Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Second-line therapy in advanced non–small cell lung cancer
CLINICAL PEARLS
-
For immunotherapy-naïve patients, second-line anti-PD-1/PD-L1 inhibitors are preferred over cytotoxic chemotherapy due to improved efficacy and tolerability.
-
For patients with disease progression on first-line immunotherapy monotherapy, second-line cytotoxic chemotherapy with platinum doublet is appropriate.
-
For patients with progression on first-line immunotherapy/platinum combinations, second-line docetaxel remains the most appropriate standard of care combination.
-
Clinical trial enrollment remains strongly encouraged where available, particularly for patients with disease progression on immunotherapy-containing regimens.
Introduction
Non−small cell lung cancer (NSCLC) is a leading cause of cancer-related death in the United States and worldwide, with the majority of cases presenting at an advanced disease stage. Following progression of the disease on or after first-line treatment, the cytotoxic chemotherapy regimens had relatively modest activity and consequently the prognosis for advanced NSCLC has been poor, with a median overall survival (OS) of <12 months. Advances in targeted therapies have improved outcomes for the small subset of patients with actionable mutations, but for the vast majority of patients without such mutations, it is only with the recent introduction of immunotherapy-containing regimens, starting with approvals in the subsequent therapy setting in 2015 and more recently expanding to the first-line setting, that outcomes have started to improve. With this shift in available systemic therapies for advanced NSCLC, management of patients with disease progression on or after first-line therapy has been rapidly evolving. We review here recent changes in the second-line therapy landscape, current best practices, and ongoing clinical investigation in this space.
Cytotoxic chemotherapy
For patients who receive immunotherapy monotherapy in the first-line setting without chemotherapy, guidelines recommend that they receive histology-based platinum doublet chemotherapy before proceeding with the subsequent agents discussed in this chapter (National Comprehensive Cancer Network [NCCN] NSCLC guidelines).
Two pivotal phase III trials published in 2000 established docetaxel as a preferred cytotoxic chemotherapy regimen for patients with disease progression on or after platinum-based first-line treatment. In the first trial, 104 patients with advanced NSCLC, progression on one or more platinum-based chemotherapy regimens, and performance status of 0 to 2 were randomized to treatment with one of two docetaxel regimens (100 vs. 75 mg/m 2 ) or best supportive care (BSC). Median OS was significantly longer in the docetaxel than the BSC arm (7 vs. 4.6 months) with a more pronounced improvement in the 75 mg/m 2 arm. Importantly, this and other trials demonstrated a significant improvement in multiple quality of life (QoL) metrics with docetaxel compared with BSC. , The second trial compared docetaxel with vinorelbine or ifosfamide in the same setting and found a significant improvement in 1-year survival rates with docetaxel compared with vinorelbine or ifosfamide (32% vs. 19%), as well as an improvement in QoL, though no difference in median OS. Pemetrexed and gemcitabine are also NCCN-recommended cytotoxic chemotherapy regimens in this setting, though with much more limited applicability than docetaxel. In 2004, a large phase III trial in the subsequent therapy setting found that pemetrexed had similar efficacy to docetaxel with lower toxicity rates in patients without prior pemetrexed exposure. However, work in the front-line setting demonstrating that the efficacy of pemetrexed is largely in patients with nonsquamous histology remains applicable here and since many, if not most, patients with nonsquamous histology receive pemetrexed as part of front-line platinum doublet therapy, the role for pemetrexed in the subsequent therapy setting is quite limited. Gemcitabine has also demonstrated modest activity in patients with disease progression on platinum therapy, but evidence for this regimen is limited to retrospective analyses and small nonrandomized phase II trials.
