A Novel Nomogram to Predict Pathological Complete Response in Breast Cancer Patients and Identify Candidates Who Might Omit Surgery: A Large Cohort Study

TL;DR

This study creates a tool to predict which breast cancer patients may achieve complete response to treatment and could potentially avoid surgery.

Contribution

A novel nomogram is developed and validated to predict pCR and identify candidates for non-surgical treatment in breast cancer.

Findings

The nomogram achieved an AUC of 0.756 in predicting pCR in the training set.

Non-surgery patients with high nomogram scores had a 70.7% 5-year OS rate.

After PSM, surgery and non-surgery groups showed significant OS differences across score groups.

Abstract

To establish and validate a nomogram for predicting pathological complete response (pCR) after neoadjuvant therapy (NAT) in breast cancer (BC) patients, aiming to identify subgroups potentially suitable for non‐surgical management.

Between 2010 and 2015, 4402 BC patients (3037 surgery, 1365 non‐surgery) were extracted from the Surveillance, Epidemiology, and End Results (SEER) database, with external validation in 339 BC patients from our hospital. Logistic regression identified pCR predictors and a nomogram model was constructed. Propensity score matching (PSM) was applied to minimize the effect of the imbalance of the prognostic factors between the surgery group and the non‐surgery group.

Univariable and multivariable analyses revealed that age, marital status, T stage, N stage, differentiation grade, hormone receptor (HR) status, human epidermal growth factor receptor 2 (HER2) status and chemotherapy were significant predictors of pCR in the surgery group (all p < 0.05). The area under the receiver operating characteristic curve (AUC) for predicting pCR was 0.756 (95% CI: 0.736–0.776), 0.717 (95% CI: 0.681–0.754), and 0.744 (95% CI: 0.687–0.800) in the training, internal validation, and external validation sets, respectively. Non‐surgery patients were stratified into high (> 375), medium (162.5–375), and low (< 162.5) nomogram score groups, with 5‐year OS rates of 70.7%, 49.2%, and 29.2% (p < 0.05). After PSM, 358 pairs of patients were included to compare 5‐year OS between the surgery group and the non‐surgery group, and the results showed that the 5‐year OS of patients stratified by high‐score group, medium‐score group and low‐score group between the surgery group and the non‐surgery group were 85.4% vs. 72.8%, 78.3% vs. 59.8% and 79.3% vs. 20.0% (all p < 0.05).

We successfully established a nomogram for pCR in BC patients. Based on this predictive model, patients with higher scores may represent potential candidates for non‐surgical therapeutic approaches, warranting further investigation in future studies.

Figures