Overall survival, Adverse events, Quality of Life

Objective response rate

Progression free survival

A systematic review that reviewed clinical outcomes after targeted therapy or chemotherapy in NSCLC patients with an ALK fusion was included in this literature analysis.

Description of studies

Peng (2023) performed a systematic review and network meta-analysis of studies to investigate the relative safety and efficacy of all first‐line treatments for patients with ALK‐positive NSCLC. The literature search covered the period 2013–2021. The review included

published/unpublished phase 2/3 randomized clinical trials in patients with histopathologically confirmed advanced (III/IV/recurrent) NSCLC with ALK gene fusions. Articles were included if there was a comparison of any two or more arms of first‐line treatments for patients with advanced NSCLC with ALK gene fusions, and reported on at least one of the following clinical outcome measures: progression‐free survival (PFS), Overall survival (OS), objective response rate (ORR), grade ≥3 treatment‐related adverse events. The following studies were excluded: Trials in which patients with ALK‐positive NSCLC were analyzed as a subgroup, trials of ALK‐TKIs not as first‐line therapy, but as maintenance or neoadjuvant therapy or sequential therapy with chemotherapy or radiotherapy; and trials comparing treatments that have not been approved by any Food and Drug Administration. In total, nine randomized controlled trials (RCTs) were included. All trials were industry sponsored or designed in cooperation with the industry. Important study characteristics of the included RCTs are described in table 1. For more details about the included studies, please see Peng (2023).

Study results

Overall survival - Critical outcome

Nine of the nine studies included in the systematic review by Peng (2023) reported on overall survival (Figure 1).

In the eXalt3 study, data about the effect of ensartinib versus crizotinib on overall survival remained immature at data cutoff. In the modified intention to treat (mITT) population, 62 patients died (24.8% in the ensartinib group and 25.4% in the crizotinib group; HR, 0.91; 95% CI, 0.54-1.54). This difference was not considered clinically relevant. The 2-year overall survival rate was 78% in both groups. Median interim OS was not reached in either group.

In the ALTA-1L study, data about the effect of brigatinib versus crizotinib on overall survival remained immature at data cutoff. At the end of the study, 92 patients had died (30% in the brigatinib arm and 37% in the crizotinib arm). The 3-year estimated OS rate was 71% (95% CI: 62%–78%) in the brigatinib arm and 68% (95% CI: 59%–75%) in the crizotinib arm without adjustment for patients who crossed over from crizotinib to brigatinib (HR = 0.81, 95% CI: 0.53–1.22, log-rank p = 0.33). This difference was not considered clinically relevant.

In the CROWN study, data about the effect of lorlatinib versus crizotinib on overall survival remained immature at data cutoff. At data cutoff 51 patients died (15% in the lorlatinib

group and 19% in the crizotinib group).

Three studies (ALEX, ALESIA, J-ALEX) reported the effect of alectinib versus crizotinib on overall survival. In the ALEX study, OS data remained immature with 37% of events recorded (stratified HR 0.67, 95% CI 0.46-0.98). At data cutoff, 113 patients died (33.6% in the alectinib group versus 41.1% in the crizotinib group). Median OS was not reached in the alectinib arm and was 57.4 months in the crizotinib arm (95% CI 34.6-NR). The 5-year OS rate was 62.5% (95% CI 54.3-70.8) with alectinib and 45.5% (95% CI 33.6- 57.4) with crizotinib. In the ALESIA study, OS data remained immature, with 8 deaths (6%) reported in the alectinib group and 13 deaths (21%) reported in the crizotinib group. In the J-ALEX study, OS data remain immature, with only nine events reported (2% in the crizotinib group and 7% in the alectinib group).

