FormalPara Key Summary Points

Why carry out this study?

The FLAURA trial demonstrated improved efficacy with osimertinib versus earlier-generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs), and osimertinib has become the recommended first-line treatment for patients with EGFR-mutated metastatic non-small cell lung cancer (mNSCLC), although there is no clear standard of care in subsequent lines of therapy once resistance develops.

This retrospective study used medical and pharmacy claims from IQVIA PharMetrics® Plus to characterize the subsequent line of therapy following discontinuation of osimertinib and real-world treatment outcomes among patients with EGFR-mutated mNSCLC.

What was learned from this study?

The median time to treatment discontinuation in the line of treatment after osimertinib discontinuation was short; regimens containing platinum-based chemotherapy (with or without immunotherapy) were commonly used (57%), with high rates of adverse events associated with hospitalization.

The most frequent adverse events associated with hospitalization were anemia, pneumonia/pneumonitis, infection/sepsis, nausea/vomiting, and fever.

These data underscore a high unmet need for new therapeutic options to address EGFR TKI resistance in patients with EGFR-mutated mNSCLC.

Introduction

Lung cancer is the leading cause of cancer-related mortality worldwide [1]. It is estimated that in the United States, 238,340 new cases and 127,070 associated deaths from lung cancer occurred in 2023 [2]. Most lung cancers (56%) are diagnosed in an advanced stage, after the disease has metastasized [2]. Accounting for 84% of all lung cancer diagnoses, non-small cell lung cancer (NSCLC) is the most common type [3]. In the United States, adenocarcinoma is the most common subtype (40%) of NSCLC, followed by squamous cell carcinoma (25–30%), and large cell carcinoma (5–10%) [4,5,6]. Epidermal growth factor receptor (EGFR) is a mutated proto-oncogene, where a sensitizing mutation can result in the activation of tyrosine kinase, thereby leading to proliferative tumor growth in patients with NSCLC [7]. EGFR sensitizing mutations affect 17% and 46% of white and Asian patients, respectively, with NSCLC [8, 9].

The National Comprehensive Cancer Network® (NCCN®) treatment guidelines for metastatic NSCLC (mNSCLC) recommend EGFR tyrosine kinase inhibitors (TKIs) for first-line treatment options in patients with common EGFR mutations (exon 19 deletions and exon 21 L858R mutations), with third-generation osimertinib the preferred choice [10]. Osimertinib was initially approved in November 2015 (USA) for the treatment of patients with T790M resistance mutations following progression on earlier-generation EGFR TKIs [11]. In April of 2018, the indication was expanded to first line based on the FLAURA trial, which demonstrated improved outcomes with osimertinib versus first-generation EGFR TKIs [12].

While osimertinib has shown considerable efficacy and tolerability, drug resistance eventually develops in most patients [13, 14]. Upon resistance, the NCCN® recommends a platinum-based chemotherapy regimen as the next line of treatment (LOT) in most circumstances. Immuno-oncology (IO) therapy  ± chemotherapy may also be considered [10]. In some circumstances, the continuation of an EGFR TKI beyond progression may be reasonable to avoid disease flare, if brain metastases are present, or if used in combination therapy (i.e., afatinib + cetuximab).

While there is considerable clinical trial data on the use of EGFR TKIs and their optimal sequencing, trial eligibility criteria often select for healthier patients who may not be representative of the full disease population. In contrast, real-world studies provide information on all patient types across multiple settings which may be of additional value to healthcare providers when caring for their patients. These data also allow for a comparison of real-world treatment decisions against those recommended in current practice guidelines.

This study aimed to characterize the subsequent LOT following osimertinib discontinuation and real-world treatment outcomes among patients with EGFR-mutated mNSCLC. To our knowledge there is only one other published study that described the full spectrum of next-line treatments after osimertinib; however, the sample was limited to 54 patients [15]. Further, we are aware of no other studies that quantify adverse event-related hospitalization rates among this population.