Notably, the activity of these cytotoxic monotherapy regimens remains quite limited in the subsequent therapy setting with overall response rates (ORRs) on the order of 5%−10% and median survival only 5 to 8 months. , Since combination chemotherapy regimens have demonstrated increased activity over monotherapy in the front-line setting, there has been extensive investigation of combination regimens in phase II and III trials in the subsequent line setting. However, no combination cytotoxic therapy regimen has shown any benefit over monotherapy while consistently increasing toxicity. This includes phase III trials of docetaxel plus carboplatin, gemcitabine, vinorelbine or capecitabine, phase II randomized trials of docetaxel plus irinotecan, pemetrexed plus carboplatin and cisplatin plus irinotecan, as well as several subsequent meta-analyses of these trials. Combinations of chemotherapy with biological agents including tyrosine kinase inhibitors (TKIs) such as erlotinib and antiangiogenic agents have also been of great interest, as have combinations of the EGFR inhibitor erlotinib with other targeted therapies including antiangiogenic agents. However, while some of these combination regimens modestly improve ORR and progression-free survival (PFS), all add significant toxicity burden, and with the exception of the combination of docetaxel with the antiangiogenic agent ramucirumab in the REVEL trial, none improve OS.
In the phase IIII REVEL trial, 1253 patients with advanced NSCLC progressing on or after first-line platinum-based chemotherapy and excellent performance status (0−1) were randomized to receive docetaxel with placebo or with the antiangiogenic agent ramucirumab, a fully humanized antivascular endothelial growth factor receptor 2 (VEGFR2) antibody. The addition of ramucirumab significantly improved ORR from 14% to 23% and PFS from 3 to 4.5 months, and importantly there was a statistically significant, though modest, improvement in OS as well, from 9.1 to 10.5 months. This benefit was preserved on exploratory subgroup analysis regardless of tumor histology or response to first-line treatment. Not unexpectedly, there was an increase in toxicity with the combination regimen, with an overall rate of grade ≥3 treatment-emergent adverse events (AEs) increasing from 71% to 79% and an increase from 23% to 33% of treatment-emergent AEs requiring dose adjustment. In particular, there was an increase in grade ≥3 febrile neutropenia (from 10% to 16%), as well as an increase from 15% to 29% in rates of all grade bleeding, though no difference in grade ≥3 bleeding events including pulmonary hemorrhage, and an increase from 5% to 11% of all grade hypertension with grade ≥3 events increased from 2% to 6%. There was no increase in the rate of AEs requiring hospitalization (42% vs. 43%) and no difference in treatment-related mortality (5% vs. 6%). A preplanned analysis also found that ramucirumab did not adversely impact QoL by standardized assessment, a particularly important consideration given the modest survival benefit, with about 75% of patients in both arms assessable. Thus the addition of ramucirumab does slightly prolong survival and response rates with a commensurate though not intolerable increase in toxicities, but this is a regimen that should only be considered in fit patients with performance status of 1 or better, who are in the minority once the subsequent therapy setting is reached.
In summary, evidence-based cytotoxic chemotherapy regimens after disease progression on or after platinum-based first-line regimens include docetaxel, pemetrexed for patients with nonsquamous histology and no prior pemetrexed exposure, consideration of docetaxel plus ramucirumab for especially fit patients, and gemcitabine for patients with contraindications to docetaxel and pemetrexed and preferably for patients with squamous histology ( Table 9.1 ).
TABLE 9.1 ■
Key Cytotoxic Chemotherapy Studies in Previously Treated Advanced Non−Small Cell Lung Cancer
| Trial | Phase | Number | Treatment Arms | Overall Survival | 1-Year Survival % | ORR % |
|---|---|---|---|---|---|---|
| Shepherd et al., 2000 | III | 104 |
|
|
|
|
| Fossella et al., 2000 | III | 373 |
|
|
|
|
| Hanna et al., 2004 | III | 571 |
|
|
|
|
| Garon et al., 2014 | III | 1253 |
|
|
|
|
| Crino et al., 1999 | II | 83 | Gemcitabine 1000 mg/m 2 IV d1/8/15 q4w | 8.5 | 45 | 19.3 |
| Sculier et al., 2000 | II | 77 | Gemcitabine 1000 mg/m 2 IV d1/8/15 q4w | 4.25 | Not reported | 5.2 |
BSC , Best supportive care; ORR , overall response rate.
