Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.

Gregory Verghese; Mengyuan Li; Fangfang Liu; Amit Lohan; Nikhil Cherian Kurian; Swati Meena; Patrycja Gazinska; Aekta Shah; Aasiyah Oozeer; Terry Chan; Mark Opdam; Sabine Linn; Cheryl Gillett; Elena Alberts; Tom Hardiman; Samantha Vaughan Jones; Selvam Thavaraj; Louise Jones; Roberto Salgado; Sarah E Pinder; Swapnil Ulhas Rane; Amit Sethi; Anita Grigoriadis

doi:10.1002/path.6088

Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.

Gregory Verghese, Mengyuan Li, Fangfang Liu, Amit Lohan, Nikhil Cherian Kurian, Swati Meena, Patrycja Gazinska, Aekta Shah, Aasiyah Oozeer, Terry Chan, Mark Opdam, Sabine Linn, Cheryl Gillett, Elena Alberts, Tom Hardiman, Samantha Vaughan Jones, Selvam Thavaraj, Louise Jones, Roberto Salgado, Sarah E PinderSwapnil Ulhas Rane, Amit Sethi, Anita Grigoriadis

Research output: Contribution to journal › Article › peer-review

6 Citations (Scopus)

Abstract

The suggestion that the systemic immune response in lymph nodes (LNs) conveys prognostic value for triple-negative breast cancer (TNBC) patients has not previously been investigated in large cohorts. We used a deep learning (DL) framework to quantify morphological features in haematoxylin and eosin-stained LNs on digitised whole slide images. From 345 breast cancer patients, 5,228 axillary LNs, cancer-free and involved, were assessed. Generalisable multiscale DL frameworks were developed to capture and quantify germinal centres (GCs) and sinuses. Cox regression proportional hazard models tested the association between smuLymphNet-captured GC and sinus quantifications and distant metastasis-free survival (DMFS). smuLymphNet achieved a Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively, and was comparable to an interpathologist Dice coefficient of 0.66 (GC) and 0.60 (sinus). smuLymphNet-captured sinuses were increased in LNs harbouring GCs (p < 0.001). smuLymphNet-captured GCs retained clinical relevance in LN-positive TNBC patients whose cancer-free LNs had on average ≥2 GCs, had longer DMFS (hazard ratio [HR] = 0.28, p = 0.02) and extended GCs' prognostic value to LN-negative TNBC patients (HR = 0.14, p = 0.002). Enlarged smuLymphNet-captured sinuses in involved LNs were associated with superior DMFS in LN-positive TNBC patients in a cohort from Guy's Hospital (multivariate HR = 0.39, p = 0.039) and with distant recurrence-free survival in 95 LN-positive TNBC patients of the Dutch-N4plus trial (HR = 0.44, p = 0.024). Heuristic scoring of subcapsular sinuses in LNs of LN-positive Tianjin TNBC patients (n = 85) cross-validated the association of enlarged sinuses with shorter DMFS (involved LNs: HR = 0.33, p = 0.029 and cancer-free LNs: HR = 0.21 p = 0.01). Morphological LN features reflective of cancer-associated responses are robustly quantifiable by smuLymphNet. Our findings further strengthen the value of assessment of LN properties beyond the detection of metastatic deposits for prognostication of TNBC patients.

Original language	English
Pages (from-to)	376-389
Number of pages	14
Journal	Journal of pathology
Volume	260
Issue number	4
Early online date	25 May 2023
DOIs	https://doi.org/10.1002/path.6088
Publication status	Published - Aug 2023

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1002/path.6088Licence: CC BY

Cite this

Verghese, G., Li, M., Liu, F., Lohan, A., Kurian, N. C., Meena, S., Gazinska, P., Shah, A., Oozeer, A., Chan, T., Opdam, M., Linn, S., Gillett, C., Alberts, E., Hardiman, T., Jones, S. V., Thavaraj, S., Jones, L., Salgado, R., ... Grigoriadis, A. (2023). Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies. Journal of pathology, 260(4), 376-389. https://doi.org/10.1002/path.6088

