What is the optimal systemic therapy for stage IV inoperable NSCLC?

Guideline contents > What is the optimal systemic therapy for stage IV inoperable NSCLC?

Guidelines commissioned by

Cite this page

Cancer Council Australia Lung Cancer Guidelines Working Party. Clinical question:What is the optimal systemic therapy for stage IV inoperable NSCLC? . In: Clinical practice guidelines for the treatment of lung cancer. Sydney: Australian Government / Cancer Australia. [Version URL: https://wiki.cancer.org.au/australiawiki/index.php?oldid=72278, cited 2023 Sep 24]. Available from: https://wiki.cancer.org.au/australia/Guidelines:Lung_cancer.

Last modified:
11 March 2014 05:51:20

What is the optimal systemic therapy for stage IV inoperable NSCLC?

Introduction

The majority of patients treated with NSCLC have stage IV disease, with common sites of metastases including lymph nodes, the pleura, liver, adrenal glands, bone and brain. Consequently, systemic therapy has been the mainstay of treatment attempting to control overall disease. A historical summary of the evolution of systemic drug treatment for stage IV NSCLC can be found here The focus of the following sections of this guideline will be based on the evidence in support of the old and new practice paradigms for stage IV NSCLC. For consistency, empirical therapy refers to therapy given to all fit patients deemed suitable without any particular restrictions, biologic therapy refers to monoclonal antibodies (MAbs) and will be summarised by the specific MAb target; molecularly targeted therapy refers to therapy given to patients selected exclusively by the presence of a particular gene or its protein product identified as the specific drug target.

Comments

Back to top

Empirical therapy in unselected patients

First-line chemotherapy

The first piece of evidence to establish a standard of practice was the meta-analysis of randomised trials until 1992 evaluating chemotherapy for non-Small Cell Lung Cancer by the Non-small Cell Lung Cancer Collaborative Group. Data from eight trials (N = 778) evaluating best supportive care versus best supportive care and cisplatin based chemotherapy showed a clear survival benefit in favour of chemotherapy with a hazard ratio of 0.73 (P<0.0001), or 27% reduction in the risk of death. This is equivalent to an absolute improvement in survival of 10% at one year, improving survival from 15% to 25%.

It is important to note that empirical chemotherapy has only been formally evaluated in "fit" patients. Patient performance status (PS) has conventionally been used to standardise and quantify cancer patient’s general well-being and activities of daily life. The simplest of such scores in widespread use is the ECOG/WHO/ZUBROD score

WHO/ECOG/ZUBROD Performance Status Scale

0 – Asymptomatic (Fully active, able to carry on all pre-disease activities without restriction)

1 – Symptomatic but completely ambulatory (Restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature. For example, light housework, office work)

2 – Symptomatic, <50% in bed during the day (Ambulatory and capable of all self care but unable to carry out any work activities. Up and about more than 50% of waking hours)

3 – Symptomatic, >50% in bed, but not bed-bound (Capable of only limited self-care, confined to bed or chair 50% or more of waking hours)

4 – Bed-bound (Completely disabled. Cannot carry on any self-care. Totally confined to bed or chair)

5 – Death

.

By Convention, “fit” patients have a low PS and in most chemotherapy trials, the predominant patient group included is that with PS 0 or 1, with a minority being PS 2 or greater (referred to as poor performance status and described separately in the section below). Furthermore, chemotherapy trials have usually only included patients with adequate organ function and excluded patients with medically unstable co-morbidities and uncontrolled brain metastases.

A large number of randomised controlled studies and subsequent meta-analyses have been reported addressing questions such as, which platinum agent is best (carboplatin versus cisplatin)?; which new agent paired with a platinum agent is best (often referred to as “third generation (3G)” regimens)?; is monotherapy with new (“3G”) agents as effective as platinum combination therapy?;are three chemotherapy agents (“triplet regimens”) better than two (“doublet regimens”)?; are non-platinum doublet chemotherapy regimens as effective as platinum doublet regimens?; what is the optimal duration of chemotherapy?; and is chemotherapy and a “biologic” or “targeted“ therapy superior to chemotherapy alone?

Recommendations

Evidence summary	Level	References
Platinum-based chemotherapy improves survival in stage IV NSCLC compared with best supportive care. Note that this evidence is based on clinical trials conducted in fit patients, with predominant performance status 0-1, no unstable co-morbidities, adequate organ function and without uncontrolled brain metastases.	I	,

Evidence-based recommendation	Grade
Platinum-based chemotherapy can be used to palliate disease symptoms and extend survival in newly diagnosed patients with stage IV NSCLC	A

Practice point
The decision to undertake empirical platinum-based chemotherapy in a given patient should consider factors such as patient performance status (0,1 versus 2 or more) and co-morbidities, their disease extent and symptoms, proposed treatment toxicity and their individual preferences for benefit from specific treatment(s) and toxicities.

Back to top

Carboplatin versus cisplatin

Three meta-analyses have addressed the question of whether carboplatin based chemotherapy is as effective as cisplatin based, which collectively confirm that cisplatin based regimens are associated with a slightly higher response rate than carboplatin regimens, with no definite survival difference.The first meta-analysis by Hotta et al, evaluated 2948 patients from eight randomised controlled trials (RCTs) from 1990-2004.. Cisplatin-based chemotherapy produced a higher response rate (RR), but overall survival (OS) was not significantly different. The second, by Ardizzoni et al, was an individual patient data meta-analysis of 2968 patients from nine RCTs from 1990 to 2004. This study found that objective RR was higher for patients treated with cisplatin than for patients treated with carboplatin (30% versus 24%, respectively; Odds ratio (OR) = 1.37; 95% CI = 1.16 to 1.61; P <.001). There was no overall difference in mortality, however, as in the Jiang meta-analysis, a subset analysis of survival in five trials evaluating “new” agents (gemcitabine, docetaxel, paclitaxel and vinorelbine) found OS with carboplatin slightly inferior to cisplatin (hazard ratio (HR) = 1.12; 95% CI = 1.01 to 1.23). Cisplatin-based chemotherapy was associated with more severe nausea and vomiting and nephrotoxicity; severe thrombocytopenia was more frequent during carboplatin-based chemotherapy. Jiang et al, evaluated published data from 6906 patients from 18 RCTs from 1190-2006. This study confirmed the findings of Hotta and Arziddoni with regard to RR in favour of cisplatin, however it did not find any survival difference in eight studies evaluating the new agents above.

The question of whether to use cisplatin versus carboplatin is of lower significance today especially given the new information arguing in favour of selecting specific treatments for greater benefit by histology and the presence of activating gene mutations.

Recommendations

Evidence summary	Level	References
First-line chemotherapy involving cisplatin results in a slightly higher likelihood of tumour response than the same chemotherapy with carboplatin.	I	, ,
There is no definite overall survival difference between cisplatin or carboplatin based first-line chemotherapy.	I	, ,
Cisplatin-based chemotherapy is associated with more severe nausea and vomiting and nephrotoxicity; severe thrombocytopenia is more frequent during carboplatin-based chemotherapy.	I	, ,

Evidence-based recommendation	Grade
In patients with high tumour burden and symptoms from stage IV NSCLC cisplatin based chemotherapy may be used in preference to carboplatin for the purpose of inducing a response, however, this benefit may be offset by its greater risk of toxicity.	B

Practice point
The choice of cisplatin versus carboplatin in a given patient may consider the balance between perceived benefit (in tumour response) versus known toxicity, whilst considering patient preferences.

Back to top

Which new agent/ platinum combination (“third generation (3G)” regimen) is best?

Several meta-analyses and numerous RCTS have evaluated this question either as their primary endpoint or as part of secondary analyses. New agents making up so – called “third generation” regimens include gemcitabine, vinorelbine, docetaxel, paclitaxel and irinotecan.

Baggstrom et al, meta-analysed results from twelve RCTs from 1994 – 2004 (n= 3995 patients) comparing RR and OS with 3G combination regimens including platinum-based compounds with second generation (2G) platinum-based regimens. The estimated absolute risk difference (RD) in RR in favour of 3G regimens was 12% (95% CI: 10 -15%), corresponding to a number need to treat (NNT) of eight for one patient to benefit. Owing to a high degree of heterogeneity across the studies, analysis of OS could not be undertaken.

Grossi et al, evaluated the relative impact of different 3G drugs (vinorelbine, gemcitabine, paclitaxel, docetaxel) on the activity of first-line chemotherapy in advanced NSCLC by considering RR and progressive disease (PD), in 45 RCTs (N = 11,867 patients). They found the odds of obtaining an objective response to treatment similar across the different regimens. Different rates of disease control were observed, with gemcitabine chemotherapy associated with a significant 14% lower risk for immediate progression, whereas patients receiving paclitaxel-based treatment appear to be at a higher risk for having PD as their best response. However, OS was not assessed in this meta-analysis.

Gao et al, examined whether platinum plus gemcitabine or vinorelbine are equally effective in the treatment of advanced NSCLC. This publication only meta-analysis evaluated nine RCTs involving 2186 patients, and found that no differences in RR or one-year OS. Vinorelbine plus platinum regimens led to more frequent grade 3 or 4 neutropenia, nephrotoxicity, constipationand phlebitiswhile gemcitabineplus platinum chemotherapy was associated with more grade 3 or 4 thrombocytopenia.

