Real-World Clinical Outcomes and Treatment Patterns in Advanced/Metastatic EGFR-Mutant NSCLC After Progression on First-Line Osimertinib for French Patients

Abstract

Introduction

Osimertinib, a third-generation tyrosine kinase inhibitor (TKI), is a standard first-line (1L) treatment for advanced/metastatic EGFR non-small cell lung cancer (NSCLC). Despite improved overall survival (OS) and progression-free survival (PFS) compared to earlier generation TKIs, resistance and disease progression are common. No standardized second-line (2L) treatment exists for patients who progress after 1L osimertinib. This study assessed real-world (rw) treatment patterns and clinical outcomes in French patients with advanced/metastatic NSCLC who received 2L therapy post-osimertinib progression.

Methods

This retrospective, observational study used the Epidemio-Strategy and Medical Economics (ESME) LC database (NCT03848052) (January 1, 2015–2023; data extraction September 2023). Patients ≥ 18 years with confirmed advanced/metastatic NSCLC, treated with 1L osimertinib, and documented 2L therapy were included. The study assessed patient characteristics, rw treatment patterns, rwOS, and rwPFS.

Results

Among 284 patients (71.1% female, median age 67), 67% initiated 2L treatment in 2021 or later. The median number of metastases at 2L initiation was three, with bone (65.1%) and brain (56.6%) as the most common sites; 34% of brain metastases were symptomatic. The 2L treatments included platinum-based chemotherapy (50%), EGFR TKI-based therapy (29.9%), immunotherapy (9.9%), and other regimens (10.2%). Median rwOS from 2L initiation was 10.1 (95% CI 9.2–12.1) months, and median rwPFS was 4.1 (95% CI 3.3–4.8) months. Among patients with brain metastases, median rwOS was 8.7 (95% CI 7.1–11.9) months for asymptomatic and 10.7 (95% CI 9.4–12.9) months for symptomatic patients vs 11.8 (95% CI 9.1–17.4) months in patients without brain metastases. Patients with liver metastases had a median rwOS of 8.2 (95% CI 6.4–9.6) months vs 12.3 (95% CI 10.2–14.3) months in those without.

Conclusions

Clinical outcomes for patients receiving 2L therapy post-osimertinib progression remain poor, highlighting an unmet need, particularly for those with brain and liver metastases.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-025-03465-4.

Key Summary Points

Why carry out the study?

Osimertinib was the first third generation tyrosine kinase inhibitor (TKI) approved by the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for patients with EGFR non-small cell lung cancer (NSCLC) and has demonstrated efficacy in progression-free and overall survival compared to earlier generation EGFR TKIs.

Despite the improved efficacy, patients with advanced/metastatic NSCLC still progress after treatment with first-line osimertinib and experience poor clinical outcomes.

Real-world (rw) insights into what happens to patients who receive second-line therapy after progression on osimertinib, including subsequent treatment patterns and clinical outcomes, can highlight and elucidate clinical gaps.

What was learned from the study?

Platinum-based chemotherapy + pemetrexed was the predominant treatment regimen for second-line treatment after progression with osimertinib as first-line therapy, reflecting the absence of approved targeted therapies available after progression with osimertinib.

Results of this analysis underscore the poor prognosis post first-line osimertinib therapy with a median rwOS of 10.1 (95% CI 9.2–12. 1) months and a median rwPFS of 4.1 (95% CI 3.3–4.8) months, demonstrating the need for better treatment options.

Introduction

Activating mutations in the epithelium growth factor receptor (EGFR) have been noted in 10–15% of patients with metastatic non-small cell lung cancer (NSCLC) in Europe, with most (80–85%) sensitive to targeted EGFR-tyrosine kinase inhibitors (TKIs) such as osimertinib [1–3]. Osimertinib was the first third-generation EGFR TKI approved and was specifically designed to overcome the resistance experienced when using first- and second-generation TKIs. It selectively targets and irreversibly inhibits both EGFR-sensitizing mutations (exon 19 deletions and L858R) and T790M resistance mutations, helping to overcome resistance [4]. Osimertinib was first approved by the European Medicines Agency (EMA) for first-line (1L) EGFR mutated (EGFRm) NSCLC in July 2018 based on the results of the FLAURA trial (NCT02296125) [5, 6]. It is currently one of the historical standards of care for 1L treatment in patients with advanced or metastatic EGFRm NSCLC globally and in France based on European Society for Medical Oncology (ESMO) guidelines [7, 8]. In 2024, the approval was expanded to use in combination with pemetrexed + platinum-based chemotherapy (PT-CT) in patients with an exon 19 deletion or L858R mutation based on the FLAURA2 clinical trial (NCT04035486) results [6, 9].

