Prognostic factors influencing overall survival in stage IV EGFR-mutant NSCLC patients treated with EGFR-TKIs

Abstract

Background

Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) are the established first-line treatment for non-small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR) mutations, but survival advantages in advanced-stage cases remain modest and show interpatient variability. This study seeks to delineate real-world survival outcomes in stage IV EGFR-mutant NSCLC patients undergoing treatment with EGFR-TKIs and to identify independent prognostic factors impacting overall survival (OS). The findings are expected to be a reference for clinicians in managing advanced lung cancer patients.

Methods

Clinical characteristics from advanced EGFR-mutant NSCLC patients who received EGFR-TKIs were retrospectively analyzed to investigate the association between clinical factors and overall survival.

Results

The median OS was 40.60 months (95% confidence interval: 34.72–46.48 months), with expected 1-, 3-, and 5-year OS of 87.2%, 57.8%, and 34.0%, respectively. Patients exhibiting specific characteristics, such as being under 60 years old, having stage IVA disease, adenocarcinoma histology, no liver or brain metastases, negative for neuron-specific enolase (NSE), carcinoembryonic antigen (CEA), cytokeratin fragment 19 (CYFRA 21 − 1), progastrin-releasing peptide (ProGRP), and those who received chemotherapy and radiotherapy exhibited significantly longer survival times compared to patients with the opposite characteristics (being 60 years or older, having stage IVB or IVC disease, non-adenocarcinoma histology, presence of liver or brain metastases, positive for CEA, NSE, ProGRP, CYFRA 21 − 1, and not receiving chemotherapy or radiotherapy) (P < 0.05). Cox univariate and multivariate analyses identified pathological type, liver metastasis, brain metastasis, NSE, CYFRA 21 − 1, EGFR-TKIs type, radiotherapy, and chemotherapy as independent prognostic factors for patients with stage IV NSCLC treated with EGFR-TKIs.

Conclusions

Pathological type, liver metastasis, brain metastasis, NSE, and CYFRA 21 − 1 were identified as independent risk factors for stage IV NSCLC patients treated with EGFR-TKIs, while chemotherapy and radiotherapy were determined to be independent protective factors. Taking icotinib or gefitinib, as opposed to osimertinib, was an independent risk factor for advanced NSCLC patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-025-03569-1.

Introduction

Non-small cell lung cancer (NSCLC) is the predominant form of lung cancer and is typically linked to a poor prognosis [1]. At initial diagnosis, 46.8–61.2% of NSCLC patients identified as having stage IV disease [2]. Stage IV NSCLC is generally considered incurable, with palliative care serving as the cornerstone of management [2]. However, advancements in precision medicine for advanced NSCLC have introduced a range of tyrosine kinase inhibitor regimens. Despite the expansion of treatment options, the prognosis for NSCLC remains unfavorable [3].

Gefitinib, the first clinically approved epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI), demonstrates efficacy in patients with 19del and L858R mutations, but resistance emerges rapidly [4]. Icotinib, China’s first independently developed EGFR-TKI, shares structural similarities with gefitinib but demonstrates lower toxicity and improved tolerability [5]. Osimertinib, known as a third-generation EGFR-TKI, is now regarded as the standard initial treatment for metastatic NSCLC with EGFR mutations, demonstrating high selectivity towards both sensitive EGFR mutations and T790M resistance [6].

Previous studies have demonstrated that, although advanced patients harbor specific EGFR mutations, the survival benefits may still be limited for some. Moreover, the degree of benefit from EGFR-TKI treatment varies among individuals, suggesting the influence of additional prognostic factors in EGFR-TKI-treated patients [7, 8].

Therefore, this research was performed to identify factors influencing the prognosis of individuals with advanced EGFR-mutant NSCLC receiving EGFR-TKIs, aiming to provide more reliable reference information for clinical diagnosis and prognosis assessment.

Methods

Patients

This research analyzed advanced NSCLC patients with EGFR mutations, diagnosed at Tongji Hospital, from November 1, 2014, to November 1, 2023, and patients were collected consecutively. Eligible patients were diagnosed with stage IV NSCLC through endobronchial ultrasound-guided transbronchial needle aspiration biopsy, fibreoptic bronchoscopy, percutaneous lung biopsy, or postoperative pathological examination. Genetic testing confirmed EGFR mutations, and patients were treated with first-line EGFR-TKIs, including osimertinib (80 mg daily), gefitinib (250 mg daily), or icotinib (125 mg three times daily). Exclusion criteria included patients with other primary malignancies, loss to follow-up, or incomplete medical records, resulting in a final sample size of 461. Up to the end of the research period (June 12, 2024), the survival or death statuses of the patients were recorded, and the patients who were still alive were recorded as censored data.

