Combining Osimertinib With Chemotherapy in EGFR-Mutant NSCLC at Progression

Abstract

Background:

Osimertinib is a third-generation EGFR tyrosine kinase inhibitor (TKI) that has improved survival and central nervous system (CNS) outcomes in patients with non-small cell lung cancer (NSCLC) with activating EGFR mutations. However, little is known about the efficacy and safety of combining osimertinib with chemotherapy.

Methods:

This was a retrospective study performed at three institutions. Patients with advanced EGFR-mutated NSCLC who received concurrent osimertinib with chemotherapy in the third-line or beyond were identified by chart review. Efficacy outcomes including duration on treatment (DOT), overall survival (OS), and CNS outcomes were assessed. Safety outcomes were also evaluated.

Results:

A total of 44 patients met inclusion criteria. Median DOT with osimertinib plus platinum doublet (n=28) was 6.1 months (95% CI 4.1 months-Not Reached), and with osimertinib plus single-agent chemotherapy (n=29) was 2.6 months (95% CI 1.8–4.8 months). Median OS from the start of osimertinib plus chemotherapy was 10.4 months (95% CI 7.0-13.2 months). At initiation of osimertinib plus chemotherapy, 37 patients (84%) had CNS metastases; 9 of these (24%) had CNS disease progression on osimertinib plus chemotherapy. Chemotherapy was delayed or dose reduced due to toxicity in 8 patients (18%); osimertinib was discontinued in 1 patient (2%) for reduced cardiac ejection fraction, and dose reduced in 2 patients (5%).

Conclusions:

The combination of osimertinib plus chemotherapy appeared safe and showed favorable control of CNS disease in this cohort of patients who had progressed systemically with multiple prior lines of therapy, with DOT and survival outcomes similar to historical chemotherapy controls.

MicroAbstract

Little is known about the efficacy and safety of combining osimertinib with chemotherapy in EGFR-mutated non-small cell lung cancer. We retrospectively examined outcomes in patients who received this combination after progression on EGFR-targeted monotherapy, and found that the combination appeared safe, had favorable CNS outcomes, and had comparable efficacy outcomes to historical controls who received chemotherapy alone after osimertinib.

Introduction

Osimertinib is a third-generation EGFR TKI that is now standard of care first-line therapy for advanced NSCLC harboring a sensitizing EGFR mutation, based on the results of the phase III FLAURA trial, which showed that osimertinib improves overall survival compared with first-generation EGFR TKIs.^{1, 2} Osimertinib has excellent CNS penetration, with both better response rates in the CNS and reduced risk of CNS progression compared with first-generation EGFR TKIs^{1, 3} and with platinum chemotherapy.^4–6

Osimertinib is administered as monotherapy, and thus there are limited data, and no randomized controlled trials, describing efficacy and safety of combining osimertinib with chemotherapy. There are data, however, on the combination of first-generation EGFR TKIs with chemotherapy, both in the front-line setting and after progression on initial TKI therapy. The Japanese phase III NEJ009 trial of gefitinib plus platinum doublet chemotherapy versus gefitinib as first-line therapy for advanced NSCLC with EGFR mutations showed improved progression-free survival in the combination arm, and exploratory ad hoc analysis suggested possible overall survival benefit.⁷ A similar phase III trial conducted at a large specialty cancer center in India found improved progression-free survival and overall survival in the gefitinib plus platinum doublet arm compared with gefitinib alone.⁸ Both studies noted an increase in Common Terminology Criteria for Adverse Events (CTCAE) grade 3 or higher toxicity with the addition of chemotherapy to gefitinib. The phase III IMPRESS trial compared continuation of gefitinib plus chemotherapy with placebo plus chemotherapy in patients with EGFR-mutated NSCLC who had progressed on first-line gefitinib.^{4, 6} In contrast to the other studies, in IMPRESS continuation of gefitinib with chemotherapy was surprisingly found to be detrimental to overall survival when compared with placebo plus chemotherapy. Based on these results, the combination of first-generation EGFR TKI and chemotherapy after progression on first-line TKI is not recommended. Of note, toxicity was similar in both arms, with 46% of patients in the gefitinib plus chemotherapy and 42% of patients in the placebo plus chemotherapy arms experiencing grade 3 or higher adverse events.