While osimertinib shows efficacy in progression-free survival (PFS) and overall survival (OS) compared to first-generation EGFR TKIs, nearly all patients eventually experience disease progression on EGFR TKI treatment, including combination with chemotherapy [10]. No clear treatment pathway exists for patients with advanced/metastatic NSCLC who progress after treatment with 1L osimertinib [8]. Treatment options involve clinical trials or chemotherapy regimens [and recent advances with the addition of amivantamab to chemotherapy and amivantamab + lazertinib to chemotherapy based on the MARIPOSA-2 study (NCT04988295)] [11], but management depends on factors such as disease characteristics including site(s) of progression, genomic testing, known mechanisms of resistance, and access to treatment and/or clinical trials [8, 12]. Patients who progress on 1L osimertinib experience poor clinical outcomes, particularly regarding PFS and intracranial disease control, indicating that an unmet need remains for more efficacious and targeted treatments for this population, including actual access to subsequent lines of therapy [13].

Real-world (rw) insights into what happens to patients who receive 2L therapy after progression on osimertinib, including subsequent treatment patterns and clinical outcomes, can highlight and elucidate clinical gaps. The purpose of this study was to identify and evaluate patients in France with advanced or metastatic NSCLC who received a subsequent treatment (2L) after progressing on osimertinib in 1L. The study assessed patient characteristics at initiation of 2L treatment, rw treatment patterns, rwOS, rwPFS, and rw time to discontinuation (rwTTD).

Methods

Study Design, Data Source, and Patient Sample

A retrospective, observational study using the Epidemio-Strategy and Medical Economics (ESME) lung cancer (LC) database (NCT03848052) was conducted. The ESME-LC is a multicenter rw registry database that centralizes existing data from patient medical files using a retrospective data collection process via a pre-specified electronic data collection tool. Launched in 2017, the database involves 38 contributing hospitals across France, including Comprehensive Cancer Centers (FCCCs), universities, and general hospitals. Patients eligible for enrollment are ≥ 18 years old and initiated treatment for lung cancer at participating sites from 2015 onwards. In accordance with the authorizations issued by the French authorities for the database, no formal informed consent is required. All patients were informed about the ESME program and have approved the use of their data.

The period of patient identification for this study was from January 1, 2015, to December 31, 2022. Data from individual patient records were last updated throughout 2023, with database extraction in September 2023 (potential follow-up period). The database contained 51,067 patients with LC at the time this analysis was conducted.

Male and female patients > 18 years old who had histologically confirmed advanced (IIIC) or metastatic (IV) NSCLC, treated with osimertinib as 1L therapy, and with documented 2L treatment after osimertinib 1L usage were eligible for inclusion in the study analysis. Patients who died before 2L therapy or opted not to pursue further treatment were not included in the study. The 2L treatment was defined as the index treatment, with the treatment start date as the index date. The main variables available in the ESME-LC database include gender, age, smoking status, Eastern Cooperative Oncology Group-Performance Status (ECOG-PS), NSCLC histological subtype, biomarker testing, metastatic sites, clinical events (regional progression, new metastatic sites, progression of existing metastatic sites), dates of treatments, and vital status. Antineoplastic treatments administered after the advanced or metastatic NSCLC diagnosis date were classified into lines of therapy (LOT) based on treatment records and using data on progression events. A new LOT was defined under two conditions: (1) a progression event and (2) discontinuation of or change in therapy. Progression events included diagnosis of regional tumor or lymph node progression, diagnosis of a new metastasis, progression of a previously diagnosed metastatic site, or end of a therapy with termination due to progression as indicated in the clinical notes on the patient’s record. Progression could have been based on clinical examination, imaging, or histology and was not assessed through a centralized review process. To allow for at least 6 months of potential follow-up, patients were required to have their index date occurring at least 6 months prior to the September 2023 data extraction, irrespective of subsequent outcomes (e.g., of their date of death or loss to follow-up). Presence of a positive EGFR mutation test was not required for inclusion as, per clinical input, in France receipt of osimertinib is a strong proxy for this mutation status as per guidelines for the use of osimertinib. It is possible that co-mutational types or off-label use could be included in our cohort, but the benefit to including patients who received osimertinib with the assumption that an EGFRm was present outweighed the risk of a few patients being included who may not fit this criteria. See Fig. 1 for all inclusion and exclusion criteria.

