The Japanese Lung Cancer Society Guideline for non-small cell lung cancer, stage IV

According to rapid development of chemotherapy in advanced non-small cell lung cancer (NSCLC), the Japan Lung Cancer Society has been updated its own guideline annually since 2010. In this latest version, all of the procedure was carried out in accordance with grading of recommendations assessment, development and evaluation (GRADE) system. It includes comprehensive literature search, systematic review, and determination of the recommendation by multidisciplinary expert panel which consisted of medical doctors, pharmacists, nurses, statisticians, and patients from patient advocacy group. Recently, we have had various types of chemotherapeutic drugs like kinase inhibitors or immune-checkpoint inhibitors. Thus, the guideline proposes to categorize patients into three entities: (1) driver oncogene-positive, (2) PD-L1 ≥ 50%, and (3) others. Based on this subgroup, 31 clinical questions were described. We believe that this attempt enables clinicians to choose appropriate treatment easier. Here, we report an English version of the Japan Lung Cancer Society Guidelines 2018 for NSCLC, stages IV.


Process of guideline development
Guideline committee (expert panel) consisted of medical doctors, pharmacists, nurses, statisticians, and patients from patient advocacy group. Initially, guideline committee members developed clinical questions (CQs) and collected evidences as described above. Evidences were systematically reviewed for every CQ and committee members decided the strength of evidence (Table 1). Considering this strength of evidence and other factors (i.e. risk-benefit balance and social values), we determined the final recommendation. In this guideline, we have two levels of recommendations (1; strong or 2; weak) in two directions (to do or not to do) ( Table 2). Recommendation was accepted when at least 60% of expert panel members reached an agreement. Whole process of developing guideline was carried out in accordance with Grading of Recommendations Assessment, Development and Evaluation (GRADE) system (http://www.grade worki nggro up.org/).

Outline
Cytotoxic chemotherapy has long been a mainstream component in the treatment of stage IV NSCLC. A meta-analysis showed that cytotoxic chemotherapy significantly prolonged overall survival (OS) compared with best supportive care (BSC) [1]. Cytotoxic chemotherapy accounted for 9% increase in 1-year survival (from 20 to 29%), or 1.5-month prolongation of OS. Another analysis, using third-generation cytotoxic chemotherapy, showed that monotherapy led to about 7% improvement in 1-year survival compared with BSC [2]. Regarding toxicity, another meta-analysis regarding NSCLC showed that treatment-related deaths with cytotoxic chemotherapy comprised 1.26%, and consisted of febrile neutropenia, cardiovascular events, and pulmonary toxicities [3]. Third-generation cytotoxic chemotherapy alone also showed better quality-of-life (QOL) scores than BSC [4]. Platinum agent plus third-generation cytotoxic chemotherapy showed similar QOL compared with thirdgeneration cytotoxic chemotherapy alone, while progression-free survival (PFS) and overall survival (OS) were better in combination chemotherapy in a phase III trial [5].
Since the 2000s, novel agents such as molecular-targeted drugs and immune-checkpoint inhibitors (ICIs) have been demonstrating better outcomes compared with cytotoxic chemotherapy.
Most of the molecular-targeted drugs used in the treatment of NSCLC inhibit specific targets that induce cancer evolution [so-called "oncogenic drivers (i.e., EGFR mutation, ALK translocation, ROS1 translocation, and BRAF mutation)"]. Among those patients who harbor such oncogenic drivers and have good PS, these inhibitors demonstrated both ORR and PFS improvement. Phase III trials were conducted in EGFR-and ALK-altered patients, and kinase inhibitors have been shown to be more effective compared with cytotoxic chemotherapy [6][7][8][9][10][11][12]. For other types of genetic alterations, phase III trials have not been conducted, due to their rarity. Instead, similar efficacy results were observed in phase II trials or subset analyses of phase III trials. Molecular-targeted drugs usually showed milder toxicity than cytotoxic chemotherapy [13][14][15][16]. It is also important that they showed beneficial results in patients with poor PS in relatively small, but prospective studies [17,18].
Since 2015, ICIs, which has a novel mode of action compared with other chemotherapeutic drugs, have been approved for administration in Japan. ICIs target immunecheckpoint molecules such as PD-1/L1, which are negative regulators in tumor immunity. A phase III trial, KEY-NOTE-024, compared pembrolizumab (PD-1 inhibitor) with platinum-doublet chemotherapy in EGFR/ALK-negative, advanced NSCLC patients with tumor positive for PD-L1 ≥ 50% [19]. Pembrolizumab monotherapy demonstrated significant improvement of ORR, PFS, and OS, with tolerability. In addition to this, several phase III studies showed higher efficacy of platinum-based chemotherapy plus ICI than platinum-based chemotherapy alone in advanced NSCLC [20][21][22][23].
In summary, chemotherapy (cytotoxic chemotherapy, molecular-targeted drugs, and ICIs) demonstrated prolongation of OS and improvement of QOL compared with BSC among advanced NSCLC patients with good PS. To determine optimal treatment regimen, it is essential to check appropriate biomarkers such as driver oncogenes for molecular-targeted drugs and PD-L1 status for ICIs before their use. Therefore, while diagnosing advanced NSCLC patients, it is also important to categorize them into three entities: (1) driver oncogene-positive, (2) PD-L1 ≥ 50%, and (3) others ( Fig. 1). Treatment strategy of each of these subgroups is described below (Fig. 2).
1. Driver oncogene-positive [CQ 1 -17] As previously described, kinase inhibitors for their specific driver oncogenes demonstrated improvement of ORR and PFS. OS was not improved, because most of the patients in standard arm received kinase inhibitors after progression. A large observational study in EGFR-mutated patients demonstrated that PFS with erlotinib did not differ, regardless of treatment line [24]. A firm conclusion regarding the treatment sequence between kinase inhibitors and cytotoxic chemotherapy cannot be drawn. However, a prospective observational study in the U.S. analyzed 10 genes in 733 patients, and oncogenic drivers were detected among 466 patients (64%). The study also showed that the patients who had oncogenic drivers and received kinase inhibitors lived longer than those who had oncogenic drivers, but did not receive inhibitors (3.5 years versus 2.4 years, propensity score-adjusted hazard ratio: 0.69, 95% CI: 0.53-0.90, p = 0.006) [25]. This evidence supports the notion that patients with driver oncogene should not lose the chance of receiving kinase inhibitors, and this guideline recommends that clinicians should administer kinase inhibitors prior to cytotoxic chemotherapy in this subgroup. In the second-line setting, cytotoxic chemotherapy is recommended in accordance with the patient's general condition: CQ 20-23.
