Lung-molGPA may stratify the prognostic impact of TP53 co-mutation in EGFR-mutant lung adenocarcinoma with brain metastases: a multi-center retrospective analysis

Abstract

Background:

The optimal approach to treatment intensification for epidermal growth factor receptor (EGFR)-mutant lung adenocarcinoma with brain metastases remains a topic of debate, especially in the context of high-risk molecular features such as TP53 co-mutation. The extent to which baseline clinical risk influences treatment efficacy is yet to be determined.

Objectives:

We aim to systematically evaluate the prognostic significance of EGFR/TP53 co-mutations in treatment-naïve patients with lung adenocarcinoma and newly diagnosed brain metastases.

Design:

This study enrolled 218 treatment-naïve patients with EGFR-mutant lung adenocarcinoma and brain metastases who received first-line third-generation EGFR-tyrosine kinase inhibitors (TKIs). Treatment effects were evaluated using interaction models stratified by Lung-molGPA scores. Stratification: Group A (Lung-molGPA 1–2) versus Group B (Lung-molGPA 2.5–4).

Methods:

Within each group, the influence of TP53 co-mutations on survival was analyzed. Additionally, among patients with TP53 co-mutations, the effects of cranial radiotherapy (CRT) and treatment with either third-generation EGFR-TKIs monotherapy or third-generation EGFR-TKIs combined with chemotherapy on overall survival (OS) were further investigated, taking into account the Lung-molGPA scores.

Results:

The Lung-molGPA significantly influenced the prognostic impact of TP53 mutations and the survival benefits of treatment intensification strategies (log-rank test p = 0.044, hazard ratio (HR) = 2.1832, 95% confidence interval (CI): 1.023–4.661). In patients with low Lung-molGPA scores, intensification with CRT did not correlate with a survival advantage (32.2 vs 28.0 m, log-rank test p = 0.5088, HR = 1.417, 95% CI: 0.4768–4.211). Conversely, in patients with high Lung-molGPA scores, both CRT and systemic intensification were associated with improved outcomes (44.0 vs 26.3 m, log-rank test p = 0.0054, HR = 0.3199, 95% CI: 0.1585–0.6456; 20.1 vs 32.3 m, log-rank test p = 0.0074, HR = 2.547, 95% CI: 1.113–5.831).

Conclusion:

Baseline clinical risk, as delineated by the Lung-molGPA, serves as a crucial determinant of therapeutic benefit in cases of EGFR-mutant lung adenocarcinoma with cerebral metastases. Implementing risk-adapted treatment intensification strategies could potentially prevent overtreatment in patients classified as low-risk, while simultaneously optimizing clinical outcomes in high-risk cohorts.

Plain language summary

Personal treatment for Lung adenocarcinoma with Brain Metastases: The Role of Lung-molGPA and TP53 Mutations

Many patients with EGFR-mutant NSCLC experience a complication where cancer spreads to the brain, known as brain metastases. While third-generation EGFR-TKIs work very well, doctors often debate whether adding brain radiation therapy helps these patients live longer. Furthermore, some patients carry a secondary genetic defect called a TP53 co-mutation, which typically makes the tumor more aggressive. This study was conducted to understand which patients truly benefit from brain radiation and how the TP53 mutation affects their survival based on their overall clinical health status. We carefully analyzed the medical records of 218 lung cancer patients who had brain metastases at the time of their initial diagnosis and received third-generation EGFR-TKIs. We divided the patients into two groups using a standard clinical scoring tool called Lung-molGPA, which measures a patient’s baseline fitness, age, and disease severity. We then compared survival outcomes and how fast the cancer progressed in the brain between patients who received brain radiation and those who did not, while also tracking their TP53 mutation status. Our study revealed that a “one-size-fits-all” approach does not work. For patients in relatively good clinical health (High Lung-molGPA scores): if these patients also carried the aggressive TP53 mutation, their survival outcomes were drastically shortened, indicating they need closer monitoring. However, adding brain radiation significantly extended their survival. For patients in poor clinical health or with more severe disease (Low Lung-molGPA scores): Neither adding brain radiation nor the presence of the TP53 mutation made a significant difference in survival, as their baseline physical frailty dominated their prognosis. These findings help doctors move closer to personalized medicine for lung cancer.

Introduction

Epidemiological studies have shown that approximately 15.9%–28.6% of patients with non-small cell lung cancer (NSCLC) present with brain metastases at the time of initial diagnosis.^1–3 Notably, the incidence of brain metastases is marginally higher in patients with epidermal growth factor receptor (EGFR)-mutated lung adenocarcinoma compared to those with wild-type EGFR, with an occurrence rate of approximately 60%. ⁴ A study of 262 patients diagnosed with EGFR-mutant NSCLC showed that 69% presented with brain metastases at the time of initial diagnosis, whereas 31% developed brain metastases during the course of treatment. ⁵ Among patients with NSCLC, the median survival duration following the onset of brain metastasis is only 4–8 months, contributing to poor prognosis, diminished quality of life, recurrence, and mortality.^6,7

Targeted therapy and radiotherapy (RT) represent the primary therapeutic strategies for patients with EGFR-mutated lung adenocarcinoma and associated brain metastases. ⁸ In recent years, third-generation EGFR tyrosine kinase inhibitors (EGFR-TKIs), such as osimertinib, have emerged as a frontline treatment option due to their ability to penetrate the blood-brain barrier, thereby offering enhanced efficacy relative to first- and second-generation EGFR-TKIs.^9–11 Additionally, osimertinib exhibits significantly lower toxicity compared to the second-generation inhibitor afatinib, effectively addressing the limitations associated with earlier generations of TKIs.^12,13 Cranial radiotherapy (CRT) continues to serve as the foundational treatment for patients with brain metastases.^14,15 Recent studies have demonstrated that concurrent stereotactic radiotherapy (SRT) and consolidative SRT significantly prolong intracranial progression-free survival (iPFS) and overall progression-free survival (PFS) compared to EGFR-TKI monotherapy. Moreover, concurrent SRT confers substantial overall survival (OS) benefits relative to EGFR-TKI monotherapy, corroborating findings from our prior research.^16,17