These meta-analyses collectively confirm better RR with 3G regimens compared with 2G but with differing toxicity profiles across the regimens and uncertainty or no difference in OS. A RCT of 1155 patients, evaluating four commonly used 3G platinum based regimens (vinorelbine, docetaxel, paclitaxel and gemcitabine) similarly failed to demonstrate superiority (in OS and RR) of one regimen over another although toxicity differences were observed.

In the setting of first-line empirical chemotherapy, the study by Scagliotti et al compared the effectiveness of cisplatin and pemetrexed to cisplatin and gemcitabine in a RCT of 1,725 patients. This study confirmed non-inferiority of cisplatin/pemetrexed compared with cisplatin/gemcitabine for the overall population, but also confirmed (in pre-planned analyses), superiority of cisplatin/pemetrexed for OS compared with cisplatin/gemcitabine in patients with non-SCC histology (HR 0.81, 95% CI 0.70 - 0.94), with median OS 12.6 versus 10.9 months for adenocarcinoma histology (n = 847, and 10.4 versus 6.7 months for large cell carcinoma (n = 153). Conversely, in patients with SCC, there was a significant improvement in survival with cisplatin/gemcitabine versus cisplatin/pemetrexed (n = 473; median OS 10.8 versus 9.4 months, respectively, HR 1.23 (95% CI 1.00 – 1.51, p = 0.05)). For cisplatin/pemetrexed, rates of grade 3/4 neutropenia, anemia, and thrombocytopenia (p = 0.001); febrile neutropenia (p = 0.002); and alopecia (p = 0.001) were significantly lower, whereas grade 3 or 4 nausea (p = 0 .004) was more common.

Gronberg et al compared carboplatin/pemetrexed to carboplatin/gemcitabine in a RCT of 436 patients with the primary endpoint of health-related quality of life. Compliance with completion of health-related QOL questionnaires was 87%. There were no significant differences for the primary health-related QOL endpoints, or in OS between the two treatment arms (pemetrexed/carboplatin, 7.3 months; gemcitabine/carboplatin, 7.0 months; P=0.63). Multivariate analyses and interaction tests did not reveal any significant associations between histology and survival. As in the Scagliotti study, rates of Grade ¾ haematologic toxicity were less with carboplatin/pemetrexed.

Recommendations

Evidence summary	Level	References
3G platinum-based chemotherapy (vinorelbine, paclitaxel, docetaxel or gemcitabine) is associated with higher response ratio than older 2G platinum-based chemotherapy.	I	, ,
No 3G platinum-based chemotherapy regimen (vinorelbine, paclitaxel, docetaxel or gemcitabine) has been shown to be superior to another.	I	, ,
In first-line empirical treatment of advanced NSCLC, chemotherapy with cisplatin and pemetrexed is superior to cisplatin/gemcitabine in patients with non-squamous cell carcinoma histology.	II
In first-line empirical treatment of advanced NSCLC, chemotherapy with cisplatin and pemetrexed is inferior to cisplatin/gemcitabine in patients with SCC histology.	II

Evidence-based recommendation	Grade
3G platinum-based chemotherapy (with vinorelbine, paclitaxel, docetaxel or gemcitabine) is a standard of care as first-line chemotherapy in fit patients with stage IV NSCLC.	A

Evidence-based recommendation	Grade
In the first-line setting, chemotherapy with cisplatin and pemetrexed is recommended in preference to cisplatin and gemcitabine in patients with non-squamous cell carcinoma histology.	B

Evidence-based recommendation	Grade
In the first-line setting, chemotherapy with cisplatin and gemcitabine is recommended in preference to cisplatin and pemetrexed in patients with squamous cell carcinoma histology.	B

Practice point
The choice of first-line platinum combination chemotherapy in a given patient may consider patient performance status and co-morbidities, the proposed treatment toxicity, treatment scheduling and individual patient preferences.

Back to top

Is monotherapy with new (3G) agents as effective as platinum combination therapy?

A meta-analysis by Hotta et al, examined the question of how treatment with single agent 3G agents (vinorelbine, paclitaxel, docetaxel, gemcitabine and irinotecan) compares with the same agent and a platinum agent. This meta-analysis evaluated 2374 patients from eight RCTs between 1994 – 2003. A greater than two-fold higher overall RR was seen with platinum combination than the new agent alone [odds ratio = 2.32; 95% CI 1.68–3.20]. Platinum-based doublet therapy was associated with a 13% prolongation of OS (HR = 0.87; 95% CI = 0.80–0.94,P <0.001). Despite significant increases in the frequencies of various toxicities in patients receiving platinum-based doublets, no significant difference in treatment-related mortality was observed.

Baggstrom et al in their meta-analsysis examined the effectiveness of 3G agents (vinorelbine, paclitaxel, docetaxel and gemcitabine) as first-line monotherapycompared with best supportive care in five RCTS of 1029 patients from 1996 – 2000. One trial used 5-fluorouracil (5FU)/leucovorin as the control arm. RR for the 3G regimens ranged from 12-20%. One-year survival favored the 3G agents over best supportive care with a summary absolute risk difference of 7% (95% CI: 2 - 12%). They calculated that the NNT for one patient to realise a benefit in the probability of one-year survival was 14.

Delbaldo et al examined the effectiveness of two-drug platinum combination chemotherapy compared with single agent therapy. This study evaluated 7175 patients from 29 RCTs but also included studies using older agents such as etoposide, vindesine and mitomycin C, as well as the modern 3G agents previously listed. Some of the studies included used a non-platinum combination in the comparator arm. Two-drug combination therapy was found to have a higher RR (OR, 0.42; 95% CI 0.37-0.47; p <.001). The absolute benefit was 13%, which corresponds to a two-fold increase in RR from 13% with a single-agent regimen to 26% with a doublet regimen. The benefit was higher when the control arm was an older drug (OR, 0.35) than when it was a newer drug (OR, 0.52) (P=.001). Two-drug combination therapy was associated with a significant increase in one-year survival (OR, 0.80; 95% CI, 0.70-0.91; P<.001) The absolute benefit was 5%, which corresponds to an increase in one-year survival from 30% with a single agent regimen to 35% with a doublet regimen. The benefit was higher when the control arm was an older drug than newer drug for both one-year survival rate (p=.03) and median survival (p=.007).

Recommendations

Evidence summary	Level	References
3G platinum-based combination chemotherapy (vinorelbine, paclitaxel, docetaxel, irinotecan or gemcitabine) is superior to 3G agent monotherapy.	I	,
3G platinum-based monotherapy (vinorelbine, paclitaxel, docetaxel, or gemcitabine) improves survival compared with best supportive care.	I

Evidence-based recommendation	Grade
Patients fit for chemotherapy should be offered 3G platinum-based combination chemotherapy (vinorelbine, paclitaxel, docetaxel, irinotecan or gemcitabine) in preference to 3G agent monotherapy, as it is more effective.	A

Evidence-based recommendation	Grade
Patients unfit for combination chemotherapy could be considered for 3G monotherapy with vinorelbine, paclitaxel, docetaxel or gemcitabine.	A

Back to top

Are three chemotherapy agents (“triplet regimens”) better than two (“doublet regimens”)?

Delbaldo et al also examined the effectiveness of three-drug combination chemotherapy compared with two-drug combination chemotherapy. This study evaluated 4814 patients from 28 RCTs.Adding a third drug to a doublet regimen was associated with a significantly increased RR (OR, 0.66; 95%CI, 0.58-0.75; p <.001). The absolute benefit was 8%, which corresponds to an increase in tumour RR from 23% (doublet regimen) to 31% (triplet regimen). There was no difference in RR whether the doublet regimens contained older or newer (3G) drugs (p=0.33). Adding a third drug to a doublet regimen did not improve one-year survival (OR, 1.01;95% CI, 0.85-1.21; P=0.88) and there was no significant difference according to the type of control regimens used (older drugs versus newer (3G) drugs) for both one-year survival rate (p =.28) and median survival (p =.36). However, grade ¾ toxicity was more common in triplet regimens than in doublet regimens with ORs ranging from 1.4 to 2.9,except for neurological, renal, auditory and gastrointestinal toxic effects.

Recommendations

Evidence summary	Level	References
Triplet chemotherapy regimens are associated with higher response rate, but no improvement in survival.	I
Triplet chemotherapy regimens are associated with greater grade 3 /4 toxicities.	I

Evidence-based recommendation	Grade
Triplet chemotherapy regimens are not recommended, as benefit in response rate does not outweigh extra toxicity.	A

Back to top

Are non-platinum doublet chemotherapy regimens as effective as platinum doublet regimens?