Regarding ICI for patients with oncogenic drivers, refer to CQ 5 and 6. Regarding squamous cell carcinoma with oncogenic drivers, refer to CQ 17.
2. PD-L1 ≥ 50% [CQ 18 and 19] As data suggest that this subgroup can receive much benefit from ICI, pembrolizumab monotherapy is recommended. Recommendation: Kinase inhibitors targeting each oncogene are strongly recommended for patients who have driver oncogene with good PS (0-1).
Recommendation: 1 Evidence level: A Agreement rate 100% Comments: In this guideline, we generically name EGFR mutation, ALK translocation, ROS1 translocation, and BRAF mutation as driver oncogenes that might be the direct cause of cancer evolution. Among those patients who harbor with such oncogenic drivers and have good PS, these inhibitors demonstrated both ORR and PFS improvement. Phase III trials were conducted in EGFR-and ALKmutated patients, and kinase inhibitors have been shown to be more effective compared with cytotoxic chemotherapy [6-10, 12, 26]. For other types of genetic alterations, phase III trials have not been conducted due to their rarity [i.e., EGFR uncommon mutation (see CQ 10), ROS1 translocation (see CQ 15), and BRAF mutation (see CQ 16)]. Instead, similar efficacy results were observed in phase II trials or subset analyses of phase III trials (i.e., afatinib in EGFR uncommon mutation, crizotinib in ROS1 mutation) [13][14][15][16]27]. A prospective observational study in the US analyzed 10 genes in 733 patients, and oncogenic drivers were detected among 466 patients (64%). The study also showed that the patients who had oncogenic drivers and received kinase inhibitors lived longer than those who had oncogenic drivers, but did not receive inhibitors (3.5 years versus 2.4 years, propensity score-adjusted hazard ratio: 0.69, 95% CI: 0.53-0.90, p = 0.006) [25]. As there are limited data regarding ICIs in patients with oncogenic drivers, current data do not show superiority of ICIs to efficacy of TKIs (see CQ 7).
OS was not improved in phase III trials that compared TKIs with cytotoxic chemotherapy, because most of the patients in cytotoxic chemotherapy arm received kinase inhibitors after progression. A large observational study in EGFR-mutated patients demonstrated that PFS with erlotinib did not differ, regardless of treatment line [24]. A firm conclusion regarding the treatment sequence between EGFR-TKI and cytotoxic chemotherapy cannot be drawn. Moleculartargeted drugs usually showed milder toxicity than cytotoxic chemotherapy, but AE profiles and severity are divers.
Based on this evidence, kinase inhibitors targeting each oncogene are recommended treatment for patients who have  Comments: (a) In the phase III trials in EGFR-mutated or ALK-rearranged patients, about 5-10% of the participants had PS of 2. Their efficacy was relatively similar to those with PS of 0-1 [8,9,12]. Regarding gefitinib and alectinib, prospective trials were reported in EGFR-mutated or ALK-rearranged patients with poor PS [17,18]. Although data are limited in patients who have other types of driver oncogenes, based on the results of patients with EGFR or ALK alteration, we strongly recommend administration of kinase inhibitors in this population. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) On the other hand, AE profiles and severity are divers among the drugs. Some of the agents require higher frequencies of drug discontinuation or decrease, even in patients with PS 0-1, and more caution should be paid in case of patients with PS 3-4.
We judged quality of evidence as C, and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below. In a phase II trial in elderly (≥ 75 years) patients with EGFR mutation, gefitinib demonstrated comparable efficacy (ORR of 74% and mPFS of 12.3 months) and safety compared with prior trials that included younger patients [28]. Erlotinib demonstrated similar efficacy regardless of age in a Japanese phase II trial [29].
No data are available regarding elderly (≥ 75 years) patients with other genetic alterations (i.e., EGFR uncommon mutation, ROS1 translocation, BRAF mutation), and kinase inhibitors usually showed milder toxicity than cytotoxic chemotherapy. Thus, administration of kinase inhibitors in the elderly is relatively feasible. Based on the results of EGFR-TKIs in the elderly, we recommend the use of any kinase inhibitor for elderly patients who have driver oncogene. We judged quality of evidence as C and strength of recommendation as 1. Note that more attention to adverse events should be paid in this population than in younger patients.
The results of voting by committee members to determine this recommendation are described below.

Comments:
Kinase inhibitors are the key drugs in patients with driver oncogenes. However, most of the patients received cytotoxic chemotherapy before or after kinase inhibitors in the phase III trials. Post hoc analyses represented that those patients who received cytotoxic chemotherapy had better prognosis compared with those who did not in prospective trials [30,31], and similar tendency was shown in a large-scale Japanese observational study [32]. Among EGFR-mutated patients, efficacy of chemotherapy may not be different from that in EGFR wild-type patients, although no prospective trial has been conducted. Thus, previous evidence, including mixed types of EGFR mutation, can apply to this CQ (also see CQ 5,20,21 and 23).
Based on this evidence, we strongly recommend the use of cytotoxic chemotherapy even in patients who have driver oncogene during their lines of treatment. We judged quality of evidence as A and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) In the first-line setting, a phase III trial (IMpower150) was conducted to compare carboplatin (CBDCA) + paclitaxel (PTX) + bevacizumab with or without atezolizumab in non-squamous, non-small cell lung cancer. In the subgroup analysis of EGFR-or ALK-altered population, PFS was significantly prolonged (median 9.7 months versus 6.1 months (HR 0.89, 95%CI: 0.37-0.94) in atezolizumab arm, while, in interim analysis, OS was not prolonged (median not reached versus 17.5 months (HR 0.54, 95%CI: 0.29-1.03) [21,33]. However, this result should be interpreted with caution, because this was not a preplanned analysis and mutation status was not defined as an allocation factor. Atezolizumab arm showed more AEs (≥ grade 3) than control arm (58.5% vs. 50.0%), including immune-related AEs such as diarrhea (20.6% vs. 15.2%) and skin rash (13.3% vs. 5.1%). The guideline committee determined that there is insufficient evidence to draw a firm conclusion based on this subset analysis at this time.