Two studies (PROFILE1029 and PROFILE1014) reported the effect of crizotinib versus chemotherapy on overall survival. In the PROFILE1029 study, 72 patients died 35 (33.7%) in the crizotinib arm and 37 (35.9%) in the chemotherapy arm). Median OS was 28.5 months (95% CI: 26.4–not reached) with crizotinib and 27.7 months (95% CI: 23.9–not reached) with chemotherapy. In the PROFILE1014 study, median OS was not reached (NR; 95% CI, 45.8 months to NR) for the crizotinib group and was 47.5 months (95% CI, 32.2 months to NR) for the chemotherapy group.

In the ASCEND-4 study, data about the effect of ceritinib versus chemotherapy on overall survival remained immature at data cutoff. At data cutoff, 48 events occurred in the

ceritinib group and 59 events in the chemotherapy group, representing 42% of the required events for final OS analysis. The median overall survival was not reached in the ceritinib group (95% CI 29∙3 to not estimable) and was 26∙2 months (22∙8 to not estimable)

in the chemotherapy group (HR 0∙73 [95% CI 0∙50–1∙08]; p=0∙056). This difference was not considered clinically relevant. The 2-year estimated OS rates were 70∙6% (95% CI 62.2–77.5) in the ceritinib group and 58.2% (47.6–67.5) in the chemotherapy group.

Figure 1. Forest plot Overall survival newer generation targeted therapy versus chemotherapy or crizotinib (when compared with next generation TKI)

Progression free survival (PFS) - Important outcome

Nine of the nine studies included in the systematic review by Peng (2023) reported on progression free survival (Figure 2).

In the eXalt3 study, 119 PFS events occurred in the mITT population (62.6% of the 190 expected PFS events). The median PFS was not reached in the ensartinib group (95% CI, 20.2 months to not reached) and 12.7 months in the crizotinib group (95 CI%: 8.9-16.6; HR: 0.45 95%CI: 0.30-0.66; log-rank P < .001). This difference was considered clinically relevant.

In the ALTA-1L study, 166 PFS events occurred, of which 73 in the brigatinib arm (53%) and 93 in the crizotinib arm (67%). The estimated 3-year PFS was 43% (95% CI: 34%–51%) in the brigatinib arm and 19% (95% CI: 12%–27%) in the crizotinib arm (HR = 0.48, 95% CI: 0.35–0.66). This difference was considered clinically relevant. Blinded independent review showed a 4-year PFS 36% (26%–46%) in the brigatinib arm and 18% (11%–26%) in the crizotinib arm. However, this result was limited by a high rate of censoring and few patients at risk in each group.

In the CROWN study, 127 PFS events occurred at data cutoff, of which 41 in the lorlatinib group (28%) and 86 in the crizotinib group (59%). The 12-month PFS was 78% (95% CI: 70–84) in the lorlatinib group and 39% (95% CI: 30–48) in the crizotinib group. HR for disease progression or death= 0.28 (95% CI, 0.19–0.41). This difference was considered clinically relevant

Three studies (ALEX, ALESIA, J-ALEX) reported the effect of alectinib versus crizotinib on PFS. In the ALEX study, 203 patients experienced a PFS event, of which 81 (53.3%) in the alectinib group and 122 (80.8%) in the crizotinib group. Median PFS was 34.8 months (95% CI: 17.7- not evaluable) in the alectinib group and 10.9 months (95% CI: 9.1-12.9) in the crizotinib group. In the ALESIA study, investigator assessed PFS showed that 63 PFS events occurred, of which 26 (21%) in the alectinib group and 37 (60%) in the crizotinib group. Median PFS was not evaluable (95% CI 20·3–NE) for the alectinib group and 11·1 months (95%CI: 9·1–13·0) for the crizotinib group. In the J-ALEX study, 83 PFS events occurred at the second preplanned interim analysis (25 in the alectinib group and 58 in the crizotinib group). Median PFS was not evaluable (95% CI 20·3–NE) in the alectinib group versus 10·2 months (8·2–12·0) with crizotinib (HR 0·34 [99·7% CI 0·17–0·71). The funder decided to release the results immediately, as advised by the independent data monitoring committee. The pooled results of the three RCTS concerning alectinib versus crizotinib showed a hazard ratio of 0.33 (95% CI: 0.22 to 0.49) for PFS favoring alectinib. This difference was considered clinically relevant.