Immunotherapy
Given the modest efficacy of cytotoxic chemotherapy regimens in previously treated advanced NSCLC, the emergence of immune checkpoint inhibitors (ICIs) in this space has been a practice-changing development anticipated to significantly improve patient outcomes. During the process of tumorigenesis, accumulation of genetic and epigenetic alterations leads to the expression of neoantigens on tumor cells that can be recognized by the host immune system as foreign. Tumors evade this immune surveillance by maintaining an immune-suppressive microenvironment through many different mechanisms. The PD-1/PD-L1 immune checkpoint pathway represents one such mechanism that is co-opted by tumor cells to evade immune surveillance. PD-1 is a negative costimulatory receptor primarily present on activated T cells that, when stimulated by ligand binding, inhibits T-cell proliferation and cytokine secretion. Peripheral tissue expression of the PD-1 ligands, PD-L1 and PD-L2, thus acts as a “brake” on inflammatory T-cell responses and is important for preventing autoimmunity. Tumors co-opt this mechanism by upregulating PD-L1 protein expression to evade antitumor T-cell responses, and antibodies that inhibit the PD-1/PD-L1 interaction can thus promote antitumor immune responses to control disease.
This strategy has proven to be effective in NSCLC with three ICIs approved in the subsequent therapy setting for advanced NSCLC: the anti-PD-1 monoclonal antibodies nivolumab and pembrolizumab and the anti-PD-L1 monoclonal antibody atezolizumab. Immune checkpoint blockade has now also been extended to first-line therapy of advanced NSCLC, where it has significantly improved outcomes over prior standard of care cytotoxic chemotherapy regimens as well, and this does complicate the use of ICIs in the subsequent therapy setting. We review here data supporting the approval of nivolumab, pembrolizumab, and atezolizumab as subsequent therapy ( Table 9.2 ) and discuss how the movement of ICIs to the first-line setting impacts the subsequent therapy management landscape.
TABLE 9.2 ■
Phase III Studies of Immune Checkpoint Inhibitors (ICIs) in Previously Treated Advanced Non−Small Cell Lung Cancer (NSCLC)
| Trial | Population | Number | Treatment Arms | OS (ITT) | ORR % (ITT) | Effect of PD-L1 on Outcome |
|---|---|---|---|---|---|---|
| Checkmate 017 (Brahmer et al., 2015) | Squamous NSCLC; progression after first-line platinum doublet | 272 |
|
|
|
No association with ICI benefit |
| Checkmate 057 (Borghaei et al., 2015) | Nonsquamous NSCLC; progression after first-line platinum doublet | 582 |
|
|
|
High PD-L1 associated with greater ICI benefit |
| KEYNOTE-010 (Herbst et al., 2016) | PD-L1 ≥1% NSCLC; progression on ≥1 line of systemic therapy including platinum doublet | 1034 |
|
|
|
High PD-L1 associated with greater ICI benefit |
| OAK (Rittmeyer et al., 2017) | NSCLC; progression on 1–2 lines of systemic therapy including platinum doublet | 1225 |
|
|
|
High PD-L1 associated with greater ICI benefit |
ITT, Intention to treat; ORR, overall response rate; OS, overall survival.
Nivolumab
Nivolumab was the first ICI to receive approval as subsequent therapy for advanced NSCLC. Nivolumab is a humanized IgG4 anti-PD-1 monoclonal antibody that inhibits PD-1/PD-L1 interaction in vitro, promoting T-cell responses. In an initial phase IB study across multiple solid tumors, a heavily pretreated cohort of 129 metastatic NSCLC patients were found to have an ORR of 17% with an unprecedented average response duration of 17 months across 3 dose levels of nivolumab. At the 3 mg/kg dose that was ultimately selected for further development, median OS was 14.9 months with 1-, 2- and 3-year OS rates of 56%, 42%, and 27% in this non–biomarker-selected population, comparing favorably with the 1-year OS rate of 30%–40% seen in the phase III docetaxel trials. , Remarkably, of the 18 patients who discontinued treatment for reasons other than progression, 50% had responses that lasted >9 months after their final dose. In the most recent follow-up of this study, the 5-year OS rate was reported to be 16% compared with historical 5-year OS rates of 5% or less. A subsequent phase II single-arm trial of nivolumab monotherapy for patients with advanced squamous NSCLC with progression on two prior lines of therapy enrolled 117 non–biomarker-selected patients and demonstrated similar outcomes, with an ORR of 15% and 1-year OS rate of 42%.