@article{f8db21cfeead449d949142a6aac14a2c,

title = "Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.",

abstract = "The suggestion that the systemic immune response in lymph nodes (LNs) conveys prognostic value for triple-negative breast cancer (TNBC) patients has not previously been investigated in large cohorts. We used a deep learning (DL) framework to quantify morphological features in haematoxylin and eosin-stained LNs on digitised whole slide images. From 345 breast cancer patients, 5,228 axillary LNs, cancer-free and involved, were assessed. Generalisable multiscale DL frameworks were developed to capture and quantify germinal centres (GCs) and sinuses. Cox regression proportional hazard models tested the association between smuLymphNet-captured GC and sinus quantifications and distant metastasis-free survival (DMFS). smuLymphNet achieved a Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively, and was comparable to an interpathologist Dice coefficient of 0.66 (GC) and 0.60 (sinus). smuLymphNet-captured sinuses were increased in LNs harbouring GCs (p < 0.001). smuLymphNet-captured GCs retained clinical relevance in LN-positive TNBC patients whose cancer-free LNs had on average ≥2 GCs, had longer DMFS (hazard ratio [HR] = 0.28, p = 0.02) and extended GCs' prognostic value to LN-negative TNBC patients (HR = 0.14, p = 0.002). Enlarged smuLymphNet-captured sinuses in involved LNs were associated with superior DMFS in LN-positive TNBC patients in a cohort from Guy's Hospital (multivariate HR = 0.39, p = 0.039) and with distant recurrence-free survival in 95 LN-positive TNBC patients of the Dutch-N4plus trial (HR = 0.44, p = 0.024). Heuristic scoring of subcapsular sinuses in LNs of LN-positive Tianjin TNBC patients (n = 85) cross-validated the association of enlarged sinuses with shorter DMFS (involved LNs: HR = 0.33, p = 0.029 and cancer-free LNs: HR = 0.21 p = 0.01). Morphological LN features reflective of cancer-associated responses are robustly quantifiable by smuLymphNet. Our findings further strengthen the value of assessment of LN properties beyond the detection of metastatic deposits for prognostication of TNBC patients.",

author = "Gregory Verghese and Mengyuan Li and Fangfang Liu and Amit Lohan and Kurian, {Nikhil Cherian} and Swati Meena and Patrycja Gazinska and Aekta Shah and Aasiyah Oozeer and Terry Chan and Mark Opdam and Sabine Linn and Cheryl Gillett and Elena Alberts and Tom Hardiman and Jones, {Samantha Vaughan} and Selvam Thavaraj and Louise Jones and Roberto Salgado and Pinder, {Sarah E} and Rane, {Swapnil Ulhas} and Amit Sethi and Anita Grigoriadis",

note = "Funding Information: The authors would like to thank all members of the Cancer Bioinformatics team at King's College London (London, UK) for their helpful suggestions. We thank Harry Chinque for digitising the WSIs and Tristan Clark for support with high-performance computing on the CRUK City of London Centre Computing. We thank the Breast Cancer Research Trust, Breast Cancer Now (and their legacy charity Breakthrough Breast Cancer), the Medical Research Council (MRC) [MR/X012476/1], and Cancer Research UK [CRUK/07/012, KCL-BCN-Q3] for funding this project. We thank King's Health Partners Cancer Biobank for supply of materials and clinical data and the Molecular Pathology and Biobanking of the Netherlands Cancer Institute for their lab support. The authors wish to acknowledge the role of the Breast Cancer Now Tissue Bank in collecting and making available the samples used in the generation of this publication and the patients who donated to the bank. The team at the Tata Memorial Centre and IIT-B are in part supported by the funding awarded to Swapnil Rane and Amit Sethi by the Department of Biotechnology (BT/PR32348/AI/133/25/2020). Mengyuan Li is funded by a China Scholarship Council PhD scholarship. Fangfang Liu was awarded a KC Wong Postdoctoral Fellowships to study at King's College London. Roberto Salgado is supported by the Breast Cancer Research Foundation (BCRF, Grant 17-194). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was supported by the CRUK City of London Centre Award [CTRQQR-2021/100004]. Funding Information: The authors would like to thank all members of the Cancer Bioinformatics team at King's College London (London, UK) for their helpful suggestions. We thank Harry Chinque for digitising the WSIs and Tristan Clark for support with high‐performance computing on the CRUK City of London Centre Computing. We thank the Breast Cancer Research Trust, Breast Cancer Now (and their legacy charity Breakthrough Breast Cancer), the Medical Research Council (MRC) [MR/X012476/1], and Cancer Research UK [CRUK/07/012, KCL‐BCN‐Q3] for funding this project. We thank King's Health Partners Cancer Biobank for supply of materials and clinical data and the Molecular Pathology and Biobanking of the Netherlands Cancer Institute for their lab support. The authors wish to acknowledge the role of the Breast Cancer Now Tissue Bank in collecting and making available the samples used in the generation of this publication and the patients who donated to the bank. The team at the Tata Memorial Centre and IIT‐B are in part supported by the funding awarded to Swapnil Rane and Amit Sethi by the Department of Biotechnology (BT/PR32348/AI/133/25/2020). Mengyuan Li is funded by a China Scholarship Council PhD scholarship. Fangfang Liu was awarded a KC Wong Postdoctoral Fellowships to study at King's College London. Roberto Salgado is supported by the Breast Cancer Research Foundation (BCRF, Grant 17‐194). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was supported by the CRUK City of London Centre Award [CTRQQR‐2021/100004]. Publisher Copyright: {\textcopyright} 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.",