While continuation of gefitinib with platinum doublet chemotherapy did not improve outcomes over chemotherapy alone in IMPRESS, there is reason to suspect outcomes may be different with osimertinib, which has significantly greater CNS efficacy than earlier generation EGFR TKIs; continuation of osimertinib at the time of progression may provide greater CNS control and improve outcomes. Around 20-40% of patients with NSCLC develop brain metastases throughout the course of their disease, and in patients harboring EGFR mutations, rates are higher; approximately 25% of these patients are found to have brain metastases at time of diagnosis of advanced disease, and 50-60% of patients develop brain metastases at some point during the course of their disease.^9–11 This highlights the importance of developing strategies for enhanced CNS disease control in this patient population.

Optimal subsequent therapy after progression on osimertinib must be individualized based on sites of progression and mechanisms of resistance. In patients with limited sites of progression, either in the brain or systemically, there are data supporting continuation of osimertinib along with treatment of sites of oligo-progression with definitive local therapy, such as stereotactic ablative radiotherapy (SABR) or surgery.^{12, 13} Patients with significant systemic progression typically discontinue osimertinib and, unless found on a repeat biopsy to have a resistance mechanism amenable to a different targeted therapy, switch to a chemotherapy-based regimen, with or without bevacizumab and/or a PD-1/PD-L1 checkpoint inhibitor.¹⁴ However, the option of continuing osimertinib with chemotherapy may be of particular benefit for patients on osimertinib who are progressing extracranially and also need to maintain disease control in the brain. While a change in systemic therapy is indicated, continuation of osimertinib in this situation may be beneficial due to its CNS activity. For this reason, the strategy of combining osimertinib with chemotherapy has selectively been employed in clinical practice at our institutions. In this multicenter retrospective analysis, we examined efficacy and safety of osimertinib plus chemotherapy in patients who were treated with this combination in the third line or beyond.

Methods

Study Population

Patients were identified from three participating institutions: Massachusetts General Hospital (n=25), Stanford Cancer Institute (n=10), and Georgetown Lombardi Comprehensive Cancer Center (n=9). Patients were eligible for inclusion if they had a diagnosis of advanced NSCLC with a sensitizing EGFR mutation identified by polymerase chain reaction (PCR) testing or next generation sequencing (NGS), and had received EGFR TKI therapy as initial systemic therapy for metastatic disease, plus at least one additional line of therapy, which could have been another EGFR TKI, chemotherapy, or immunotherapy. Eligible patients subsequently received osimertinib in combination with any single-agent or combination chemotherapy regimen. Patients were not required to have received osimertinib or another 3^rd generation EGFR TKI immediately preceding initiation of osimertinib plus chemotherapy. The institutional review board at each participating institution approved the study.

Data Collection

Patient demographics, treatment history, and toxicity were collected by retrospective chart review through November 2018 (MGH and Stanford) and May 2019 (Georgetown). Only patients who started osimertinib on or prior to August 1, 2018 were included in this study.

Efficacy outcomes and safety data were evaluated for each patient. Duration on treatment (DOT; first dose to last dose of therapy) was examined as a surrogate measure for radiographic progression or unacceptable toxicities. For patients who received multiple lines of chemotherapy in combination with osimertinib, each line of chemotherapy was considered a separate “episode” for assessment of duration on treatment. Overall survival from the time of diagnosis of metastatic disease, from the start of osimertinib, and from the start of osimertinib in combination with chemotherapy was examined for each patient. For safety analyses, the highest-grade toxicity (using CTCAE v5.0) during any episode of osimertinib with chemotherapy was recorded for each patient.

Statistical Analysis

Kaplan-Meier analyses were used to estimate median DOT and median overall survival from the date of diagnosis of metastatic disease, from the date of first exposure to osimertinib, and from the date of first osimertinib plus chemotherapy. Treatment episodes were stratified by chemotherapy regimen for subgroup analysis. Ninety-five percent confidence intervals (CI) were calculated using the log-log transformation. Data analyses were performed using R statistical software (Version 3.5.2).