Patient staging was conducted in accordance with the ninth edition of the TNM classification system for lung cancer. The study was approved by the Ethics Committee and Institutional Review Board of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (TJ-IRB202408023).

Data collection

Clinical data were obtained through patient charts, including age, sex, staging, smoking history, hypertension, diabetes, coronary artery disease and pathological type. Details on metastasis were also recorded, encompassing pleural, brain, hepatic, adrenal, bone, intrapulmonary metastasis, malignant pleural effusion, and other metastases. In addition, lung cancer biomarkers such as neuron-specific enolase (NSE), carcinoembryonic antigen (CEA), cytokeratin fragment 19 (CYFRA 21 − 1), squamous cell carcinoma antigen (SCC), and progastrin-releasing peptide (ProGRP), along with genetic testing results and treatment-related variables, were recorded. Treatments included using EGFR-TKIs, immunotherapy, lung surgery, metastasis surgery, radiotherapy, and chemotherapy. Follow-up information was collected through telephone interviews or medical records. The cut-off values for tumor markers were set as follows: NSE at 16.3 µg/L, CEA at 5.0 ng/mL, CYFRA 21 − 1 at 3.3 µg/L, SCC at 1.5 ng/mL, and ProGRP at 65.7 pg/mL, based on the upper limits of normal test values at Tongji Hospital.

Data analysis

The chi-square test was employed to analyze categorical variables. OS was assessed, and survival curves were generated using the Kaplan-Meier method, with comparisons made via the log-rank test. The time-dependent covariate method was used to test the equal proportional risk of all variables (see Additional file 1). Univariate and multivariate Cox proportional hazards models were developed. Variables with a P-value below 0.05 in the univariate analysis were included in the multivariate model. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated. Statistical significance was defined as a P-value < 0.05. All analyses were conducted using SPSS 27.0 (SPSS Inc., Chicago, IL, USA), and a forest plot was created with GraphPad Prism 8 (GraphPad Software, CA, USA).

Results

Basic characteristics of patients

The study enrolled 461 patients with advanced NSCLC, of which 257 (55.7%) were female. The cohort’s median age was 58, with 259 patients (56.2%) classified as stage IVA. The histological analysis identified 444 cases (96.3%) as adenocarcinoma and 17 cases (3.7%) as non-adenocarcinoma, including seven adenosquamous, seven squamous, and three pulmonary sarcomatoid carcinomas. Genetic analysis revealed that 242 patients (52.5%) harbored 19del mutations, 183 (39.7%) carried L858R mutations, and 36 (7.8%) exhibited other types of EGFR mutations. Comprehensive details regarding the patients’ clinical information are provided (Table 1).

Table 1.

Basic characteristics of patients

Characteristics	N
Sex
Female	257(55.7%)
Male	204(44.3%)
Age (years)
< 60	260(56.4%)
≥ 60	201(43.6%)
Stage
IVA	259(56.2%)
IVB or IVC	202(43.8%)
Smoking status
No	317(68.8%)
Yes	144(31.2%)
Hypertension
No	361(78.3%)
Yes	100(21.7%)
Diabetes
No	439(95.2%)
Yes	22(4.8%)
Coronary disease
No	441(95.7%)
Yes	20(4.3%)
Pathology type
Adenocarcinoma	444(96.3%)
Non-adenocarcinoma	17(3.7%)
EGFR mutation
19del	242(52.5%)
L858R	183(39.7%)
Other mutations	36(7.8%)
Pleural metastasis
No	329(71.4%)
Yes	132(28.6%)
Liver metastasis
No	433(93.9%)
Yes	28(6.1%)
Brain metastasis
No	363(78.7%)
Yes	98(21.3%)
Bone metastasis
No	285(61.8%)
Yes	176(38.2%)
Adrenal gland metastasis
No	428(92.8%)
Yes	33(7.2%)
Intrapulmonary metastasis
No	319(69.2%)
Yes	142(30.8%)
Malignant pleural effusion
No	395(85.7%)
Yes	66(14.3%)
Other metastasis
No	420(91.1%)
Yes	41(8.9%)
CEA
Negative	157(34.1%)
Positive	304(65.9%)
SCC
Negative	398(86.3%)
Positive	63(13.7%)
NSE
Negative	286(62.0%)
Positive	175(38.0%)
ProGRP
Negative	336(72.9%)
Positive	125(27.1%)
CYFRA 21 − 1
Negative	238(51.6%)
Positive	223(48.4%)
EGFR-TKIs
Osimertinib	208(45.1%)
Icotinib	200(43.4%)
Gefitinib	53(11.5%)
Immunotherapy
No	423(91.8%)
Yes	38(8.2%)
Lung surgery
No	453(98.3%)
Yes	8(1.7%)
Metastasis surgery
No	455(98.7%)
Yes	6(1.3%)
Radiotherapy
No	284(61.6%)
Yes	177(38.4%)
Chemotherapy
No	257(55.7%)
Yes	204(44.3%)