D’Addario et alevaluated this question in a meta-analysis of 7633 patients from 37 RCTs between 1983 and 2002. Platinum-based therapy was associated with a 62% increase in theodds ratio (OR) for RR (OR, 1.62; 95% CI,1.46 =1.8; P <.0001). The one-year OS was increased by 5% with platinum-basedregimens (34% versus 29%; OR, 1.21; 95% CI, 1.09 to 1.35; P _ .0003). However, no statistically significant increase in one-year survival was found when platinum therapies were compared to 3G –based combination regimens (OR, 1.11; 95% CI, 0.96 to 1.28; P _ .17). The toxicity of platinum-based regimens was significantly higher for hematologic toxicity, nephrotoxicity, and nausea and vomiting, but not for neurotoxicity, febrile neutropenia rate,or toxic death rate.

Rajeswaran et al also evaluated this question in a meta-analysis of 4920 patients from 17 RCTs. Platinum based doublet regimens were associated with a slightly higher one-year survival (RR = 1.08, 95% CI 1.01—1.16, p = 0.03), a greater response rate (RR = 1.11, 95% CI 1.02—1.21, p = 0.02), but with a higher risk of anemia, nausea, and neurotoxicity. Cisplatin-based doublet regimens improved one-year survival (RR = 1.16, 95% CI 1.06-1.27, p = 0.001), complete response. (RR = 2.29, 95% CI 1.08-4.88, p = 0.03), and partial response (RR = 1.19, 95% CI 1.07-1.32, p = 0.002), but with an increased risk of anaemia, neutropenia, neurotoxicity and nausea. Conversely, carboplatinbaseddoublet regimens did not increase one-year survival (RR = 0.95, 95% CI 0.85—1.07,p = 0.43). However, although carboplatin-based doublet regimens were associated with higher risk of anemia and thrombocytopenia, there was no increased nausea and/or vomiting.

Li et al compared the activity, efficacy, and toxicity of gemcitabine plus paclitaxel versus carboplatin plus either gemcitabine or paclitaxel in 2186 patients with untreated advanced NSCLC from four RCTSs. A significant difference in RR favouring gemcitabine plus paclitaxel over carboplatin-based doublets was observed [OR = 1.20; 95% CI 1.02–1.42; P = 0.03], whereas the trend toward an improved one-year OS was not significant (OR = 1.07; 95% CI = 0.91–1.26; P = 0.41). An increased risk of grade 3/4 toxicities for patients receiving carboplatin-based chemotherapy was demonstrated.

Recommendations

Evidence summary	Level	References
Platinum-based doublet 3G chemotherapy is associated with a higher response rate and slightly higher one-year survival than non-platinum doublet chemotherapy.	I	, ,
Platinum-based doublet 3G chemotherapy is associated with greater risk of anemia and thrombocytopenia than non-platinum combination therapy.	I	, ,
Gemcitabine and paclitaxel improves response ratio without added toxicity, compared with gemcitabine or paclitexel and carboplatin combinations.	I

Evidence-based recommendation	Grade
Non-platinum 3G doublet chemotherapy is an effective alternative option for patients unsuitable for platinum-based therapy.	B

Back to top

What is the optimal duration of first-line chemotherapy?

By convention, many clinical trials evaluating chemotherapy in stage IV NSCLC capped treatment to a maximum of six cycles, often being limited due to toxicity. Efficacy assessments usually occurred after the second or third chemotherapy cycle at 6-8 weekly intervals. Although several small RCTs have been conducted addressing the question of duration of treatment, there is a great deal of heterogeneity in the design of these studies in terms of the treatment regimens used, the scheduling and duration of chemotherapy being explored. Two systematic reviews have attempted to address the optimal duration of chemotherapy .

The study by Soon et al was designed to determine the effects of extending chemotherapy beyond a standard number of cycles. It evaluated 3,027 patients from 13 RCTs comparing a defined number of cycles with continuation of the same chemotherapy until disease progression, a larger defined number of cycles of identical chemotherapy, RCTs comparing a defined number of cycles of identical initial chemotherapy followed by additional cycles of an alternative chemotherapy.

The key findings were that extending chemotherapy appeared to significantly improve progression free survival (PFS; HR0.75; 95% CI: 0.69 -0.81; p < .00001) whereas the effect on OS was modest and less certain (HR, 0.92; 95% CI: 0.86 - 0.99; P < .03). Subgroup analysis revealed that the effects on PFS were greater for trials extending chemotherapy with 3G regimens rather than older regimens (P < .003). Extending chemotherapy was associated with more frequent adverse events in all trials where it was reported and impaired health related quality of life (QOL) in two of seven trials.

The study by Lima et al was designed to determine the effects of continuing first-line chemotherapy. It evaluated 1559 patients from 7 RCTs (included in the Soon meta-analysis) comparing different durations of first-line treatment of advanced NSCLC. Treatment for more than 4 cycles was not associated with a decrease in mortality relative to shorter treatment (HR = 0.97; 95% CI = 0.84 - 1.11; P = 0.65). Patients receiving more chemotherapy had significant longer progression-free survival (HR = .75; 95% CI = 0.60 – 0.85; P < 0.0001) than the group with shorter duration of treatment, but there was no difference in RR and longer treatment was associated with more severe leucopaenia, although non-haematological toxicities were not significantly increased.

The study by Lima et al more closely addressed the question of duration of first line chemotherapy, whereas the study by Soon et al, focused more on whether more chemotherapy is better than a fixed amount. It however contains a more heterogeneous mix of studies with a greater variety of regimens including regimens not in use (involving alkylating agents). However the overall study findings are not changed with the inclusion of these individual studies. Both studies agree in the finding that PFS is prolonged with longer chemotherapy however a consistent improvement in overall survival was not observed. Given the toxicity associated with standard first-line chemotherapy, it appears reasonable to stop after 4 cycles of treatment. Continuing the same first line treatment beyond this should be individually based and consider the evidence for continuation or switch maintenance therapy discussed in detail in the section below.

Recommendations

Evidence summary	Level	References
Extending the duration of first-line combination chemotherapy beyond four cycles of chemotherapy, in non-progressive patients, improves progression free survival but not overall survival, and at the expense of increased toxicity and potentially reduced quality of life.	I	,

Evidence-based recommendation	Grade
First-line combination chemotherapy should in most cases be stopped at disease progression or after four cycles in patients with advanced NSCLC.	B

Practice point
The duration of first-line chemotherapy in a given patient in practice may be based on the benefit being obtained in terms of tumour response, the desire to delay tumour progression and improve or maintain quality of life balanced against treatment toxicity. In practice maximum benefit from first-line chemotherapy has usually been obtained by four cycles of treatment.

Back to top

Is chemotherapy and a “biologic” or “targeted“ therapy superior to chemotherapy alone in unselected patients?

Numerous trials have been reported over the years exploring the benefit of adding novel drug therapy to standard chemotherapy. This section will review the evidence for benefit or lack thereof, of the addition of modern biologic or targeted therapy to standard first-line chemotherapy in “unselected patients”.

Back to top

Chemotherapy and AIs (anti-VEGF Mab (bevacizumab) or anti-VEGF TKIs)

There have been two phase III and one phase II RCT of chemotherapy +/- bevacizumab as first-line therapy in patients with stage IV NSCLC. The first study, a randomised phase II study by Johnston et al showed promising activity with bevacizumab but found an unexpectedly high incidence of pulmonary haemorrhage in patients with SCC. The study by Sandler et al examined carboplatin and paclitaxel +/- bevacizumab, whilst the study by Reck et al examined cisplatin and gemcitabine +/- bevacizumab. Consequently both subsequent PIII studies excluded patients with the following: SCC histologic type, brain metastases, clinically significant hemoptysis,inadequate organ function, ECOG PS of 1, therapeutic anticoagulation, clinically significant cardiovascular disease, or medically uncontrolled hypertension. The overall safety and efficacy of chemotherapy and bevacizumab has been summarised in a meta-analysis of four trials with 2101 patients by Yang et al. Bevacizumab has been studies at high dose (HD: 15 mg/kg) or low dose (LD: 7.5 mg/kg) every three weeks with chemotherapy.

Yang et al found that neither HD or LD bevacizumab improved one-year survival when added to chemotherapy. However, the addition of HD bevacizumab increased two-year OS (RR 1.24; 95% CI 1.04 – 1.49) and tumour response rate (RR 1.69; 95% CI 1.21-2.35). However in an independent systematic review by Botrel et al, although an OS benefit was observed with HD bevacizumab (HR 0.89, 95% CI 0.8 – 1.0, p =0.04), there was moderate statistical heterogeneity (Chi2 = 5.09, 3df, p = 017; I2 = 41%), making this finding less certain. PFS was improved with both LD bevacizumab (HR 0.76; 95%; CI 0.64-0.90) and HD bevacizumab (HR 0.73; 95%CI 0.65-0.81). However, HD bevacizumab was associated with an increase in treatment related deaths (RR 2.07, 95%; CI 1.19-3.59). Patients treated with HD bevacizumab experienced more hypertension, headaches, haemoptysis, neutropenia and rash than patients on chemotherapy alone. In the phase III trials bevacizumab was continued if tolerated until disease progression.