Methods

Data Source

This study was a retrospective analysis of adjudicated medical and pharmacy claims from IQVIA PharMetrics® Plus. Administrative claims are generated for the billing of healthcare services and generally include attributes related to care such as diagnosis and procedure codes, pharmacy information, locations and sources of care, admission and discharge dates, patient demographic information, and cost of care. The database used for this study includes more than 150 million unique enrollees across the United States, with more than 95 million having both medical and pharmacy benefits (40 million active lives in recent years) and a diverse representation of geographic zones, employers, payers, providers, and therapy areas. This study did not involve human subjects. All analyses were conducted using de-identified data from insurance claims databases and are exempt from IRB review and approval.

Patient Selection

Included in the analysis were adult patients (≥ 18 years of age) with ≥ 1 medical claim indicating a diagnosis of lung cancer between January 2010 and September 2019 (index period), and ≥ 1 medical claim for a secondary malignancy after the lung cancer diagnosis date (metastatic lung cancer [mLC] diagnosis date). Patients were also required to have ≥ 1 pharmacy claim for an EGFR TKI after the mLC diagnosis date (proxy for EGFR-mutated NSCLC). Patients were further screened to select only those who received osimertinib monotherapy after the mLC diagnosis date and who initiated a subsequent LOT (index date) between November 2015 and September 2019 after osimertinib discontinuation. Continuous medical and pharmacy enrollment was required 6 months before the mLC diagnosis date (baseline period) and ≥ 1 month after the initiation of the subsequent LOT. Patients were excluded if they had prior medical claims for other primary cancer 6 months prior to their mLC diagnosis. Patient study selection is presented in Table 1.

Table 1 Patient selection
Full size table

Study Design

Antineoplastic use after the mNSCLC diagnosis date was captured using medical and pharmacy claims and included EGFR TKIs, vascular endothelial growth factor inhibitors, IO, chemotherapy, and other targeted therapy (e.g., alectinib, crizotinib that may have been prescribed off-label). To determine LOTs, an algorithm was developed based on previously published literature. A combination regimen was defined as any medication filled or administered within 14 days of a previous drug fill or administration beginning at the initiation of each new LOT. LOT discontinuation occurred when all medications in a regimen were discontinued for ≥ 60 days or when a new agent was added (new LOT). For infused agents, the treatment gap was measured from date of last administration plus 1 day until the subsequent administration date or end of data period. For oral agents, the treatment gap was measured from last day of supply plus 1 day until a subsequent fill date or end of data period. For oral agents with missing days’ supply, a 30-day prescription was assumed. The treatment discontinuation date was defined as the last infused treatment plus 1 day or the last of oral agent supply plus 1 day. LOTs did not advance if an agent was dropped from an existing combination regimen to account for maintenance therapy or tolerability issues; however, the addition of a new agent (beyond the 14-day combination period) did advance the LOT.

After osimertinib discontinuation, index LOT regimens (i.e., the subsequent LOT after osimertinib discontinuation) were categorized into five mutually exclusive groups: (1) EGFR TKI monotherapy, (2) EGFR TKI combination (all other EGFR TKI-containing regimens), (3) platinum-based chemotherapy regimen (any platinum-based regimen with or without IO therapy), (4) IO therapy alone, and (5) other. Patients were followed from the index date to the end of continuous insurance eligibility or end of data availability (September 2019) (Fig. 1).

Fig. 1
figure 1

Study design. NSCLC histology for patients with an ICD-9-CM or ICD-10-CM diagnosis for LC is inferred from the use of an EGFR TKI, while date of metastatic disease is based on ICD-9-CM or ICD-10-CM secondary malignancy diagnosis codes. EGFR epidermal growth factor receptor, ICD-9-CM International Classification of Diseases, 9th Revision, Clinical Modification, ICD-10-CM International Classification of Diseases, 10th Revision, Clinical Modification, LC lung cancer, LOT line of therapy, mLC metastatic lung cancer, NSCLC non-small cell lung cancer, TKI tyrosine kinase inhibitor

Full size image

Outcomes and Statistical Analysis

Patient demographics and clinical characteristics were captured at baseline. Descriptive statistics were presented as mean and standard deviation (SD), median, and interquartile range (IQR) for continuous variables and frequencies and percentages for categorical variables.

The observed mean (SD) duration of the index treatment was calculated as the time elapsed between treatment initiation and the day of discontinuation. In addition, the median time to discontinuation (TTD) of the index LOT was estimated using Kaplan–Meier analysis. Median TTD was defined as the time from index LOT initiation until discontinuation (event) or end of study follow-up period (censoring for patients without observed discontinuation).

