Patients characteristics for the entire cohort in the study

In total, 603 patients met the inclusion criteria (September 2023). Median age at diagnosis was 55 years (range: 22–84 years), 373 females (61.9%) and 492 non-smokers (81.6%). As first-line treatment, EGFR-TKIs were administered to 65.3% patients (394/603), while platinum-based therapy was administered to 34.7% (209/603). Mutations in exons 19 and 21 of EGFR were detected in approximately 42.5% (256/603) and 44.6% (269/603) of patients, respectively. The majority of the Lung-molGPA scores of the patients fell within the range of 2.5–4 (413, 68.5%).

Effect of EGFR-TKIs combined with craniocerebral RT on the prognosis of EGFR-mutant lung adenocarcinoma patients with brain metastasisPatients characteristics for first- or second-generation EGFR-TKIs combined with craniocerebral RT

Before PSM, 603 patients were enrolled in this part of the study after excluding patients who used third-generation EGFR-TKIs in the first-line setting. The patients were divided into two groups based on whether they received craniocerebral RT: the EGFR-TKI + craniocerebral RT group (186 patients) and the EGFR-TKI alone group (417 patients). (Table 2).

Table 2 Clinical features of the EGFR-TKI alone group and EGFR-TKIs combined with cerebral radiotherapy group with chi-square test for Categorical Variables in patients who used first- or second-generation EGFR-TKIs cohort

According to chi-square test results (Table 2), there were differences between the two groups in terms of age (p = 0.0443) and third-generation EGFR-TKIs use (p < 0.001). Univariate and multivariable analyses (Table 3) showed that smoking history, lung-molGPA, time of craniocerebral metastasis, and use of third-generation EGFR-TKIs affected patients’ prognosis. We performed a 1:1PSM to eliminate the impact of these factors. The grouping index was whether it was combined with craniocerebral RT, and predictive factors were sex (male or female), age (when first diagnosed), gene mutation site (19, 21 EGFR mutants), lung-molGPA, and whether EGFR-TKIs were first-line treatment, timing of craniocerebral metastasis (first diagnosis), and use of third-generation EGFR-TKIs. The tendency score was calculated using logistic regression. The nearest neighbor algorithm (caliper: 0.2) was used to match the propensity score. After PSM, 346 patients were successfully matched, with 173 patients in each group. (Table 2).

Table 3 Univariable and Multivariable Analyses of Covariable Associated with OS of the EGFR-TKI alone group and EGFR-TKIs combined with cerebral radiotherapy group in patients who used first- or second-generation EGFR-TKIs cohortSurvival analysis for first- or second- generation EGFR-TKIs combined with craniocerebral RT

Before PSM, 381 deaths (63.2%) were recorded as of October 2023 with median duration of follow-up time 64.5 months. The median OS was 48.8 months, while the median iPFS was 14.2 months (Fig. 1A, B). In total, 264 patients developed iPFS after EGFR-TKI or craniocerebral RT (EGFR-TKI + craniocerebral RT group, n = 249; EGFR-TKI only group, n = 15). Median OS of patients in EGFR-TKI + craniocerebral RT group and EGFR-TKI only group was 50.4 months and 44.0 months, respectively (Fig. 1C). There was no significant difference (log-rank test p = 0.1623; HR, 0.8457; 95% CI, 0.6685 to 1.070; Fig. 1C).

Fig. 1

In patients who used first- or second-generation epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) cohort: A Overall survival (OS) and (B) intracranial progression-free survival (iPFS) of the cohort before propensity score matching (PSM). C Comparison of OS between the EGFR-TKIs combined with craniocerebral RT group and EGFR-TKI alone group before PSM. (a) OS and (b) iPFS of the cohort after PSM. (c) Comparison of OS between the EGFR-TKIs combined with craniocerebral RT group and EGFR-TKI alone group after PSM

After PSM, 206 deaths were recorded in successfully matched patients. The median follow-up period was 60.5 months. In total, 117 patients developed intracranial progression after EGFR-TKI treatment or craniocerebral RT (EGFR-TKI + craniocerebral RT group, n = 102; EGFR-TKI alone group, n = 15). The median OS was 47.3months, and the median iPFS was 13.1 months (Fig. 1a, b). From PSM analysis, median OS of EGFR-TKIs + craniocerebral RT group and EGFR-TKI alone group was 52.0 months and 43.2 months respectively, with significant difference between the two groups (log-rank test p = 0.0363; HR,0.7395;95%CI,0.5574 to 0.9809. (Fig. 1c).