Two studies (PROFILE1029 and PROFILE1014) reported the effect of crizotinib versus chemotherapy on PFS. In the PROFILE1029 study, 166 patients experienced a PFS event (77 in the crizotinib arm and 89 treated with chemotherapy). Median PFS was 11.1 months (95% CI: 8.3–12.6) in the crizotinib group and 6.8 months (95% CI: 5.7–7.0) in the chemotherapy group. In the PROFILE1014 study, median PFS was 10.9 months (95% CI: 8.3 to 13.9) in the crizotinib group and 7.0 months (95% CI: 6.8 to 8.2) in the chemotherapy group. The pooled results of the two RCTS concerning crizotinib verus chemotherapy showed a hazard ratio of 0.43 (95% CI: 0.35 to 0.53) for PFS favoring crizotinib. This difference was considered clinically relevant.

In the ASCEND-4 study, median PFS as assessed by the blinded independent review committee was 16∙6 months (95% CI: 12∙6–27∙2) in the ceritinib group versus 8∙1 months

(5∙8–11∙1) in the chemotherapy group (HR 0.55; 95% CI: 0.42–0.73). This difference was considered clinically relevant.

Figure 2. Forest plot Progression free survival of newer generation targeted therapy versus chemotherapy or crizotinib (when compared with next generation TKI)

Objective response rate (ORR) - Important outcome

Nine of the nine studies included in the systematic review by Peng (2023) reported on ORR (Figure 3). The HR for ORR in patients treated with ensartinib, brigatinib, and alectinib compared to patients treated with crizotinib was 1.11 (95% CI: 0.95 to 1.31), 1.18 (95% CI: 0.99 to 1.40), and 1.14 (95% CI: 1.07 to 1.23) respectively. These differences were not considered clinically relevant. The ORR was higher in patients treated with lorlatinib compared to patients treated with crizotinib (HR 1.31; 95% CI: 1.11 to 1.55). This differences was considered clinically relevant. The ORR was higher in patients treated with crizotinib (pooled HR 1.75 ; 95% CI: 1.51 to 2.04) or ceritinib (HR 2.71; 95% CI: 2.11 to 3.49) compared to patients treated with chemotherapy. These differences were considered clinically relevant.

Figure 3. Forest plot objective response rate of newer generation targeted therapy versus chemotherapy or crizotinib (when compared with next generation TKI)

Adverse events (AEs) Grade ≥3 - Important outcome

Nine of the nine studies included in the systematic review by Peng (2023) reported on adverse events grade ≥3. Most studies reported a high number of patients with adverse events grade ≥3.

The Risk difference for AEs grade ≥3 in patients treated with ensartinib, brigatinib, and lorlatinib compared to patients treated with crizotinib was 0.08 (-0.04, 0.19), 0.14 (0.03, 0.24), and 0.17 (0.06, 0.28) respectively. These differences were not considered clinically relevant.

Combining the data from the three RCTs (ALEX, ALESIA, J-ALEX), the occurrence of grade≥ 3 adverse events was 142 in 380 patients treated with alectinib (37%) and 169 events in 317 patients treated with crizotinib (53%). The pooled risk difference was -0.16 (95% CI: -0.30 to -0.03). This difference was not considered clinically relevant.

The Risk difference for AEs grade ≥3 in patients treated with crizotinib compared to patients treated with chemotherapy was -0.03 (-0.14, 0.08). The Risk difference for AEs grade ≥3 in patients treated with ceritinib compared to patients treated with chemotherapy was 0.17 (0.07, 0.26). These differences were not considered clinically relevant.

Figure 4 Forest plot of newer generation targeted therapy versus chemotherapy or crizotinib (when compared with next generation TKI) outcome safety, defined as number of grade ≥3 adverse events 

Quality of life (QoL) - Important outcome

Five of the nine studies included in the systematic review by Peng (2023) reported on quality of life (QoL).