These encouraging early-phase results led to two randomized phase III trials of nivolumab versus docetaxel in patients with advanced NSCLC with either squamous (CheckMate 017) or nonsquamous (CheckMate 057) histology after progression on or after first-line platinum-based chemotherapy. CheckMate 017 randomized 272 patients with advanced squamous NSCLC and progression on or after first-line platinum-based chemotherapy to either nivolumab 3 mg/kg IV every 2 weeks ( n = 135 patients) or docetaxel 75 mg/m 2 IV every 3 weeks ( n = 137 patients). Performance status of ≤1 was required, and patients with prior ICI or docetaxel treatment, autoimmune disease, symptomatic interstitial lung disease, or systemic immunosuppression were excluded. The primary endpoint was OS, with secondary endpoints including PFS, ORR, and efficacy according to tumor PD-L1 expression. The population was predominantly current or former smokers and predominantly male, and 76% had an Eastern Cooperative Oncology Group performance status of 1. About half of the population had received gemcitabine in combination with platinum as first-line therapy, and 35% had received paclitaxel with platinum; 30% of patients had disease refractory to first-line therapy, and 45% had progression within 3 months of most recent prior therapy, overall consistent with a relatively treatment-refractory population. The study was terminated early when the prespecified stopping boundary for superiority of nivolumab was met in 2014, though planned enrollment was complete at that time. With a minimum of 11 months of follow-up, the median OS was 9.2 months for nivolumab versus 6 months for docetaxel, with a 1-year OS of 42% versus 24%. ORR was improved in the nivolumab arm (20% vs. 9%), and as seen in the early-phase trials, duration of response was significantly longer, not reached in the nivolumab arm versus 8.4 months in the docetaxel arm. Interestingly, a preplanned subgroup analysis of tumor PD-L1 expression found that it was neither prognostic nor predictive of benefit. Nivolumab was better tolerated than docetaxel, with a lower rate of any grade AEs at 58% versus 86% with docetaxel, and an even more striking difference in grade ≥3 AEs at 7% versus 55% with docetaxel. Only 3% of patients in the nivolumab arm discontinued treatment due to treatment-associated AEs compared with 10% in the docetaxel arm, and no deaths on trial were attributed to nivolumab, while three were attributed to docetaxel (one each due to interstitial lung disease, pulmonary hemorrhage, and complications of sepsis).
CheckMate 057 investigated the same question in patients with nonsquamous NSCLC, randomizing 582 patients with advanced nonsquamous NSCLC and progression on or after first-line platinum containing regimens to either 3 mg/kg IV every 2 weeks ( n = 287 patients) or docetaxel 75 mg/m 2 IV every 3 weeks ( n = 268 patients). Relevant inclusion/exclusion criteria and endpoints were similar to CheckMate 017. The population was again predominantly current or former smokers and 70% had a performance status of 1, though as expected with this histology, this trial did have a higher proportion of women (45%). Forty percent of patients had progression of disease as best response on most recent prior line of therapy. Patients with targetable mutations were not excluded, and indeed about 30% had an EGFR or KRAS mutation or ALK translocation. As with CheckMate 017, the study was terminated early due to significant improvement in primary endpoint of OS at the time of the interim analysis in 2015. With a minimum follow-up of 13.2 months, the median OS was 12.2 months with nivolumab versus 9.4 months with docetaxel, with a 1-year OS rate of 51% versus 39%, though interestingly there was no improvement in PFS. ORR was improved at 19% versus 12% with docetaxel, and median duration of response was 17.2 months versus 5.6 months. In an exploratory prespecified subgroup analysis, the benefit of nivolumab was preserved with the exception