All-cause healthcare resource utilization  (HCRU) and adverse events (AEs) associated with hospitalization were assessed for the observed duration of the index LOT plus 60 days, or prior to the initiation of the next LOT if < 60 days, to account for any treatment-related AEs that may have resulted in index treatment discontinuation. HCRU was reported per 100 patient-months of treatment exposure and included inpatient (IP) days, emergency department (ED) days, and outpatient (OP) days with and without antineoplastic infusion. Proportions of patients with at least one inpatient admission and at least one ED visit were also described. AEs commonly associated with NSCLC treatments were reported per 100 patient-months of treatment exposure (duration of index LOT plus 60 days) and identified based on diagnosis codes observed at any position in the medical claim during a hospital admission [16,17,18]. Any identified AE was counted only one time per patient to account for prolonged AE episodes that may occur over time (See Supplementary Materials Table S1 for the full list of AEs of interest). All analyses were conducted using SAS® Enterprise Guide® version 7.15.

Results

A total of 135 patients with mNSCLC met the criteria for study eligibility (Table 1). The median (IQR) age was 59 (52–64) years, and 64.4% of the patients were female (Table 2). The mean (SD) time between mLC diagnosis and initiation of index LOT was 22.5 (14.9) months, and the mean (SD) observed duration of follow-up from initiation of index LOT to end of follow-up was 8.2 (6.3) months. Most patients (76.3%) initiated the index LOT between January 2018 and September 2019. The mean (SD) number of LOTs prior to index treatment was 2.2 (1.1); 22.2% (n = 30) of patients had one prior LOT, 49.6% (n = 67) had two prior LOTs, 17.8% (n = 24) had three prior LOTs, and 10.4% (n = 14) had four or more prior LOTs (not shown).

Table 2 Baseline characteristics
Full size table

After discontinuation of osimertinib monotherapy, regimens containing platinum-based chemotherapy were most common and accounted for 57.0% (n = 77) of the index LOT, with 40.3% (n = 31) in combination with IO. EGFR TKI monotherapy was prescribed for 15.6% (n = 21) of patients, followed by IO alone (13.3% [n = 18]), EGFR TKI combination therapy (8.1% [n = 11]), and other therapy (5.9% [n = 8]; Fig. 2). Re-treatment with an EGFR TKI as monotherapy or combination therapy was observed in 23.7% (n = 32) of patients, of which osimertinib accounted for 12.6% (n = 17; n = 6 as monotherapy, n = 11 as combination therapy). The observed mean (SD) duration of the index LOT was 3.0 (4.1) months, with 2.7 (3.0) months for platinum-based chemotherapy ± IO, 3.8 (4.0) months for EGFR TKI monotherapy, 2.4 (5.4) months for IO alone, 5.1 (7.9) months for EGFR TKI combination, and 1.9 (2.2) months for other therapy. The most frequently used agents in the index LOT were carboplatin (56.3%), pemetrexed (50.4%), pembrolizumab (19.3%), osimertinib (12.6%), and nivolumab (9.6%; Fig. 3).

Fig. 2
figure 2

Distribution of index LOT regimens.a,b,c EGFR epidermal growth factor receptor, IO immuno-oncology, LOT line of therapy, TKI tyrosine kinase inhibitor, VEGFR vascular endothelial growth factor receptor. aOf the 77 patients receiving a platinum-based regimen, 31 (40.3%) included IO therapy. bAmong eight patients treated with “other” index LOT regimens, five (63%) patients were treated with non-platinum chemotherapy-based regimens (gemcitabine, pemetrexed, docetaxel + ramucirumab), two (25%) with non-EGFR targeted therapies only (dabrafenib + trametinib, crizotinib), and one (13%) with VEGFR monotherapy (bevacizumab). cObserved mean durations of index LOT were 3.0 months overall, 3.8 months for EGFR TKI monotherapy, 5.1 months for EGFR TKI combination, 2.7 months for platinum chemotherapy, 2.4 months for IO, and 1.9 months for the “other” treatment group

Full size image
Fig. 3
figure 3

Most frequently used agents at index

Full size image

The median (95% CI) TTD for index LOT was 2.4 (1.8–3.1) months (Fig. 4). Patients who received EGFR TKI monotherapy had the longest TTD (4.6 [1.9–5.1] months), and those prescribed IO alone had the shortest TTD (1.5 [0.5–2.8] months). The platinum-based chemotherapy ± IO and other treatment groups had similar TTD (2.1 [0.7–3.1] and 2.0 [0.03–not estimable (NE)] months, respectively). The median was not reached for EGFR TKI combination therapy.