Results

Patient characteristics

We identified 44 patients who received osimertinib with concurrent chemotherapy and were eligible for study inclusion, with a total of 57 individual chemotherapy plus osimertinib episodes among the 44 patients. Baseline characteristics are summarized in Table 1. The median age at diagnosis of lung cancer was 58 years (range 34–82 years). The majority of patients were female (68%). All patients had adenocarcinoma histology, and the majority had an EGFR exon 19 deletion (64%) or L858R mutation (34%). Most patients had stage IV disease at diagnosis (n=41; 93%), while three (7%) had recurrent disease. At the time of diagnosis of metastatic disease, 23 patients (52%) had CNS metastases. An additional 12 patients developed brain metastases before starting osimertinib, for a total of 35 patients (80%) who had CNS metastases prior to the start of osimertinib monotherapy. Two additional patients developed CNS metastases while on osimertinib monotherapy; thus, 37 patients (84%) had CNS metastases prior to the start of any osimertinib plus chemotherapy combination.

Table 1.

Baseline Characteristics

Characteristic	All patients (n=44)
Median age at diagnosis, years (range)	58 (34–82)
Sex – n (%)
Female	30 (68)
Male	14 (32)
Race – n (%)
Asian	20 (45)
White	17 (39)
Black	3 (7)
Other	4 (9)
Histologic type – n (%)
Adenocarcinoma	44 (100)
Stage at initial diagnosis – n (%)
Stage I-III	3 (7)
Stage IV	41 (93)
EGFR mutation – n (%)
Exon 19 deletion	28 (64)
L858R	15 (34)
G719A	1 (2)
T790M (ever detected with above) – n (%)	30 (68)
CNS metastases present – n (%)
At diagnosis of metastatic disease	23 (52)
Prior to start of first osimertinib	35 (80)
Prior to start of osimertinib + chemotherapy	37 (84)
Number of prior lines of therapy (before osimertinib + chemotherapy) – n (%)
2	22 (50)
3	10 (23)
4	9 (20)
≥5	3 (7)
Prior treatment (before osimertinib + chemotherapy) – n (%)
Prior EGFR TKI	44 (100)
Prior 3rd generation EGFR TKI	43 (98)
Prior 1st or 2nd generation EGFR TKI	42 (95)
Prior chemotherapy alone	15 (34)
Prior immunotherapy	3 (7)
Immediate prior treatment (directly preceding osimertinib + chemotherapy) – n (%)
Osimertinib monotherapy	38 (86)
Osimertinib + EGFR mAb	2 (5)
Investigational 3rd generation EGFR TKI	2 (5)
Chemotherapy alone	2 (5)

TKI, tyrosine kinase inhibitor; mAb = monoclonal antibody

Prior to the start of osimertinib plus chemotherapy, half of the patients (n=22 patients; 50%) had received two lines of therapy, 10 patients (23%) had received 3 lines of therapy, 9 patients (20%) had received 4 lines of therapy, and 3 patients (7%) had received 5 or more lines of therapy. All patients had received an EGFR TKI as first-line treatment after being diagnosed with metastatic disease. All except three patients (n=41; 93%) had received prior osimertinib monotherapy, and all except two patients (n=42; 95%) had received an additional prior EGFR TKI. Of the three patients who had not received osimertinib monotherapy, two had received an investigational third-generation EGFR TKI on a clinical trial (nazartinib, also known as EGF816). The other patient had a low allelic fraction of T790M on tumor testing, and thus there was concern for low likelihood of response to osimertinib alone, so the patient’s first exposure to osimertinib was in combination with chemotherapy; this patient had previously been treated with first-generation EGFR TKI. One third of patients (n=15; 34%) had received one or more lines of chemotherapy. Three patients (7%) had received immunotherapy.