Fig. 1.

Patient eligibility criteria and attrition. *Index date is ≥ 6 months prior to data extraction date. 1L first-line therapy, 2L second-line therapy, ESME LC Epidemio-Strategy and Medical Economics lung cancer

Outcomes of Interest

Patient and clinical characteristics assessed at index date included gender, age, ECOG-PS (− 30 to + 15 days of index date), smoking status, NSCLC histological subtype, year of index treatment, number and location of metastatic sites, and length of follow-up (defined as time from index date to date of last medical information). Resistance testing at any point after progression on 1L was also described where available and captured in the case report form. Specific tests included EGFR mutation tests for C797S, T790m, exon 20 insertion, and mesenchymal epithelial transition (MET) amplification testing. Treatment patterns were described via the following categorization: platinum-based chemotherapy regimens, EGFR TKI regimens (which could include immunotherapy drugs and/or chemotherapy drugs and/or VEGF drugs), other immunotherapy combinations, non-platinum-based chemotherapy, and ‘other’ regimens. Regimens beyond the 2L setting were described when available.

Clinical outcomes including rwOS, rwPFS, and rwTTD were also analyzed. rwOS was defined as the time from the index date to the date of death from any cause. Patients who were alive at the cut-off date were censored at the date of last medical information. rwPFS was defined as the time from index date to the date of first progression of the disease or date of death for any cause, whichever occurred first. Patients alive without progression reported after the index date or with death reported > 9 months after the date of the last medical information were censored at the date of last medical information. Patients were considered lost to follow-up after this time as > 9 months was no longer considered to be attributable to disease progression. rwTTD was defined as the time from the index date to the discontinuation (stop) date of the last treatment in 2L or the date of death, whichever occurred first. As was the case for rwPFS, patients who were alive without treatment discontinuation or had no treatment discontinuation recorded and death reported > 9 months from the date of last medical information were censored. Subgroup analyses were conducted to assess rwOS and rwPFS stratified by ECOG-PS and presence of metastases (bone, liver, and brain) at index date and among patients treated with standard of care platinum-based chemotherapy.

Descriptive statistics were used for patient demographics and clinical characteristics; rwOS, rwPFS, and rwTTD, and the duration of observed follow-up were estimated using Kaplan-Meier methods. No imputation was performed for missing data. Statistical tests were not performed to assess differences between groups, and no statistical techniques such as matching or weighting were utilized to account for differences in demographic or clinical differences. All analyses were conducted using SAS 9.4 (Cary, NC) and R Studio 4.3.1.

Ethical Approvals

In accordance with the authorizations issued by the French authorities for the database, no formal informed consent is required. All patients were informed about the ESME program and have approved the use of their data. The authors confirm they have permission to access and use the ESME database for this study. The authors also confirm that no form of IRB approval was required for this study.

Results

Overall, 284 patients met additional eligibility criteria (Fig. 1). Of these, 71.1% of patients were female, with a median age of 67.0 years at index date. ECOG-PS was recorded in 56.7% of patients at index date, with 70.2% with an ECOG-PS status of 0–1 in those with a score available. Approximately half of patients were non-smokers, and most (63%) patients initiated index treatment in 2021 or 2022. The median number of sites of metastases upon initiation of index treatment was 3. The most common site was bone (65.1%) followed by brain (56.6%), with 34% of patients with brain metastases being symptomatic (Table 1).

Table 1.

Patient characteristics

	N = 284
Gender
Male	82 (28.9%)
Female	202 (71.1%)
Age at index date (years)
Median age	67.0
< 65	120 (42.3%)
65–74	96 (33.8%)
≥ 75	68 (23.9%)
Smoking status at enrollment in center
Non-smoker	140 (49.3%)
Former smoker	103 (36.3%)
Smoker	23 (8.1%)
Not available	18 (6.3%)
NSCLC histological subtype
Adenocarcinoma	269 (94.7%)
Other	15 (5.3%)
ECOG-PS at index date
0–1	113 (39.8%)
≥ 2	48 (16.9%)
Not available	123 (43.3%)
Year of index date
2017	5 (1.8%)
2018	10 (3.5%)
2019	24 (8.5%)
2020	53 (18.7%)
2021	86 (30.3%)
2022	93 (32.8%)
2023	13 (4.6%)
Documented EGFR exon19del or L858R mutation at any time during course of disease	218 (76.8%)
Metastatic sites at index datea
Median number of sites	3.0
No metastatic sites documented	3 (1.1%)
Metastatic site location at index date (not mutually exclusive, n = 281)
Contralateral lung	108 (38.4%)
Bone	183 (65.1%)
Brain—all	159 (56.6%)
Brain—symptoms not documented	105 (66.0%)
Brain—symptomatic	54 (34.0%)
Brain radiotherapy received prior to start date	51 (32.1%)
Symptoms not documented	24 (15.1%)
Symptomatic	27 (17.0%)
Liver	94 (33.5%)
Observed follow-up time from index date (months)
Median follow-up duration (reverse KM)	19.1 (95% CI 16.8–23.9)