The results of voting by committee members to determine this recommendation are described below. In a phase III trial (KEYNOTE-024) comparing pembrolizumab with platinum-doublet chemotherapy in chemo-naïve patients whose tumor was positive for PD-L1 ≥ 50%, patients who had EGFR or ALK alterations were excluded. In a pooled analysis of phase III trials that compared ICIs (nivolumab, pembrolizumab, and atezolizumab) with docetaxel (DTX), HR of OS comparing ICIs with DTX was 1.05 (95%CI: 0.70-1.55, p < 0.81) [34]. In a single-center analysis, ORR of ICIs was only 3.8% in patients who were positive for EGFR or ALK [35]. This evidence may suggest that ICI does not function similarly in patients with driver oncogene as in those without. However, the studies contained relatively small numbers of subjects. The guideline committee finally determined that there is insufficient evidence to draw a firm conclusion on this CQ.
The results of voting by committee members to determine this recommendation are described below. Comments: EGFR Exon 19 deletion and L858R mutation consist of about 90% of EGFR mutation and are known as sensitive to EGFR-TKIs. In phase III trials comparing EGFR-TKI (gefitinib, erlotinib, or afatinib) with platinum-doublet chemotherapy, types of EGFR mutation were limited to or consisted mostly of these two mutations [6][7][8][9][10]26]. All trials demonstrated significant prolongation of PFS in EGFR-TKI arm, and several QOL items were improved [36].
(a) FLAURA is a phase III trial comparing osimertinib with first-generation EGFR-TKIs (gefitinib or erlotinib) in stage IV NSCLC patients who have EGFR mutation (exon 19 deletion or L858R mutation) and PS of 0-1.
In this trial, PFS (primary endpoint) was significantly prolonged in osimertinib arm [18.9 months versus 10.2 months, HR 0.46 (95%CI: 0.37-0.57), p < 0.001] [37]. Regarding toxicity, first-generation EGFR-TKIs showed diarrhea in 57%, skin rash in 48%, AST elevation in 25%, and interstitial lung disease in 2% of patients, while osimertinib showed milder toxicities (diarrhea 58%, skin rash 25%, AST elevation 9%, and interstitial lung disease 4%). Based on these balances between efficacy and toxicity, we strongly recommend osimertinib in patients who have EGFR mutation (exon 19 deletion or L858R mutation) with PS of 0-1. We judged quality of evidence as B and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.  [38,39]. However, dacomitinib was more toxic: diarrhea in 78%, paronychia in 54%, and skin rash in 35% of patients, while gefitinib showed diarrhea in 55%, paronychia in 19%, and skin rash in 35%. Based on these balances between efficacy and toxicity, we weakly recommend dacomitinib in patients who have EGFR mutation (exon 19 deletion or L858R mutation) with PS of 0-1. We judged quality of evidence as B and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below.  [40]. In a randomized phase II trial, afatinib demonstrated longer PFS than gefitinib, but toxicity was severe [41].
Based on these balances between efficacy and toxicity, we weakly recommend gefitinib, erlotinib, or afatinib in patients who have EGFR mutation (exon 19 deletion or L858R mutation) with PS of 0-1. We judged quality of evidence as A and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below. Comments: In the two phase III trials in metastatic NSCLC patients with EGFR mutation comparing erlotinib with platinumdoublet chemotherapy, 7% and 14% of the patients had PS of 2 [8,9]. The efficacy in those patients was relatively similar to that in those with PS of 0-1. Gefitinib showed efficacy in prospective trials in patients with poor PS [28,46]. Data are insufficient regarding efficacy and safety of afatinib and dacomitinib in this population [10,26,38]. Although the situation is similar, osimertinib can be considered, because the toxicity is generally milder than that of gefitinib or erlotinib, except for interstitial lung disease [37].
Based on this evidence, we strongly recommend the administration of gefitinib or erlotinib in patients who have EGFR mutation (exon 19 deletion or L858R mutation) with PS of 2. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) In a prospective trial of gefitinib in EGFR-mutated patients with poor PS (mostly PS 3-4), PS was improved among about 80% of the patients and efficacy was good (ORR 66%, median PFS 6.5 months and median OS 17.8 months) [46]. On the other hand, administration should be considered carefully, because poor PS is a well-known risk factor of interstitial lung disease, as are male, smoking history, pre-existing interstitial lung disease, patients who have limited normal lung region, and those with heart disease [40,41]. The guideline committee had thorough discussion regarding patients with PS of 4. For such patients, physicians must consider whether EGFR-TKI would improve PS, or symptoms that are more meaningful outcomes than good PS.
Based on this evidence, we strongly recommend the administration of gefitinib in patients who have EGFR mutation (exon 19 deletion or L858R mutation) with PS of 3-4. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. Other mutations are called "uncommon mutations," and are located between exon 18 and 21 (i.e., E709X, G719X, S768I, P848L, L861Q, or exon20 insertion). They are usually sensitive to EGFR-TKIs, but ORR was slightly lower than that observed in common mutation [50]. Most of the phase III trials completely excluded these mutations [6,8,9], or the mutations existed in only 10% of the entire population [7,10,26].
Among patients with uncommon mutation other than exon20 insertion and T790M mutation, ORR was reported as 48-71% [16,50]. Of those, ORR with afatinib in a small subset of a trial was 71.1%, which showed relatively higher tendency than other EGFR-TKIs [16]. Although this result was obtained in a prospective manner, a firm conclusion as to recommend afatinib as better than other EGFR-TKIs cannot be drawn.
Based on this evidence, gefitinib, erlotinib, or afatinib is weakly recommended in patients who have EGFR mutation of exon18-21, except for exon 19 deletion, L858R mutation, exon20 insertion, and T790M mutation, with PS of 0-1. We judged quality of evidence as C and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) There are few data regarding EGFR exon 20 insertion, and ORR in these reports was around 10% [16,51]. Thus, EGFR-TKI was not considered as a first-line treatment option.