With regard to the small molecule TKIs, numerous phase III studies have been conducted but only one study has been published in full to date. Scagliotti et al, reported the outcomes of their phase III RCT evaluating the efficacy and safety of sorafenib, in combination with carboplatin and paclitaxel in chemotherapy-naïve patients. The study was terminated after the interim analysis concluded that the study was highly unlikely to meet its primary end point for OS. A pre-specified exploratory analysis revealed that patients with squamous cell histology had greater mortality in arm A than in arm B (HR 1.85; 95%; CI 1.22 to 2.81).

Recommendations

Evidence summary	Level	References
In carefully selected** patients with advanced NSCLC, high dose bevacizumab improves tumour response rate and progression free survival. **Patients with the following criteria were excluded from the trials: SCC histologic type, brain metastases, clinically significant haemoptysis,inadequate organ function, ECOG PS of 1, therapeutic anticoagulation, clinically significant cardiovascular disease, or medically uncontrolled hypertension.	I	,
In carefully selected** patients with advanced NSCLC, treatment with high dose bevacizumab is associated with an increase in treatment related deaths.	I

Evidence-based recommendation	Grade
High dose bevacizumab (15 mg/kg three-weekly) may be considered in addition to chemotherapy (carboplatin/paclitaxel or cisplatin/gemcitabine) in carefully selected** patients with non-squamous cell carcinoma.	B

Back to top

Chemotherapy and anti-EGFR TKIs

Following the discovery of the first generation EGFR TKIs gefitinib and erlotinib, four first-line placebo controlled RCTS were undertaken, evaluating the efficacy of the addition of these agents to two commonly used chemotherapy regimens (carboplatin/paclitaxel and cisplatin/gemcitabine) In all four trials the addition of the EGFR TKIs, gefitinib or erlotinib to a standard chemotherapy regimen did not improve outcomes (OS, RR or time to progression (TTP) compared with chemotherapy alone.

Evidence summary	Level	References
The addition of the EGFR TKIs gefitinib or erlotinib to a standard chemotherapy regimen does not improve outcomes (OS, RR or time to progression (TTP)) compared with chemotherapy alone.	II	, , ,

Evidence-based recommendation	Grade
The first generation EGFR TKIs gefitinib or erlotinib should not be used in unselected patients in combination with standard chemotherapy.	A

Back to top

Chemotherapy and anti-EGFR with the Mab cetuximab

The first monoclonal antibody to EGFR to enter the clinic was cetuximab. Two meta-analyses have summarised the evidence for the addition of cetuximab to standard chemotherapy, from four RCTs with 2018 patients with advanced NSCLC (selected by the presence of EGFR-positive tumor as measured by immunohistochemistry (IHC), two of which were phase III RCTs. Both meta-analsyses concur in finding that overall survival was improved by the addition of cetuximab to chemotherapy (HR 0.87; 95%CI, 0.79–0.96; p = 0.004) and overall response rate was increased (50% increase (odds ratio (OR) = 1.48; (CI = 1.22–1.80); p < 0.0001). PFS whilst improved with the addition of cetuximab to chemotherapy was not significantly better than chemotherapy alone (HR, 0.91; 95%CI, 0.83–1.00; p = 0.06). Of the two Phase III trials, only the Pirker study which added cetuximab to cisplatin/vinoerlbine was positive for survival, whilst the Lynch study, which added cetuximab to carboplatin/paclitaxel showed improved RR but not PFS or OS. The addition of cetuximab was associated with increased grade 3/4 rash and infusion reactions. In the phase III trials cetuximab was continued if tolerated until disease progression.

Recommendations

Evidence summary	Level	References
In patients with advanced NSCLC (selected by the presence of EGFR-positive tumour as measured by immunohistochemistry), the addition of cetuximab to chemotherapy increases response rate and improves overall survival. This overall benefit was modest and observed only in the phase III trial using cisplatin/vinorelbine .	I	,

Evidence-based recommendation	Grade
In patients with whose tumours have been shown to express EGFR by immunohistochemistry, cetuximab may be considered in addition to cisplatin/vinorelbine chemotherapy to improve response rate and overall survival.	B

Back to top

Different chemotherapy regimens and health related quality of life in NSCLC

Many of the aforementioned clinical trials have formally included patient rated QOL evaluation usually as a secondary endpoint. The overall effect of common chemotherapy regimens on health related QOL in NSCLC is probably best summarised in the meta-analysis by Tanvetyanon et al. This study identified 14 RCTs from 1998 – 2005 with 6665 patients to determine differences in QOL between the regimens studies. Of these, 13 trials using a validated QOL instrument were included for review. The meta-analsysis found QOL reporting/analysis techniques were heterogeneous. Nine RCTs reported the rate of completedbaseline assessment and compliance survivors at analysis of greaterthan 50%, for data synthesis. Of these, only one trial found a significant difference in QOL between the comparator arms: paclitaxelplus cisplatin was better than teniposide plus cisplatin. However, teniposide is not used in practice today. Based on this review, it seems unlikely that a major difference exists in the global QOL associated with standard chemotherapy regimens for advanced NSCLC. Furthermore, the authors concluded that although the available QOL reporting formats are largely acceptable, a lack of uniformity in analysis and a poor compliance to QOL assessment made between-trial comparisons difficult.

A large single RCT of 926 patients (not included in the Tanvetyanon meta-analysis) comparing docetaxel and cisplatin (DC) or carboplatin (DCb) with cisplatin /vinorelbine (VC) also examined QOL using the Lung Cancer Symptom Scale (LCSS) and the general EuroQol five-dimensional questionnaire (EQ-5D). DCband DC were superior to VC in the QoL outcomes assessed except for the difference between DC and VC in LCSS ‘‘QOL today’’, which was not significant.

There does not appear to be any major difference evident in the global quality of life associated with standard chemotherapy regimens for advanced NSCLC.

Practice point
As overall quality of life does not seem to differ across the different chemotherapy regimens, the choice of chemotherapy in an individual patient may involve discussion regarding expected toxicities and the patient’s preferences.

Back to top

First-line Maintenance therapy

"Maintenance" therapy, described in detail here refers to the concept of continuing drug therapy until progression. Maintenance treatment can be further characterised by continuing part of the initial treatment regimen (usually the new generation regimen whilst stopping the platinum) - this is referred to as "continuation maintenance" or by switching after disease control with an initial combination therapy to another agent - this is referred to as "switch maintenance".

There have been nine studies conducted formally addressing the question of maintenance therapy in advanced NSCLC, eight of which were evaluated in a systematic review by Zhang et al, with the most recent RCT of maintenance pemetrexed after induction pemetrexed/platinum therapy reported in 2011 but not yet published. Zhang et al undertook a systematic review of eight trials (3, 736 patients) investigating maintenance therapy with either a continuous or a switch strategy for patients with non-progressing NSCLC compared with placebo or observation with OS as the primary outcome. Another study not included in the meta-analysis by Zhang evaluated the role of maintenance carboxyaminotriazole (CAI) in patients with advanced NSCLC with non-progression after initial chemotherapy. CAI was a novel but unproven therapy, demonstrated to modulate tumour cell motility, adhesion and angiogenesis. Unfortunately no benefit was observed compared to placebo.

Three studies were identified evaluating continuation maintenance with gemcitabine whilst the switch maintenance studies included in this meta-analysis evaluated either docetaxel, pemetrexed, erlotinib, erlotinib and bevacizumab, or gefitinib. Zhang et al found that the concept of switch maintenance therapy substantially improved OS compared with placebo or observation (hazard ratio [HR], 0.85; 95% CI, 0.79-0.92; P , .001). However, gemcitabine as continuous maintenance therapy, did not statistically improve survival (HR, 0.88; 95% CI, 0.74-1.04; P = 0.124). Subgroup analyses showed benefits of switch maintenance therapy with both cytotoxic agents (docetaxel/pemetrexed studies combined) (HR, 0.80; 95% CI, 0.69-0.93; P 5 .003) and EGFR TKI-targeted agents (gefitinib/erlotinib studies combined) (HR, 0.87; 95% CI, 0.80-0.95; P 5 .001). Clinically and statistically significant improvement in PFS was found with both maintenance strategies (switch maintenance therapy HR, 0.67; 95% CI, 0.57-0.78; continuous maintenance therapy HR, 0.53; 95% CI, 0.43-0.65; interaction P = 0.128).

The meta-analysis by Zhang et al thus confirmed in principle that there is benefit from the concept of maintenance therapy in NSCLC, with OS benefit observed in switch maintenance approach and PFS benefit from both switch and continuous maintenance approaches. Of the studies included in the meta-analysis by Zhang et al, the three positive published studies are of switch maintenance using docetaxel, pemetrexed or erlotinib. These are discussed in more detail below to determine the benefits and harms of each approach for the purposes of providing a more specific practice guideline.

The first study by Fidias et al, compared immediate with delayed docetaxel after first-line therapy with gemcitabine and carboplatin in 309 patients with advanced NSCLC. Patients were randomised to immediate versus delayed docetaxel if they were stable after four cycles of induction chemotherapy. The primary endpoint of OS was not significantly different (p = 0.085) even though there was a trend in median OS in favour of immediate docetaxel (12.3 months versus 9.7 months (delayed)). However, PFS was significantly longer for immediate docetaxel (5.7 months) than for delayed docetaxel (2.7 months) (P .0001), and QOL results were not statistically different (P= 0.76) between docetaxel groups.