Numerous studies have failed to adequately address the impact of EGFR co-mutations, particularly those involving the TP53 gene, a pivotal tumor suppressor. ¹⁸ TP53 is integral to the regulation of the cell cycle, DNA repair, and apoptosis, and TP53 mutation contributes to cancer cell proliferation, invasion, metastasis, and the development of drug resistance.^19,20 The high prevalence of TP53 mutations in tumor cells may result from selective pressures that favor the evasion of cellular surveillance mechanisms, thereby facilitating the avoidance of cell death. ²¹

Multiple clinical studies have demonstrated that TP53 mutations adversely affect the efficacy of EGFR-TKI therapy. A meta-analysis of 24 studies and a cohort of 2227 patients with NSCLC and EGFR mutations revealed that individuals with EGFR/TP53 co-mutations exhibited significantly shorter PFS and OS compared to those with wild-type TP53.^22,23 A recent study identified that among patients with NSCLC treated with third-generation EGFR-TKIs, those harboring EGFR/TP53 co-mutations exhibited significantly shorter median PFS in contrast to patients with wild-type TP53. ²⁴ Taken together, these data imply that the presence of EGFR/TP53 co-mutations may collectively diminish the therapeutic efficacy of EGFR-TKIs in patients with NSCLC.

Lung-molGPA is a prognostic scoring tool that has undergone extensive validation for patients with NSCLC and brain metastases.^25,26 However, there is currently no standardized methodology for integrating molecular mutation status (e.g., TP53, PIK3CA) into Lung-molGPA scores to devise individualized treatment strategies. In this study, we aim to systematically evaluate the prognostic significance of EGFR/TP53 co-mutations in treatment-naïve patients with lung adenocarcinoma and newly diagnosed brain metastases. In addition, we seek to assess the variations in clinical outcomes across different treatment modalities, including monotherapy, combination chemotherapy, and CRT, among patient cohorts stratified by Lung-molGPA scores.

Materials and methods

Patient cohort

Between May 2019 and April 2024, over 800 patients diagnosed with NSCLC and brain metastases were screened at the Shandong First Medical University Affiliated Cancer Hospital, the Shandong First Medical University Affiliated Provincial Hospital, and the Second Affiliated Hospital of Chongqing Medical University. The inclusion criteria for this study were as follows: (1) a pathological diagnosis of primary lung adenocarcinoma; (2) the presence of EGFR exon 19 or 21 mutation; (3) diagnosis of brain metastases at initial presentation confirmed via enhanced computed tomography or magnetic resonance imaging; (4) availability of comprehensive clinical information, including treatment regimens and clinicopathological characteristics; (5) administration of first-line treatment regimens incorporating third-generation EGFR-TKIs such as osimertinib, almonertinib, or furmonertinib; (6) availability of a complete large panel genetic testing (more than 51 genes) report; and (7) absence of other primary malignant tumors. Patients with incomplete medical records or those who did not satisfy the aforementioned criteria were excluded from the study (Figure 1).

Figure 1.

Flowchart of patient selection and study cohort stratification.

The following variables were collected for analysis: age, gender, smoking history, EGFR mutation status and co-mutation status, presence of brain metastasis at initial diagnosis, and prior treatments. Data were recorded on the commencement date of initial local treatment, initiation of EGFR-TKIs, intracranial progression, recent follow-up, and mortality. Intracranial progression was defined as either the radiological progression of pre-existing brain metastases, the emergence of new brain metastases, or both. OS was calculated from the date of pathological diagnosis of lung adenocarcinoma to the date of the last follow-up or death. The time to intracranial progression was measured from the initiation of brain radiotherapy to the occurrence of intracranial progression. PFS was defined by the time from diagnosis to the date of progression or death. The Lung-molGPA is a specific graded prognostic assessment that incorporates patient age, Karnofsky Performance Status (KPS), the number of extracranial and brain metastases, and genetic mutation status.^25,26 The scoring criteria are detailed in Table 1. The reporting of this study conforms to the ESMO Guidance for Reporting Oncology real-World evidence (GROW) statement. ²⁷

Table 1.

Calculation of the Lung-molGPA.

Prognostic	0	0.5	1.0
Age (years)	⩾70	<70	NA
KPS	<70	80	90–100
ECM	Present		Absent
Brain metastases, number	>4	1–4	NA
Gene status	EGFR neg/unk and ALK neg/unk	NA	EGFR pos or ALK pos

ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, graded prognostic assessment; KPS, Karnofsky Performance Status; NA, not applicable; neg/unk, negative or unknown; pos, positive.

EGFR genotyping

Genomic DNA was extracted from tissue samples obtained via fiberoptic bronchoscopy or biopsy. Gene mutations were identified using next-generation sequencing (NGS). All patients underwent targeted NGS. The least targeted genes for NGS list were in Table S1, and the largest targeted genes for NGS list were in Table S2.

Radiotherapy

In patients receiving RT, specifically whole-brain radiotherapy (WBRT), the median prescribed dose was 37.5 Gy (range, total dose: 30–46 Gy, 2–3 Gy/day, 10–23 fractions). For localized RT, the median prescribed dose was 45 Gy (range, total dose: 30–56 Gy, 2–6 Gy/day, 10–28 fractions). For stereotactic radiosurgery (SRS), the median prescribed dose was 20 Gy (range, total dose: 18–24 Gy, 12–20 Gy/day, 1–2 fractions). Furthermore, in the Simultaneous Integrated Boost (SIB) group, the median prescribed dose for WBRT was 37.5 Gy (range, 30–60 Gy, 2–3 Gy/day, 10–20 fractions), while the median dose for radiation to localized metastases was 50 Gy (range, 35–52.5 Gy, 2.5–4 Gy/day, 10–20 fractions).