Based on this evidence, EGFR-TKI is not strongly recommended in patients who have EGFR exon 20 insertion mutation with PS of 0-1. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. Second-line and further treatment of those patients who have EGFR mutation (Fig. 5).

CQ 11
What is the recommended second-line treatment in patients with EGFR T790M mutation after progression of EGFR-TKIs?
Recommendation: 1 Evidence level: B Agreement rate 100% Comments: Osimertinib is a third-generation EGFR-TKI that has a potent activity to both sensitive EGFR mutation and resistant T790M mutation. A phase III trial comparing cytotoxic chemotherapy with osimertinib was conducted in those patients who had EGFR T790M mutation after progression of the first-or second-generation EGFR-TKIs (gefitinib, erlotinib, or afatinib) [53]. PFS (primary endpoint) was significantly prolonged in osimertinib arm [10.1 months versus 4.4 months, HR 0.30 (95%CI: 0.23-0.41), p < 0.001]. There were fewer adverse events greater than grade 3 in osimertinib arm (6% versus 34%).
Based on this evidence, we strongly recommend the administration of osimertinib in patients with EGFR T790M mutation after progression of EGFR-TKIs. We judged quality of evidence as B and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
The results of voting by committee members to determine this recommendation are described below.

ALK-rearranged, non-squamous cell lung cancer
First-line treatment of those patients who have ALK rearrangement is shown in Fig. 6.

CQ 12
What is the recommended first-line treatment in patients who have ALK rearrangement with PS of 0-1? Recommendation: (a) Alectinib is strongly recommended.
Based on the current data, we strongly recommend the administration of alectinib in patients who have ALK rearrangement with PS of 0-1. We judged quality of evidence as A and strength of recommendation as 1. On the other hand, OS data in these trials were immature. The superiority of alectinib in terms of OS is unknown.
Thus, crizotinib can be an alternative first-line treatment. We judged quality of evidence as A and strength of recommendation as 2.
Similarly, ceritinib can be considered as an alternative, but no comparative study with other ALK-TKI has been conducted. Treatment-related adverse events were seen in 65% of the patients, and liver dysfunction was the main event [55]. We judged quality of evidence as B and strength of recommendation as 2.
Regarding elderly patients, 16% of the participants were ≥ 65 years old in the phase III trial comparing crizotinib with platinum-doublet chemotherapy, and efficacy was similar between elderly and younger patients [54]. In addition, 11% of the participants were ≥ 75 years old in the phase III trial comparing alectinib with crizotinib, and advantage of alectinib over crizotinib was similar among the elderly.
The results of voting by committee members to determine this recommendation are described below. One study reported the efficacy of alectinib in patients who have ALK rearrangement with poor PS [18]. Although the number of patients was relatively small (PS 2: 12 patients, PS 3: 5 patients, and PS 4: 1 patient), there was no severe adverse event. In this trial, ORR was 72%. In the phase III trial comparing alectinib with crizotinib, there were fewer adverse events in the alectinib arm, although only 2% of the patients had PS of 2 [56].
Based on this evidence, we strongly recommend the administration of alectinib in patients who have ALK rearrangement with PS of 2-4. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)  Comments: (a) In a phase I/II trial and phase II trials conducted overseas, crizotinib followed by alectinib in patients with ALK rearrangement showed relatively good efficacy (ORR of 48-50% and median PFS of 8.1-8.9 months) [59][60][61]. In Japan, 23 patients who had progressed with crizotinib were treated with alectinib [62]. In this trial, ORR was 65% and mPFS was 12.9 months. Although there are no comparable data of alectinib with platinum-doublet chemotherapy, these results from single-arm trials can be considered to be at least equivalent.
Based on this evidence, we strongly recommend the administration of alectinib in crizotinib-refractory, ALK-rearranged patients. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.  [65]. In a Japanese phase I trial, 5 of 9 patients who had treatment history with crizotinib had PR with higher GI toxicities [66]. Although there are no comparable data of ceritinib with platinum-doublet chemotherapy, these efficacy results can be considered equivalent, and toxicity may be higher than with alectinib. In addition, the ASCEND-5 study set cytotoxic chemotherapy monotherapy as a control, which makes the results difficult to interpret.
Based on this evidence, we weakly recommend the administration of ceritinib in crizotinib-refractory, ALKrearranged patients. We judged quality of evidence as C and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)  [13]. In an East Asian study, 127 patients showed ORR of 69.3% and mPFS of 13.4 months [14]. These results are comparable or better than those for crizotinib in ALK-rearranged patients.
Based on this evidence, we strongly recommend the administration of crizotinib in patients who have ROS1 rearrangement (also refer to CQ 17). We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)  [15]. In another phase II trial, of dabrafenib plus trametinib for BRAF-mutated patients who were chemo-naïve, 36 patients showed ORR (primary endpoint) of 64% and mPFS of 10.9 months [27]. Although there were limited numbers of Japanese patients in these trials, we strongly recommend the administration of dabrafenib plus trametinib in patients who have BRAF mutation (also refer to CQ 17). We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. Among EGFR-mutated, squamous cell lung cancer patients, no comparative study has been conducted using EGFR-TKI. However, subgroup data from prospective trials and their pooled analysis were reported. In a prospective trial, 11 patients for whom data were available demonstrated ORR of 9.1%. Although this suggests limited efficacy of EGFR-TKI in this population, caution is advised, because at least three patients had uncommon EGFR mutation in this report [69][70][71]. Several retrospective analyses with small sample size showed ORR of 25-32% and mPFS of 1.4-3.9 months [69,72,73]. No prospective trial has been conducted in patients with uncommon EGFR mutation, ALK rearrangement, ROS1 rearrangement, or BRAF mutation.