Fig. 4
figure 4

Median time to LOT discontinuation. LOT line of therapy, mTTD median time to discontinuation

Full size image

Among all patients, the mean (SD) number of all-cause inpatient admissions per 100 patient-months was 15.9 (32.3), and 31.9% of patients had ≥ 1 hospitalization during the index LOT. Inpatient admission rates per 100 patient-months were high among those treated with a platinum-based chemotherapy ± IO (19.4 [37.9]) and IO alone (15.6 [31.4]), although the mean (SD) length of stay was considerably longer for patients on a chemotherapy ± IO regimen (221 [665] days per 100 patient-months) compared with an IO alone (148 [274] days per 100 patient-months). Inpatient admission rates for those treated with EGFR TKI monotherapy were 7.1 (13.9) per 100 patient-months, with a mean (SD) length of stay of 63 (178) days per 100 patient-months.

The mean (SD) number of all-cause ED visits per 100 patient-months was 20.0 (46.6); 35.6% of patients had ≥ 1 ED visit (Table 3). The rate of ED visits was more than 60% higher for the IO alone treatment group (29.1 [57.3]) than the platinum chemotherapy ± IO group (18.2 [38.9]); however, the proportions of patients with ≥ 1 ED visit were similar (38.9% and 39.0%, respectively). For patients who received EGFR TKI monotherapy, the rate of ED visits was lower (10.3 [21.2]), and 28.6% had ≥ 1 ED visit.

Table 3 All-cause HCRU per 100 patient-months
Full size table

The average number of AE-related hospitalizations during the index LOT was 20.7 per 100 patient-months (Fig. 5 and Supplementary Materials Table S1). On average, patients treated with platinum-based chemotherapy ± IO and IO alone experienced the highest rates of hospitalization-related AEs (25.3 and 24.6, respectively), whereas patients treated with EGFR TKI monotherapy experienced the lowest rate (3.2). Among all patients, the most common AEs associated with hospitalization during the index LOT were anemia (average rate: 3.6 per 100 patient-months), pneumonia/pneumonitis (3.0), infection/sepsis (2.1), nausea/vomiting (1.9), fever (1.7), and diarrhea (1.5).

Fig. 5
figure 5

Adverse events associated with hospitalization per 100 patient-months. AE adverse event, EGFR epidermal growth factor receptor, IO immuno-oncology, IP inpatient, TKI tyrosine kinase inhibitor. AEs occurring ≥ 0.5 times per 100 patient-months in all patients on average were reported. For each of the AEs presented, only the first occurrence of the AE was counted (i.e., based on first IP admission at any diagnosis position because diagnoses observed in distinct days may be part of the same AE episode)

Full size image

Anemia was the most frequent AE associated with hospitalization among patients who received a platinum-based chemotherapy ± IO regimen (4.4 per 100 patient-months), and its occurrence was similar to those who received IO alone (4.6) (Supplementary Materials Table S1). Pneumonia/pneumonitis was the most frequent AE among patients who received IO alone (6.1), a rate more than double compared with platinum-based chemotherapy ± IO recipients (2.7). Among patients who received EGFR TKI monotherapy, only nausea/vomiting (2.1) and hepatic toxicity (1.1) were associated with hospitalization. Although skin toxicity and diarrhea are common AEs of these agents, neither was associated with hospitalization.