Twenty-seven patients (61%) received platinum doublet chemotherapy with osimertinib, including one who received two lines of platinum doublet chemotherapy, for a total of 28 episodes. The platinum doublet regimens that patients received in combination with osimertinib were: carboplatin/pemetrexed (n=24; 55%), platinum/pemetrexed/bevacizumab (n=2; 5%), carboplatin/taxane (n=2; 5%). The single-agent chemotherapy regimens that patients received with osimertinib were: gemcitabine (n=9; 20%), pemetrexed (n=6; 14%), docetaxel (n=5; 11%), irinotecan (n=3; 7%), vinorelbine (n=2; 5%), nab-paclitaxel (n=2; 5%), pemetrexed/bevacizumab (n=1; 2%), docetaxel/ramucirumab (n=1; 2%). The majority of patients (n=34; 77%) received only one line of chemotherapy in combination with osimertinib. Of the 10 patients who received multiple lines of chemotherapy with osimertinib, 8 (18%) received two lines, one (2%) received three lines, and one (2%) received four lines.

Most patients (n=41; 93%) received standard dose osimertinib (80 mg daily) in combination with first chemotherapy; however, one of these patients went on to receive higher dose osimertinib (160 mg daily) in combination with next-line chemotherapy, due to the development of new brain metastases with leptomeningeal disease. The other three patients (7%) were on non-standard osimertinib doses prior to starting chemotherapy, and were continued on these doses when started on chemotherapy: two patients (5%) were on osimertinib 160 mg daily (one because of high CNS disease burden, and one because osimertinib had been started initially on a clinical trial), and one patient (2%) was on 80 mg every other day (due to the development of pneumonitis while on osimertinib monotherapy, prompting dose reduction at that time).

Efficacy

Median follow-up time was 7.0 months (interquartile range 4.9–11.8 months). The median DOT with osimertinib plus first chemotherapy for all patients (n=44) was 5.6 months (95% CI 4.1–8.2 months) (Figure 1A). Median DOT with osimertinib plus platinum doublet chemotherapy (n=28 episodes) was 6.1 months (95% CI 4.1 months-Not Reached) (Figure 1B). Median DOT with osimertinib plus single agent chemotherapy (n=29 episodes), either as first chemotherapy line in combination with osimertinib, or subsequent chemotherapy line in combination with osimertinib, was 2.6 months (95% CI 1.8–4.8 months) (Figure 1C).

Figure 1.

Kaplan-Meier estimates of duration on treatment for (A) osimertinib plus first chemotherapy (n=44), (B) osimertinib plus platinum doublet (n=28), (C) osimertinib plus single agent chemotherapy (n=29). Median duration on treatment (mDOT) shown in months (95% confidence interval). NR, not reached.

For the full cohort of 44 patients, median overall survival from diagnosis of metastatic disease was 48.4 months (95% CI 37.2–54.4 months) (Figure 2A), from the start of first osimertinib was 20.4 months (95% CI 17.6–32.0 months) (Figure 2B), and from the start of osimertinib plus chemotherapy was 10.4 months (95% CI 7.0-13.2 months) (Figure 2C). For the subgroup of patients who received platinum doublet chemotherapy in combination with osimertinib (n=27 patients; 61% of overall cohort), median OS from diagnosis of metastatic disease was 48.4 months (95% 30.7 months – NR), from the start of first osimertinib was 27.9 months (17.6 months – NR), and from the start of osimertinib plus chemotherapy was 11.8 months (6.0 months – NR) (Supplementary Figure S1).

Figure 2.

Kaplan-Meier estimates of overall survival for full cohort (n=44 patients) (A) from diagnosis of metastatic disease, (B) from start of osimertinib therapy, (C) from start of osimertinib plus chemotherapy. Median overall survival (mOS) shown in months (95% confidence interval). NR, not reached.

Among the 37 patients with CNS metastases present at the time of initiation of concurrent chemotherapy and osimertinib, 9 (24%) had CNS disease progression while on osimertinib plus chemotherapy. Among the 7 patients who did not have CNS metastases at the initiation of concurrent chemotherapy and osimertinib, 2 (29%) developed new CNS metastases while on chemotherapy plus osimertinib – one was 2.6 months and one was 5.3 months after starting combination therapy.