year = "2023",

month = aug,

doi = "10.1002/path.6088",

language = "English",

volume = "260",

pages = "376--389",

journal = "Journal of pathology",

issn = "0022-3417",

publisher = "Pathological Society of Great Britain and Ireland",

number = "4",

}

Verghese, G , Li, M , Liu, F, Lohan, A, Kurian, NC, Meena, S, Gazinska, P, Shah, A, Oozeer, A, Chan, T, Opdam, M, Linn, S, Gillett, C, Alberts, E, Hardiman, T, Jones, SV, Thavaraj, S, Jones, L, Salgado, R, Pinder, SE, Rane, SU, Sethi, A & Grigoriadis, A 2023, 'Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.', Journal of pathology, vol. 260, no. 4, pp. 376-389. https://doi.org/10.1002/path.6088

TY - JOUR

T1 - Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.

AU - Verghese, Gregory

AU - Li, Mengyuan

AU - Liu, Fangfang

AU - Lohan, Amit

AU - Kurian, Nikhil Cherian

AU - Meena, Swati

AU - Gazinska, Patrycja

AU - Shah, Aekta

AU - Oozeer, Aasiyah

AU - Chan, Terry

AU - Opdam, Mark

AU - Linn, Sabine

AU - Gillett, Cheryl

AU - Alberts, Elena

AU - Hardiman, Tom

AU - Jones, Samantha Vaughan

AU - Thavaraj, Selvam

AU - Jones, Louise

AU - Salgado, Roberto

AU - Pinder, Sarah E

AU - Rane, Swapnil Ulhas

AU - Sethi, Amit

AU - Grigoriadis, Anita

N1 - Funding Information: The authors would like to thank all members of the Cancer Bioinformatics team at King's College London (London, UK) for their helpful suggestions. We thank Harry Chinque for digitising the WSIs and Tristan Clark for support with high-performance computing on the CRUK City of London Centre Computing. We thank the Breast Cancer Research Trust, Breast Cancer Now (and their legacy charity Breakthrough Breast Cancer), the Medical Research Council (MRC) [MR/X012476/1], and Cancer Research UK [CRUK/07/012, KCL-BCN-Q3] for funding this project. We thank King's Health Partners Cancer Biobank for supply of materials and clinical data and the Molecular Pathology and Biobanking of the Netherlands Cancer Institute for their lab support. The authors wish to acknowledge the role of the Breast Cancer Now Tissue Bank in collecting and making available the samples used in the generation of this publication and the patients who donated to the bank. The team at the Tata Memorial Centre and IIT-B are in part supported by the funding awarded to Swapnil Rane and Amit Sethi by the Department of Biotechnology (BT/PR32348/AI/133/25/2020). Mengyuan Li is funded by a China Scholarship Council PhD scholarship. Fangfang Liu was awarded a KC Wong Postdoctoral Fellowships to study at King's College London. Roberto Salgado is supported by the Breast Cancer Research Foundation (BCRF, Grant 17-194). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was supported by the CRUK City of London Centre Award [CTRQQR-2021/100004]. Funding Information: The authors would like to thank all members of the Cancer Bioinformatics team at King's College London (London, UK) for their helpful suggestions. We thank Harry Chinque for digitising the WSIs and Tristan Clark for support with high‐performance computing on the CRUK City of London Centre Computing. We thank the Breast Cancer Research Trust, Breast Cancer Now (and their legacy charity Breakthrough Breast Cancer), the Medical Research Council (MRC) [MR/X012476/1], and Cancer Research UK [CRUK/07/012, KCL‐BCN‐Q3] for funding this project. We thank King's Health Partners Cancer Biobank for supply of materials and clinical data and the Molecular Pathology and Biobanking of the Netherlands Cancer Institute for their lab support. The authors wish to acknowledge the role of the Breast Cancer Now Tissue Bank in collecting and making available the samples used in the generation of this publication and the patients who donated to the bank. The team at the Tata Memorial Centre and IIT‐B are in part supported by the funding awarded to Swapnil Rane and Amit Sethi by the Department of Biotechnology (BT/PR32348/AI/133/25/2020). Mengyuan Li is funded by a China Scholarship Council PhD scholarship. Fangfang Liu was awarded a KC Wong Postdoctoral Fellowships to study at King's College London. Roberto Salgado is supported by the Breast Cancer Research Foundation (BCRF, Grant 17‐194). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was supported by the CRUK City of London Centre Award [CTRQQR‐2021/100004]. Publisher Copyright: © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