Based on this evidence, efficacy of kinase inhibitors in squamous cell lung cancer patients who have driver oncogene is limited, but some responders existed. The guideline committee weakly recommends for kinase inhibitors in squamous cell lung cancer patients who have driver oncogene. We judged quality of evidence as D and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below. PD-L1 ≥ 50% (Fig. 10) CQ 18 What is the recommended first-line treatment in patients with PS 0-1 and whose tumor is positive for PD-L1 ≥ 50%? Recommendation: (a) Pembrolizumab monotherapy is strongly recommended.
Recommendation: 1 Evidence level: B Agreement rate 69% Comments: (a) A phase III trial comparing pembrolizumab with platinum-doublet chemotherapy was conducted in stage IV NSCLC patients with PS 0-1, whose tumor was positive for PD-L1 ≥ 50%, and who did not have EGFR mutation or ALK rearrangement (KEYNOTE-024 study) [74]. In this trial, 305 patients were randomized and 66 (43.7%) were treated with pembrolizumab after progression of platinum-doublet chemotherapy as a crossover treatment. In an interim analysis, PFS (primary endpoint) was significantly prolonged in pembrolizumab arm [ Regarding the elderly patients ≥ 75 years old, no subgroup analysis from this trial or any prospective data has been reported.
Based on this evidence, pembrolizumab monotherapy is strongly recommended in patients whose tumor is positive for PD-L1 ≥ 50% and does not have EGFR mutation or ALK rearrangement. We judged quality of evidence as B and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. A phase III trial comparing platinum-doublet chemotherapy plus pembrolizumab with platinum-doublet chemotherapy was conducted in stage IV, non-squamous NSCLC patients with PS 0-1 and who did not have EGFR mutation or ALK rearrangement (KEYNOTE-189 study) [20]. In this trial, 616 patients were randomized in 2:1 ratio and 67 (32.5%) were treated with pembrolizumab monotherapy after progression of platinum-doublet chemotherapy as a crossover treatment. In an interim analysis, PFS (primary endpoint) was significantly prolonged in chemotherapy plus pembrolizumab arm  Main adverse events were nausea, anemia, fatigue, and constipation in chemotherapy plus pembrolizumab arm, and adverse events ≥ grade 3 were similar (67.2% versus 65.8%). On the other hand, in chemotherapy plus pembrolizumab arm, acute kidney injury and immune-related adverse events ≥ grade 3 were observed in 5.2% and 8.9% of the patients, respectively. In addition, three treatment-related deaths due to interstitial lung disease were reported. Careful management should be applied.
Another phase III trial, comparing platinum-based chemotherapy plus atezolizumab with platinum-based chemotherapy, was conducted in stage IV, non-squamous NSCLC patients with PS 0-1 (IMPOWER150 study). Results of CBDCA + PTX + bevacizumab plus atezolizumab (arm B) and CBDCA + PTX + bvacizumab (arm C) were reported [21]. PFS (one of the co-primary endpoints) was significantly prolonged in chemotherapy plus atezolizumab arm . Main adverse events were appetite loss, peripheral neuropathy, nausea, and fatigue in chemotherapy plus atezolizumab arm, and adverse events ≥ grade 3 were slightly higher (58.5% versus 50.0%). In chemotherapy plus atezolizumab arm, immune-related adverse events such as skin rash, liver dysfunction, thyroid dysfunction, pneumonitis, and colitis were reported. Careful management should be applied.
A phase III trial comparing platinum-doublet chemotherapy plus pembrolizumab with platinum-doublet chemotherapy was conducted in stage IV, squamous cell lung cancer patients with PS 0-1 (KEYNOTE-407 study) [22]. In this trial, 559 patients were randomized. In an interim analysis, PFS (one of the co-primary endpoints) was significantly prolonged in chemotherapy plus pembrolizumab arm [6. . Main adverse events were anemia, appetite loss, neutropenia, and nausea in chemotherapy plus pembrolizumab arm, and adverse events ≥ grade 3 were similar (69.8% versus 68.2%). On the other hand, treatmentrelated deaths were higher in chemotherapy plus pembrolizumab arm (3.6% versus 2.1%).
Regarding the elderly patients ≥ 75 years old, these four trials [20][21][22][23] allowed registration of this population. In two trials using atezolizumab, subgroup analyses were reported. Both trials showed that PFS was better in the chemotherapy plus atezolizumab arm, even in the elderly [9.7 months versus 6.8 months (HR 0.78) in the IMpower150 study and 7.0 months versus 5.6 months (HR 0.78) in the IMpower131 study] [21,23]. However, safety data were not reported in this population. Careful management should be applied.
Based on this evidence, platinum-based chemotherapy plus PD-1/PD-L1 inhibitor is strongly recommended in patients with PS 0-1, whose tumor is positive for PD-L1 ≥ 50%, and who do not have EGFR mutation or ALK rearrangement. On the other hand, as there has been no direct comparison between chemotherapy plus PD-1/ PD-L1 inhibitor and pembrolizumab alone, the superiority has not been clarified. We judged quality of evidence as B and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) Unable to determine recommendation. Comments: (a) Refer to CQ 22. (b) As the KEYNOTE-024 study allowed enrollment of patients with PS 0-1 [74], there have been no efficacy or safety data of pembrolizumab monotherapy in stage IV NSCLC patients with PS 2. On the other hand, there is limited evidence regarding cytotoxic chemotherapy in this area. The efficacy was modest, while these patients often suffered from adverse events. Regarding the lesser toxicity of pembrolizumab, many guideline committee members supported pembrolizumab as an option for this population. Based on this evidence, after careful consideration, the guideline committee weakly recommends the administration of pembrolizumab monotherapy in patients with PS 2 and whose tumor is positive for PD-L1 ≥ 50% as an expert opinion. We judged quality of evidence as D and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below. (c) As four phase III trials comparing platinum-based chemotherapy plus PD-1/PD-L1 inhibitor with platinum-based chemotherapy allowed enrollment of patients with PS 0-1, there are no efficacy or safety data of platinum-based chemotherapy plus PD-1/ PD-L1 inhibitor in stage IV NSCLC patients with PS 2. Basically, there are some safety concerns regarding cytotoxic chemotherapy in this population. Thus, adding PD-1/PD-L1 inhibitor to platinum-based chemotherapy may not be tolerable. Based on this, there is no clear evidence to recommend the combination of platinum-based chemotherapy plus PD-1/PD-L1 inhibitor in patients with PS 2 and whose tumor is positive for PD-L1 ≥ 50%. The guideline committee finally determined that there is insufficient evidence to draw a firm conclusion on this CQ. The results of voting by committee members to determine this recommendation are described below.