Statistical methods

For statistical analysis, descriptive comparisons were made for each set of features, and categorical variables were evaluated using the Chi-square test. Survival analysis was conducted using the Kaplan–Meier method and Cox proportional-hazard modeling, with the impact of individual variables on survival assessed via the log-rank test. A p-value of less than 0.05 (two-tailed) was considered statistically significant. Statistical analyses were performed using GraphPad Prism version 8.0.1 (GraphPad Software, San Diego, CA, USA) and R version 4.5.3 (R Foundation for Statistical Computing, Vienna, Austria).

Results

The effect of EGFR/TP53 co-mutation on prognosis in patients with EGFR-mutant lung adenocarcinoma and newly diagnosed untreated brain metastases

Patients’ characteristics

The study included 218 patients who satisfied the inclusion criteria as of June 2025. The median age at diagnosis was 61 years, with an age range of 31–85 years. The cohort comprised 140 females (64.2%) and 176 non-smokers (80.7%). Mutations in EGFR exon 19 and exon 21 were identified in approximately 44.0% (96/218) and 56.0% (122/218) of patients, respectively. EGFR/TP53 co-mutations were detected in 108 patients, representing 49.5% of the cohort (Figure 2). The majority of patients had Lung-molGPA scores between 2.5 and 4, accounting for 130 patients (59.6%; Table 2).

Figure 2.

Oncoprint heatmap illustrates the landscape of genetic alterations and clinical characteristics of 218 patients with EGFR-mutated lung adenocarcinoma.

EGFR, epidermal growth factor receptor.

Table 2.

Characteristics of 218 NSCLC patients and with Chi-square test for categorical variables.

Characteristics	TP53mt (n = 108)	TP53wt (n = 110)	p Value
Age, years			0.0006
>60	40	74
⩽60	68	36
Sex			0.2181
Female	65	75
Male	43	35
Smoking status			0.7310
Never	86	90
Former/current	22	20
EGFR mutation			0.7982
Exon 19	48	47
Exon 21	60	63
ECM			0.8299
Present	78	78
Absent	30	32
Lung-molGPA			0.0081
1–2	34	54
2.5–4	74	56
Drug regimen			0.1009
Third-generation EGFR-TKIs monotherapy	61	74
Combination chemotherapy	47	36
Third TKI			0.4241
Osimertinib	59	54
Almonertinib	40	41
Furmonertinib	9	15
CRT			0.3302
Yes	36	30
No	72	80
CRT strategies			0.9838
WBRT	10	8
Local radiotherapy	16	14
SIB	10	8
Timing of CRT			0.3266
Concurrent	29	27
Sequential	7	3

CRT, craniocerebral radiotherapy; ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; NSCLC, non-small cell lung cancer; SIB, simultaneous integrated boost; TKI, tyrosine kinase inhibitors; TP53mt, TP53-mutant group; TP53wt, TP53 wild-type group; WBRT, whole-brain radiotherapy.

Univariate analyses of covariates were conducted for OS and iPFS (Tables 3 and 4). The findings indicated that gender, smoking status, and extracranial metastases did not have a statistically significant impact on prognosis. Multivariate analysis for OS identified age and Lung-molGPA as independent prognostic factors (p < 0.0001; p = 0.0212). Similarly, multivariate analysis for iPFS identified EGFR mutation status (19/21) as an independent prognostic factor (p = 0.0296). The results of PFS analysis are indicated in Figure S1. The distribution of TP53 exon mutations and copy number amplifications is in Table S3.

Table 3.

Univariable and multivariable analyses of covariable associated with OS.

Variable	Univariable analysis			Multivariable analysis
	HR	95% CI	p	HR	95% CI	p
Age (years)
⩽60 vs >60	0.4639	0.3197–0.6732	<0.0001	1.044	1.023–1.065	<0.0001
Sex
Female vs male	1.164	0.7893–1.716	0.2852
Smoking status
Never vs current/former	0.6954	0.42321–1.143	0.1047
EGFR mutation
Exon 19 vs Exon 21	0.7107	0.4873–1.037	0.0808
ECM
Present vs absent	1.004	0.6645–1.517	0.9844
Lung-molGPA
1–2 vs 2.5–4	1.640	1.121–2.400	0.0072	1.6310	1.1147–2.3864	0.0212
TP53 mutation
mt vs wt	1.623	1.083–2.433	0.4698
CRT
Yes vs no	0.8913	0.5956–1.334	0.5827
Drug regimen
Third-generation EGFR-TKIs monotherapy vs combination chemotherapy	0.9775	0.6557–1.457	0.9091
Third-generation EGFR-TKIs
Osimertinib vs Almonertinib	1.024	0.7061–1.484	0.9002

Variables with p < 0.05 in the univariable analysis were entered into the multivariable Cox regression model using the “Enter” method.

CI, confidence interval; CRT, craniocerebral radiotherapy; ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; HR, hazard ratio; mt, mutant group; OS, overall survival; TKI, tyrosine kinase inhibitors; wt, wild type.

Table 4.

Univariable and multivariable analyses of covariable associated with iPFS.

Variable	Univariable analysis			Multivariable analysis
	HR	95% CI	p	HR	95% CI	p
Age (years)
⩽60 vs >60	1.069	0.7105–1.609	0.7475
Sex
Female vs male	1.188	0.7826–1.803	0.4252
Smoking status
Never vs current/former	1.402	0.8174–2.404	0.2707
EGFR mutation
Exon 19 vs Exon 21	0.6088	0.4043–0.9166	0.0170	0.6240	0.4080–0.9543	0.0296
ECM
Present vs absent	0.9866	0.6374–1.527	0.9509
Lung-molGPA
1–2 vs 2.5–4	1.144	0.7449–1.755	0.5300
TP53 mutation
mt vs wt	1.556	1.031–2.347	0.0297	1.526	0.997–2.334	0.0515
CRT
Yes vs no	0.8432	0.5515–1.289	0.4389
Drug regimen
Third-generation EGFR-TKIs monotherapy vs combination chemotherapy	0.9672	0.6285–1.488	0.8795
Third-generation EGFR-TKIs
Osimertinib vs almonertinib	0.7063	0.4570–1.092	0.1015

Variables with p < 0.05 in the univariable analysis were entered into the multivariable Cox regression model using the “Enter” method.