Safety

Chemotherapy was delayed or dose reduced in 18% of patients (n=8 of 44); the reasons were fatigue (n=5, 11%), transaminitis (n=1, 2%), neutropenia (n=1, 2%), and thrombocytopenia (n=1, 2%). Osimertinib was dose reduced from 80 mg to 40 mg in 2 patients: for thrombocytopenia (n=1, 2%) and transaminitis (n=1, 2%). Additionally, one of the two patients who was on osimertinib 160 mg daily at the start of first chemotherapy was decreased temporarily to 80 mg daily due to fatigue, but was later able to increase back up to 160 mg daily. Osimertinib was discontinued in one patient (2%) due to reduction in cardiac ejection fraction (from 66% to 38%).

Selected toxicities are shown in Table 2. Hematologic toxicities were most common. Anemia occurred in 33 patients (75%); 4 patients (9%) had grade 3, and none had grade 4 events. Neutropenia developed in 14 patients (32%); 7 patients (16%) had grade 3-4 events. Thrombocytopenia occurred in 26 patients (59%); 2 patients (5%) had grade 3-4 events. No grade 5 events were observed attributable to therapy in these patients. Among the 12 patients who experienced grade 3-4 hematologic toxicities while on osimertinib plus chemotherapy, 6 (50%) had not had any cytopenias while on osimertinib monotherapy, and 6 (50%) had experienced lower grade cytopenias while on osimertinib monotherapy (all were grade 1). This included 3 of the 4 patients with grade 3 anemia, 2 of the 7 patients with grade 3-4 neutropenia, and 1 of the 2 patients with grade 3-4 thrombocytopenia (of note, one of the patients developed both grade 3+ thrombocytopenia and anemia; this patient had previously had grade 1 anemia but no thrombocytopenia).

Table 2.

Selected Toxicities: Highest-Grade Toxicity During Any Episode of Osimertinib with Chemotherapy (n=44 patients)

Adverse Events	Any Grade	Grade 3	Grade 4
AST/ALT elevation	15 (34%)	1 (2%)	0
Anemia	33 (75%)	4 (9%)	0
Neutropenia	14 (32%)	5 (11%)	2 (5%)
Thrombocytopenia	26 (59%)	1 (2%)	1 (2%)
Ejection fraction decreased	1 (2%)	1 (2%)	0
Pneumonitis	0	0	0

AST, aspartate aminotransferase; ALT, alanine aminotransferase

Discussion

EGFR targeted therapies have revolutionized the treatment of advanced EGFR-mutated NSCLC, and in particular, osimertinib has improved survival as well as CNS outcomes compared with earlier generation EGFR TKIs.^{1, 15} Despite these improvements, most patients ultimately progress, and those with adequate performance status typically go on to receive chemotherapy-based regimens. As data on continuing osimertinib when starting chemotherapy are lacking, we performed the current study to retrospectively evaluate efficacy and safety outcomes of patients who received osimertinib in combination with chemotherapy after initial treatment with an EGFR TKI and at least one other line of therapy.

We found that, despite the multiply pretreated status and the high prevalence of brain metastases in our cohort, median duration on treatment – 6.1 months for platinum doublet with osimertinib, and 2.6 months for single agent chemotherapy with osimertinib – was comparable to earlier-line progression-free survivals seen in historical controls, both in trials of chemotherapy given after progression on first-line targeted therapy, and trials of frontline combination chemotherapy plus first-generation EGFR TKI. In the IMPRESS trial, patients who received placebo plus platinum-pemetrexed chemotherapy after progression on first-line gefitinib had PFS of 5.3 months (T790M mutation-positive patients) and 5.4 months (T790M mutation-negative patients), and those who received continuation gefitinib with platinum-pemetrexed chemotherapy had PFS of 4.6 months (T790M mutation-positive patients) and 6.7 months (T790M mutation-negative patients). A post hoc analysis based on the presence of brain metastases at baseline showed significantly worse PFS (as well as OS) in patients with brain metastases, with HR 1.82 (95% CI 1.14–2.82, p=0.0089).⁴ In the AURA3 trial, patients who received platinum-pemetrexed after frontline gefitinib, erlotinib, or afatinib, had PFS of 4.4 months (versus osimertinib arm, which had PFS of 10.1 months).¹⁶ In a Chinese phase III trial comparing osimertinib versus docetaxel-bevacizumab as third-line treatment in EGFR T790M mutated NSCLC, the PFS in the docetaxel-bevacizumab arm was 2.95 months (versus osimertinib arm, which had PFS of 10.2 months).¹⁷