ECOG-PS Eastern Cooperative Oncology Group Performance Status

^aPatients can have multiple metastatic sites (different organs involved) recorded

While all patients in the study received osimertinib as 1L therapy per eligibility criteria, most patients (88.4%) received osimertinib monotherapy as 1L treatment compared to 11.6% who received osimertinib as combination therapy. One patient received osimertinib plus a clinical trial drug (Fig. 2). In the 2L setting, treatments included platinum-based chemotherapies (50.0%), EGFR TKI- based therapy (29.9%; 1st–3rd-generation TKIs), immunotherapy (IO) alone or in combination (9.9%), and other regimens (10.2%). Specifically, 59 of 284 patients (20.8%) received osimertinib in 2L: 27 of these patients received osimertinib monotherapy, and 32 patients received osimertinib in combination with another treatment (Fig. 2, Table 2).

Fig. 2.

Treatment sequencing. 1L first-line therapy, combo combination therapy, CT chemotherapy, IO immunotherapy, mono monotherapy, PT-CT platinum-based chemotherapy (including cisplatin, carboplatin, oxaliplatin, derived-platinum), TKI tyrosine-kinase inhibitor

Table 2.

Treatments received in 2L setting after progressing on 1L osimertinib

Treatment category and name	N = 284
Platinum-based chemotherapy (PT-CT)a	142 (50.0%)
PT-CT + pemetrexed	65
PT-CT + paclitaxel	19
PT-CT + bevacizumab + pemetrexed	46
Other PT-CT regimens	12
EGFR TKIb	85 (29.9%)
EGFR TKI 1st/2nd-generation mono	20
EGFR TKI 1st/2nd-generation combo	3
EGFR TKI 3rd-generation mono	24
EGFR TKI 3rd-generation combo	38
Immunotherapy comboc	26 (9.2%)
Immunotherapy alone	2 (0.7%)
Non-PT-CTd	18 (6.3%)
Other regimense	11 (3.9%)
Experimental study drugf	0 (0.0%)

^aPT-CT included regimens with cisplatin, carboplatin, oxaliplatin, or derived-platinum. Other PT-CT regimens included PT-CT ± bevacizumab ± other agents such as paclitaxel, gemcitabine, or etoposide

^bThe 1st- or 2nd-generation EGFR TKI drugs are afatinib, dacomitinib, erlotinib, gefitinib, nazaritinib, and sapitinib. The 3rd-generation EGFR TKI drugs are lazertinib, mobocertinib, naquotinib, osimertinib, poziotinib, rociletinib, and sunvozertinib. Mono regimens contain EGFRTKI drugs only. Combo regimens contain EGFR TKI drugs and (IO drugs and/or CT drugs and/or VEGF drugs)

^cThe IO drugs are PD-1/PD-L1 inhibitors: atezolizumab, avelumab, budigalimab, cemiplimab, dostarlimab, durvalumab, ezabenlimab, inupadenant, lodapolimab, monalizumab, nivolumab, pembrolizumab, spartalizumab, and tislelizumab. Immunotherapy combo does not contain any experimental study drugs or EGFR TKI drugs. This line contains IO drugs and CT or VEGF drugs

^dAll chemotherapy regimens not including cisplatin or carboplatin (e.g., pemetrexed alone, paclitaxel alone, gemcitabine alone, etc.)

^eAny regimen not included in the other classes

^fAny drug received in a blinded clinical trial

1L first-line therapy, combo combination therapy, CT chemotherapy, IO immunotherapy, mono monotherapy, PT-CT platinum-based chemotherapy (including cisplatin, carboplatin, oxaliplatin, derived-platinum), TKI tyrosine-kinase inhibitor

After progressing on 1L osimertinib, 147 patients (51.8%) underwent additional testing for EGFR mutations at any point, as recorded in the database. MET amplification testing was documented in 49 patients (17.3%). Additional information on testing was not available.

