Predicting prognosis of breast cancer patients with brain metastases in the BMBC registry – comparison of three different prognostic scores

Purpose Several scores have been developed in order to estimate the prognosis of brain metastases (BM) by objective criteria. The original-GPA includes age, Karnofsky-Index (KI), number of BM and extracranial metastases (ECM), the breast-GPA age, KI and tumor-subtype whereas the updated breast-GPA adds subtype and ECM. Aim of this analysis was to validate all three GPA-scores in a large cohort of patients with BM in the Brain Metastases in Breast Cancer (BMBC)-registry.

Methods Retrospective data of 882 patients deriving from the BMBC-registry has been analysed.

Results Median age at BM diagnosis was 57 years. 22.3 % (n = 197) had triple-negative, 33.4 % (n = 295) luminal A, 25.1 % (n = 221) luminal B and 19.2 % (n = 169) HER2-positive breast cancer. Most patients had a good KI (KI > 50 %: 88.4 %, < 50 %: 11.6 %).

All scores were associated with OS. Although the original-GPA had a high sensitivity in predicting 3-months and 12-months-survival (96.2, 92.2 %, score values > vs. ≤3), specificity was low (17.2, 21.8 %), resulting in a low positive predictive value (PPV) (29.7, 62.6 %). Adding more parameters to the score calculation resulted in a higher specificity (53.5 %, 68.7 % for breast-GPA and 35.7, 48.1 % for updated breast-GPA) and PPV (36.1, 75.6 for breast-GPA and 34.2, 69.9 for updated breast-GPA).

Conclusions In this analysis, several clinical parameters and all GPA-scores were significantly associated with OS. All GPA-scores showed only a moderate diagnostic accuracy in predicting OS. Eventhe updated breast-GPA with most of the statistically significant parameters included did not result ina better performance.

Publication History

Article published online:
07 October 2020