Driver oncogene-negative and PD-L1 < 50%, or unknown
First-line treatment in patients who are driver oncogene-negative and PD-L1 < 50%, or unknown is shown in Fig. 11 CQ 20 Is cytotoxic chemotherapy recommended as a firstline treatment in patients with PS 0-1 and younger than 75 years old, when their tumor is driver oncogene-negative and PD-L1 is < 50%, or unknown?
Recommendation: Combination of platinum agent and third-or later-generation drug is strongly recommended.
Recommendation: 1 Evidence level: A Agreement rate 93%

Comments:
A meta-analysis demonstrated that platinum-based chemotherapy significantly prolonged OS compared with BSC [1]. In another meta-analysis, comparing secondgeneration agent with third-generation agent as a part of a platinum-based regimen, the latter was found to be superior, with ORR of 12% and 1-year survival rate of 6% [2]. In a Japanese phase III trial (FACS study), four regimens of platinum agents plus third-generation agents were compared, and their efficacies were similar [75].
Newer agents demonstrated the efficacy in several phase III studies, but some showed their efficacy on specific histology types. PEM is one typical example and is approved for non-squamous cell lung cancer. In study JMDB, comparing cisplatin (CDDP) + PEM with CDDP + gemcitabine (GEM), efficacy results were similar in overall population. However, CDDP + PEM showed better OS in non-squamous cell lung cancer [ [76]. Even though this was a subset analysis, CDDP + PEM are considered one of the preferred regimens for non-squamous cell lung cancer in terms of efficacy and safety. Regarding CBDCA + PEM, although there has been no phase III trial to investigate its OS benefit, it is commonly used due to its milder non-hematologic toxicity than CDDP. In randomized studies comparing CBDCA + PEM with CBDCA + GEM, CBDCA + DTX, or CBDCA + PTX + bevacizumab, superiority in OS or safety was not demonstrated [77][78][79]. On the other hand, Kaplan-Meier curves of OS seemed to be similar between CBDCA + PEM and CBDCA + PTX + bevacizumab [79], and PFS of CBDCA + PEM + bevacizumab tended to be better than CBDCA + PTX + bevacizumab [80]. Based on these results, CBDCA + PEM can be considered as a firstline treatment regimen.
Based on this evidence, combination of platinum agent and third-or later-generation drug is strongly recommended in patients with PS 0-1 and younger than 75 years old. We judged quality of evidence as A and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. In a post hoc analysis of a phase III trial of first-line treatment and adjuvant chemotherapy, efficacy was similar between those who were younger than 65 years and who were 65 years of age or older [85]. They also reported that instrumental ADL, not chronological age, was associated with better prognosis. Another report mentioned that octogenarian patients did not show worse OS than younger patients (7 months versus 11 months, p = 0.20) and they also had similar safety result, even when they had PS of 0-1 [86]. These reports suggest that elderly patients should not be excluded from anti-cancer treatment due to their chronological age.
(a) Phase III trials demonstrated that vinorelbine (VNR) prolonged OS compared with BSC, and also showed that GEM had similar efficacy to VNR [87,88]. Based on this evidence, monotherapy using a third-generation drug is strongly recommended in patients with PS 0-1 and older than 75 years old. We judged quality of evidence as A and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. However, these results were not more satisfactory than DTX monotherapy in a Japanese trial, and treatment-related deaths were relatively high in combination arm (4.4%). In addition, dosing of combination arm was unfamiliar in Japanese clinical practice. These results should, therefore, be interpreted with caution.
Based on this evidence, CBDCA doublet chemotherapy is weakly recommended in patients with PS 0-1 and older than 75 years old. We judged quality of evidence as A and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)

CQ 22
Is cytotoxic chemotherapy recommended as a firstline treatment in patients with PS 2, when their tumor is driver oncogene-negative and PD-L1 is < 50%, or unknown? Recommendation: (a) Monotherapy using third-generation drug is strongly recommended.
Recommendation: 2 Evidence level: B Agreement rate 100% Comments: Because PS 2 comprises heterogeneous patients, standard treatment has not been established. In a subset of a meta-analysis comparing chemotherapy with BSC, OS was prolonged by chemotherapy regardless of PS [6% improvement in 1-year OS rate in patients with PS ≥ 2 (14% versus 8%)] [1].
(a) A meta-analysis comparing the third-generation drugs (DTX, PTX, VNR, or GEM) with BSC demonstrated 7% of improvement in 1-year OS rate [2]. In this study, about 30% of patients had PS 2. From the three trials that were included in this meta-analysis, subset analyses of patients with PS 2 were reported, and each of the OS results tended to be prolonged [92]. Based on this evidence, monotherapy using a thirdgeneration drug is strongly recommended in patients with PS 2. We judged quality of evidence as A and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) Regarding toxicity, anemia and neutropenia were common in the combination arm and treatmentrelated death was observed in 3.9% of the patients. Based on this evidence, platinum-doublet chemotherapy is weakly recommended in patients with PS 2. We judged quality of evidence as B and strength of recommendation as 2. The guideline committee calls attention to the clinicians that this evidence is limited, and most regimens in these trials were CBDCA-based or reduced-dosed. The results of voting by committee members to determine this recommendation are described below.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) CQ 23 Is addition of PD-1/PD-L1 inhibitor to platinumbased chemotherapy recommended in patients with PS 0-2, when their tumor is driver oncogene-negative and PD-L1 is < 50%, or unknown? Recommendation: (a) Addition of PD-1/PD-L1 inhibitor to platinum-based chemotherapy is strongly recommended in patients with PS 0-1.
Recommendation: 1 Evidence level: B Agreement rate 78% (b) There is no clear evidence to recommend the addition of PD-1/PD-L1 inhibitor to platinum-containing chemotherapy in patients with PS 2.
Unable to determine recommendation. Comments: (a-1) Non-squamous cell lung cancer.