Median rwOS from index date was 10.1 (95% CI 9.2–12.1) months (Fig. 3). The median rwPFS was 4.1 (95% CI 3.3–4.8, Fig. 4) months, and 211 patients (74.3%) had documented progression during the index treatment. rwTTD was 4.2 (95% CI 3.6–4.8) months. Almost half of patients [n = 140 (49.3%)] started a subsequent LOT (3L) during the follow-up period.

Fig. 3.

Real-world overall survival from index date. CI confidence interval, rwOS real-world overall survival

Fig. 4.

Real-world progression-free survival from index date. CI confidence interval, rwPFS real-world progression-free survival

rwOS by subgroups of interest is presented in Fig. 5 and Table 3. Patients with an ECOG-PS score of 0–1 had a longer median rwOS at 12.3 (95% CI 10.2–14.9) months vs 7.5 (95% CI 5.6–9.7) months in patients with a score of ≥ 2. Median rwOS varied by metastatic site. Patients with bone metastases had a median rwOS of 9.4 (95% CI 8.2–10.7) months compared to 15.1 (95% CI 10.6–18.9) months in those without. Among patients with brain metastases, median rwOS was 8.7 (95% CI 7.1–11.9) months for those who were asymptomatic and 10.7 (95% CI 9.4–12.9) months for patients without symptoms documented, compared to 11.8 (95% CI 9.1–17.4) months in patients without brain metastases. Patients with liver metastases had a median rwOS of 8.2 (95% CI 6.4–9.6) months vs 12.3 (95% CI 10.2–14.3) months in those without. Patients with ≥ 4 metastatic sites had the shortest median rwOS at 7.3 (95% CI 6.2–9.6) months, while patients with one metastatic site had a median rwOS of 18.9 (95% CI 15.1–25.5) months. Patients who received PT-CT as their index treatment had a median rwOS of 10.1 (95% CI 8.5–12.3) months (data not shown).

Fig. 5.

rwOS in subgroups of interest. a By ECOG status, b by presence of bone metastatic sites, c by presence of brain metastases, d by presence of liver metastases, e by number of metastatic sites. CI confidence interval, ECOG-PS Eastern Cooperative Oncology Group Performance Status, NR not reached, rwOS real-world overall survival

Table 3.

Real-world overall survival and progression-free survival in subgroups of interest

	Number of patients	Median rwOS in months [95% CI]	Median rwPFS in months [95% CI]
ECOG status
0–1	113	12.3 (10.2–14.9)	4.7 (3.4–5.5)
2+	48	7.5 (5.6–9.7)	3.7 (1.5–5.6)
Presence of bone metastases
Yes	183	9.4 (8.2–10.7)	3.7 (2.9–4.8)
No	101	15.1 (10.6–18.9)	4.7 (3.5–5.9)
Presence of brain metastases
Yes—symptomatic	54	10.7 (9.4–12.9)	3.7 (1.5–4.4)
Yes—symptoms not documented	105	8.7 (7.1–11.9)	4.0 (2.5–5.2)
No	125	11.8 (9.1–17.4)	4.8 (3.5–5.7)
Presence of liver metastases
Yes	94	8.2 (6.4–9.6)	2.8 (1.8–4.8)
No	190	12.3 (10.2–14.3)	4.6 (3.7–5.3)
Number of metastatic sites
1	49	18.9 (15.1–25.5)	5.2 (3.5–7.6)
2–3	128	10.9 (9.4–13.8)	4.8 (4.1–5.6)
4+	104	(6.2–9.6)	2.8 (1.9–4.0)

CI confidence interval, ECOG-PS Eastern Cooperative Oncology Group Performance Status, rwPFS real-world progression-free survival, rwOS real-world overall survival

Median rwPFS was 4.7 (95% CI 3.4–5.5) months in patients with an ECOG of 0–1 and 3.7 (95% CI 1.5–5.6) months in patients with an ECOG of ≥ 2. Patients with bone metastases and brain metastases had similar median rwPFS estimates [3.7 (95% CI 2.9–4.8) months for bone and 3.7 (95% CI 1.5–4.4) months for symptomatic brain metastases and 4.0 (95% CI 2.5–5.2) months for patients with metastatic symptoms not documented], whereas patients with liver metastases had the shortest median rwPFS at 2.8 (95% CI 1.8–4.8) months vs 4.6 (95% CI 3.7–5.3) months for patients without liver metastases. Like in rwOS, patients with ≥ 4 metastatic sites had a shorter median rwPFS [2.8 (95% CI 1.9–4.0) months] than patients with 2–3 metastases [median rwPFS 4.8 (95% CI 4.1–5.6) months] or 1 metastatic site (median rwPFS 5.2 (95% CI 3.5–7.6) months] (Fig. 6, Table 3). Among patients who received PT-CT as their index treatment, the median rwPFS was 3.7 (95% CI 2.8–4.8) months (data not shown).