In the ALTA-1L study, median time to worsening of GHS/QoL was 26.7 months in the brigatinib arm versus 8.3 months in the crizotinib arm (HR = 0.69, 95% CI: 0.49–0.98). The authors also reported that the time to worsening of emotional and social functioning and symptoms of fatigue, nausea and vomiting, appetite loss, and constipation was significantly delayed by brigatinib.

In the CROWN study, mean baseline scores of global quality of life were 64.6 (SE 1.82) in the lorlatinib group and 59.8 (SE 1.90) in the crizotinib group. A significantly greater overall improvement was observed among patients in the lorlatinib group compared to patients in the crizotinib group (estimated mean difference, 4.65; 95% CI: 1.14 to 8.16). The authors state that this difference was not considered clinically meaningful.

In the PROFILE1029 study, significantly greater improvement in global QoL scores from baseline was observed in patients in the crizotinib group compared to patients in the chemotherapy group(p < 0.001). Statistically significantly greater improvements were also observed with regard to physical functioning, role functioning, social functioning, and cognitive functioning (assessed by the QLQC30) and in general health status scores (assessed by the EQ-5D-3L visual analog scale). In patients treated with crizotinib significantly greater reductions from baseline were observed compared to patients in the chemotherapy group for symptoms of fatigue, pain, insomnia, appetite loss, alopecia, pain in the arm/shoulder, pain in other body parts, coughing, dyspnea, and pain in the chest. Differences in mean changes in overall quality of life score were not reported.

In the PROFILE1014 study, Solomon (2014) suggested a greater reduction from baseline in the symptoms of pain, dyspnea, and insomnia (assessed by the QLQ-C30) and in the symptoms of dyspnea, cough, chest pain, arm or shoulder pain, and pain in other parts of the body as assessed with the use of the QLQ-LC13 with crizotinib than with chemotherapy. The results, however, were only plotted in a figure. The authors reported that significantly greater improvement in global QoL scores from baseline was observed in patients in the crizotinib group compared to patients in the chemotherapy group(p < 0.001). Statistically significantly greater improvements were also observed with regard to physical functioning, social, emotional and role functioning domains. Differences in mean changes in overall quality of life score were not reported.

In the Ascend-4 study, lung cancer-specific symptoms and quality of life measures with the EuroQol (EQ-5D-5L) were significantly improved for ceritinib versus chemotherapy. No mean quality of life scores per treatment group were reported; results were only plotted in a figure.

In the eXalt3 study, the ALEX study, and the ALESIA study, and the J-ALEX study QoL outcomes were not reported.

Level of evidence

We did not assess the certainty of the evidence for each outcome separately based on the individual studies by using GRADE, but described the main limitations of the systematic reviews or the studies included in the systematic reviews to provide an indication of the overall certainty of the evidence. The risk of bias assessment for the individual included studies was adopted from the systematic review by Peng (2023).

There are four levels of evidence: high, moderate, low, and very low. RCTs start at a high level of evidence.

Targeted therapy (ensartinib, brigatinib, lorlatinib, or alectinib) versus crizotinib

The levels of evidence for OS, PFS, ORR, AEs, and QoL were downgraded with three levels because of limitations in the study design (risk of bias: no blinding of patients and physicians and unclear whether outcome data was complete and unclear if the study was free of other bias), inconsistency of results (heterogeneity), imprecision (OIS not reached).

Crizotinib or ceritinib versus chemotherapy:

The levels of evidence for OS, PFS, ORR, AEs, and QoL were downgraded with three levels because of limitations in the study design (risk of bias: no blinding of patients and physicians and limited information on concealment of allocation), inconsistency of results (heterogeneity), and imprecision (OIS not reached).

The search and selection methods can be found in the main module ‘Behandeling incurabel NSCLC met (zeldzame) mutaties’.