The median overall survival from time of confirmed metastatic disease in our patient cohort was 48.4 months (95% CI 37.2–54.4 months), even though the majority of patients (52%) had CNS metastases at the time of diagnosis of metastatic disease. By comparison, median OS in the FLAURA trial in the front-line osimertinib arm, in which only 19% of patients had CNS metastases at baseline, was 38.6 months (95% CI 34.5–41.8 months).¹⁵ Based on these observations, there doesn’t seem to be a detriment in overall survival using the approach of chemotherapy plus osimertinib, which was a potential concern given the findings of the IMPRESS trial. The median overall survival from the start of osimertinib in our cohort was 20.4 months (95% CI 17.6–32.0 months), which is slightly lower than the median overall survival of 26.8 months (95% CI 23.5–31.5 months) observed in the osimertinib arm of the AURA3 trial, in which osimertinib was given in second line or beyond after a prior EGFR TKI. Of note, in AURA3, 96% of patients had only received one prior line of therapy, whereas our cohort was more heavily pre-treated. The presence of a T790M mutation was a requirement for AURA3 but only 68% of patients in our study had this detected at any point. Only 33% of patients in AURA3 had brain metastases prior to the start of osimertinib, compared with 80% of our cohort.^{16, 18} Additionally, all of the patients enrolled on AURA3 were eligible for platinum doublet chemotherapy, which was not the case for our cohort. In the subgroup of patients from our cohort who received platinum doublet chemotherapy with osimertinib – which is likely more comparable to the AURA3 patient population – the median survival was 27.9 months, which is very similar to AURA3.

Control of CNS disease appears favorable with osimertinib and chemotherapy in this cohort of patients heavily enriched for brain metastases (84% at baseline), with only 24% of patients who had brain metastases at the start of osimertinib plus chemotherapy experiencing CNS disease progression on this combination. This is comparable to what was observed on the osimertinib arm of the FLAURA study, in which 19% of patients with known CNS metastases at trial entry experienced CNS progression, compared with 43% on first-generation TKI.¹ Prior research has shown that rates of CNS progression in patients with EGFR mutations receiving platinum-based chemotherapy regimens, which are known to have relatively modest CNS penetration, are similar to that seen with first-generation TKI.¹⁹ Thus, in our cohort, it is likely that the continuation of osimertinib, rather than the addition of chemotherapy, principally contributed to the ongoing CNS control. Our findings are especially notable because almost all of the patients in this cohort had prior systemic progression on osimertinib monotherapy, but nevertheless most maintained good CNS control after switching to chemotherapy plus osimertinib. Thus, continuation of osimertinib with initiation of chemotherapy appears to be a promising approach, and may specifically preserve CNS control among patients with brain metastases.

Adding osimertinib to chemotherapy did not appear to add unacceptable toxicities to chemotherapy, across the multiple different chemotherapy combinations. Rates and severity of myelosuppression were comparable to those observed in the platinum/pemetrexed arm of the AURA3 trial²⁰; cytopenia rates for our cohort and both arms of the AURA3 trial are shown in Table 3. Of note, however, while the majority of patients in our cohort (n=28; 64%) received a platinum doublet in combination with osimertinib, some patients only received single agent chemotherapy in combination with osimertinib. Additionally, our numbers are small, limiting conclusions about true toxicity rates of the combination. Importantly, only 8 patients (18%) required delay or dose reduction of chemotherapy due to toxicity, and two patients (5%) required dose reduction of osimertinib. Only one patient (2%) required discontinuation of osimertinib, and this was due to reduced cardiac ejection fraction; this side effect has been observed in osimertinib monotherapy arms in the prior clinical trials, which have reported rates of 1-3%.^{1, 16}

Table 3.