Fig. 6.

rwPFS in subgroups of interest. a By ECOG status; b by presence of bone metastatic sites, c by presence of brain metastases, d by presence of liver metastases, e by number of metastatic sites. CI confidence interval, ECOG-PS Eastern Cooperative Oncology Group Performance Status, NR not reached, rwPFS real-world progression-free survival

Discussion

This retrospective real-world analysis using the ESME-LC database provides valuable insights into patients with EGFRm advanced and metastatic NSCLC in France from 2015–2023. The database provides comprehensive data from mostly specialized cancer care centers, thus allowing for an in-depth view into treatment patterns and outcomes. Results of this analysis underscore the poor prognosis post-1L osimertinib therapy with a median rwOS of 10.1 (95% CI 9.2–12.1) months and a median rwPFS of 4.1 (95% CI 3.3–4.8) months.

The analysis spans the time period from 2015–2023, incorporating several years prior to the approval of osimertinib in 1L treatment. From 2015–2018, 5.3% of patients received osimertinib, which could be indicative of early access and compassionate use. EMA approval was in 2018, and reimbursement in France occurred in 2020 for 1L use. In this analysis, 86.3% of patients received osimertinib in 2020 or later.

When evaluating by patient characteristics, patients with an ECOG-PS of 0–1 had longer rwOS than patients with an ECOG-PS of ≥ 2; although this was not formally compared statistically, it is consistent with trends linking poorer functional status to higher disease burden, comorbidities, and potentially reduced treatment tolerance. However, rwPFS did not appear to differ greatly between EGOG-PS groups, which may suggest the impact of access and effectiveness of post-osimertinib treatment. The ECOG ≥ 2 group represents a clinically important but underrepresented population in trials, where eligibility often favors ECOG-PS 0–1. This study highlights the frequency of such patients in real life and the need for treatment strategies to optimize management in patients with poor functional status.

When considering metastatic burden, patients with ≥ 4 metastatic sites had poorer outcomes in terms of rwOS and rwPFS than those with 2–3 or a single metastatic site, although no statistical testing was performed. Notably, median rwOS and rwPFS with extensive metastases were less than half those of patients with only one metastatic site, underscoring the strong link between higher tumor burden and worse prognosis. Among metastatic sites, liver metastases had a short median rwOS at 8.2 months, reinforcing the aggressive nature of liver-involved disease. Notably, brain metastases were present in more than half (56.6%) of the cohort at index, highlighting the high CNS disease burden in this population beyond a clinical trial setting. Within this group, patients with brain metastases had worse rwOS than those without brain metastases, although this was not statistically tested. However, rwPFS was low in patients with symptomatic brain metastases and patients with brain metastases but symptoms not documented (3.7 and 4.0 months, respectively), suggesting that intracranial disease management remains a significant challenge for current treatments regardless of symptom status. Patients with high metastatic burden require more aggressive and earlier combination therapy while accounting for previous therapy and resistance mechanisms [8].

Half (50%) of patients were treated with PT-CT post-osimertinib progression, which aligns to treatment guidelines at the time the study was conducted. Of the patients who received a TKI, 62 of 85 received a third-generation EGFR TKI as either monotherapy or combination therapy. Patients could have continued EGFR TKIs beyond progression if they exhibited slow or indolent disease progression where continuing EGFR blockade could still provide disease control, although this has been challenged by clinical trials [14]. Some patients may have experienced oligoprogression and as such remained on osimertinib therapy but also received localized radiation or surgery to treat progressive lesions. It is also possible that patients may have received osimertinib beyond progression because of maintained clinical benefit or limited treatment options at the time the cohort was analyzed. Ultimately, recent results from the COMPEL trial indicate the benefit of osimertinib continuation when initiating platinum-based chemotherapy at the time of acquired resistance to osimertinib [15].

