Predictive value of MRI-based deep learning model for lymphovascular invasion status in node-negative invasive breast cancer.

Publisher: Nature Publishing Group Country of Publication: England NLM ID: 101563288 Publication Model: Electronic Cited Medium: Internet ISSN: 2045-2322 (Electronic) Linking ISSN: 20452322 NLM ISO Abbreviation: Sci Rep Subsets: MEDLINE

Original Publication: London : Nature Publishing Group, copyright 2011-

Breast Neoplasms*/pathology ; Breast Neoplasms*/diagnostic imaging ; Deep Learning* ; Magnetic Resonance Imaging*/methods ; Lymphatic Metastasis*/diagnostic imaging ; Lymph Nodes*/pathology ; Lymph Nodes*/diagnostic imaging ; Neoplasm Invasiveness*; Humans ; Female ; Middle Aged ; Retrospective Studies ; Adult ; Aged ; Predictive Value of Tests

To retrospectively assess the effectiveness of deep learning (DL) model, based on breast magnetic resonance imaging (MRI), in predicting preoperative lymphovascular invasion (LVI) status in patients diagnosed with invasive breast cancer who have negative axillary lymph nodes (LNs). Data was gathered from 280 patients, including 148 with LVI-positive and 141 with LVI-negative lesions. These patients had undergone preoperative breast MRI and were histopathologically confirmed to have invasive breast cancer without axillary LN metastasis. The cohort was randomly split into training and validation groups in a 7:3 ratio. Radiomics features for each lesion were extracted from the first post-contrast dynamic contrast-enhanced (DCE)-MRI. The Least Absolute Shrinkage and Selection Operator (LASSO) regression method and logistic regression analyses were employed to identify significant radiomic features and clinicoradiological variables. These models were established using four machine learning (ML) algorithms and one DL algorithm. The predictive performance of the models (radiomics, clinicoradiological, and combination) was assessed through discrimination and compared using the DeLong test. Four clinicoradiological parameters and 10 radiomic features were selected by LASSO for model development. The Multilayer Perceptron (MLP) model, constructed using both radiomic and clinicoradiological features, demonstrated excellent performance in predicting LVI, achieving a high area under the curve (AUC) of 0.835 for validation. The DL model (MLP-radiomic) achieved the highest accuracy (AUC = 0.896), followed by DL model (MLP-combination) with an AUC of 0.835. Both DL models were significantly superior to the ML model (RF-clinical) with an AUC of 0.720. The DL model (MLP), which integrates radiomic features with clinicoradiological information, effectively aids in the preoperative determination of LVI status in patients with invasive breast cancer and negative axillary LNs. This is beneficial for making informed clinical decisions.
(© 2024. The Author(s).)

BMC Surg. 2023 Jul 26;23(1):209. (PMID: 37495945)
J Magn Reson Imaging. 2017 Oct;46(4):1037-1044. (PMID: 28370761)
Ann Oncol. 2019 Oct 1;30(10):1541-1557. (PMID: 31373601)
Breast Cancer. 2015 Jan;22(1):66-70. (PMID: 25336185)
Front Oncol. 2022 Jun 06;12:876624. (PMID: 35734595)
Ann Surg Oncol. 2003 Dec;10(10):1140-51. (PMID: 14654469)
Curr Oncol. 2024 Jan 10;31(1):403-424. (PMID: 38248112)
J Magn Reson Imaging. 2019 Sep;50(3):847-857. (PMID: 30773770)
Diagn Interv Imaging. 2023 Jun;104(6):275-283. (PMID: 36739225)
Eur Radiol. 2021 Jun;31(6):3703-3711. (PMID: 33270144)
Cancer Med. 2024 Feb;13(3):e6932. (PMID: 38230837)
Acad Radiol. 2022 Jan;29 Suppl 1:S126-S134. (PMID: 34876340)
Breast Care (Basel). 2013 Jun;8(3):221-9. (PMID: 24415975)
Eur Radiol. 2023 Sep;33(9):6204-6212. (PMID: 37017702)
Contrast Media Mol Imaging. 2018 May 3;2018:8329041. (PMID: 29853811)
Eur Radiol. 2022 Jun;32(6):4079-4089. (PMID: 35050415)
PLoS One. 2020 Jun 25;15(6):e0235278. (PMID: 32584887)
Radiology. 2019 Sep;292(3):520-536. (PMID: 31361209)
J Magn Reson Imaging. 2024 Jun;59(6):2238-2249. (PMID: 37855421)
Eur J Radiol. 2024 Jun;175:111457. (PMID: 38640824)
Br J Radiol. 2019 Nov;92(1103):20190417. (PMID: 31398071)
Breast J. 2013 Jul-Aug;19(4):374-81. (PMID: 23714006)
J Surg Oncol. 2014 Jun;109(8):764-9. (PMID: 24647874)
Ann Surg Oncol. 2017 Jan;24(1):38-51. (PMID: 27646018)
J Natl Compr Canc Netw. 2021 Jan 06;19(1):77-102. (PMID: 33406487)
Ann Surg. 2004 Aug;240(2):306-12. (PMID: 15273556)
JAMA Oncol. 2023 Nov 1;9(11):1557-1564. (PMID: 37733364)
Eur J Radiol. 2019 Jan;110:30-38. (PMID: 30599870)
Front Oncol. 2023 Nov 17;13:1269971. (PMID: 38053656)
J Breast Imaging. 2022 Apr 15;4(2):168-176. (PMID: 38422427)
BMC Cancer. 2017 May 17;17(1):335. (PMID: 28514957)
Breast Cancer. 2004;11(2):169-74. (PMID: 15550863)
Cancer Med. 2017 Dec;6(12):2775-2786. (PMID: 29072365)
Oncotarget. 2017 Jun 27;8(26):42455-42465. (PMID: 28418916)
Cancer Imaging. 2022 Oct 22;22(1):61. (PMID: 36273200)
Eur Radiol. 2024 Feb 9;:. (PMID: 38334760)
Eur J Radiol. 2018 Oct;107:111-118. (PMID: 30292254)
BMC Cancer. 2022 Mar 11;22(1):259. (PMID: 35277131)
Eur Radiol. 2022 Mar;32(3):1676-1687. (PMID: 34767068)
Cancers (Basel). 2023 Sep 26;15(19):. (PMID: 37835431)
Acad Radiol. 2023 Aug;30(8):1620-1627. (PMID: 36414494)
Jpn J Clin Oncol. 2023 Jan 6;53(1):35-45. (PMID: 36156086)
World J Surg Oncol. 2021 Mar 15;19(1):76. (PMID: 33722246)
Radiology. 2018 Apr;287(1):68-75. (PMID: 29315062)

82172025 National Natural Science Foundation of China; 2021YFC2500400 Chinese National Key Research and Development Project

Keywords: Breast cancer; Deep learning; Lymphovascular invasion; Magnetic resonance imaging

Date Created: 20240713 Date Completed: 20240713 Latest Revision: 20240801

20240801

PMC11246470

10.1038/s41598-024-67217-0

39003325