CEA, carcinoembryonic antigen; CYFRA 21 − 1, cytokeratin-19 fragments; EGFR, epidermal growth factor receptor; EGFR-TKIs, epidermal growth factor receptor tyrosine kinase inhibitors; NSE, neuron-specific enolase; ProGRP, progastrin-releasing peptide; SCC, squamous cell carcinoma antigen

Survival of stage IV EGFR-mutant NSCLC patients

Across the follow-up duration, 233 patients (50.5%) died. The overall cohort had a median OS of 40.60 months (95% CI: 34.72–46.48 months), with estimated 1-, 3-, and 5-year OS of 87.2%, 57.8%, and 34.0%, respectively (Fig. 1).

Fig. 1.

Survival of stage IV EGFR-mutant NSCLC patients. mOS, median overall survival

The study results demonstrated that patients with specific characteristics (age < 60 years, stage IVA, adenocarcinoma, absence of liver metastasis, absence of brain metastasis, CEA-negative, NSE-negative, ProGRP-negative, CYFRA 21-1-negative, receiving radiotherapy, receiving chemotherapy, and receiving osimertinib) had significantly longer survival than those with opposite characteristics (age ≥ 60 years, stages IVB or IVC, non-adenocarcinoma, presence of liver and brain metastasis, CEA-positive, NSE-positive, ProGRP-positive, CYFRA 21-1-positive, not receiving radiotherapy, not receiving chemotherapy, and receiving icotinib or gefitinib) (Table 2), (Fig. 2). Kaplan-Meier curves with no significant differences in other subgroups are provided as additional files (P ≥ 0.05) (Additional file 2).

Table 2.