Effect of intervention timing of craniocerebral RT combined with first- or second-generation EGFR-TKIs on prognosis of patients with EGFR mutant lung adenocarcinoma complicated with brain metastasisPatients characteristics and univariate and multivariable analyses of covariates with the OS results

In total, 417 patients who underwent craniocerebral RT were enrolled in this study. Based on the timing of craniocerebral RT combined with first- or second-generation EGFR-TKIs, patients were divided into two groups: upfront (who were treated simultaneously with craniocerebral RT and EGFR-TKIs) and delayed (those who were treated with craniocerebral RT after developing EGFR-TKI resistance). Patient characteristics are shown in Table 4.

Table 4 Characteristics of 417 NSCLC Patients and with chi-square test for Categorical Variables in patients who used first- or second-generation EGFR-TKIs cohort

Univariate analysis of covariates with OS (Table 5) showed that the effects of sex, age, smoking status, EGFR mutation, and timing of EGFR-TKI treatment were not statistically significant. Multivariable analysis showed that lung molGPA, brain metastasis timing, and third-generation EGFR-TKIs were independent predictive factors (p < 0.0001; p = 0.0022, p < 0.0001, respectively).

Table 5 Univariable and Multivariable Analyses of Covariable Associated with OS in patients who used first- or second-generation EGFR-TKIs cohortEffects of intervention timing of craniocerebral RT on iPFS

Of the 417 patients who underwent craniocerebral RT, 249 showed intracranial progression after radiotherapy (upfront group, n = 108; delayed group, n = 141). iPFS analysis was performed in the upfront and treatment-resistant groups. The median iPFS of upfront-group and delay-group was 19 and 11 months, respectively (Fig. 2A), and the difference between the two groups was statistically significant (log-rank test, p < 0.0001; HR: 0.4644; 95% CI: 0.3577–0.6030; Fig. 2A).

Fig. 2

In patients used first- or second-generation epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) cohort: A Comparison of iPFS between the upfront-group and delay-group. B iPFS of patients in group A stratified according to intervention timing of craniocerebral RT. C iPFS of patients in group B stratified according to intervention timing of craniocerebral RT. D Comparison of OS between the upfront-group and delay-group. E OS of patients in group A stratified according to intervention timing of craniocerebral RT. F OS of patients in group B stratified according to intervention timing of craniocerebral RT

In addition, Lung-molGPA had a significant impact on the prognosis of patients; thus, we performed iPFS analysis on the two groups stratified by Lung-molGPA score. For group A (Lung-molGPA score 1–2), the median iPFS was 13.8 and 8.9 months respectively (Fig. 2B), the difference was statistically significant (log-rank test p = 0.0178; HR: 0.5778; 95% CI: 0.3671 to.0.9094; Fig. 2B).

For patients in group B (Lung-molGPA score 2.5–4), the median iPFS of the two groups was 22.1 and 11.7 months respectively (Fig. 2C.). A significant difference was observed between the two groups (log-rank test, p < 0.0001; HR, 0.4209; 95% CI, 0.3058–0.5794; Fig. 2C).

Effects of intervention timing of craniocerebral RT on OS

For the OS analysis of patients in the upfront and delay groups, the median OS was 44.3 and 58.9 months respectively (Fig. 2D). There was a significant difference in the median OS between upfront-group and delay-group (log-rank test, p = 0.0234; HR, 1.320; 95% CI, 1.038–1.678; Fig. 2D).

In addition, patients were divided into groups A and B according to their Lung-molGPA scores. For group A, the median OS of the upfront and delay groups was 27 and 42.1 months, respectively. There was significant difference between upfront-group and delay-group (log-rank test p = 0.0019; HR: 1.813; 95% CI: 1.246 to 2.639; Fig. 2E).

For patients in group B, there was no significant difference in OS between the two groups (log-rank test, p = 0.9642; HR, 0.9926; 95% CI, 0.7172–1.374; Fig. 2F). The median OS of the two groups was 59.40 and 67.70 months, respectively.