Discussion

Treatment in the index LOT was fragmented, suggesting a lack of an established standard of care after osimertinib discontinuation. Although approximately half of the patients who discontinued osimertinib received a platinum-based chemotherapy ± IO in the subsequent LOT, nearly one-quarter of patients were re-treated with EGFR TKI monotherapy or combination therapy, and more than one-third were treated with IO either alone or in combination. The use of platinum-based chemotherapy in this population is in line with NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®); however, a considerable number of patients were treated with platinum in combination with IO even though the benefit may be unclear. Given recent results from the CHECKMATE-722 and KEYNOTE-789 trials which showed no efficacy benefit when adding IO to chemotherapy in EGFR TKI-resistant patients, it is possible that these agents will be prescribed less frequently in the future [19, 20].

NCCN Guidelines® do suggest continuing or restarting EGFR TKIs beyond progression in some circumstances (i.e., if a flare phenomenon should occur upon discontinuation of these agents or continuing/restarting treatment with osimertinib for patients with progressive central nervous system disease). In our study, reutilization of EGFR TKIs as monotherapy after osimertinib discontinuation occurred in 16% (n = 21) of patients. Of these, six patients continued osimertinib after a gap in treatment, and 15 patients were treated with afatinib, erlotinib, or gefitinib.

Rechallenge with EGFR TKIs has been noted in published reports [21,22,23,24,25,26]. One review highlighted that several mutations promoting resistance to osimertinib remain sensitive to earlier-generation EGFR inhibitors, suggesting that rechallenging patients with these agents could be an effective approach [21]. Furthermore, due to their improved tolerability when compared with IO or cytotoxic therapies, retreatment with EGFR TKIs might also be considered [16, 27]. A few published case studies and small cohort studies have also touched upon rechallenging with EGFR TKIs, showing some potential as effective treatment [21, 23, 24]. However, rechallenge may also reflect a lack of alternative treatments in later LOTs as well as the limited knowledge of the optimal EGFR TKI sequence among patients with metastatic EGFR-mNSCLC [28]. Research is currently ongoing to further understand mechanisms of resistance to osimertinib, optimal treatment sequencing based on genomic testing at the time of disease progression, and the potential for improvement in patient outcomes [29].

To the best of our knowledge, only one previous study has examined real-world treatment patterns after discontinuation of osimertinib [15]. In a retrospective analysis using Flatiron electronic health records, Winfree et al. [15] reported findings from a cohort of patients with EGFR-mutated mNSCLC treated with first-line osimertinib between December 2017 and July 2019. Similar to our study, the majority of patients with a second-line treatment received platinum-based chemotherapy (± IO) (32%), and there was considerable use of EGFR TKI monotherapy (20%) and IO monotherapy (19%). The use of EGFR TKI and IO monotherapies following osimertinib suggests that clinicians and patients may prefer to try other treatment options first and delay the use of platinum chemotherapy for later LOTs where possible. Future studies are needed to further explore reasons for clinical decision-making, as these data are not available in the claims.

The duration of treatment in the LOT following osimertinib was short, underscoring the limited clinical benefit of existing treatment options for patients with EGFR TKI resistance. For nearly half of the patients who received a platinum-based chemotherapy ± IO, the mean duration of treatment was only 2.7 months. In the AURA3 trial, patients with a T790M mutation and disease advancement after first-line EGFR TKI were randomized to either osimertinib or platinum chemotherapy + pemetrexed [30]. The mean duration of treatment was 4.8 months for the chemotherapy arm. This difference may be attributable to the timing of treatment and the composition of the therapeutic regimens. In the AURA3 trial, treatment duration was assessed in the second LOT. In our study, 28.2% of patients had received three or more LOTs prior to the index LOT, suggesting more advanced disease in this population that may have reduced time on treatment. Further, patients in the AURA3 trial were treated with platinum + pemetrexed only, whereas in our study 40.3% of patients received IO in combination with a platinum-based chemotherapy regimen. It is possible that toxicities from this combination contributed to shorter time on treatment for some patients [30].