Comparison of overall survival in patients with stage IV EGFR-mutant NSCLC

Characteristics	Death	Total	mOS(months)	95% CI	P
Sex
Female	129	257	43.13	34.81—51.45	0.416
Male	104	204	39.10	34.29—43.91
Age (years)
< 60	128	260	42.73	34.93—50.53	0.028
≥ 60	105	201	38.57	31.87—45.27
Stage
IVA	126	259	47.17	38.61—55.74	<0.001
IVB or IVC	107	202	32.17	26.67—37.67
Smoking status
No	150	317	43.13	34.59—51.67	0.123
Yes	83	144	39.10	33.95—44.25
Hypertension
No	182	361	40.53	36.53—44.53	0.540
Yes	51	100	42.77	25.17—60.37
Diabetes
No	220	439	40.90	34.68—47.12	0.260
Yes	13	22	27.80	14.00—41.61
Coronary disease
No	224	441	40.40	35.25—45.55	0.520
Yes	9	20	59.93	39.95—79.91
Pathology type
Adenocarcinoma	221	444	41.60	35.47—47.73	0.010
Non-adenocarcinoma	12	17	22.67	0.53—44.81
EGFR mutation
19del	123	242	47.17	38.90—55.44	0.302
L858R	94	183	37.93	33.32—42.54
Other mutations	16	36	37.10	15.86—58.34
Pleural metastasis
No	158	329	38.57	30.96—46.18	0.310
Yes	75	132	42.73	34.00—51.46
Liver metastasis
No	217	433	40.90	34.84—46.96	0.015
Yes	16	28	28.83	18.35—39.31
Brain metastasis
No	185	363	44.10	37.51—50.69	<0.001
Yes	48	98	29.30	26.21—32.39
Bone metastasis
No	143	285	40.63	35.03—46.24	0.269
Yes	90	176	40.40	28.54—52.26
Adrenal gland metastasis
No	215	428	42.03	35.59—48.47	0.053
Yes	18	33	28.00	20.39—35.61
Intrapulmonary metastasis
No	164	319	40.40	36.13—44.67	0.430
Yes	69	142	44.10	32.78—55.42
Malignant pleural effusion
No	201	395	40.63	34.82—46.44	0.177
Yes	32	66	34.50	13.60—55.40
Other metastasis
No	206	420	40.63	34.13—47.13	0.455
Yes	27	41	40.40	25.54—55.26
CEA
Negative	65	157	51.37	39.19—63.55	0.003
Positive	168	304	38.97	35.16—42.78
SCC
Negative	200	398	41.60	35.03—48.17	0.352
Positive	33	63	38.57	31.65—45.49
NSE
Negative	137	286	49.90	41.47—58.33	<0.001
Positive	96	175	32.00	27.01—36.99
ProGRP
Negative	153	336	43.97	37.11—50.83	0.023
Positive	80	125	33.20	26.97—39.43
CYFRA 21 − 1
Negative	95	238	56.73	45.91—67.55	<0.001
Positive	138	223	29.17	26.28—32.06
EGFR-TKIs
Osimertinib	48	208	NR	NR	0.021
Icotinib	151	200	38.53	32.98—44.08
Gefitinib	34	53	52.53	32.59—72.47
Immunotherapy
No	216	423	40.40	36.40—44.41	0.131
Yes	17	38	53.20	30.21—76.19
Lung surgery
No	230	453	40.53	35.38—45.68	0.052
Yes	3	8	71.13	NR
Metastasis surgery
No	229	455	40.53	35.41—45.65	0.682
Yes	4	6	54.97	19.86—90.08
Radiotherapy
No	143	284	36.77	30.76—42.78	<0.001
Yes	90	177	51.40	40.95—61.85
Chemotherapy
No	128	257	32.53	27.75—37.32	<0.001
Yes	105	204	52.20	44.41—59.99

P-values were derived from Kaplan-Meier survival analysis using the Log-Rank test. CEA, carcinoembryonic antigen; CI, confidence interval; CYFRA 21 − 1, cytokeratin-19 fragments; EGFR, epidermal growth factor receptor; EGFR-TKIs, epidermal growth factor receptor tyrosine kinase inhibitors; mOS, median overall survival; NR, not reach; NSE, neuron-specific enolase; ProGRP, progastrin-releasing peptide; SCC, squamous cell carcinoma antigen

Fig. 2.

Kaplan-Meier survival curves for overall survival in different subgroups of stage IV NSCLC patients. CEA, carcinoembryonic antigen; CYFRA 21 − 1, cytokeratin-19 fragments; EGFR-TKIs, epidermal growth factor receptor tyrosine kinase inhibitors; NSE, neuron-specific enolase; ProGRP, progastrin-releasing peptide

COX univariate and multivariate analyses

Cox univariate analysis identified age, stage, pathology type, liver metastasis, brain metastasis, CEA, NSE, ProGRP, CYFRA 21 − 1, EGFR-TKIs, radiotherapy, and chemotherapy as significant factors (P < 0.050). Based on these findings, a multivariate Cox model was constructed, revealing that pathology type, liver metastasis, brain metastasis, NSE, CYFRA 21 − 1, EGFR-TKIs, radiotherapy, and chemotherapy remained significant. These results suggest that pathology type, liver metastasis, brain metastasis, NSE, and CYFRA 21 − 1 serve as distinct risk factors for advanced NSCLC patients. Radiotherapy and chemotherapy serve as independent protective factors. Additionally, using icotinib or gefitinib was identified as a distinct risk factor compared to using osimertinib for advanced NSCLC patients (Table 3).

Table 3.