The second study by Ciuleanu et al, randomised 663 patients who had not progressed after four cycles of first-line platinum-based chemotherapy to pemetrexed or placebo. The primary endpoint of the study was PFS. Pemetrexed was shown to significantly improve PFS (4.3 months [95% CI 4.1–4.7] versus 2.6 months [1.7–2.8]; HR 0.50, 95% CI 0.42–0.61, p<0.0001) and also improve OS (13.4 months [11.9–15.9] versus 10.6 months [8.7–12.0]; HR 0.79, 0.65–0.95, p=0.012) compared with placebo. These benefits were mainly in patients with non- SCC histology (PFS HR 0.44; (95% CI 0.36–0.55); and OS HR 0.70; (95% CI 0.56–0.88). compared with squamous histology (PFS HR 0・69, 95% CI 0.49–0.98); and OS HR 1.07, (95% CI 0.77–1.50). In the non-SCC population, the benefit with pemetrexed was most certain in the adenocarcinoma (PFS HR 0.51 (0.38–0.68); <0.0001, and OS HR 0.73 (0.56–0.96); 0.026) and other NSCLC sub-groups. This observation was confirmed by a significant treatment-by-histology interaction with both PFS (p=0・036) and OS (p=0・033). Drug-related grade three or higher toxic effects were higher with pemetrexed than with placebo (mainly fatigue and neutropenia).

The final study, by Capuzzo et al, randomised 884 patients who had not progressed after four cycles of first-line platinum-based chemotherapy were randomised to erlotinib 150 mg daily or placebo. The co-primary endpoints were PFS in all patients irrespective of EGFR status, and PFS in patients whose tumours had EGFR protein over-expression, as determined by IHC. Median PFS was signifi cantly longer with erlotinib than with placebo: 12.3 weeks (erlotinib) versus 11.1 weeks (placebo) group (HR 0.71, 95% CI 0.62–0.82; p<0.0001). PFS was also significantly longer in patients who were EGFR IHC positive and treated with erlotinib (n=307) compared with EGFR-positive patients given placebo (n=311; median PFS 12.3 weeks (erlotinib) versus 11.1 weeks (placebo); HR 0.69, 0.58–0.82; p<0.0001). A pre-planned analysis of PFS in patients with EGFR-activating mutations confirmed a substantial benefit (HR 0.10, 95% CI 0.04 – 0.25; p<0.0001; ) but also benefit, albeit more modest, in patients with wild-type EGFR (HR 0・78, 95% CI 0.63–0.96; p=0.0185). Overall survival was significantly prolonged with erlotinib versus placebo in the intention-to-treat population (median 12.0 versus 11.0 months; HR 0.81, 95% CI 0.70–0.95; p=0.0088). The commonest grade 3 or higher toxicities associated with erlotinib were rash (9%) and diarrhea (2%).

Taken collectively and allowing for differences in study design, all three studies indicated benefit in unselected patients from this “switch maintenance” approach of immediate alternative treatment with single agent docetaxel, pemetrexed or erlotinib, by significantly delaying PFS, and in the cases of erlotinib and pemetrexed in non-SCC histology, also improving OS. In each study, the treatment switch was evaluated only in patients with stable disease or response after four cycles of standard first-line chemotherapy. In the case of pemetrexed, the benefit appears to be in the non-SCC histology sub-group.

Recommendations

Evidence summary	Level	References
In patients with stable or responsive advanced NSCLC after initial platinum doublet chemotherapy, the principle of switch maintenance therapy to either chemotherapy or anti-EGFR TKI improves overall survival.	I
In patients with stable or responsive advanced NSCLC after four cycles of initial platinum doublet chemotherapy, both approaches of switch maintenance and continuation maintenance improves progression free survival.	I
In patients with stable or responsive advanced NSCLC after four cycles of initial carboplatin/gemcitabine chemotherapy, immediate docetaxel prolongs progression free survival compared with delaying treatment for relapse, without decreasing quality of life.	II
In patients with stable or responsive advanced NSCLC after four cycles of initial platinum doublet chemotherapy, in patients with non-SCC histology, “switch maintenance” chemotherapy with pemetrexed improves progression free survival and overall survival.	II
In patients with stable or responsive advanced NSCLC after four cycles of initial platinum doublet chemotherapy, “switch maintenance” therapy with erlotinib improves progression free survival and overall survival.	II

Evidence-based recommendation	Grade
In unselected patients with stable or responsive advanced NSCLC after four cycles of initial platinum doublet chemotherapy, “switch maintenance” therapy to an alternative agent is recommended to delay tumour progression. Options in unselected patients include erlotinib or docetaxel, whilst in patients with non-squamous cell carcinoma histology, pemetrexed.	A

Back to top

Second-line therapy

Monotherapy in unselected patients

Several RCTS have been reported examining the role of second line systemic therapy in unselected patients. The first studies examined docetaxel, establishing it as a standard of care in suitably fit patients. Subsequent studies examined different schedules of docetaxel, or examined the efficacy of new agents using it as the reference standard.

In 2000, two key RCTs were reported evaluating the efficacy of single agent docetaxel in previously treated NSCLC. Shepherd et al evaluated the efficacy of docetaxel versus best supportive care in 104 patients previously treated with platinum-based chemotherapy. Compared with best supportive care, docetaxel 75 mg/m2 Q three-weekly, improved one-year survival (37% versus 11%; P = 0 .003). Fossella et al randomised 373 previously treated patients with advanced NSCLC to two dose regimens of docetaxel compared with control arm of vinorelbine or ifosfamide. one-year survival was significantly greater with docetaxel 75 mg/m2 than with the control treatment (32% versus 19%; P = 0.025,). Based on these two studies, docetaxel became the standard of care as second-line treatment of advanced NSCLC. Further supporting the clinical value of docetaxel was the results of the QOL analysis in the Shepherd study, which indicated less deterioration in QOL for docetaxel treated patients compared with best supportive care.

Bria et al, compared the efficacy of weekly docetaxel with the reference standard of three-weekly, by evaluating data from 1018 patients from six RCTs. No significant differences in OS or RR in favour of the weekly schedule were found, however weekly docetaxel was associated with fewer grade ¾ neutropenic events.

Hanna et al, then compared single agent pemetrexed to three-weekly docetaxel as second line monotherapy of advanced NSCLC. This study of 571 patients, randomised to three-weekly pemetrexed or docetaxel, showed equivalent efficacy outcomes (PFS, one-year survival) but significantly fewer side effects in favour of pemetrexed. Consequently, pemetrexed was soon registered as an alternative second-line agent in NSCLC. Scagliotti et al in a post hoc analysis of data from two RCTS of pemetrexed, subsequently showed that pemetrexed increased OS in patients with non-SCC histology (p = 0.047), whereas OS was decreased with pemetrexed in SCC histology (p = 0.018). A subsequent systematic review has confirmed this treatment-by-histology interaction effect with pemetrexed treatment showing greatest benefit in non-SCC histology.

Two studies evaluated the effectiveness of the single agent first generation EGFR TKIs gefitinib or erlotinib compared with placebo in previously treated patients with NSCLC. Thatcher et al, reported the effect of gefitinib as second or third-line therapy in 1692 patients with NSCLC refractory to or intolerant of previous treatment (ISEL study). Median survival was not significantly different between gefitinib and placebo treated patients in the overall population or the pre-specified adenocarcinoma subgroup. However, in pre-planned subgroup analyses, OS was longer with gefitinib in never-smokers (n = 375, OS HR 0.67 (95% CI 0.49-0.92), p = 0.012) and patients of Asian origin (n = 342, OS HR 0.66 (95% CI 0.48-0.91)). This sub-group effect most likely can be attributed to the greater incidence of activating EGFR gene mutations in this population.

Shepherd et al, randomised 731 patients, previously treated for advanced NSCLC, to receive erlotinib 150 mg daily or placebo as second or third-line treatment (BR21 study). A RR of 8.9 % was observed with erlotinib, which was shown to prolong OS (median 6.7 months (erlotinib) versus 4.7 months (placebo); HR 0.70, p <0.001). More patients receiving erlotinib had improvements in cough, pain, dyspnoea and in the overall physical function domain of QOL.

Kim et al randomised 1433 patients previously treated for advanced NSCLC to receive gefitinib 250 mg daily or three-weekly docetaxel chemotherapy (INTEREST study). The primary objective was to compare OS and to assess non-inferiority of gefitinib in the overall population and superiority in patients with high EGFR gene copy number. Non- inferiority of gefitinib compared with docetaxel was confirmed for OS (HR 1.02, 95% CI 0.905 – 1.15). Superiority of gefitinib in patients with high EGFR gene copy number was not proven. Skin rash and diarrhea were more common with gefitinib, whilst neutropenia, asthenia and alopecia were more common with docetaxel.