CI, confidence interval; CRT, craniocerebral radiotherapy; ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; HR, hazard ratio; iPFS, intracranial progression-free survival; mt, mutant group; OS, overall survival; TKI, tyrosine kinase inhibitors; wt, wild type.

Effects of EGFR/TP53 co-mutations on OS

As of June 2025, a total of 97 deaths (48.3%) were documented, with a median follow-up duration of 30.4 months. The median OS was determined to be 32.3 months, while the median iPFS was 17.7 months (Figure 3(a) and (b)). Median OS and iPFS in Group A (Lung-molGPA scores 1–2) and Group B (Lung-molGPA scores 2.5–4) were in Figure 3(c) and (d). Median OS in the TP53-mutant group (TP53mt) was 30.6 months compared to 35.0 months for the TP53 wild-type group (TP53wt; log-rank test p = 0.3903; HR = 1.174; 95% CI: 0.8096–1.702; Figure 4(a)). We next performed an OS analysis stratifying patients according to their Lung-molGPA scores. In Group A (Lung-molGPA scores 1–2), the median OS was 30.2 and 22.2 months, respectively, with no statistically significant difference observed (log-rank test p = 0.5241; HR = 0.8427; 95% CI: 0.4968–1.429; Figure 4(b)). In contrast, for patients in Group B (Lung-molGPA scores 2.5–4), the median OS for the two subgroups was 32.0 and 35.0 months, respectively, with a significant difference noted between the groups (log-rank test p = 0.0316; HR = 1.746; 95% CI: 1.032–2.952; Figure 4(c)).

Figure 3.

In patients who were EGFR-mutant lung adenocarcinoma and newly diagnosed untreated brain metastases cohort: (a) OS and (b) iPFS of the cohort. (c) Comparison of OS between Group A (Lung-molGPA scores 1–2) and Group B (Lung-molGPA scores 2.5–4). (d) Comparison of iPFS between Group A and Group B.

EGFR, epidermal growth factor receptor; iPFS, intracranial progression-free survival; OS, overall survival.

Figure 4.

(a) Comparison of OS between the TP53mt group and TP53wt group. (b) Comparison of OS of patients in Group A (Lung-molGPA scores 1–2) between the TP53mt group and TP53wt group. (c) Comparison of OS of patients in Group B (Lung-molGPA scores 2.5–4) between the TP53mt group and TP53wt group. (d) Comparison of iPFS between the TP53mt group and TP53wt group. (e) Comparison of iPFS of patients in Group A (Lung-molGPA scores 1–2) between the TP53mt group and TP53wt group. (f) Comparison of iPFS of patients in Group B (Lung-molGPA scores 2.5–4) between the TP53mt group and TP53wt group.

EGFR, epidermal growth factor receptor; GPA, graded prognostic assessment; iPFS, intracranial progression-free survival; OS, overall survival; TP53mt, TP53-mutant group; TP53wt, TP53 wild-type group.

We performed an interaction test subsequently. A Cox proportional hazards model applied to the entire cohort, including an interaction term between TP53 mutation status and the Lung-molGPA score group (TP53 Status × Lung-molGPA Group). The p-value for this interaction term was 0.044 (HR = 2.1832; 95% CI: 1.023–4.661).

Effects of EGFR/TP53 co-mutations on iPFS

iPFS analysis was performed for both the TP53mt and TP53wt groups. The median iPFS was 17.0 m for the TP53mt group and 19.1 m for the TP53wt group, with the difference reaching statistical significance (log-rank test p = 0.0297; HR = 1.556; 95% CI: 1.031–2.347; Figure 4(d)).

Next, patients were stratified into Group A and Group B according to their Lung-molGPA scores. In Group A, the median iPFS for the TP53mt and TP53wt groups was 17.0 and 16.5 months, respectively, with no statistically significant difference observed (log-rank test p = 0.8483; HR = 0.9371; 95% CI: 0.4785–1.835; Figure 4(e)). Conversely, in Group B, the median iPFS was 15.6 months for the TP53mt group and 25.7 months for the TP53wt group, with a statistically significant difference noted (log-rank test p = 0.0034; HR = 2.153; 95% CI: 1.279–3.624; Figure 4(f)).

The effect of CRT on prognosis in patients with lung adenocarcinoma with untreated brain metastases and EGFR/TP53 mutations

Patient characteristics

In this segment of the study, a cohort of 108 patients with newly diagnosed brain metastases originating from EGFR/TP53-mutated lung adenocarcinoma were enrolled. Participants were stratified into two distinct groups: the CRT group, comprising 36 patients, and the non-CRT group, consisting of 72 patients. The demographic and clinical characteristics of the patients are detailed in Table 5. Thirty-six patients who underwent CRT: 16 patients (44.4%) received targeted local radiotherapy for limited lesions. Ten patients (27.8%) received standard WBRT. And 10 patients (27.8%) received WBRT combined with a SIB to macroscopic intracranial lesions. Radiotherapy timing breakdown: 29 patients (80.6%) received concurrent EGFR-TKI and CRT. And seven patients (19.4%) received sequential treatment.

Table 5.

Characteristics of 108 NSCLC patients and with Chi-square test for categorical variables.

