Characteristics of recurrence, predictors for relapse and prognosis of rapid relapse triple-negative breast cancer

Abstract

Background: Triple-negative breast cancer (TNBC) patients who recur at different times are associated with distinct biological characteristics and prognoses. Research on rapid-relapse TNBC (RR-TNBC) is sparse. In this study, we aimed to describe the characteristics of recurrence, predictors for relapse, and prognosis in rrTNBC patients.

Methods: Clinicopathological data of 1584 TNBC patients from 2014 to 2016 were retrospectively reviewed. The characteristics of recurrence were compared between patients with RR-TNBC and slow relapse TNBC(SR-TNBC). All TNBC patients were randomly divided into a training set and a validation set to find predictors for rapid relapse. The multivariate logistic regression model was used to analyze the data of the training set. C-index and brier score analysis for predicting rapid relapse in the validation set was used to evaluate the discrimination and accuracy of the multivariate logistic model. Prognostic measurements were analyzed in all TNBC patients.

Results: Compared with SR-TNBC patients, RR-TNBC patients tended to have a higher T staging, N staging, TNM staging, and low expression of stromal tumor-infiltrating lymphocytes (sTILs). The recurring characteristics were prone to appear as distant metastasis at the first relapse. The first metastatic site was apt to visceral metastasis and less likely to have chest wall or regional lymph node metastasis. Six predictors (postmenopausal status, metaplastic breast cancer,≥pT3 staging,≥pN1 staging, sTIL intermediate/high expression, and Her2 [1+]) were used to construct the predictive model of rapid relapse in TNBC patients. The C-index and brier score in the validation set was 0.861 and 0.095, respectively. This suggested that the predictive model had high discrimination and accuracy. The prognostic data for all TNBC patients showed that RR-TNBC patients had the worst prognosis, followed by SR-TNBC patients.

Conclusion: RR-TNBC patients were associated with unique biological characteristics and worse outcomes compared to non-RR-TNBC patients.

Keywords: biological characteristics analysis; prognostic analysis; rapid relapse; slow relapse; triple negative breast cancer.

Figure 1

Constitution of the study population.BC, breast cancer; HR+,hormone receptor positive; HER2+, human epidermal growth factor receptor 2 positive; TNBC,triple negative breast cancer; RR-TNBC:rapid relapse-triple negative breast cancer; SR-TNBC:slow relapse-triple negative breast cancer; DCIS, ductal carcinoma in situ; LCIS, lobular carcinoma in situ.

Figure 2

Nomogram for predicting rapid relapse of TNBC patients in the training set.

Figure 3

ROC analysis for predicting rapid relapse in the validation set. ROC, receiver operating characteristic; AUC, area under the curve.

Figure 4

Calibration curve analysis for predicting rapid relapse in the validation set.

Figure 5

(A) K-M overall survival analysis of relapsed patients and no relapsed patients; (B) K-M overall survival analysis of different relapsed types in 1584 TNBC patients.

Figure 6

K-M overall survival analysis of all TNBC patients in six predictors for rapid relapse (A) Menopausal status at diagnosis; (B) Pathological pattern; (C) Tumour size staging; (D) Nodal staging; (E) Stromal tumor-infiltrating lymphocytes(sTIL); (F) Her2 expression levels).

Figure 7

Different expressions of sTILs [(A) Low, (B) intermediate, (C) high] and different expression levels of Her2 [(D) IHC 0, (E) IHC 1+, (F) IHC 2+/FISH-negative] were detected in tumors by using hematoxylin and eosin staining (H&E×200).