Ciuleanu et al, randomised patients that progressed after first line platinum doublet chemotherapy to recive either erlotinib 150 mg daily or chemotherapy (pemetrexed or docetaxel by investigator choice). This study (TITAN) was originally designed to test for superiority of erlotinib versus chemotherapy, in patients progressing after the induction chemotherapy phase of the switch maintenance erlotinib study by Capuzzo et al. However when the Capuzzo study closed, so to did recruitement to the TITAN study. Nonetheless, for the 424 patients randomised, median overall survival was 5.3 months (95% CI 4.0–6.0) with erlotinib and 5.5 months (4.4–7.1) with chemotherapy (HR 0.96, 95% CI 0.78–1.19; log-rank p=0.73). The adverse-event profile of each group was in line with previous studies, with more skin rash and diarrhea with erlotinib, and alopecia associated with chemotherapy (mainly due to docetaxel).

Recommendations

Evidence summary	Level	References
In previously treated patients with advanced NSCLC, single agent docetaxel 75 mg/m2 improves survival compared with best supportive care or vinorelbine and ifosfamide.	II	,
In previously treated patients with advanced NSCLC, single agent pemetrexed has similar efficacy but fewer side effects than three-weekly docetaxel.	II
In previously treated patients with advanced NSCLC, compared with docetaxel, pemetrexed appears to have greater efficacy in non-squamous cell carcinoma histology, and inferior efficacy in squamous cell carcinoma.	I

Evidence-based recommendation	Grade
In unselected patients previously treated for advanced NSCLC, chemotherapy with docetaxel or pemetrexed may be used as second-line therapy. Pemetrexed is preferred in non-squamous cell carcinoma histology, and docetaxel is preferred in squamous cell carcinoma.	B

Evidence summary	Level	References
In unselected previously treated patients with advanced NSCLC single agent erlotinib150 mg per day orally as second-line therapy improves survival compared with placebo.	II
In unselected previously treated patients with advanced NSCLC, single agent gefitinib 250 mg per day orally does not improve survival compared with placebo.	II
In unselected previously treated patients with advanced NSCLC, gefitinib 250 mg per day orally is equivalent to three-weekly docetaxel chemotherapy.	II
In unselected patients with advanced NSCLC, progressing after first-line platinum-based chemotherapy, there is no difference in survival between erlotinib 150 mg daily or chemotherapy (either pemetrexed or docetaxel).	II

Evidence-based recommendation	Grade
In unselected patients previously treated for advanced NSCLC, gefitinib 250 mg per day orally or erlotinib 150 mg per day orally can be used as second-line therapy, instead of chemotherapy.	B

Back to top

Combination therapy in unselected patients

Di Maio et al, examined whether doublet chemotherapy is more effective than single agent chemotherapy as second-line treatment of advanced NSCLC in 847 patients from six RCTS from 1999 – 2005. Single agents evaluated include docetaxel (three studies), irinotecan, cisplatin, or pemetrexed. Response rate was greater for doublet therapy (15 % versus 7.3 %, p = 0.0004), as was PFS (HR 0.79, 95% CI 0.68 – 0.91). However, there was no significant difference in OS between single agent and doublet chemotherapy and there were significantly more grade ¾ haematologic and non-haematologic toxicities with doublet chemotherapy.

Qi et al, examined whether doublet pemetrexed based therapy is more effective than single agent pemetrexed as second-line treatment of advanced NSCLC in 1,186 patients from five RCTS from 1999 – 2005. Only one of these studies was a phase III RCT, that of the dual targeted TKI vandetanib (anti-VEGF and anti EGFR). Here doublet therapy was associated with a greater RR, but did not improve PFS ). The other four phase II RCTS evaluated the addition of carboplatin, and the new agents enzastorurin, matuzumab and bortezomib to pemetrexed. Overall, there was improvement in RR and PFS with doublet therapy but not survival. Furthermore, there was more grade ¾ neutropenia and thrombocytopenia with the doublet therapy.

Herbst et al, also evaluated the efficacy of vandetanib. In their double blind RCT, the effect of Vandetanib plus docetaxel was compared with docetaxel as second-line treatment for patients with advanced NSCLC, on PFS in 1391 patients. Vandetanib plus docetaxel was shown to be an active regimen with significant improvement in PFS versus placebo plus docetaxel (HR 0.79, 97.58% CI 0.70–0.90; p<0.0001)., however, the size of the effect on median PFS was small (4.0 months (vandetanib) versus 3.2 months (placebo), and therefore of questionable clinical significance, and survival benefit not shown.

Recommendations

Evidence summary	Level	References
Doublet therapy as second-line treatment of advanced NSCLC increases response ratio and performance status, but is more toxic and does not improve survival compared with single agent chemotherapy.	I	,

Evidence-based recommendation	Grade
Doublet therapy is not recommended as second-line treatment of advanced NSCLC .	A

Back to top

Third-line therapy

Few RTCs have evaluated third line therapy in unselected patients with advanced NSCLC. The aforementioned negative RCT (ISEL) of gefitinib v placebo in 1692 patients included 847 patients (50%) that had received two previous lines of therapy. The positive RCT (BR21) of erlotinib versus placebo in 731 patients included approximately 50% of patients having received two previous lines of therapy. Univariate analysis of OS by number of prior regimens found OS remained in favour of erlotinib (compared with placebo) by similar magnitude to the overall study population results (HR 0.80, p = 0.02). The study by Kim et al, comparing gefitinib to docetaxel in previously treated advanced NSCLC, only included 235 (16%) patients that had received two previous lines of therapy. Analysis of OS number of prior regimens found OS more in favour of docetaxel. But as this is a post hoc analysis with small patient numbers, it is not appropriate to draw conclusions.

Recommendations

Evidence summary	Level	References
In unselected previously treated patients with advanced NSCLC who have received two lines of therapy, single agent erlotinib 150 mg per day orally as third-line therapy improves survival compared with placebo.	II

Evidence-based recommendation	Grade
In unselected patients having previously received two lines of treatment for advanced NSCLC, erlotinib 150 mg per day orally can be used as third- line therapy.	B

Back to top

Systemic therapy of specific patient clinical sub-groups

Poor performance status patients

Most studies with cytotoxic chemotherapy have been evaluated in “fit” patients, predominantly with PS 0 or 1. Patients with PS 2 are generally considered a poor prognostic group and at higher risk of toxicity, particularly from cytotoxic chemotherapy. Attempts to improve outcomes in this poor performance group population (PS 2) of patients with advanced NSCLC have been challenging with trials focused on the use of less toxic regimes or monotherapy with 3G agents or anti-EGFR TKIs.

Liu et al undertook a systematic review of phase II and II studies to examine the safety and efficacy of EGFR TKI monotherapy versus single-agent chemotherapy using third-generation cytotoxics as first-line treatment for patients with advanced non-small cell lung cancer and poor performance status. No RCTs were identified. Fifteen single arm phase II studies (1425 patients) were evaluated to determine pooled estimates for RR and safety. The pooled RR (95% CI) to EGFR TKIs for unselected populations was 6%(3–8%), which compares with 9% (6–13%) reported by single-agent 3G chemotherapy trials. By summary comparison only, toxicity profiles were more favourable for the EGFR TKIs than chemotherapy. This study confirms the feasibility of treatment in the poor PS population but does not provide information on the overall benefit of such treatment.

Baggstrom et al reported a meta-analysis of five trials (n =1029 patients) compared 3G single agents with BSC. Four of the trials included a BSC control arm, and one trial included 5-fluorouracil (5FU)/ leucovorin as the control arm. Response rates for the 3G agents ranged from 12% to 20%. One-year survival favored the 3G agents over BSC with risk difference of 7% (95% CI: 2% to 12%). The number needed to treat for one patient to realize a benefit in the probability of one-year survival was 14. These five trials evaluated single agent vinorelbine, paclitaxel, docetaxel and gemcitabine. The study by Crawford et al of single agent vinorelbine included 50% of patients with low PS, the vinoerlbine study by Gridelli et al in patients over 70 included 24% of patients with PS 2, the paclitaxel study by Ranson et al included 15% PS 2 patients, the docetaxel study by Roszkowski et al, included 20% PS 2 patients whilst the gemcitabine study by Anderson et al was mainly in low PS patients. The study by Anderson et al of gemcitabine versus best supportive care evaluated QOL as its primary endpoint and confirmed better QOL and reduced disease-related symptoms compared with those receiving best supportive care alone, although breathlessness was least well palliated and OS was no different. Quality of life was also in favour of paclitaxel, docetaxel and vinorelbine (versus best supportive care) in the respective studies.