Characteristics	CRT (n = 36)	Non-CRT (n = 72)	p Value
Age, years			0.8879
>60	13	27
⩽60	23	45
Sex			0.4871
Female	20	45
Male	16	27
Smoking status			0.4992
Never	30	56
Former/current	6	16
EGFR mutation			0.2178
Exon 19	13	35
Exon 21	23	37
ECM			0.3621
Present	28	50
Absent	8	22
Lung-molGPA			0.2412
1–2	14	20
2.5–4	22	52
Drug regimen			0.9386
Third-generation EGFR-TKIs monotherapy	20	41
Combination chemotherapy	16	31
Third-generation EGFR-TKIs			0.3133
Osimertinib	20	39
Almonertinib	15	25
CRT strategies
WBRT	10
Local radiotherapy	16
SIB	10
Timing of CRT
Concurrent	29
Sequential	7

CRT, craniocerebral radiotherapy; ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; NSCLC, non-small cell lung cancer; SIB, Simultaneous Integrated Boost; TKI, tyrosine kinase inhibitors; WBRT, whole-brain radiotherapy.

Analysis using the Chi-square test revealed no statistically significant differences in the variables between the two groups. The results of PFS analysis are indicated in Figure S2.

Effects of CRT on OS

Analysis of OS for patients in the CRT group and non-CRT group revealed median OS values of 36.0 and 28.0 months, respectively (Figure 5(a)); however, the difference in median OS between the CRT and non-CRT groups was not statistically significant (log-rank test p = 0.3268; HR = 0.7469; 95% CI: 0.4253–1.312; Figure 5(a)).

Figure 5.

In patients who were EGFR/TP53 co-mutant lung adenocarcinoma and newly diagnosed untreated brain metastases cohort: (a) Comparison of OS between the CRT group and non-CRT group. (b) Comparison of OS of patients in Group A (Lung-molGPA scores 1–2) between the CRT group and non-CRT group. (c) Comparison of OS of patients in Group B (Lung-molGPA scores 2.5–4) between the CRT group and non-CRT group. (d) Comparison of iPFS between the CRT group and non-CRT group. (e) Comparison of iPFS of patients in Group A (Lung-molGPA scores 1–2) between the CRT group and non-CRT group. (f) Comparison of iPFS of patients in Group B (Lung-molGPA scores 2.5–4) between the CRT group and non-CRT group.

CRT, cranial radiotherapy; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; iPFS, intracranial progression-free survival; OS, overall survival.

Patients were categorized into Group A and Group B based on their Lung-molGPA scores. In Group A, the median OS for the CRT group and non-CRT group was 18.2 and 30.6 months, respectively. Among patients with lower Lung-molGPA scores, a difference was observed between the CRT and non-CRT groups (log-rank test p = 0.0506; HR = 2.213; 95% CI: 0.8298–5.900; Figure 5(b)).

For patients in Group B, there was a significant difference in OS between the two groups (log-rank test p = 0.0054; HR = 0.3199; 95% CI: 0.1585–0.6456). The median OS for the CRT group and non-CRT group was 44.0 and 26.3 months, respectively (Figure 5(c)).

Similarly, a single Cox proportional hazards model for the entire cohort was used to test for formal interaction. The model included an interaction term (Lung-molGPA Group × CRT Status) between the prognostic score and the treatment. The p-value of interaction term was 0.0032 (HR = 0.2671; 95% CI: 0.1111–0.6425).

Effects of CRT on iPFS

In the cohort of 108 patients with EGFR/TP53 co-mutations, iPFS analysis was conducted for the CRT group and the non-CRT group. Median iPFS in the CRT group and the non-CRT group was 17.5 and 15.0 months, respectively (Figure 5(d)), with no statistically significant difference observed (log-rank test p = 0.3912; HR = 0.7768; 95% CI: 0.4427–1.363; Figure 5(d)).

We next performed iPFS analysis based on Lung-molGPA scores. For Group A (Lung-molGPA scores 1–2), the median iPFS was 11.7 and 22.0 months, respectively (Figure 5(e)), with the difference not reaching statistical significance (log-rank test p = 0.1291; HR = 2.204; 95% CI: 0.7477–6.496; Figure 5(e)).

For patients in Group B (Lung-molGPA scores 2.5–4), the median iPFS for the two groups was 21.3 and 13.6 months, respectively (Figure 5(f)), which was reached statistical significance (log-rank test p = 0.0470; HR = 0.4826; 95% CI: 0.2500–0.9318; Figure 5(f)).

The effect of third-generation EGFR-TKIs and chemotherapy on the prognosis of patients with EGFR/TP53-mutated lung adenocarcinoma and brain metastases who did not undergo CRT

Patients’ characteristics

Overall, there were 72 patients enrolled who did not undergo CRT. Based on whether chemotherapy was combined, the patients were divided into two groups: the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. The characteristics of patients are shown in Table 6. The results of PFS analysis are indicated in Figure S3.

Table 6.

Characteristics of 72 NSCLC patients and with Chi-square test for categorical variables.

Characteristics	Third TKI monotherapy (n = 41)	Combination chemotherapy (n = 31)	p Value
Age, years			0.7586
>60	16	11
⩽60	25	20
Sex			0.8537
Female	26	19
Male	15	12
Smoking status			0.2795
Never	30	26
Former/current	11	5
ECM			0.8072
Present	28	22
Absent	13	9
Lung-molGPA			0.7454
1–2	12	8
2.5–4	29	23
Third-generation EGFR-TKIs			0.1446
Osimertinib	22	17
Almonertinib	12	13
Furmonertinib	7	1

ECM, extracranial metastases; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; NSCLC, non-small cell lung cancer; TKI, tyrosine kinase inhibitors.

Effects of combination chemotherapy on OS

For patients in the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group, the analysis revealed median OS of 26.3 and 35.0 months, respectively (Figure 6(a)). A significant difference in median OS was observed between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group (log-rank test p = 0.0271; HR = 2.065; 95% CI: 1.068–3.992; Figure 6(a)).

Figure 6.

In patients who were EGFR/TP53 co-mutant lung adenocarcinoma and newly diagnosed untreated brain metastases and without CRT cohort: (a) Comparison of OS between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. (b) Comparison of OS of patients in Group A (Lung-molGPA scores 1–2) between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. (c) Comparison of OS of patients in Group B (Lung-molGPA scores 2.5–4) between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. (d) Comparison of iPFS between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. (e) Comparison of iPFS of patients in Group A (Lung-molGPA scores 1–2) between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. (f) Comparison of iPFS of patients in Group B (Lung-molGPA scores 2.5–4) between the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group.

