Applying the QUARTZ Trial Results in Clinical Practice: Development of a Prognostic Model Predicting Poor Outcomes for Non-small Cell Lung Cancers with Brain Metastases

• Published in: Clinical oncology (Royal College of Radiologists (Great Britain)) Vol. 30; no. 6; pp. 382 - 390
• Main Authors: Agarwal, J.P., Chakraborty, S., Laskar, S.G., Mummudi, N., Patil, V.M., Upasani, M., Prabhash, K., Noronha, V., Joshi, A., Purandare, N., Tandon, S., Arora, J., Badhe, R.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-06-2018
• Subjects: Brain metastases; non-small cell lung cancer; prognostic model; QUARTZ; short-term survival; non-small cell lung cancer; Brain metastases; short-term survival; prognostic model; QUARTZ
• ISSN: 0936-6555; 1433-2981
• DOI: 10.1016/j.clon.2018.02.002

The role of whole brain radiotherapy (WBRT) in patients with brain metastases from non-small cell lung cancers (NSCLC) has been questioned. However, no reliable criteria exist to identify patients who do not benefit from WBRT. The objective of the current study was to develop a prognostic model to identify such patients whose survival matches that of the Quality of Life after Treatment for Brain Metastases (QUARTZ) study.
Outcome data of patients with NSCLC with brain metastases undergoing WBRT enrolled in a prospective observational study in a tertiary cancer centre were used to develop a prognostic model. Baseline clinico-radiological factors were used for development of the model. The model was internally validated and calibration accuracy was checked for prediction of 70 day mortality. The generated prognostic model was presented as a nomogram.
The median overall survival of 140 patients enrolled in the study was 166 days (95% confidence interval 108–242 days). The prognostic model identified gender, Karnofsky performance status and epidermal growth factor receptor activating mutation status as significant factors influencing overall survival. The model showed a modest discriminative ability with an optimism-corrected C-index of 0.64. However, model calibration error did reveal a moderate degree of calibration error. The high-risk subgroup identified by the model had a median overall survival of 67 days (95% confidence interval 56–101 days), which was similar to that observed in the QUARTZ trial.
This prognostic model derived from traditional clinico-radiological features had a modest ability to identify patients with poor prognosis who may not benefit from WBRT. However, the high-risk subgroup identified using this prognostic model had a survival similar to that observed for patients in the QUARTZ trial.
•Prognostic model for identifying a high-risk subgroup of NSCLC with brain metastases.•Model can identify patients with survival less than 70 days.•Nomogram generated for routine clinical use.•Survival of the poor prognostic group similar to patients from the QUARTZ trial.