In the second-line setting, several of the key RCTs that evaluated the efficacy of EGFR TKIs have included PS 2 or greater patients. Both the placebo controlled trials of gefitinib and erlotinib enrolled > 30 % of patients with PS 2, whilst the study by Kim et al comparing gefitinib to docetaxel included 11% of PS 2 patients. In the BR21 study, analysis of benefit by the PS 2 and 3 subgroups that received erlotinib versus placebo demonstrated a benefit in OS (HR 0.8; 95% CI 05-1.1 (PS 2); 0.4-1.3 (PS 3)), which compares with OS HR 0.7 for the overall population. (0.6-0.9). Thatcher et al, demonstrated the direction of benefit to be in favour of gefitinib over placebo in the OS analysis by sub-populations (30% of patients with PS2).. In the small PS2 sub-population in the study by Kim et al comparing gefitinib with docetaxel, the direction of benefit favoured gefitinib but the confidence limits were wide. Overall. confident conclusions cannot be made for benefit from gefitinib in unselected PS 2 or more patients. However, given the magnitude of benefit observed with gefitinib in first line patients with activating EGFR gene mutations (GMT+, ,described in the section below), it would be reasonable to expect that EGFR GMT + "selected" patients may still potentially benefit from an EGFR TKI , even if of poor performance status, given the size of the observed benefit and relatively low toxicity.

Recommendations

Evidence summary	Level	References
In patients with poor performance status (PS 2), first-line monotherapy with 3G chemotherapy (vinorelbine, gemcitabine, paclitaxel or docetaxel) may improve survival and/or quality of life.	I, II	, , , , ,

Evidence-based recommendation	Grade
First-line monotherapy with 3G chemotherapy could be offered to selected patients with PS2 for symptom improvement and possible survival gain, who are willing to accept treatment toxicity.	B

Evidence summary	Level	References
There is evidence for benefit with erlotinib 150 mg daily as second or third-line therapy in unselected poor performance status patients (PS2 or 3) .	II

Evidence-based recommendation	Grade
Poor performance status patients having received 1 or 2 lines of prior therapy, may be offered erlotinib 150 mg daily.	B

Practice point
Decision-making on treatment in poor performance status patients may weigh up benefits against toxicity and patient preferences. Whilst a single agent 3G chemotherapy is an option in unselected patients, patients with known activating EGFR MTs should be considered for first line EGFR TKIs as the magnitude of benefit is greater and toxicity profile more favourable.

Back to top

The Elderly

The age criterion for designation of “elderly” has varied somewhat across NSCLC studies with the elderly groups commonly defined as those patients either 65 or 70 years of age or older. Several RCTs have been conducted within this subgroup. As a group elderly patients are considered at higher risk of treatment related toxicity, due to possible age physiologic effects on drug handling and high proportion of co-morbidities. Gridelli et al first reported findings to indicate benefit from monotherapy with vinorelbine in patients over 70, with improvement seen in OS 0.65 (95% CI = 0.45–0.93) and fewer reported lung cancer related symptoms in a RCT of 161 patients Kudoh et al, subsequently compared docetaxel 60 mg/m2 (day one) to vinorelbine 25 mg/m2 (days one and eight) every 21 days for four cycles, in a RCT of 182 Japanese patients over 70 years of age. There was no statistical difference in the primary endpoint of median OS with docetaxel versus vinorelbine (14.3 months versus 9.9 months; HR 0.780; 95% CI 0.561 - 1.085; P = 0.138). However, median PFS (5.5 months versus 3.1 months; P = 0.001), RR (22.7% versus 9.9%; P = 0.019) and disease-related symptoms favoured docetaxel over vinorelbine (odds ratio, 1.86; 95% CI, 1.09 - 3.20). Docetaxel was associated with more grade 3/4 neutropenia (82.9% for docetaxel; 69.2% for vinorelbine; P = 0.031).

Hainsworth et al, randomised 350 patients over 65 years of age to first line single-agent weekly docetaxel versus the combination of docetaxel and gemcitabine. There was no difference in OS with the combination treatment compared with single agent weekly docetaxel. Russo et al reported a literature-based meta-analysis of RCTs that compared a gemcitabine based doublet regimen with a 3G single agent in elderly patients (> 65). This meta-analysis included the study by Hainsworth et al. Four trials evaluating 1436 patients were included in the meta-analysis. A significant difference in RR was seen favoring gemcitabine doublet therapy over single 3G agents (OR 0.65; 95% CI 0.51-0.82, p < .001), whereas one-year survival rate was not significantly different (OR, 0.78; 95% CI, 0.57-1.06, P = 0.169). Only Grade ¾ thrombocytopenia was greater with combination therapy (OR, 1.76; 95% CI, 1.12-2.76, P= 0.014).

More recently, Quoix et al reported findings from a RCT of that compared a carboplatin and paclitaxel doublet chemotherapy regimen with 3G monotherapy in 451 elderly patients (age 70-89) with advanced NSCLC. Patients were treated with carboplatin AUC 6 on day one and 90 mg/m. paclitaxel on days 1, 8, and 15 Q4 weekly or 3G monotherapy with either 25 mg/m2. vinorelbine on days one and eight or 1150 mg/m2 gemcitabine on days one and eight, Q3 weekly. Overall survival was in favour of the combination (median 10.3 months for doublet chemotherapy versus 6.2 months for 3G monotherapy (HR 0.64, 95% CI 0.52–0.78; p<0.0001)). Toxicity was more frequent in the doublet chemotherapy group than in the monotherapy group (neutropenia (48.4% vs 12.4%); asthenia (10.3% versus 5.8%)

Recommendations

Evidence summary	Level	References
First-line single agent vinorelbine (30 mg/m2 on days one and eight, Q3 weekly) in patients over 70 years of age improves survival and reduces disease related symptoms.	II
In patients over 70 years of age, first line single agent docetaxel 60 mg/m2 (day one) compared to vinorelbine 25 mg/m2 (days one and eight) every 21 days, improves RR, PFS and disease related symptoms, but not overall survival and is associated with more G3/4 neutropenia.	II
In patients over 65 years of age, gemcitabine doublet chemotherapy improves response rate compared with single agent 3G chemotherapy but does not improve survival and is associated with greater thrombocytopenia.	I
In patients over 70 years of age, first-line carboplatin/weekly paclitaxel combination improves survival compared with 3G monotherapy (weekly vinorelbine or gemcitabine) but, is associated with more neutropenia.	II

Evidence-based recommendation	Grade
Suitably fit patients over 65 years of age, can be offered first-line mono-chemotherapy with a 3G single agent (vinorelbine (25-30 mg/ m2 Day one, eight Q3 weekly), docetaxel (60 mg/m2 (day one, Q3 weekly) or gemcitabine (1150 mg/m2 days one and eight, Q3 weekly).	B

Evidence-based recommendation	Grade
In elderly patients, first-line gemcitabine doublet chemotherapy is not recommended.	B

Evidence-based recommendation	Grade
In fit elderly patients, first-line carboplatin/weekly paclitaxel may be offered instead of 3G monotherapy but at the expense of greater neutropenia.	B

Back to top

Therapy in selected patients, including “targeted” therapy

Selection by histology

Stanfield et al conducted a systematic review of prospective, RCTs to examine whether histology had a treatment modifying effect (TME) on the efficacy outcomes (OS and PFS) of chemotherapeutic agents in patients with advanced NSCLC. A total of 17 systematic reviews, five individual patient data (IPD) meta-analyses, and 165 potentially relevant primary studies were identified for full review. Four of the five IPD meta-analyses investigated TME of histology and one did not, but none found a significant TME by histology. One hundred and twenty two (74%) of the 165 primary publications retrieved for full review did not report data in a way in which the TME of histology could be determined. Data from three pemetrexed RCTs,comparing (i) second-line pemetrexed versus docetaxel, (ii) first-line pemetrexed and cisplatin versus gemcitabine and cisplatin, and (iii) switch maintenance pemetrexed versus placebo, showed a statistically significant TME by histology for OS and PFS.

A fourth RCT comparing pemetrexed and carboplatin versus gemcitabine and carboplatin found no significant association between histology and OS. Patients with non-SCC appear to gain the greatest benefit from treatment with pemetrexed, whilst patients with SCC appear to have poorer OS when pemetrexed is compared with other active treatments, and similar OS when compared with placebo. A reproducible pattern of TME effect by histology was not seen clearly with other chemotherapeutic agents.

Histology has also been shown to be a predictor for toxicity with the anti-VEGF Mab, bevacizumab, with higher incidence of pulmonary haemorrhage observed in SCC. Is histology also associated with a treatment modifying effect with bevacizumab? Sandler et al, in a post hoc analysis of their pivotal phase III RCT of first-line carboplatin/paclitaxel (PC) +/- bevacizumab (PCB) study in 878 carefully selected patients with non-SCC, reported their findings by histologic subgroups. The largest histologic subgroup in the study was adenocarcinoma (68.8% of patients), whilst not-otherwise specified represented 18.9% of patients. For adenocarcinoma, median OS was 10.3 months for PC treatment (n= 302) and 14.2 months for PCB (n = 300), HR 0.69 (95%CI: 0.58–0.83). Sample sizes for other specific histologic subtypes were considered too small for meaningful comparisons.

The TME of histology in predicting benefit from pemetrexed, the observation of greater toxicity with bevacizumab and possibly other anti-VEGF therapies in SCC, and the finding of activating EGFR gene mutations (EGFR GMTs and other mutations) in adenocarcinomas has led to a great clinical need for diagnostic accuracy in the sub-classification of NSCLC on diagnostic specimens. Consequently, the International Association for the Study of Lung Cancer (IASLC) undertook a systematic literature review of the adenocarcinoma histologic classification. In their review, Travis et al describe a revised classification system for diagnosing and reporting NSCLC with guidance for small biopsies, cytology and resected specimens, to enable classifying NSCLC primarily into adenocarcinoma or SCC due to the therapeutic implications of this distinction.