CRT, cranial radiotherapy; EGFR, epidermal growth factor receptor; GPA, Graded Prognostic Assessment; iPFS, intracranial progression-free survival; OS, overall survival; TKI, tyrosine kinase inhibitors.

Patients were stratified into Group A and Group B based on Lung-molGPA scores. In Group A, the median OS for the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group was 32.2 and 28.0 months, respectively. In the low Lung-molGPA subgroup, no significant difference was found between the two groups (log-rank test p = 0.5088; HR = 1.417; 95% CI: 0.4768–4.211; Figure 6(b)). For patients in Group B, a significant difference in OS was observed between the two groups (log-rank test p = 0.0074; HR = 2.547; 95% CI: 1.113–5.831). The median OS for the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group was 20.1 and 32.3 months, respectively (Figure 6(c)).

We also performed a single Cox proportional hazards model for the entire cohort. The model included an interaction term (Lung-molGPA Group × Drug Regimen) between the prognostic score and the treatment. The p-value of interaction term was 0.1760 (HR = 0.5642; 95% CI: 0.2462–1.2926).

Effects of combination chemotherapy on iPFS

An analysis of iPFS was conducted for the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group. The median iPFS for the third-generation EGFR-TKI monotherapy group and the combination chemotherapy group was 14.0 and 21.0 months, respectively (Figure 6(d)). The difference between the two groups was not statistically significant (log-rank test p = 0.1792; HR = 1.608; 95% CI: 0.8197–3.154; Figure 6(d)).

Subsequently, iPFS analysis was performed for the two groups stratified by Lung-molGPA scores. For Group A (Lung-molGPA score 1–2), the median iPFS was 22.0 months for the third-generation EGFR-TKI monotherapy group and not reached or the combination chemotherapy group (Figure 6(e)). The difference was not statistically significant (log-rank test p = 0.8801; HR = 1.173; 95% CI: 0.1525–9.017; Figure 6(e)).

For Group B patients (Lung-molGPA score 2.5–4), the median iPFS for the two groups was 13.6 and 21.0 months, respectively (Figure 6(f)). The difference between the two groups was significant (log-rank test p = 0.0271; HR = 2.183; 95% CI: 1.029–4.630; Figure 6(f)).

Discussion

In the overall cohort, TP53 mutation was not uniformly associated with survival outcomes. However, when stratified by Lung-molGPA score, a clear effect modification was observed. Among patients with low Lung-molGPA scores (1–2), TP53 mutation showed limited prognostic impact, whereas in patients with high Lung-molGPA scores (2.5–4), TP53 mutation was associated with significantly worse OS and iPFS. Formal interaction testing confirmed a significant interaction between TP53 mutation status and Lung-molGPA score, indicating that the prognostic relevance of TP53 mutation is conditional on baseline clinical risk. It is worth noting that the difference in OS between the two prognostic groups was less pronounced than initially expected. This phenomenon can be largely explained by the fact that our entire cohort consisted exclusively of patients with baseline brain metastases. The profound, dominant negative prognostic impact of intracranial disease likely diluted the stratification power of general clinical markers (such as age and KPS), leading to an overlapping survival trajectory. Therefore, the study exclusively enrolled patients who received an initial diagnosis of brain metastasis during their first consultation. The standardization of first-line treatment protocols further augments the reliability of the findings. To the best of our knowledge, the present study represents the largest sample size and the most detailed patient stratification among related studies.

More importantly, the prognostic impact of EGFR/TP53 co-mutations is profoundly related to the Lung-molGPA score, highlighting the necessity of stratified precision treatment approaches. In patients with poor clinical conditions and low Lung-molGPA scores of 1–2, EGFR/TP53 co-mutation status may also affect the OS or iPFS outcomes of this subgroup. However, due to the relatively high tumor burden of brain metastases and the fact that survival time is limited in these patients, the main treatment focus should be symptom relief and ensuring quality of life while extending survival time through the use of third-generation EGFR-TKI monotherapy, which has fewer side effects.^28–30 Moreover, a single Cox proportional hazards model that includes Lung-molGPA Group × TP53 Status results provides stronger, more direct evidence the effect of TP53 co-mutation depends on the Lung-molGPA score(p = 0.044). This phenomenon strongly supports the hypothesis that TP53 status is closely tied to baseline clinical status, with TP53-mutant patients inherently presenting with worse initial Lung-molGPA profiles. More importantly, it underscores that TP53 can only fully exert its negative prognostic effect when general clinical prognostic markers are favorable. In patients with an already compromised clinical status, the profound immediate disease burden serves as a dominant prognostic driver that overshadows subtle molecular influences. However, no clear OS benefit was demonstrated, and potential harm cannot be excluded, but residual confounding precludes firm conclusions.