Findings from our analysis are broadly aligned with results of clinical trials, recognizing patient populations may be somewhat different as our analysis included patients with ECOG-PS scores ≥ 2. The MARIPOSA-2 study evaluated amivantamab + carboplatin-pemetrexed ± lazertinib in patients with advanced/metastatic NSCLC after progression on 1L or 2L osimertinib. In the study, PFS was significantly longer for amivantamab chemotherapy and amivantamab-lazertinib chemotherapy versus chemotherapy. The median PFS in the chemotherapy arm for all patients was 4.2 months and was 4.1 (95% CI 3.3–4.3) months in patients with prior 1L osimertinib [11]. This finding is similar to the rwPFS observed in our analysis of 4.1 months for all patients and 3.7 months for patients who received PT-CT as 2L therapy. In contrast, rwOS was shorter in our analysis at 10.1 (95% CI 9.2–12.1) months compared to 15.3 (95% CI 13.7–16.8) months in the chemotherapy arm of the MARIPOSA-2 trial, a likely reflection of the stringent selection criteria in effect selecting for healthier patients in a trial setting [16].

A real-world analysis of patients in the US in patients with EGFRm NSCLC sought to characterize the subsequent LOT following osimertinib therapy. Using the IQVIA PharMetrics Plus database, data from November 2015 to September 2019 were analyzed. After osimertinib was discontinued, most patients were treated with a PT-CT regimen (57%), of which 40.3% included immunotherapy. Reuse of EGFR TKIs occurred in 24% of patients. TTD was noted to be short at 2.4 months [17]. These findings are in line with our analysis, in which 50% of patients received PT-CT after osimertinib 1L therapy, and 29.9% received EGFR TKI therapy. In a 2015–2021 US-based study of patients receiving therapy after osimertinib and PB-CT, a wide range of subsequent treatments was noted, including non-PT-CT chemotherapy (31.1%), immunotherapy alone or in combination (32.6%), EGFR TKI alone or in combination (27.5%), and PT-CT (11.7%), underscoring the lack of a clear therapeutic path following progression on osimertinib.[18]. More recently, new options based on bispecifics, such as ivonescimab [19], and antibody-drug conjugates, such as datopotamab deruxtecan [20], are being investigated in this setting.

PT-CT outcomes are often suboptimal for patients progressing on osimertinib, as demonstrated by this and other rw studies, which have consistently shown limited rwPFS and rwOS benefits. While PT-CT remains the standard of care in the 2L setting, median rwPFS is consistently shown to be between 3 to 4 months. This demonstrates a need for more effective treatment options for patients with widespread disease progression or those who have developed resistance mechanisms not amenable to further EGFR TKI therapy after 1L progression, who would otherwise be advised to transition to PT-CT.

There are several limitations to this study that should be noted. The study uses data up to the ESME-LC database cutoff date of September 2023, which may not capture changes in treatment practices after this period (specifically changes in treatment patterns that may be attributable to the MARIPOSA-2 and FLAURA2 study findings). Given gaps in data availability on biomarker testing, our ability to describe the uptake of re-biopsy after progression on 1L osimertinib and resulting treatment patterns is limited, and related findings should be interpreted with caution. The rw nature of the data and heterogeneity of the rw cohort (e.g., ECOG-PS status) should be considered when comparing results against trials. Importantly, due to the varying nature of assessments and clinicians' subjective judgment in determining rw progression, rwPFS is a fundamentally different outcome than trial-based PFS and may vary across rw settings as well. This should be interpreted in the context of the remaining outcomes that are less subject to misclassification. Notably, ECOG-PS status was missing for 43.3% of patients. ECOG-PS can significantly influence the interpretation of outcomes and treatment efficacy, resulting in confounding issues in comparative studies (as ECOG status is a strong prognostic factor), limitations in generalizability if the study population is not reflective of the broader NSCLC EGFRm population, and understanding treatment decisions as ECOG-PS is often considered in treatment decisions. Comparisons between subgroups was not conducted because of potential confounding that could not be accounted for, thus making comparisons between groups inappropriate and not informative.

Despite these limitations, this analysis provides valuable insights into EGFRm advanced and metastatic NSCLC patients treated with a subsequent line after progression on 1L osimertinib.