In the AURA3 trial, the proportion of patients with grade 3 or higher AEs was lower with osimertinib (23%) than with platinum-based chemotherapy (47%) [30]. Similarly, the overall rate of AEs associated with hospitalization in our study was considerably lower for patients treated with EGFR TKI monotherapy versus other treatments. Only nausea/vomiting and hepatic toxicity were identified as being associated with an inpatient stay among patients treated with EGFR TKI monotherapy. Although skin disorders and diarrhea are quite common in some EGFR TKI-treated patients, they may not necessitate visits to a tertiary care facility. The overall rates of AEs associated with hospitalization were similarly high across platinum chemotherapy ± IO and IO alone treatment groups. However, platinum chemotherapy ± IO appeared to have a less tolerable AE profile, including nausea/vomiting, hepatic and renal toxicities, and neutropenia, which may have contributed to longer inpatient stays and more frequent hospital readmissions versus IO alone. As IO therapy is associated with an increased risk of immune-related AEs, pneumonia/pneumonitis was more frequent in the group of patients treated with IO alone than in those treated with platinum chemotherapy, which had fewer patients receiving IO.

Similarly, we found that patients treated with platinum-based chemotherapy ± IO and/or IO alone had higher rates of all-cause HCRU than patients who received EGFR TKI monotherapy in the LOT after osimertinib discontinuation [13]. Although the proportions of patients with ≥ 1 IP admission and ≥ 1 ED visit were similar across platinum chemotherapy ± IO and IO alone treatment groups, patients who received platinum chemotherapy ± IO had numerically higher mean rates of IP admissions, IP days, and ED visits, suggesting more frequent and lengthier HCRU on a per-patient basis compared with those treated with IO alone. This finding is in line with previous studies reporting that IO was associated with lower HCRU and AEs compared with chemotherapy [31, 32].

Although current treatment options after progression on EGFR TKI therapy offer limited clinical benefit, promising new treatments are under investigation in this setting. HERTHENA-Lung02 is a phase III study comparing patritumab deruxtecan (an antibody–drug conjugate) with a platinum-based chemotherapy in patients with EGFR-mutated locally advanced or mNSCLC after progression on a third-generation EGFR TKI [33]. Likewise, MARIPOSA-2 is a phase III trial assessing the safety and efficacy of the EGFR-MET bispecific antibody amivantamab + platinum-based chemotherapy and amivantamab + lazertinib + platinum-based chemotherapy versus platinum-based chemotherapy in patients with EGFR-mutated advanced NSCLC who progressed on or after osimertinib monotherapy [34]. These agents may offer improved treatment options for patients with EGFR TKI resistance.

This study has several strengths and limitations. Results from this study fill a gap in the literature by characterizing the subsequent LOT and real-world outcomes among patients with mNSCLC harboring an EGFR-sensitizing mutation who were previously treated with osimertinib monotherapy. The study was conducted using medical and pharmacy claims data from a patient sample representing commercial plans across diverse US geographic regions. For purposes of interpretation, it should be noted that this study examined the post-osimertinib LOT between November 2015 and September 2019; because osimertinib was initially approved to address first- and second-generation EGFR TKI resistance (November 2015) and then was subsequently approved for first-line treatment of EGFR mNSCLC (April 2018), patients included in the study may have used osimertinib as either first- or later-line treatment. Although claims data allow for a large and representative patient sample, detailed clinical information is not recorded. For example, because histology and biomarker information could not be identified via medical claims, patients treated with an EGFR TKI were assumed to have non-small cell histology and a common EGFR-sensitizing mutation. Further, reasons for treatment discontinuation were not available, and therefore it cannot be determined whether patients discontinued due to disease progression, tolerability issues, personal choice, or other reasons. AE rates may have been underestimated, as this study presented only those that were associated with hospital admissions. Due to the small sample size, results for specific treatment regimens should be interpreted with caution. Additionally, claims data are subject to inaccuracies in coding and may lead to misclassification of diagnoses, treatment, and outcomes. Moreover, complete mortality information is not available in claims data. In February of 2024, the Food and Drug Administration approved the use of first-line osimertinib combined with a platinum doublet based on improved progression-free survival versus osimertinib alone [35]. As this approval may change the current treatment paradigm, future studies are warranted to further characterize real-world treatment sequencing and outcomes among patients with EGFR TKI resistance.

Conclusions

Among patients with mNSCLC and common EGFR mutations, this real-world study suggests that the effectiveness of the subsequent LOT following osimertinib was limited given its short duration. Patients were most frequently treated with platinum-based chemotherapies ± IO, which were prone to tolerability issues requiring hospitalization or ED visits. These data underscore the high unmet need for new therapeutic options to address EGFR TKI resistance in this patient population.