Subsequently, we investigated the impact of baseline clinical risk on survival outcomes following intracranial radiotherapy. In patients presenting with lower Lung-molGPA scores, intracranial radiotherapy did not correlate with enhanced survival outcomes and, in fact, demonstrated a trend toward inferior results. Conversely, among patients with elevated Lung-molGPA scores, intracranial radiotherapy was significantly correlated with extended OS. The study identified a significant interaction between intracranial radiotherapy and Lung-molGPA scores, suggesting that the benefit–risk profile of the treatment varies across different clinical risk strata. Moreover, a single Cox proportional hazards model that includes Lung-molGPA Group × CRT Status results provide stronger, more direct evidence the effect of CRT truly depends on the Lung-molGPA score (p = 0.0032). However, this finding must be interpreted with caution due to inherent selection bias (confounding by indication). Patients with higher Lung-molGPA scores often present with a limited number of intracranial lesions (i.e., an oligometastatic state) and a favorable performance status. In our cohort, 16 patients underwent localized/ablative CRT targeted at a limited number of brain metastases (Table 5), the majority of whom (10 out of 16) were clustered in the high Lung-molGPA score group. Clinicians are naturally more inclined to offer aggressive local treatments, such as SRS or localized fractionated RT, to these oligometastatic patients. This approach achieves excellent, durable local control while sparing the neurocognitive toxicities associated with WBRT. Therefore, the superior OS observed in the high Lung-molGPA group likely reflects a synergistic effect of a lower baseline systemic tumor burden and the high efficacy of localized ablative therapy, rather than a generalized benefit of radiotherapy across all patients with good prognosis. The incorporation of CRT in lower Lung-molGPA scores patients may result in a deterioration of neurocognitive function, consequently diminishing their quality of life. ³¹ Our study results also showed that CRT is associated with a shorter median OS (18.2 months) compared to non-CRT (30.6 months; p = 0.0506), which suggests that the toxicity or immunosuppressive effects associated with CRT may outweigh the benefits for patients with typically poor prognosis and low Lung-molGPA scores. Furthermore, when interpreting the therapeutic efficacy of CRT, the clinical impact of baseline numerical imbalances cannot be ignored. In our cohort, patients in the CRT group presented with a numerically higher proportion of the EGFR exon 21 mutation (64% vs 51%) and a greater prevalence of poor Lung-molGPA scores (39% vs 28%) compared to the non-CRT group. Although these variations did not achieve statistical significance, they reflect an inherently worse prognostic profile in the CRT cohort at baseline. A recent real-world study by emphasized that EGFR-mutated lung adenocarcinoma carries a unique biology leading to distinct brain metastasis patterns and compromised OS; notably, the exon 21 L858R mutation has been widely associated with less durable responses to EGFR-TKIs and worse clinical outcomes than exon 19 deletions. ³² Consequently, this disproportionate baseline risk in the CRT group likely acted as a negative confounding factor, potentially masking the true survival benefits of intensive cranial intervention.

In patients with relatively better clinical conditions and elevated Lung-molGPA scores (2.5–4), the presence of EGFR/TP53 co-mutations significantly worsens OS (32.0 vs 35.0 months, p = 0.0316) and, notably, iPFS (15.6 vs 25.7 months, p = 0.0072) compared to those with TP53 wild-type. This pronounced disparity underscores the pivotal role of TP53 mutations in disease progression and metastasis within this patient cohort. Additionally, research suggests that within the mechanisms and roles associated with the intrinsic sensitivity of NSCLC cells and the development of acquired resistance to EGFR-TKIs, TP53 mutations facilitate epithelial–mesenchymal transition in H1975 cells. This process activates EGFR mutations and augments resistance to osimertinib.^33,34 Consequently, there is a necessity for aggressive combination treatment strategies to mitigate the adverse effects associated with EGFR/TP53 co-mutations. Given the generally favorable clinical status of these patients, their therapeutic objectives diverge from those with lower Lung-molGPA scores. With a relatively low tumor burden in brain metastases, the primary aim for this group should be achieving long-term remission, warranting consideration of more aggressive combination therapies to enhance local control and extend survival.^6,17,35 The findings indicate that CRT was significantly associated with a prolonged OS for patients with TP53-mutant tumors within this higher scoring group (mOS of 44.0 months compared to 26.3 months without CRT, p = 0.0054). Moreover, studies suggest that CRT may enhance the permeability of the blood-brain barrier, thereby increasing the concentration of intracranial drugs and improving the local control rate of intracranial lesions.^36,37 Similarly, third-generation EGFR-TKIs increase radio-sensitivity, which improves the local control rate of radiotherapy. ³⁸ An in vitro study revealed that following the transplantation of tumor cells into nude mice, the cohort subjected to sequential radiotherapy and osimertinib exhibited a statistically significant inhibition of tumor growth during and post-treatment (p < 0.05). This combined in vivo and in vitro investigation substantiates the synergistic interaction between osimertinib and radiotherapy. ³⁸ This evidence strongly advocates for the early incorporation of CRT in the management of these patients.

Owing to the insufficient availability of radiotherapy facilities in numerous domestic hospitals, we conducted an analysis on TP53-mutant patients who did not undergo CRT. ³⁹ Our findings demonstrate that, in comparison to third-generation EGFR-TKI monotherapy, the addition of chemotherapy significantly enhanced OS, particularly within the subgroup characterized by a higher Lung-molGPA score. Furthermore, this combination therapy markedly improved iPFS within the same subgroup. Prior research has established that in populations exhibiting EGFR/TP53 co-mutations, treatment protocols integrating EGFR-TKIs and chemotherapy may yield enhanced patient outcomes. The findings of the FLAURA2 study suggest that the integration of osimertinib with chemotherapy results in enhanced PFS by approximately 9 months, while maintaining safety within manageable parameters. Furthermore, for patients presenting with brain metastases, this combined therapeutic approach may constitute a more effective treatment alternative.^40,41 A retrospective analysis of 95 patients with EGFR/TP53 co-mutated NSCLC revealed that, in comparison to EGFR-TKI monotherapy, the combination of EGFR-TKI with platinum-based doublet chemotherapy significantly improved both the objective response rate (ORR) and time to progression (TTP) in patients with advanced NSCLC harboring these co-mutations. Specifically, the ORR was 55.9% in the combination therapy cohort versus 34.4% in the monotherapy cohort (p = 0.042), and the TTP was extended to 16.1 months from 11.1 months (p = 0.002). ⁴² Another retrospective investigation examined the potential benefits of combining EGFR-TKI and chemotherapy for NSCLC patients with EGFR-sensitive mutations, while also assessing the influence of EGFR/TP53 co-mutations on the molecular characteristics of the treatment regimen. The findings demonstrated that in patients harboring EGFR/TP53 co-mutations, combination therapy yielded a significantly greater PFS advantage relative to monotherapy, with median PFS durations of 19.1 versus 11.4 months, respectively (p = 0.001). ⁴³ These results suggest that systemic treatment intensification may effectively mitigate the intrinsic resistance associated with TP53 dysfunction. Moreover, osimertinib is a highly potent and selective EGFR-TKI, whereas chemotherapy drugs have a nonselective antitumor effect. Thus, it is possible that the combination overcomes intratumor heterogeneity by eliciting an additive effect by means of the killing of different tumor cell populations to improve clinical outcomes.^40,44–47 For patients with relatively better clinical conditions and high Lung-molGPA scores, combination chemotherapy may offer substantial benefits when CRT is not a viable option. Owing to the Cox proportional hazards model indicated that combination chemotherapy was not dependent on a patient’s baseline Lung-molGPA score in the present study (p = 0.1760), this lack of statistical significance may be attributable to the limited sample size and insufficient statistical power. Furthermore, the FLAURA2 trial unequivocally demonstrated that this combination regimen was associated with a higher incidence of grade ⩾3 adverse events. ³⁶ Integrating our preliminary findings with those of the FLAURA2 trial provides a more comprehensive clinical perspective: In patients harboring EGFR/TP53 co-mutations who did not undergo CRT, the intensification of systemic treatment through chemotherapy was correlated with enhanced survival outcomes in the high Lung-molGPA subgroup. Conversely, no discernible benefit was observed in patients classified as low risk. These results imply that systemic treatment intensification may partially mitigate the effects of adverse tumor biology in certain high-risk patients when local treatment intensification is not a viable option. Overall, our study offers a novel perspective on personalized precision therapy for patients harboring EGFR/TP53 co-mutations with brain metastases. Future research should focus on identifying reliable biomarkers beyond Lung-molGPA to more accurately define the patient population most likely to derive maximal benefit with minimal risk, thereby advancing the pursuit of genuine precision medicine.