COX univariate and multivariate analyses in patients with EGFR-mutant stage IV NSCLC

Variables	Univariate analysis		Multivariate analysis
	HR(95%CI)	P	HR(95%CI)	P
Sex
Female	Reference
Male	1.11(0.86—1.44)	0.416
Age (years)
< 60	Reference		Reference
≥ 60	1.34(1.03—1.73)	0.029	1.20(0.91—1.58)	0.191
Stage
IVA	Reference		Reference
IVB or IVC	1.55(1.20—2.01)	<0.001	0.94(0.69—1.29)	0.698
Smoking status
No	Reference
Yes	1.24(0.94—1.62)	0.124
Hypertension
No	Reference
Yes	1.10(0.81—1.51)	0.540
Diabetes
No	Reference
Yes	1.40(0.78—2.50)	0.262
Coronary disease
No	Reference
Yes	0.80(0.41—1.57)	0.521
Pathology type
Adenocarcinoma	Reference		Reference
Non-adenocarcinoma	2.11(1.18—3.77)	0.012	3.57(1.95—6.54)	<0.001
EGFR mutation		0.304
19del	Reference
L858R	1.24(0.94—1.62)	0.126
Other mutations	1.04(0.62—1.75)	0.882
Pleural metastasis
No	Reference
Yes	0.87(0.66—1.14)	0.311
Liver metastasis
No	Reference		Reference
Yes	1.87(1.12—3.12)	0.016	2.15(1.23—3.77)	0.007
Brain metastasis
No	Reference		Reference
Yes	1.94(1.40—2.70)	<0.001	1.94(1.30—2.88)	0.001
Bone metastasis
No	Reference
Yes	1.16(0.89—1.51)	0.270
Adrenal gland metastasis
No	Reference
Yes	1.61(0.99—2.60)	0.055
Intrapulmonary metastasis
No	Reference
Yes	0.89(0.67—1.18)	0.430
Malignant pleural effusion
No	Reference
Yes	1.29(0.89—1.88)	0.178
Other metastasis
No	Reference
Yes	1.17(0.78—1.74)	0.455
CEA
Negative	Reference		Reference
Positive	1.55(1.16—2.07)	0.003	1.24(0.91—1.70)	0.180
SCC
Negative	Reference
Positive	1.19(0.82—1.72)	0.352
NSE
Negative	Reference		Reference
Positive	1.82(1.40—2.38)	<0.001	1.43(1.06—1.92)	0.019
ProGRP
Negative	Reference		Reference
Positive	1.37(1.04—1.79)	0.024	1.21(0.89—1.65)	0.227
CYFRA 21 − 1
Negative	Reference		Reference
Positive	2.35(1.80—3.06)	<0.001	1.77(1.31—2.41)	<0.001
EGFR-TKIs		0.022		<0.001
Osimertinib	Reference		Reference
Icotinib	1.54(1.10—2.16)	0.012	2.49(1.71—3.60)	<0.001
Gefitinib	1.14(0.72—1.79)	0.581	1.89(1.13—3.16)	0.015
Immunotherapy
No	Reference
Yes	0.69(0.42—1.12)	0.133
Lung surgery
No	Reference
Yes	0.34(0.11—1.06)	0.064
Metastasis surgery
No	Reference
Yes	0.81(0.30—2.19)	0.682
Radiotherapy
No	Reference		Reference
Yes	0.58(0.44—0.76)	<0.001	0.62(0.47—0.83)	<0.001
Chemotherapy
No	Reference		Reference
Yes	0.52(0.40—0.68)	<0.001	0.51(0.38—0.67)	<0.001

P-values were calculated using the Wald test in the Cox regression analysis. CEA, carcinoembryonic antigen; CI, confidence interval; CYFRA 21 − 1, cytokeratin-19 fragments; EGFR, epidermal growth factor receptor; EGFR-TKIs, epidermal growth factor receptor tyrosine kinase inhibitors; HR, hazard ratio; mOS, median overall survival; NSE, neuron-specific enolase; ProGRP, progastrin-releasing peptide; SCC, squamous cell carcinoma antigen

Differences in sex, age, and staging in the association between EGFR-TKIs and death

To further investigate potential differences in the association between EGFR-TKIs and mortality among advanced NSCLC patients by sex, age, and stage, separate subgroup analyses were performed according to these variables. The results indicated that among female patients, treatment with icotinib (HR = 2.16, 95% CI: 1.33–3.50, P = 0.002) and gefitinib (HR = 2.00, 95% CI: 1.07–3.74, P = 0.029) was significantly related to a higher risk of mortality than treatment with osimertinib. This association was insignificant among male patients receiving gefitinib or icotinib compared to those taking osimertinib. However, among male patients, treatment with icotinib was associated with a significantly higher risk of death compared to osimertinib (HR = 2.99, 95% CI: 1.65–5.42, P < 0.001).

