Purpose: Identifying vulnerabilities in older patients through comprehensive geriatric assessment is crucial but resource-intensive. This prospective study validated shorter tools- Onco-multidimensional prognostic index (MPI) and Geriatric-8 (G8) for predicting 2-year mortality in breast cancer patients and developed a nomogram using significant clinical predictors.
Methods: The geriatric assessment was done in newly diagnosed treatment-naïve breast cancer patients with age ≥65 years. Overall survival was analysed by Kaplan-Meier test; univariate and multivariate analyses done to determine predictors of survival using Cox-proportional hazards model. A nomogram was constructed based on multivariate model and validated using calibration curve.
Results: Among 300 patients (median age 70 years, IQR 67–74), 95.6% had ECOG-PS 0–1. High-grade histology and lympho-vascular invasion (LVI) positivity were seen in 77% and 30%, respectively. Most patients (77.6%) were non-metastatic. Luminal, HER2+ and triple negative subtype was present in 70%, 10.3% and 18.7% tumours. Rate of non-compliance to treatment was 25%. At median follow-up of 28.8 months, 2-year survival was 78%. The 2-year mortality for Onco-MPI categories was 13.36% (low-risk), 25.9% (medium-risk) and 29.8% (high-risk); high versus low-risk showed statistical difference (HR 2.68, 95% CI 1.13, 6.37; p = 0.026). G8 scores ≤14 and ≤12 both significantly predicted survival (p = 0.001; p = 0.005). On multivariate analysis, higher nodal stage, metastatic status, triple negative breast cancer subtype and LVI positivity independently predicted poorer survival. G8 score remained a significant predictor at both cut-offs (≤14 and ≤12) compared to Onco-MPI. A nomogram integrating nodal stage, metastatic status, LVI and G8 predicted 1- and 2-year survival with a C-index of 0.75 (95% CI: 0.65–0.84) and demonstrated good calibration at 12 and 24 months.
Conclusion: In elderly breast cancer patients, G8 score was a stronger predictor of survival, outperforming onco-MPI. The nomogram combining conventional prognostic factors with G8 showed good discrimination for pre-treatment prognostication.
