Pembrolizumab (9) – Keytruda®

• Clinical trials
  • KEYNOTE 021G (Cohort G of the KEYNOTE 021 study) is an open-label study comparing pembrolizumab in combination with pemetrexed and carboplatin against pemetrexed and carboplatin alone.
  • KEYNOTE 189 is a blinded trial comparing pembrolizumab in combination with pemetrexed and cisplatin/carboplatin against pemetrexed and cisplatin/carboplatin.

a) Adult patients receiving first-line treatment for metastatic non-squamous NSCLC without EGFR- or ALK-positive tumour mutations and with PD-L1 expression of < 50% (TPS)

• For adult patients receiving first-line treatment for metastatic non-squamous NSCLC without EGFR- or ALK-positive tumour mutations and with a PD-L1 expression of < 50 % (TPS), pembrolizumab in combination with pemetrexed and platinum-based chemotherapy offers a non-quantifiable additional benefit compared with pemetrexed in combination with platinum-based chemotherapy.
• Taking a holistic view, and given that the extent of the observed additional benefit in terms of the overall survival endpoint cannot be quantified for the entire patient population, the G-BA concludes that there is an additional benefit whose extent is non-quantifiable.
• In a balancing decision, the G-BA classifies the certainty of the evidence, based on the available data, as ‘hint’.
• mortality
  • In the meta-analysis of the two studies KEYNOTE 021G and KEYNOTE 189, overall survival was statistically significantly prolonged by the administration of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for the TPC patient population relevant to the assessment, with a PD-L1 expression < 50% (TPS) (HR: 0.55; [95% CI: 0.38; 0.77]; p-value = 0.001).
  • The subgroup analysis for the overall survival endpoint provided proof of an effect modification by the characteristic ‘sex’.
  • For these reasons, a separate assessment of additional benefit by sex is not undertaken. Nevertheless, these are relevant findings from the present assessment, which must be taken into account when interpreting the results for the entire patient population with regard to the endpoint ‘overall survival’. Consequently, an additional benefit is identified for the endpoint ‘overall survival’, the extent of which is non-quantifiable.
• Morbidity – Symptoms (constipation)
  • In the KEYNOTE 189 study, there is a statistically significant advantage in favour of pembrolizumab in combination with platinum-based chemotherapy compared with platinum-based chemotherapy alone for the endpoint of constipation, as assessed by the EORTC QLQ-C30 (HR: 0.59; 95% CI [0.38; 0.90]; p = 0.013).
  • The endpoint-specific risk of bias for the symptom-related endpoints in KEYNOTE 189 is high, as a high proportion of patients in the relevant patient population were excluded from the analysis due to missing baseline data, and response rates declined significantly over the course of the study.
• Morbidity – Health status
  • The responder analyses show no significant difference between the two treatment arms, whether based on a MID of 7 or 10 points.
• Health-related quality of life
  • In each case, the time to the first clinically relevant deterioration is considered, defined as a decrease in the score of at least 10 points from baseline; however, no statistically significant difference between the treatment arms was observed for this endpoint.
  • This endpoint category was not assessed in the KEYNOTE 021G study.
  • The endpoint-specific risk of bias for the health-related quality of life endpoints in KEYNOTE 189 is high, as a high proportion of patients in the relevant patient population were excluded from the analysis due to missing baseline data, and response rates declined significantly over the course of the study.
• Side effects – severe adverse events (CTCAE grade ≥ 3)
  • In the meta-analysis of the KEYNOTE 021G and KEYNOTE 189 trials, there is a statistically significant advantage for pembrolizumab in combination with platinum-based chemotherapy compared with platinum-based chemotherapy alone for the endpoint of severe adverse events (CTCAE grade ≥ 3) (HR: 0.74; 95% CI [0.55; 0.9957]; p = 0.047).
  • The endpoint-specific risk of bias for the endpoints ‘severe AEs (CTCAE grade ≥ 3)’ (in both studies) and immune-mediated AEs as well as severe immune-mediated AEs (CTCAE grade ≥ 3) (in KEYNOTE 189 only) is assessed as high, as no information is available on the duration of follow-up or the temporal distribution of therapy discontinuations.
• Side effects – therapy discontinuations due to adverse events
  • For the endpoint ‘therapy discontinuations due to adverse events’, no statistically significant difference was observed between the treatment arms.
  • In the KEYNOTE 021G trial, there is a high potential for bias regarding this endpoint due to the open-label study design.
• Overall assessment
  • In the mortality endpoint category, the meta-analysis of the two studies shows a statistically significant advantage of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for the overall survival endpoint.
  • Subgroup analyses provide proof of an effect modification by the characteristic ‘gender’, showing a statistically significant difference in favour of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for women. For men, therefore, there is no statistically significant difference.
  • In the morbidity (symptoms) endpoint category, the KEYNOTE 189 study showed a statistically significant advantage from treatment with pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for the endpoint of constipation.
  • The results on health-related quality of life show no statistically significant differences between the treatment arms.
  • With regard to side effects, the meta-analysis of the KEYNOTE 021G and KEYNOTE 189 studies shows a statistically significant difference in favour of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for the endpoint of severe side effects (CTCAE grade ≥ 3), a statistically significant difference in favour of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy. For the endpoint ‘therapy discontinuations due to adverse events’, no statistically significant difference was observed between the treatment arms. For the endpoints ‘serious adverse events (SAEs)’ and ‘immune-mediated SAEs’ (from the two studies KEYNOTE 021G and KEYNOTE 189) and the endpoints ‘immune-mediated adverse events’ and ‘immune-mediated severe adverse events (CTCAE grade ≥ 3)’ (from KEYNOTE 021G), no usable analyses are available.