PY - 2023/8

Y1 - 2023/8

N2 - The suggestion that the systemic immune response in lymph nodes (LNs) conveys prognostic value for triple-negative breast cancer (TNBC) patients has not previously been investigated in large cohorts. We used a deep learning (DL) framework to quantify morphological features in haematoxylin and eosin-stained LNs on digitised whole slide images. From 345 breast cancer patients, 5,228 axillary LNs, cancer-free and involved, were assessed. Generalisable multiscale DL frameworks were developed to capture and quantify germinal centres (GCs) and sinuses. Cox regression proportional hazard models tested the association between smuLymphNet-captured GC and sinus quantifications and distant metastasis-free survival (DMFS). smuLymphNet achieved a Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively, and was comparable to an interpathologist Dice coefficient of 0.66 (GC) and 0.60 (sinus). smuLymphNet-captured sinuses were increased in LNs harbouring GCs (p < 0.001). smuLymphNet-captured GCs retained clinical relevance in LN-positive TNBC patients whose cancer-free LNs had on average ≥2 GCs, had longer DMFS (hazard ratio [HR] = 0.28, p = 0.02) and extended GCs' prognostic value to LN-negative TNBC patients (HR = 0.14, p = 0.002). Enlarged smuLymphNet-captured sinuses in involved LNs were associated with superior DMFS in LN-positive TNBC patients in a cohort from Guy's Hospital (multivariate HR = 0.39, p = 0.039) and with distant recurrence-free survival in 95 LN-positive TNBC patients of the Dutch-N4plus trial (HR = 0.44, p = 0.024). Heuristic scoring of subcapsular sinuses in LNs of LN-positive Tianjin TNBC patients (n = 85) cross-validated the association of enlarged sinuses with shorter DMFS (involved LNs: HR = 0.33, p = 0.029 and cancer-free LNs: HR = 0.21 p = 0.01). Morphological LN features reflective of cancer-associated responses are robustly quantifiable by smuLymphNet. Our findings further strengthen the value of assessment of LN properties beyond the detection of metastatic deposits for prognostication of TNBC patients.

AB - The suggestion that the systemic immune response in lymph nodes (LNs) conveys prognostic value for triple-negative breast cancer (TNBC) patients has not previously been investigated in large cohorts. We used a deep learning (DL) framework to quantify morphological features in haematoxylin and eosin-stained LNs on digitised whole slide images. From 345 breast cancer patients, 5,228 axillary LNs, cancer-free and involved, were assessed. Generalisable multiscale DL frameworks were developed to capture and quantify germinal centres (GCs) and sinuses. Cox regression proportional hazard models tested the association between smuLymphNet-captured GC and sinus quantifications and distant metastasis-free survival (DMFS). smuLymphNet achieved a Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively, and was comparable to an interpathologist Dice coefficient of 0.66 (GC) and 0.60 (sinus). smuLymphNet-captured sinuses were increased in LNs harbouring GCs (p < 0.001). smuLymphNet-captured GCs retained clinical relevance in LN-positive TNBC patients whose cancer-free LNs had on average ≥2 GCs, had longer DMFS (hazard ratio [HR] = 0.28, p = 0.02) and extended GCs' prognostic value to LN-negative TNBC patients (HR = 0.14, p = 0.002). Enlarged smuLymphNet-captured sinuses in involved LNs were associated with superior DMFS in LN-positive TNBC patients in a cohort from Guy's Hospital (multivariate HR = 0.39, p = 0.039) and with distant recurrence-free survival in 95 LN-positive TNBC patients of the Dutch-N4plus trial (HR = 0.44, p = 0.024). Heuristic scoring of subcapsular sinuses in LNs of LN-positive Tianjin TNBC patients (n = 85) cross-validated the association of enlarged sinuses with shorter DMFS (involved LNs: HR = 0.33, p = 0.029 and cancer-free LNs: HR = 0.21 p = 0.01). Morphological LN features reflective of cancer-associated responses are robustly quantifiable by smuLymphNet. Our findings further strengthen the value of assessment of LN properties beyond the detection of metastatic deposits for prognostication of TNBC patients.

UR - http://www.scopus.com/inward/record.url?scp=85160345384&partnerID=8YFLogxK

U2 - 10.1002/path.6088

DO - 10.1002/path.6088

M3 - Article

SN - 0022-3417

VL - 260

SP - 376

EP - 389

JO - Journal of pathology

JF - Journal of pathology

IS - 4

ER -

Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies.

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this