Imaging- and Tumor Biomarker-Based Multivariable Model for Early Prediction of Pathologic Complete Response to Neoadjuvant Systemic Therapy in Triple-Negative Breast Cancer

Authors

Beatriz E Adrada
Mary S Guirguis
Lei Huo
Clinton Yam
Debu Tripathy
Rosalind Candelaria
Wei Yang
Miral Patel
Tanya Moseley
Gary J Whitman
Jessica W T Leung
Huong Le-Petross
Deanna L Lane
Nour Abuhadra
Jennifer Litton
Vicente Valero
Kelly K Hunt
Banu Arun
Rania Mohamed
Jia Sun
Anil Korkut
Sanaz Pashapoor
Jason White
Alastair Thompson
Peng Wei
Jingfei Ma
Stacy Moulder
Gaiane M Rauch

Language

English

Publication Date

10-1-2025

Journal

JCO Precision Oncology

DOI

10.1200/PO-25-00410

PMID

41411609

PMCID

PMC12716372

PubMedCentral® Posted Date

12-20-2025

PubMedCentral® Full Text Version

Author MSS

Abstract

Purpose: The response of triple-negative breast cancer (TNBC) to neoadjuvant therapy (NAT) varies widely. This study aimed to determine the performance of clinicopathologic biomarkers and volumetric changes on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in predicting pathologic complete response (pCR) to NAT in patients with TNBC.

Patients and methods: This study included 264 patients with stage I to III TNBC enrolled in a prospective clinical trial. These patients underwent DCE-MRI at baseline and after two and/or after four cycles of dose-dense anthracycline and cyclophosphamide. Tumor volume (TV) was calculated by measuring three tumor dimensions at each time point. Clinicopathologic markers were analyzed. Treatment response at surgery (pCR v non-pCR) was documented. The patients were randomly assigned to discovery and validation cohorts. Multiple logistic regression and receiver operating characteristic analysis were used to assess associations and build predictive models.

Results: Of the 264 patients, 124 (47%) achieved a pCR. The optimal thresholds for TV reduction (TVR) on DCE-MRI were ≥60% after two cycles and ≥90% after four cycles. TVR, Ki-67, and stromal tumor-infiltrating lymphocytes (sTILs) were independently associated with pCR on univariable analysis. A combined model including TVR ≥60% after two cycles, sTILs, and Ki-67 predicted pCR with an AUC of 0.84 (90% CI. 0.76 to 0.92) in the discovery cohort and 0.80 (95% CI, 0.71 to 0.88) in the validation cohort. A combined model including TVR ≥90% after four cycles, sTILs, and Ki-67 predicted pCR with an AUC of 0.79 (90% CI, 0.71 to 0.86) in the discovery cohort and 0.80 (95% CI, 0.72 to 0.88) in the validation cohort.

Conclusion: A model incorporating imaging and clinicopathologic variables showed good performance in predicting pCR.

Keywords

Adult, Aged, Female, Humans, Middle Aged, Biomarkers, Tumor, Cyclophosphamide, Magnetic Resonance Imaging, Neoadjuvant Therapy, Pathologic Complete Response, Prospective Studies, Treatment Outcome, Triple Negative Breast Neoplasms, TNBC, NAT, pCR, DCE-MRI volumetric changes, clinico-pathological biomarkers

Published Open-Access

yes

Recommended Citation

Adrada, Beatriz E; Guirguis, Mary S; Huo, Lei; et al., "Imaging- and Tumor Biomarker-Based Multivariable Model for Early Prediction of Pathologic Complete Response to Neoadjuvant Systemic Therapy in Triple-Negative Breast Cancer" (2025). Faculty, Staff and Student Publications. 5293.
https://digitalcommons.library.tmc.edu/uthgsbs_docs/5293