b) Adult patients receiving first-line treatment for metastatic non-squamous NSCLC without EGFR- or ALK-positive tumour mutations and with PD-L1 expression of ≥ 50% (TPS)

• For adult patients receiving first-line treatment for metastatic non-squamous NSCLC without EGFR- or ALK-positive tumour mutations and with PD-L1 expression of ≥ 50 % (TPS), pembrolizumab in combination with pemetrexed and platinum-based chemotherapy offers a non-quantifiable additional benefit compared with pembrolizumab alone.
• On balance, the G-BA concludes that, given that the extent of the identified additional benefit in terms of the overall survival endpoint cannot be quantified for the entire patient population and that an assessment of symptoms and health-related quality of life is not possible, the G-BA determines that there is an additional benefit, the extent of which is non-quantifiable.
• In a balancing decision, the G-BA classifies the certainty of the evidence, based on the available data, as ‘hint’.
• mortality
  • Consequently, in an indirect comparison according to Bucher for the TPC patient population relevant to the assessment, with PD-L1 expression ≥ 50% (TPS), the administration of pembrolizumab in combination with platinum-based chemotherapy results in a statistically significant prolongation of overall survival (HR: 0.40; [95% CI: 0.20; 0.79]; p-value = 0.008).
  • The subgroup analysis for the endpoint of overall survival revealed, based on the adjusted indirect comparison, an effect modification by the characteristic ‘gender’.
  • For these reasons, a separate assessment of the additional benefit by sex is not undertaken. Nevertheless, these are relevant findings from the present assessment, which must be taken into account when interpreting the results for the entire patient population with regard to the endpoint ‘overall survival’. Consequently, an additional benefit is identified for the endpoint ‘overall survival’, the extent of which is non-quantifiable.
• Morbidity – Symptoms
  • It is therefore not possible to make statements regarding additional benefit based on the results of the indirect comparison.
• Morbidity – Health status
  • It is therefore not possible to make any statements regarding additional benefit based on the results of the indirect comparison.
• Health-related quality of life
  • It is therefore not possible to draw conclusions regarding additional benefit based on the results of the indirect comparison.
• Side effects – severe adverse events (CTCAE grade ≥ 3)
  • In the indirect comparison according to Bucher, there was no statistically significant difference in the assessment-relevant TPC patient population with PD-L1 expression ≥ 50 % (TPS) that the administration of pembrolizumab in combination with platinum-based chemotherapy did not result in a statistically significant difference in the incidence of severe AEs (AEs) of CTCAE grade 3 or 4.
  • The subgroup analysis for the endpoint of severe AEs revealed, on the basis of the adjusted indirect comparison, an effect modification by the characteristic ‘gender’.
  • As outlined in the discussion of the overall survival endpoint, the G-BA considers the interpretation of the available subgroup analyses for the characteristic ‘gender’ to be subject to significant uncertainties. It is therefore deemed justified to make a statement on additional benefit without stratifying by gender. Nevertheless, these are relevant findings from the present assessment which must be taken into account when interpreting the results for the entire patient population, including for the endpoint of severe AEs.
• Side effects – therapy discontinuations due to adverse events
  • For the endpoint ‘discontinuation due to AEs’, no statistically significant difference was observed between the treatment arms.
  • Furthermore, due to their open-label design, the KEYNOTE 021G, KEYNOTE 024 and KEYNOTE 042 studies present a high potential for bias regarding the endpoint ‘discontinuation due to AEs’.
• Side effects – serious adverse events (SAEs) and immune-mediated SAEs
  • No usable data are available for SAE and immune-mediated SAE (all four studies), or for immune-mediated AEs and immune-mediated severe AEs (CTCAE grade ≥ 3) (KEYNOTE 021G, KEYNOTE 024 and KEYNOTE 042), no usable data are therefore available.
  • In the G-BA’s view, the chosen operationalisation is problematic for the assessment, and the present result of this operationalisation could have been avoided by specifying a standardised duration of follow-up.
• Overall assessment
  • In the mortality endpoint category, the indirect comparison for the overall survival endpoint shows a statistically significant advantage of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy.
  • Subgroup analyses provide proof of an effect modification by the characteristic ‘gender’, showing a statistically significant difference in favour of pembrolizumab in combination with pemetrexed and platinum-based chemotherapy for women. For men, however, there is no statistically significant difference.
  • No conclusions can be drawn from the indirect comparison regarding the endpoint categories of morbidity and health-related quality of life.
  • With regard to side effects, the indirect comparison shows no statistically significant difference between the treatment groups for the endpoint of severe adverse events (CTCAE grade ≥ 3).
  • For the endpoint ‘discontinuation due to AEs’, no statistically significant difference was observed between the treatment arms.
  • For the endpoints ‘serious adverse events (SAEs)’ and ‘immune-mediated SAEs’ (from all four studies) and the endpoints ‘immune-mediated adverse events’ and ‘immune-mediated severe adverse events (CTCAE grade ≥ 3)’ (from KEYNOTE 021G, KEYNOTE 024 and KEYNOTE 042), no usable analyses are available.

Courtesy translation only, please refer to the German original.