+ patients. In multivariate analysis for OS, tumour size (P =0.004), grade (P< 0.0001), and LVI (P <0.0001) retained significance. However, for DFI, only LVI remained significant (P<0.0001). Lowest LVI positivity was in tubular carcinoma (5.6%) and highest in invasive carcinoma of no special type (26%). Basal phenotype tumours were significantly associated with higher LVI (29%) compared with non-basal (19%, P =0.002).
In the LVI+ specimens; 203 (76%) had peri-tumoural and 58 (24%) intra-tumoural LVIs (9 specimens had LVIs in the intra-tumoural vessels only and 49 specimens had an LVI in both intra- and peri-tumoural vessels). The frequency of LVI ranged from 1 to 79: 167specimens (79%) had <5 LVIs, 23 (11%) had 6-10 LVIs and 22 (10%) had >10LVIs. There was no significant difference between the LVI frequency and development of recurrence or death from the disease.
Acknowledgements
The authors gratefully acknowledge Cancer Research UK for funding this work