Cytopenia Rates in the Present Cohort and in Both Arms of AURA3 Trial

Adverse Events	Present cohort: chemotherapy + osimertinib (n=44)	AURA3: platinum + pemetrexed arm (n=131)	AURA3: osimertinib arm (n=279)
Anemia
Any grade	75%	79%	43%
Grade ≥ 3	9%	3%	0%
Neutropenia
Any grade	32%	49%	27%
Grade ≥ 3	16%	12%	2%
Thrombocytopenia
Any grade	59%	48%	46%
Grade ≥ 3	5%	7%	<1%

The limitations of our study are inherent to retrospective data collection. While this was a multi-institution study, the numbers of patients were modest, and because the three participating centers are academic institutions, the patient population may not be representative of patients with EGFR-mutant lung cancer in general. There may have been selection bias when considering patients for osimertinib plus chemotherapy combination strategy; for example, the median age in this cohort (58 years) is quite young. We did not include a comparison cohort of patients who received chemotherapy alone after progression on osimertinib, but rather compared with historical controls published in the literature. However, there are no suitable historical controls to which to compare overall survival data from the start of chemotherapy after progression on osimertinib, so we compared OS from other time points (diagnosis of metastatic disease and start of osimertinib) and DOT. Additionally, we used DOT as a surrogate measure of PFS. However, DOT has been shown to correlate closely with PFS, and may even underestimate PFS, as some patients discontinue treatment for toxicity rather than due to progression.²¹ Thus our DOT estimates may in fact be conservative surrogate estimates of PFS. Because osimertinib had received FDA approval in the first-line setting only shortly before the defined cut-off date for start of osimertinib therapy in our cohort, patients who received osimertinib in the first-line setting were not included in this analysis, and all of our patients were in their third line of therapy or beyond. We would expect patients who receive osimertinib as frontline therapy followed by chemotherapy with osimertinib in the second line to have as good – and likely better – DOT than that observed in our cohort.

Conclusion

In summary, our data suggest that continuation of osimertinib with chemotherapy may improve CNS outcomes, and does not appear to worsen survival outcomes or cause worsened toxicity compared with historical controls of patients receiving chemotherapy regimens alone in the second line or beyond. Thus, patients to consider for this strategy include those who have progressed systemically but have maintained CNS disease control on osimertinib monotherapy. A randomized trial of continuation of osimertinib with chemotherapy versus placebo with chemotherapy is being conducted in the second-line setting, after progression on first-line osimertinib (COMPEL; ClinicalTrials.gov identifier pending). However, it is also possible that use of chemotherapy and osimertinib in the first-line setting may improve survival of patients. A phase III clinical trial of osimertinib plus platinum/pemetrexed versus osimertinib monotherapy for front-line treatment of EGFR-mutated, advanced NSCLC, is currently recruiting (FLAURA2; ClinicalTrials.gov identifier: NCT04035486).²² These prospective studies are needed to further examine outcomes of osimertinib given in combination with chemotherapy and potentially move toward a new standard of care.

Clinical Practice Points.

Osimertinib is now standard-of-care first-line therapy for patients with advanced non-small cell lung cancer with sensitizing EGFR mutations. Because of its excellent CNS penetration, there may be benefit to continuing osimertinib with chemotherapy in patients who have had extracranial disease progression but have maintained ongoing disease control in the CNS on osimertinib monotherapy. This strategy has been used selectively in practice at our institutions. This study analyzed 44 patients who received treatment with osimertinib plus chemotherapy; some patients received multiple lines of chemotherapy in combination with osimertinib, thus there were 57 individual chemotherapy plus osimertinib episodes among the 44 patients. Despite our cohort being more heavily pre-treated and having a higher rate of brain metastases at baseline, survival outcomes were comparable to those seen with historical controls of patients who received chemotherapy alone after EGFR-targeted therapy. Most patients with brain metastases maintained CNS disease control while on the combination. Importantly, adding osimertinib to chemotherapy did not appear to add significant toxicities to chemotherapy; cytopenias were the most common side effects, with rates comparable to those in historical chemotherapy-only controls. The data we present may support a clinician’s decision to continue osimertinib with chemotherapy after progression on osimertinib. Patients who may benefit especially with this strategy include those who have progressed systemically but have maintained CNS disease control on osimertinib monotherapy.