Conclusions

Limited options were available after treatment with osimertinib in patients with advanced/metastatic NSCLC with an EGFR mutation who progressed, with most treated with platinum-based chemotherapy. Outcomes in this patient population remain poor, and there is a high unmet need, particularly for patients who have brain and liver metastases. A need for better treatment options remains, although the treatment landscape is evolving, both in the post-osimertinib setting and in the novel regimens being investigated in the 1L setting, which may have an impact on subsequent LOTs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Nicolas Girard, Danalyn Byng, Jie Meng, Friso Coerts, Alessandria Struebing, M. Inaam Haq, and Olga Tymejczyk contributed to the study concept and design, interpretation, manuscript drafting, and manuscript review. Pauline Macouillard and Lise Bosquet contributed to the study concept and design, analysis, interpretation, and manuscript review. Maurice Pérol, Shannon Hunter, Yan Xiong, Christos Chouaid, Clarisse Audigier-Valette, Didier Debieuvre, and Xavier Quantin contributed to the study concept and design, interpretation, and manuscript review. Sarah Park contributed to interpretation, manuscript drafting, and manuscript review.

Funding

Open access funding provided by Université de Versailles Saint-Quentin-en-Yvelines. Funding for this study was provided by Daiichi Sankyo, GmbH. The journal’s Rapid Service Fee was funded by Daiichi Sankyo, GmbH.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available due to ethical and legal requirements related to the ESME data warehouse which does not allow sharing individual patient level data. However, requests for access may be considered on a case-by-case basis and may be made available by UNICANCER upon reasonable request.

Declarations

Conflict of Interest

Authors Pauline Macouillard, Sarah Park, Didier Debievre, Xavier Quantin, and Lise Bosquet have no conflicts to disclose. Authors Danalyn Byng, Jie Meng, Friso Coerts, Shannon Hunter, Yan Xiong, M. Inaam Haq, Olga Tymejczyk are employees of Daiichi Sankyo. Jie Meng, Alessandria Struebing, and Yan Xiong also report holding stock for Daiichi Sankyo. Christos Chouaid has received research grants, travel, and honoraria from AZ, BI, GSK, Roche, Sanofi Aventis, BMS, MSD, Lilly, Novartis, Pfizer, Takeda, Bayer, Janssen, Viatris, Chugai, and Amgen. Clarisse Audigier-Valette has served in a consulting or advisory role for Roche, Janssen Oncology, Lilly, BMS GmbH and Co. KG, MSD Oncology, AstraZeneca, Pfizer, Sanofi, Takeda, and Abbvie. Maurice Pérol has been a consultant for Roche, Eli Lilly, Pfizer, Boehringer Ingelheim, Merck Sharp & Dohme, Bristol-Myers Squibb, Novartis, AstraZeneca, Takeda, Gritstone, Sanofi, GlaxoSmithKline, Amgen, Abbvie, Janssen, Ipsen, Pierre Fabre, Esai, AnHeart Therapeutics, Nuvation Bio, Daiichi-Sankyo, Revolution Medicine, Regeneron, and Novocure; received grants from Roche, AstraZeneca, Takeda, and Boehringer Ingelheim; received lecture fees from Eli Lilly, Roche, AstraZeneca, Pfizer, Amgen, Boehringer Ingelheim, Bristol-Myers Squibb, Takeda, Merck Sharp & Dohme, Chugai, and Illumina; and travel support from Roche, Pfizer, Merck Sharp & Dohme, Bristol-Myers Squibb, AstraZeneca, Takeda, Amgen, and Janssen. Nicolas Girard has received research grants/support from Abbvie, Amgen, AstraZeneca, BeOne, Boehringer Ingelheim, Bayer, BioNTEch, Bristol Myers Squibb, Daiichi-Sankyo, Gilead, J&J, Lilly, Sivan, and Tahio; consulted for Abbvie, Accord, Amgen, AstraZeneca, BeOne, Bristol Myers Squibb, Daiichi-Sankyo, Ipsen, J&J, Lilly, Merck Sharp & Dohme, Natera, Pfizer, Pharmamar, Pierre Fabre, Regeneron, and Summit; served as speaker for Abbvie, Amgen, AstraZeneca, BeOne, Bristol Myers Squibb, Daiichi-Sankyo, J&J, Merck Sharp & Dohme; hospitality from Bristol Myers Squibb, J&J, Pierre-Fabre; and reports employment of a family member with AstraZeneca.

Ethical Approval

In accordance with the authorizations issued by the French authorities for the database, no formal informed consent is required. All patients were informed about the ESME program and have approved the use of their data. The authors confirm they have permission to access and use the ESME database for this study. The authors also confirm that no form of IRB approval was required for this study.