Among patients under 60, those treated with icotinib had a significantly higher risk of death compared to those receiving osimertinib (HR = 2.61, 95% CI: 1.56–4.36, P < 0.001). This association was not significant for patients treated with gefitinib. In patients aged 60 or above, treatment with icotinib was related to a higher risk of death than osimertinib (HR = 2.46, 95% CI: 1.39–4.35, P = 0.002), while gefitinib use was also linked to a higher risk of death than osimertinib (HR = 2.74, 95% CI: 1.39–5.39, P = 0.004).

Among patients with stage IVA, treatment with icotinib (HR = 3.19, 95% CI: 1.78–5.71, P < 0.001) or gefitinib (HR = 2.30, 95% CI: 1.14–4.63, P = 0.020) was strongly linked to a higher risk of mortality than osimertinib. In stage IVB or IVC patients, icotinib use was also markedly related to an increased risk of death compared to osimertinib (HR = 2.34, 95% CI: 1.43–3.82, P < 0.001) (Fig. 3).

Fig. 3.

Differences in sex, age, and staging between osimertinib and other EGFR-TKIs in the Cox proportional hazard model. CI, confidence interval; HR, hazard ratio

Discussion

With advancements in precision and personalized medicine, it is essential to research stage IV NSCLC patients with EGFR mutations. Assessing the prognostic impact of patients’ clinical characteristics on EGFR-TKI treatment can provide more reliable reference information for clinical diagnosis, treatment, and prognosis assessment, supporting clinicians in making more informed treatment choices.

Patients with non-adenocarcinomas exhibit a lower prevalence of EGFR mutations than those with lung adenocarcinomas, with rates typically between 0% and 20% [9]. Among patients harboring EGFR mutations, non-adenocarcinoma exhibited a shorter PFS (Progression-free survival) than adenocarcinoma (4.60 vs. 11.00 months, P < 0.001), while OS was not assessed [10]. Adenocarcinoma histology is a distinct predictor of improved outcomes for EGFR-mutant NSCLC undergoing EGFR-TKI therapy [11, 12]. In our study, EGFR-TKI treatment resulted in longer OS for adenocarcinoma patients than non-adenocarcinoma patients.

CYFRA 21 − 1 released by cancer cells has been identified as a clinically valuable marker for detecting specific squamous lung cancers [13]. Evidence suggests that increased CYFRA 21 − 1 levels before treatment were linked to an unfavorable prognosis, regardless of the treatment regimen [14]. Takeuchi A et al. demonstrated that among individuals diagnosed with advanced NSCLC receiving EGFR-TKIs, patients exhibiting increased CYFRA 21 − 1 levels had significantly lower PFS and OS than those with normal levels [7]. Normal levels of CYFRA 21 − 1 may predict better PFS, which can be achieved by reintroducing first- and second-generation EGFR-TKIs in patients resistant to osimertinib [15]. In studies of EGFR-TKIs for EGFR-mutant NSCLC, NSE is commonly used as a marker to predict small-cell lung cancer transformation [16–18]. Suh KJ et al. analyzed 151 patients with EGFR-mutant NSCLC receiving gefitinib or erlotinib and found significantly shorter PFS and OS in patients with elevated NSE levels [19]. Doctors can predict patients’ better PFS based on their negative NSE before taking the drug [20]. Our study suggests that NSE negativity is a prognostic factor that predicts patients getting longer OS, but further studies may be needed to prove it.

Identifying organ metastases that impact the prognosis of individuals undergoing EGFR-TKI treatment is crucial for patient education and guiding treatment decisions. We analyzed pleural, brain, bone, adrenal, liver, and other metastases, finding that liver and brain metastases were significant prognostic factors affecting OS in advanced NSCLC. Consistent with previous reports, liver metastases were linked to a poor prognosis, with significantly lower survival compared to patients without liver metastases [21, 22]. Several researchers propose that the HGF/MET signaling pathway may contribute to this phenomenon. Increased levels of HGF have been reported in liver diseases, including various malignancies, and are frequently elevated in patients with liver metastases. As HGF-mediated c-MET activation can intensify ErbB3/PI3K/AKT signaling, liver metastases could theoretically enhance resistance to EGFR-TKIs, leading to a worse prognosis and diminished treatment response. Nonetheless, further investigations are required to confirm this hypothesis [8]. On the other hand, the lack of routine local consolidation therapy in clinical practice results in significantly deteriorated outcomes for patients presenting with liver metastases [21]. Brain metastases frequently occur in NSCLC patients, affecting approximately 25–40% of individuals, usually emerging within two years following the identification of the primary malignancy. Taniguchi Y et al. also demonstrated that brain metastases are linked to worse OS among advanced patients [23].

