Risk Factors for Brain Metastases in Patients With Small Cell Lung Cancer: A Systematic Review and Meta-Analysis

The use of prophylactic cranial irradiation (PCI) for small cell lung cancer (SCLC) patients is controversial. Risk factors for brain metastasis (BM) development are largely lacking, hampering personalized treatment strategies. This study aimed to identify the possible risk factors for BM in SCLC.We systematically searched the Pubmed database (1 January 1995 to 18 January 2021) according to the PRISMA guidelines. Eligibility criteria: studies reporting detailed BM data with an adequate sample size (randomized clinical trials [RCTs]: N ≥50; non-RCTs: N ≥100) in patients with SCLC. We summarized the reported risk factors and performed meta-analysis to estimate the pooled hazard ratios (HR) if enough qualified data (i.e., two or more studies; the same study type; the same analysis method; and HRs retrievable) were available. In total, 61/536 records were eligible (18 RCTs and 39 non-RCTs comprising 13,188 patients), in which 57 factors were reported. Ten factors qualified BM data for meta-analysis: Limited stage disease (LD) (HR = 0.34, 95% CI: 0.17-0.67; P = 0.002) and older age (≥65) (HR = 0.70, 95% CI: 0.54-0.92; P = 0.01) were associated with less BM; A higher T stage (≥T3) (HR = 1.72, 95% CI: 1.16-2.56; P = 0.007) was a significant risk factor for BM. Male sex (HR = 1.24, 95% CI: 0.99-1.54; P = 0.06) tended to be a risk factor, and better PS (0-1) (HR = 0.66, 95% CI: 0.42-1.02; P = 0.06) tended to have less BM. Smoking, thoracic radiotherapy dose were not significant (P >0.05). PCI significantly decreased BM (P <0.001), but did not improve OS in ED-SCLC (P = 0.81). A higher PCI dose did not improve OS (P = 0.11). The impact on BM was conflicting between Cox regression data (HR = 0.59, 95% CI: 0.26-1.31; P = 0.20) and competing risk regression data (HR = 0.74, 95% CI: 0.55-0.99; P = 0.04). Compared to M0-M1a, M1b was a risk factor for OS (P = 0.01) in ED-SCLC, but not for BM (P = 0.19). As regular brain imaging is rarely performed, high-quality data is lacking. Other factors such as N-stage and blood biomarkers had no qualified data to perform meta-analysis. In conclusion, younger age, higher T stage, and ED are risk factors for BM, suggesting that PCI should be especially discussed in such cases. Individual patient data (IPD) meta-analysis and well-designed RCTs are needed to better identify more risk factors and further confirm our findings. Systematic Review Registration: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021228391, identifier CRD42021228391.