King's College London

Research portal

Histological scoring of immune and stromal features in breast and axillary lymph nodes is prognostic for distant metastasis in lymph node-positive breast cancers

Research output: Contribution to journalArticlepeer-review

Anita Grigoriadis, Patrycja Gazinska, Trupti Pai, Sheeba Irshad, Yin Wu, Rosemary R Millis, Kalnisha Naidoo, Julie Elizabeth Owen, Cheryl Elaine Gillett, Andrew Nicholas James Tutt, ACC Coolen, Sarah Elizabeth Pinder

Original languageEnglish
Pages (from-to)39-54
JournalThe journal of pathology. Clinical research
Issue number1
Early online date8 Jan 2018
Accepted/In press4 Nov 2017
E-pub ahead of print8 Jan 2018
Published24 Jan 2018


King's Authors


The prognostic importance of lymph node (LN) status and tumour-infiltrating lymphocytes (TILs), is well
established, particularly TILs in triple negative breast cancers (TNBCs). So far, few studies have interrogated
changes in involved and uninvolved LNs and evaluated if their morphological patterns add valuable information
for the prediction of disease progression in breast cancer. In a cohort of 309 patients enriched for
TNBCs (170/309), we histologically characterised immune and stromal features in primary tumours and
associated involved and uninvolved axillary LNs on routine haematoxylin and eosin stained sections. Of the
309 patients, 143 had LN-positive disease. Twenty-five histopathological features were assessed, including
the degree of TIL presence, quantitative and qualitative assessment of germinal centres (GCs) and sinus histiocytosis.
Multivariate and cross-validated proportional hazard regression analyses were used to identify
optimal covariate sets for prediction of distant metastasis-free survival (DMFS). The degree of intratumoural
and peritumoural immune infiltrate was associated with architectural changes in both uninvolved and
involved LNs. By including clinicopathological characteristics as well as tumour and LN histopathological
features in L2-regularised proportional hazard models, the prediction of 5-year DMFS was improved by
3–15% over the baseline in all cancers and in TNBCs. In LN-positive cancers, the combination of Salgado’s
classification, lymphocytic lobulitis, size and number of GCs in the uninvolved LNs and location of GCs in
the involved LNs carried significant prognostic information. From these features, a multivariate crossvalidation-
stable risk signature was constructed, which identified low-risk groups within both LN-positive
breast cancers and the LN-positive TNBCs group with a 10-year DMFS probability of 78 and 87%, respectively.
This study illustrates that, by incorporating histopathological patterns of involved and uninvolved
LNs combined with primary tumour immune and stromal features, the prediction of developing distant
metastasis in LN-positive breast cancers can be estimated more accurately.

Download statistics

No data available

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454