Recommendations

Evidence summary	Level	References
Histology (non-squamous cell carcinoma versus squamous cell carcinoma) is associated with a significant treatment modifying effect for patients treated with pemetrexed based chemotherapy, with superior survival effect of pemetrexed observed in non-squamous cell carcinoma histology and inferior survival effect observed in squamous cell carcinoma histology, compared with other standard regimens when pemetrexed is used first-line, as switch maintenance or as second-line treatment.	I

Evidence-based recommendation	Grade
Due to the therapeutic implications, it is important to classify the histologic subtype of NSCLC on diagnostic specimens as accurately as possible, particularly to enable accurate distinction between the key histologic subtypes: adenocarcinoma and squamous cell carcinoma.	A

Practice point
Given the importance of accurate histologic diagnosis and the potential need to have sufficient tissue for subsequent molecular testing, it is important to obtain as much tissue as possible at initial diagnosis in patients suspected to have NSCLC. A multidisciplinary team discussion may be required in order to decide on the most appropriate diagnostic method to obtain adequate tissue.

Back to top

Selection by clinical phenotype

The early single arm and RCTs evaluating the first generation EGFR TKIs gefitinib and erlotinib identified that benefit from EGFR TKIs appeared to be greatest in certain NSCLC patient sub-populations: never smokers with adenocarcinoma, and especially, but not exclusively, in women,and Asian background.

Consequently, Mok et al, undertook a first-line RCT to compare gefitinib versus carboplatin/paclitaxel chemotherapy. They randomly assigned previously untreated patientsin East Asia who had advanced pulmonary adenocarcinoma and who were nonsmokers or former light smokers to receive gefitinib (250 mg per day) or carboplatin/paclitaxel chemotherapy. The study met its primary objective of showing noninferiority of gefitinib and also showed its superiority, as compared with carboplatin– paclitaxel, with respect to PFS (HR 0.74; 95% CI 0.65 - 0.85; P<0.001). In the subgroup of 261 patients who were EGFR GMT + PFS was significantly longer with gefitinib than chemotherapy (HR 0.48; 95% CI 0.36 - 0.64; P<0.001), whereas in the subgroup of 176 patients who were negative for EGFR GMT, PFS was significantly longer among those who received chemotherapy than gefitinib (HR for progression or death with gefitinib 2.85; 95% CI,2.05 - 3.98; P<0.001). The most common adverse events in the gefitinib group were rash or acne (in 66.2% of patients) and diarrhea (46.6%).

Back to top

Selection by molecular testing of tumours

Several randomised controlled trials have been published with the first generation EGFR TKIs (gefitinib and erlotinib) in patients selected for treatment by the presence of an activating EGFR gene mutation. All have compared first-line treatment with an EGFR TKI with standard chemotherapy on PFS. Similarly designed trials are in progress with newer generation EGFR targeting agents and inhibitors to other known driving molecular changes (eg. EML4-Alk gene fusion).

Bria et al has reported a literature-based meta-analysis undertaken to quantify the magnitude of benefit with upfront EGFR TKI in Asian patients with activating EGFR mutation (exon-19 deletions or exon-21 point mutations,(EGFR-GMT+) They report findings from five RCTs involving 805 Asian patients, with results for efficacy in patients with activating EGFR mutations reported prospectively (three RCTS) or retrospectively (two RCTs). Four trials evaluated the efficacy of gefitinib and one trial erlotinib, compared with a standard platinum based 3G chemotherapy regimen. EGFR TKI therapy significantly increased PFS (HR) 0.45, 95% CI: 0.36–0.58, P < 0.0001),and overall RR (HR 2.08, 95% CI 1.75–2.46, P < 0.0001) over chemotherapy, with significantly lower neutropenia. The absolute difference in PFS was 26%, corresponding to three to four patients needed to treat for one to benefit, whilst the absolute difference in RR was 36.5%, which translating into two to three patients needed to treat for one to benefit. No significant difference was observed in overall survival, thought largely to be due to treatment crossover with most patients initially treated with chemotherapy going on to receive EGFR TKIs at progression. The rate of exon-19 mutations, female gender, and nonsmoking status were identified as additional predictors of outcome in a meta-regression analysis.

In a Caucasan population, Rosell et al, randomised 174 patients with advanced NSCLC and EGFR mutations (exon 19 deletion or L858R mutation in exon 21) to receive either first-line erlotinib 150 mg daily or a choice of a platinum based 3G doublet regimen. (cisplatin and gemcitabine or docetaxel). The study met its primary endpoint of improved PFS at its pre-planned interim analysis, with median PFS in the erlotinib group of 9.7 months (95% CI 8.4−12.3), compared with 5.2 months (95% CI 4.5−5.8) in the standard chemotherapy group (HR 0.37, 95% CI 0.25−0.54; p<0.0001). Response rate was also in favour of erlotinib (58% versus 15%).

These studies evaluating first line EGFR TKIs in EGFR GMT + patients, which demonstrate dramatic improvements in RR and PFS but not OS, have added to the debate regarding whether OS should remain the the most important therapeutic objective of first line studies in advanced NSCLC. As this guideline has demonstrated, there is evidence for improvement in PFS and OS beyond first line therapy with the use of first line maintenance, second line and even third line therapy. Survival post progression (SPP) on first line therapies has been evaluated in a systematic review by Hotta et al who reviewed 70 phase III trials initiated between 1988 and 2007 involving 38,721 patients with advanced NSCLC.

This review also included studies evaluating molecularly targeted agents but did not report results according to each agent nor whether these studies were only conducted inpatients with an identified molecular target. Nonetheless, Hotta et al observed a stronger association between median survival time (MST) and SPP (r² = 0.8917) than MST and median PFS time (r² = 0.2563), finding that SPP and MPFS can account for 89% and 25% of the variation in MST, respectively. This association between MST and SPP became closer over the years from 1988 to 2007, leading to the conclusion that a PFS advantage from first line treatment is unlikely to be associated with an OS advantage due to this increasing impact of SPP on OS, and that prolongation of SPP might impact on the ability for OS to assessing true efficacy from early-line chemotherapy in future clinical trials. In simple terms, this review highlights the impact of cross over at the completion of initial study treatment to other active drug therapy. How does this relate to anti-EGFR TKIs? Assuming a majority of patients commenced in initial chemotherapy do get to cross over to anti-EGFR TKIs at progression then OS does not appear to be compromised for the population, as found in the Bria and Rossell studies. However for an individual patient there is the potential risk that second line treatment may not occur. The study by Fidias et al of immediate versus delayed docetaxel in non progressing patients after first - line platinum based chemotherapy, demonstrated an attrition rate of 37% i.e. 58 of 156 patients allocated to receive docetaxel at progression did not end up getting treated, 43% (25/58) due to progressive disease . Whilst this may not be the case for the less toxic EGFR TKIs, it would be unreasonable for any patient to miss out on receiving treatment that can result in such a large effect on RR and PFS.

Recommendations

Evidence summary	Level	References
In Asian patients with advanced NSCLC and known common activating EGFR GMs (exon-19 deletions or exon-21 point mutations), first-line therapy with a first generation EGFR TKI (gefitinib or erlotinib) significantly prolongs progression free survival and increases overall response rate, compared with standard platinum-based chemotherapy.	I
In regards to progression free survival, first-line gefitinib is not inferior to carboplatin/paclitaxel chemotherapy in Asian patients, particularly females, with adenocarcinoma, who have never smoked.	II
In caucasian patients with advanced NSCLC and known activating EGFR GMs (exon-19 deletions or exon-21 point mutations), first-line therapy with erlotinib significantly prolongs progression free survival and increases overall response rate, compared with standard platinum based chemotherapy.	II

Evidence-based recommendation	Grade
Patients with known activating gene mutations (exon-19 deletions or exon-21 point mutations) to EGFR should be treated with an EGFR TKI.	A

Evidence summary	Level	References
Progression free survival is significantly longer among patients treated with initial chemotherapy, than those treated with gefitinib in patients known not to have EGFR mutations.	II

Evidence-based recommendation	Grade
Where EGFR mutation status is negative or unknown, patients should be treated with standard chemotherapy.	B

Practice point

The evidence in support of large treatment benefits with first line EGFR TKIs in RR and PFS argues for consideration of obtaining adequate tumour tissue where possible, to enable molecular testing for the presence of activating EGFR gene mutations. This will enable clinicians to offer patients initial EGFR TKIs versus empirical therapy, bearing in mind that overall survival for EGFT GMT + patients does not appear to be compromised as long they go on to receive EGFR TKIs after chemotherapy.

Back to top

Evidence summary and recommendations

See each sub-heading.

Back to top

References

Back to top

Appendices

View recommendation components

View pending evidence

View body of evidence

View all comments

View literature search

Back to top