of patients receiving third-line therapy, patients with intracranial metastases, patients with EGFR mutations and never-smokers, while PD-L1 expression was highly predictive of outcome across different expression cutoffs. As was seen in CheckMate 017, nivolumab was better tolerated than docetaxel, with all grade AEs in 69% versus 88% of patients and grade ≥3 AEs in 10% versus 54% of patients. Discontinuation due to a treatment-associated AE was observed in 5% of patients with nivolumab versus 15% with docetaxel, with one death in each group attributed to study treatment (one case of encephalitis in the nivolumab arm and one from complications of febrile neutropenia in the docetaxel arm). Longer-term follow-up from these trials is now emerging and has been consistent with the primary analyses. The 4-year survival outcomes from a pooled analysis of all of the subsequent therapy nivolumab trials including CheckMate 003, 063, 017, and 057 demonstrated a 4-year OS rate for patients receiving nivolumab of 14% for all patients ( n = 664), 19% for patients with tumor PD-L1 expression ≥1%, and 11% for PD-L1 <1%. The pooled analysis of CheckMate 017 and 057 also confirmed the OS benefit versus docetaxel, with a 4-year OS rate of 14% for patients who received nivolumab versus 5% for patients who received docetaxel. On the basis of the primary results of CheckMate 017 and 057, nivolumab received US Food and Drug Administration (FDA) approval as subsequent therapy after progression on or after platinum doublet for both squamous and nonsquamous advanced NSCLC regardless of PD-L1 status in 2015.
Pembrolizumab
Pembrolizumab is a humanized IgG4 kappa anti-PD-1 monoclonal antibody that showed initial promise in advanced NSCLC in the phase I setting. KEYNOTE-001 was a phase I trial of pembrolizumab monotherapy in advanced solid tumors that included patients with NSCLC. A total of 495 patients with NSCLC, both treatment-naïve and pretreated, received pembrolizumab at three different doses, including 2 mg/kg and 10 mg/kg IV every 3 weeks and 10 mg/kg IV every 2 weeks. Endpoints included safety, side-effect profile, and efficacy of pembrolizumab. After early data indicated an association between tumor PD-L1 expression and pembrolizumab efficacy in the NSCLC cohort, a coprimary endpoint was added, assessing efficacy specifically in patients with previously treated NSCLC and high tumor PD-L1 expression, with PD-L1 cutoff determined by analysis of initial data from a training cohort of 182 patients. Most (75%) of the people were current or former smokers, 65% had a performance status of 1 (performance status of ≥2 was excluded), and 47% were female. Though all histologies were permitted, 80% of the patients had nonsquamous histology; 15% had EGFR mutation, 2% had an ALK translocation, and 26% had KRAS mutation. For the most part, this was a heavily pretreated patient population, with 19% of patients being treatment naïve, 15% having received one prior line of treatment, and the remaining 65% having received two or more prior therapies. ORR was 19.4% in the overall population and 18% in the subsequent therapy group ( n = 394) versus 24.8% in the first-line therapy group ( n = 101). Patients who had PD-L1 expression of ≥50% had an ORR of 45.2% ( n = 73). As with nivolumab, responses were durable, with a median duration of 10.4 months in the subsequent therapy group; median OS in this group was 9.3 months. Recently updated 5-year survival data from the KEYNOTE-001 NSCLC cohort, with a median of 60.6 months of follow-up, demonstrated an updated median OS in pretreated patients of 10.5 months, with a remarkable 5-year OS rate of 25%. Overall, pembrolizumab monotherapy was well tolerated with any grade AEs occurring in 71% of patients but grade ≥3 AEs in 9.5% of patients. The most common immune-related adverse events (IRAEs) included infusion reactions in 3%, hypothyroidism in 7%, and pneumonitis in 3.6% including 1.8% grade ≥3 and one pneumonitis-related death.