A phase III trial comparing platinum-doublet chemotherapy plus pembrolizumab with platinum-doublet chemotherapy was conducted in stage IV, non-squamous NSCLC patients with PS 0-1 and who did not have EGFR mutation or ALK rearrangement (KEYNOTE-189 study) [20]. 616 patients were randomized in 2:1 ratio and 67 (32.5%) were treated with pembrolizumab monotherapy after progression of platinum-doublet chemotherapy as a crossover treatment. In an interim analysis, PFS (primary endpoint) was significantly prolonged in chemotherapy plus pembrolizumab arm [ . Main adverse events were nausea, anemia, fatigue, and constipation in chemotherapy plus pembrolizumab arm, and adverse events ≥ grade 3 were similar (67.2% versus 65.8%). On the other hand, in chemotherapy plus pembrolizumab arm, acute kidney injury and immunerelated adverse events ≥ grade 3 were observed in 5.2% and 8.9% of the patients, respectively. In addition, three treatment-related deaths due to interstitial lung disease were reported. Careful management should be applied.
Another phase III trial comparing platinum-based chemotherapy plus atezolizumab with platinum-based chemotherapy was conducted in stage IV, non-squamous NSCLC patients with PS 0-1 (IMPOWER150 study). Results of CBDCA + PTX + bevacizumab plus atezolizumab (arm B) and CBDCA + PTX + bevacizumab (arm C) were reported [21] PFS (one of co-primary endpoints) was significantly prolonged in chemotherapy plus ate- . Main adverse events were appetite loss, peripheral neuropathy, nausea, and fatigue in chemotherapy plus atezolizumab arm, and adverse events ≥ grade 3 were slightly higher (58.5% versus 50.0%). In chemotherapy plus atezolizumab arm, immune-related adverse events such as skin rash, liver dysfunction, thyroid dysfunction, pneumonitis, and colitis were reported. Careful management should be applied.
A phase III trial comparing platinum-doublet chemotherapy plus pembrolizumab with platinum-doublet chemotherapy was conducted in stage IV squamous cell lung cancer patients with PS 0-1 (KEYNOTE-407 study) [22]. In this trial, 559 patients were randomized. In an interim analysis, PFS (one of the co-primary endpoints) was significantly prolonged in chemotherapy plus pembrolizumab arm . Main adverse events were anemia, appetite loss, neutropenia, and nausea in chemotherapy plus pembrolizumab arm, and adverse events ≥ grade 3 were similar (69.8% versus 68.2%). On the other hand, treatment-related deaths were higher in chemotherapy plus pembrolizumab arm (3.6% versus 2.1%).
Another phase III trial, comparing platinum-doublet chemotherapy plus atezolizumab with platinum-doublet chemotherapy, was conducted in stage IV, squamous cell lung cancer patients with PS 0-1 (IMPOWER131 study). Results of CBDCA + nab-PTX plus atezolizumab (arm B) and CBDCA + nab-PTX (arm C) were reported [23]. PFS (one of the co-primary endpoints) was significantly prolonged in chemotherapy plus atezolizumab arm Regarding elderly patients ≥ 75 years old, these four trials [2,[75][76][77] allowed registration of this population. In two trials using atezolizumab, subgroup analyses were reported. Both trials showed that PFS was better in the chemotherapy plus atezolizumab arm, even in the elderly [9.7 months versus 6.8 months (HR 0.78) in the IMpower150 study; and 7.0 months versus 5.6 months (HR 0.78) in the IMpower131 study] [75,77]. However, safety data were not reported in this population. Careful management should be applied.
Based on this evidence, addition of PD-1/PD-L1 inhibitor to platinum-containing chemotherapy is strongly recommended in patients with PS 0-1 when their tumor is driver oncogene-negative and PD-L1 is < 50%, or unknown. We judged quality of evidence as B and strength of recommendation as 1. (b) As four phase III trials comparing platinum-based chemotherapy plus PD-1/PD-L1 inhibitor with platinum-based chemotherapy allowed enrollment of patients with PS 0-1 [20][21][22][23], there has been no efficacy or safety data of platinum-based chemotherapy plus PD-1/PD-L1 inhibitor in stage IV NSCLC patients with PS 2. Basically, there are some safety concerns regarding cytotoxic chemotherapy in this population. Thus, adding PD-1/PD-L1 inhibitor to platinum-based chemotherapy may not be tolerable. Based on this, there is no clear evidence to recommend the addition of PD-1/PD-L1 inhibitor to platinum-based chemotherapy in patients with PS 2, when their tumor is driver oncogene-negative and PD-L1 is < 50%, or unknown. The guideline committee finally determined that there is insufficient evidence to draw a firm conclusion on this CQ. The results of voting by committee members to determine this recommendation are described below. Two phase III trials comparing three or four courses of combination chemotherapy consisting of platinum and third-generation agent with six courses of the same regimen both showed that 1-year survival and OS were similar, while toxicity was milder with the former [97,98]. An individual-patient data meta-analysis that containing these two studies compared six courses with five or fewer courses demonstrated that PFS was significantly prolonged by six courses of treatment, but OS was similar [99].
Recent phase III trials mostly defined the maximum number of courses of platinum-based chemotherapy as four to six. In a phase III trial comparing CDDP + PEM with CDDP + GEM (JMDB study), the median number of courses of CDDP treatment was five in both arms [76].
Based on this evidence, six or fewer courses of platinumbased chemotherapy are strongly recommended. We judged quality of evidence as C and strength of recommendation as 1.  [103]. ORR was also improved. On the other hand, in another phase III trial comparing bevacizumab and CDDP + GEM with CDDP + GEM (AVAiL study), PFS was prolonged, but OS was similar [104]. In a Japanese phase II trial (JO19907 study) using the same regimen as in ECOG4599, ORR [105]. In China, a phase III trial (BEYOND study) also using the same regimen as in ECOG4599 was conducted [106]. Based on this evidence, addition of bevacizumab to platinum-doublet chemotherapy is weakly recommended in patients with PS 0-1 and < 75 years old. We judged quality of evidence as A and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below.