Previous research has indicated that osimertinib, which targets 19del, L858R, and T790M mutations, achieves superior OS and PFS compared to gefitinib and icotinib [24–27]. In summary, osimertinib is superior to gefitinib and icotinib. Our study indicated that icotinib and gefitinib were linked to poorer OS in patients with advanced-stage cancer compared to osimertinib. Subgroup analyses by sex, age, and stage revealed that the risk associated with icotinib was significantly greater than that of osimertinib. It is important to note that this study did not stratify patients who developed the T790M resistance mutation following treatment with gefitinib or icotinib and subsequently received osimertinib therapy. Therefore, the difference in OS between patients with stage IV NSCLC who initially received first-generation versus third-generation TKIs should be interpreted with caution. Longer follow-up studies and more detailed subgroup analyses are required for further evaluation.

EGFR-TKIs have demonstrated significant efficacy in oligometastatic EGFR-mutant NSCLC, but additional local radiotherapy can still provide further benefits. The phase III SINDAS trial results indicated that administering local radiotherapy to this patient group was safe and led to enhancements in PFS and OS [28]. Numerous investigations have demonstrated that integrating chemotherapy with EGFR-TKIs yields greater efficacy in advanced NSCLC [24, 29, 30]. However, this combination therapy had more adverse effects [24].

This research has certain limitations. First, it is a retrospective analysis conducted at a single institution, limited by a relatively modest sample size. Second, this study focused exclusively on OS as the sole endpoint, while PFS and therapeutic response were not investigated. Third, detailed treatment regimens of the patients were not reviewed. Fourth, some patients received third-generation EGFR-TKIs following disease progression on first-generation EGFR-TKIs. The absence of these data makes it impossible to address the concern that, had osimertinib been appropriately administered following the failure of gefitinib or icotinib, the prognosis might not have differed regardless of the initial EGFR-TKI used. However, our findings offer new insights, and additional extensive research is needed to confirm and deepen understanding of our results.

Conclusions

Unlike previous studies, the patients included were all EGFR mutant stage IV NSCLC, and we included a larger number of study populations with a longer follow-up time. So the results in this study are more appropriate for guiding the treatment of stage IV patients. This study identified pathology type, liver metastasis, brain metastasis, CYFRA 21 − 1, and NSE as independent risk factors for advanced NSCLC patients receiving EGFR-TKIs. In contrast, radiotherapy and chemotherapy were identified as independent protective factors. Regarding long-term survival, the first-line use of osimertinib seems to offer a superior prognosis in stage IV patients; however, further clinical evidence and studies are required for validation. Subgroup analyses by sex, age, and disease stage revealed that the risk linked to icotinib was notably higher than with osimertinib.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CEA

Carcinoembryonic antigen

CI

Confidence interval

CYFRA 21 − 1

Cytokeratin fragment 19

EGFR

Epidermal growth factor receptor

EGFR-TKIs

Epidermal growth factor receptor tyrosine kinase inhibitors

HGF

Hepatocyte growth factor

HR

Hazard ratio

MET

Mesenchymal-epithelial transition

NSE

Neuron-specific enolase

NSCLC

Non-small cell lung cancer

OS

Overall survival

PFS

Progression-free survival

ProGRP

Progastrin-releasing peptide

SCC

Squamous cell carcinoma antigen

Author contributions

(I) Conception and design: L.W. Kuang; (II) Administrative support: All authors; (III) Provision of study materials or patients: All authors; (IV) Collection and assembly of data: All authors; (V) Data analysis and interpretation: L.W. Kuang; (VI) Manuscript writing: L.W. Kuang; (VII) Final approval of manuscript: All authors.

Funding

None.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. The study was approved by the Medical Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology(TJ-IRB202408023). Given that this was a retrospective study, the written informed consent was waived and the verbal informed consent has been approved by the Medical Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology. We informed all participants during the telephone follow-up and obtained their verbal informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.