On the basis of these encouraging early results, pembrolizumab was compared with docetaxel in the subsequent therapy setting for advanced NSCLC in the phase II/III KEYNOTE-010 trial. In this trial, 1034 patients with advanced NSCLC of any histology and progression on or after platinum-doublet chemotherapy were randomized to pembrolizumab 2 mg/kg IV ( n = 345), pembrolizumab 10 mg/kg IV ( n = 346), or docetaxel 75 mg/m 2 IV ( n = 343) every 3 weeks. Performance status of 0 or 1 and tumor PD-L1 expression of ≥1% were required. In addition to prior ICI or docetaxel treatment, active autoimmune disease requiring systemic steroids, interstitial lung disease, or history of pneumonitis were key exclusion criteria. Primary endpoints were PFS and OS in patients with PD-L1 expression ≥1% and ≥50%. The majority (80%) of the population were current or former smokers, 60% were male, and 65% had a performance status of 1. Most (70%) of the patients had nonsquamous histology, 8% had EGFR mutation, and 1% had ALK translocation. Forty percent had PD-L1 ≥50%, and distribution of PD-L1 expression was similar between arms. Compared with KEYNOTE-001, this was a less heavily pretreated population, with 70% of patients having received one prior line of therapy. At the primary analysis, with a median follow-up of 13.1 months, median OS was significantly improved in the pembrolizumab arms at 10.4 months for 2 mg/kg and 12.7 months for 10 mg/kg compared with the docetaxel arm at 8.5 months. The improvement in OS was more pronounced in patients with tumor PD-L1 ≥50% ( n = 633), with median OS of 14.9 months for 2 mg/kg pembrolizumab and 17.3 months for 10 mg/kg pembrolizumab versus 8.2 months for docetaxel. While there was no significant improvement in PFS in the overall population (4 months in all three arms), there was a small but statistically significant improvement with pembrolizumab in the PD-L1 ≥50% population, at 5 and 5.2 months with 2 and 10 mg/kg pembrolizumab versus 4.1 months for docetaxel. ORR was significantly improved in the whole population, at 18% for both pembrolizumab doses versus 9% for docetaxel, and was even higher in the PD-L1 ≥50% population at 30% for both pembrolizumab doses versus 8% for docetaxel. Finally, as with nivolumab, duration of response was significantly longer in the pembrolizumab arms, with median duration not reached versus 8 months with docetaxel. A prespecified subgroup analysis of OS found preserved benefit across histologies and PD-L1 expression (though with HR of 0.53 in PD-L1 ≥50% vs. 0.76 in 1%−49%), but notably, there was not a significant benefit in the EGFR mutant subgroup, though this was a small subset and the confidence interval was wide. The PFS subgroup analysis yielded similar findings. In terms of toxicity, pembrolizumab was better tolerated than docetaxel, with rates of any grade AEs 60% in both pembrolizumab arms versus 80% with docetaxel, and rates of grade ≥3 AEs 13% for 2 mg/kg pembrolizumab, 16% for 10 mg/kg, and 35% for docetaxel. IRAEs occurred in 20% of the patients in each pembrolizumab arm, with the most common being hypothyroidism, hyperthyroidism, and pneumonitis, and grade 3 to 5 IRAEs including pneumonitis, colitis, and severe skin reactions in 2% of patients. Toxicity-related drug discontinuation occurred in 4% of patients receiving 2 mg/kg pembrolizumab, 5% of patients receiving 10 mg/kg, and 10% of patients receiving docetaxel. Treatment-related deaths were rare and occurred at a similar rate between arms, with three in the 2 mg/kg pembrolizumab group (two cases of pneumonitis and one of pneumonia); three in the 10 mg/kg pembrolizumab group (myocardial infarction, pneumonitis, and pneumonia); and five in the docetaxel group (acute cardiac failure, dehydration, febrile neutropenia, interstitial lung disease, and pneumonia). A prespecified analysis of patient-reported outcomes (PROs) also demonstrated that pembrolizumab was associated with improved QoL and delayed deterioration in QoL compared with docetaxel in this trial, with high rates of PRO instrument completion (>85%) across arms. On the basis of the favorable efficacy compared with docetaxel in the overall population of KEYNOTE-010, pembrolizumab received FDA approval as subsequent therapy for advanced NSCLC with PD-L1 ≥1% in 2015.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here