In a subset analysis of ECOG4599 study, the elderly population (> 70 years old) did not benefit from addition of bevacizumab, but showed increase of neutropenia, bleeding, and proteinuria ≥ grade 3 [107]. In another combined analysis, of ECOG4599 and the PointBreak study, benefits of OS and PFS were less, especially in patients who were > 75 years old [108]. In a retrospective study conducted in the US (ARIES study), efficacy was similar between patients age below 65, between 65 and 75, and those above 75 years, but arterial thrombosis ≥ grade 3 tended to be increased in the elderly subset (1.5% < 65 years, 2.9% ≥ 65 years, and 3.5% ≥ 75 years) [109]. In a cohort study conducted in Europe (SAiL study), efficacy was similar between patients aged below and above 70 years, but hemorrhagic events tended to be increased in the elderly subset (3.5% < 70 years, 5.3% ≥ 70 years). In Japan, there are no sufficient data of bevacizumab-containing chemotherapy in the elderly.
Based on this evidence, not to add bevacizumab to platinum-doublet chemotherapy is weakly recommended in patients ≥ 75 years old. We judged quality of evidence as C and strength of recommendation as 2. The results of voting by committee members to determine this recommendation are described below. Based on this evidence, in patients who received four courses of platinum-based chemotherapy without disease progression and with tolerable toxicity, not to administer continuation or switch maintenance is strongly recommended. We judged quality of evidence as C and strength of recommendation as 1. The results of voting by committee members to determine this recommendation are described below. Cytotoxic chemotherapy was usually not indicated in patients with PS 3-4. Regarding PD-1/PD-L1 inhibitor, clinical trials enrolled mainly patients with PS 0-1. Efficacy and safety data were unknown in patients with PS 3-4. In a phase III trial, erlotinib was compared with BSC in patients with poor PS or concomitant disease (TOPICAL study [117]). Median age of participants was 77, 30% had PS 3, and 52% were EGFR wild-type. OS (primary endpoint) was not prolonged [3.7 months versus 3.6 months, HR 0.94 (95%CI: 0.81-1.10), p = 0.46].
Based on this evidence, not to administer any chemotherapy is strongly recommended. We judged quality of evidence as D and strength of recommendation as 1. Second-or further-line treatment in patients who are driver oncogene-negative and PD-L1 is < 50% or unknown (Fig. 12) CQ 28 Is chemotherapy recommended as a second-or further-line treatment in patients with PS 0-2 who progressed with first-line treatment other than PD-1/PD-L1 inhibitor? Recommendation: (a) PD-1/PD-L1 inhibitor is strongly recommended.  [118][119][120]), while a study of pembrolizumab was conducted only in patients whose tumor expressed PD-L1 > 1% (KEYNOTE-010 study [121]). These trials enrolled patients with PS 0-1; thus, efficacy and safety data of PD-1/PD-L1 inhibitor in patients with PS 2 in the second-line setting were unclear. However, considering milder toxicity of these agents compared with cytotoxic drugs, many guideline committee members supported PD-1/PD-L1 inhibitor as an option for this population. On the other hand, ORR of these drugs was 10-20% in the previous reports. Once they are not effective, the guideline committee also recommends a prompt change to cytotoxic chemotherapy.
Based on this evidence, PD-1/PD-L1 inhibitor is strongly recommended in patients who progressed with the first-line treatment. We judged quality of evidence as A and strength of recommendation as 1.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)  [118][119][120][121][122], several studies showed a cross in Kaplan-Meier curves of PFS or OS. This may suggest that some patients benefited from DTX. In addition, DTX + RAM significantly prolonged OS compared to DTX alone, but this combination has not been compared with any PD-1/PD-L1 inhibitors. Based on this evidence, cytotoxic chemotherapy is weakly recommended in patients who progressed with first-line treatment. We judged quality of evidence as A and strength of recommendation as 2.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients)  12 Second-line and further treatment in patients with stage IV NSCLC, who were driver oncogene-negative and PD-L1 was < 50%, or unknown. PD-1 programed cell death-1, PD-L1 programed death-ligand 1, PS performance status 1 3

CQ 29
What is the recommended chemotherapy as the second-or further-line treatment in patients with PS 0-2?
Recommendation: 1 Evidence level: A Agreement rate 100% Comment: Two phase III trials of DTX in patients who progressed with first-line treatment were reported. In TAX-320 study, DTX 75 mg/m 2 showed improvement in ORR, PFS rate at 26 weeks, and OS rate at 1 year compared with VNR or IFM, although they were not statistically significant [123]. In another study, comparing DTX 75 mg/m 2 with BSC, MST and OS rate at 1 year were superior in DTX arm (7.5 months versus 4.6 months, 37% versus 19%, respectively), and QOL was also improved [124]. In a Japanese phase II trial, 60 mg/ m 2 of DTX showed relatively similar efficacy (ORR 18.2% and OS 7.8 months) [125]. Regarding evidence of DTX + RAM, refer to CQ 30.
Based on this evidence, DTX+/-RAM, PEM, or S-1 is strongly recommended in patients with PS 0-2 as their second-or further-line treatment. We judged quality of evidence as A and strength of recommendation as 1.
Based on this evidence, addition of RAM to DTX is weakly recommended in patients with PS 0-1. We judged quality of evidence as B and strength of recommendation as 2.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) There are no efficacy data of DTX + RAM, because the REVEL study and JVCG study excluded these patients. Considering the rate of febrile neutropenia in patients with PS 0-1, clinicians should heed concern regarding the safety in patients with PS 2.
Based on this evidence, not to add RAM with DTX is weakly recommended in patients with PS 2. We judged quality of evidence as D and strength of recommendation as 2.
Involved members regarding this voting: sub-committee on chemotherapy (medical doctors, pharmacists, nurses, statisticians, and patients) Recommendation: Not to administer erlotinib is weakly recommended in patients who are EGFR wild-type or unknown.
Recently, PD-1/PD-L1 inhibitors and DTX + RAM demonstrated superiority to DTX, and S-1 demonstrated noninferiority to DTX. Considering this situation, efficacy of erlotinib is relatively low. In addition, those patients who were EGFR wild-type had some clinical risk factors of interstitial lung disease with EGFR-TKI.