This study addresses a critical issue in contemporary clinical practice: the pursuit of precision therapy. By integrating TP53 mutation status with Lung-molGPA score to predict treatment response, we aimed to inform personalized treatment decisions for patients newly diagnosed with brain metastases resulting from EGFR-mutant lung adenocarcinoma. Furthermore, retrospective studies, in contrast to prospective studies, can rapidly integrate real-world cases to investigate efficacy variations across diverse populations, examine the relationships between various factors and prognostic outcomes, and offer guidance for future prospective research. However, there are several limitations to our study. Firstly, due to the relatively recent introduction of third-generation EGFR-TKIs, a significant proportion of patients have not yet reached survival endpoints. Secondly, due to the retrospective nature of our analysis, which carries inherent risks of selection bias and unmeasured confounding variables and the specificity of the study population resulted in a limited sample size. Consequently, the long-term survival differences and sub-cohort estimates reported here should be interpreted as exploratory trends. To address this limitation, our team is actively and continuously maintaining the long-term follow-up of this patient cohort, and updated survival analyses with a higher density of events and greater data maturity will be presented in future publications to confirm the absolute long-term survival trajectories of these molecular subgroups. Secondly, the study grouped all forms of CRT (WBRT, local radiotherapy, and SIB), as well as different treatment timings (concurrent vs sequential with TKI), under a single category termed “CRT.” This represents a methodological limitation, as these modalities are applied to clinically distinct patient populations. However, given the retrospective nature, this is highly likely due to selection bias (i.e., patients in Group A with worse symptoms or larger tumor burden were selected for WBRT/CRT as a salvage or palliative measure, leading to shorter OS). Moreover, WBRT is generally reserved for patients with extensive metastatic burden or poorer prognosis, whereas SRS is preferred for those with a limited number of lesions and better clinical status. Combining these heterogeneous treatment modalities into a single analytical group may introduce confounding bias. Thirdly, the sample sizes for certain specialized treatment combinations within specific prognostic strata were highly limited (for instance, only eight patients received combination chemotherapy in Group A). Consequently, the statistical power for these granular sub-analyses was severely restricted, as reflected by the non-significant interaction test (p = 0.1760). Finally, although we successfully characterized the baseline presence of TP53 co-mutations, our analysis treated it as a binary variable due to sample size constraints. Emerging translational frameworks have elegantly proven that TP53 variants are not functionally uniform; specific sub-classes—such as disruptive mutations within the structural DNA-binding domain (Exons 5–8), Exon 8 hotspots, truncating alterations, or co-amplifications—harbor distinct, tiered prognostic weights regarding early failure on third-generation EGFR-TKIs.^18,48 Furthermore, the dynamics of variant allele frequency and parallel genomic bypasses involving RB1, PIK3CA, or CDKN2A remain critical modifiers of clinical progression. Although the predominance of DBD mutations (74.1%) in our cohort underscores a highly pathogenic phenotype, future large-scale collaborative prospective validation is strictly required to dissect the independent effect size of each molecular sub-class. In light of these limitations, we propose to conduct large-scale, multi-institutional prospective studies to validate the current findings and further elucidate the optimal sequencing and combination of treatment strategies.

Conclusion

The integration of Lung-molGPA scores with EGFR/TP53 co-mutation status serves as an effective framework for tailoring treatment strategies in patients with untreated EGFR/TP53 co-mutation lung adenocarcinoma and brain metastases. The present study highlights that treatment benefit in EGFR/TP53 co-mutation lung adenocarcinoma with brain metastases is not uniform but strongly dependent on baseline clinical risk. Lung-molGPA emerges as a key effect predictor that determines whether treatment intensification confers benefit or potential harm. In patients with low Lung-molGPA scores, aggressive treatment intensification, particularly CRT, may represent overtreatment with limited survival benefit and potential neurotoxicity. Conversely, in patients with high Lung-molGPA scores, failure to intensify treatment may represent a missed opportunity, as both local and systemic intensification strategies appear to improve outcomes in this subgroup. The findings of our study underscore the importance of integrating molecular characteristics and prognostic scoring systems into clinical decision-making processes to facilitate personalized treatment regimens.