Correspondence on ‘Pembrolizumab hybrid dosing is non-inferior to flat dosing in advanced non-small cell lung cancer: a real-world, retrospective bicenter cohort study’ by Smeenk et al

The article by Michiel M Smeenk and colleagues, published in the Journal for ImmunoTherapy of Cancer, has the potential to influence immunotherapy administration. However, the results should be interpreted with caution.¹ The authors compared the efficacy of hybrid versus flat dosing of pembrolizumab in advanced non-small cell lung cancer. Although the study showed that the survival outcomes of hybrid-dosed patients were non-inferior to those of flat-dosed patients, and that the hybrid regimen results in reduced drug and associated costs, these findings should be viewed cautiously before informing clinical practice and warrant further validation in future studies.

A well-designed non-inferiority study should incorporate several essential elements.² Among these, defining an acceptable non-inferiority margin is critical, as its size can directly impact the validity of the study’s conclusions. In this study, the justification for selecting a non-inferiority margin of a HR of 1.15 appeared insufficient.

Patients in this study received pembrolizumab with or without chemotherapy, as first-line, second-line, or later-line treatment. It is important to recognize that patients undergoing third-line or later-line therapy typically have significantly shorter overall survival (OS) compared with newly diagnosed patients. Therefore, progression-free survival (PFS) might serve as a more appropriate primary endpoint than OS in the context of this non-inferiority study.

The authors attempted to minimize bias between patients in the two treatment groups. However, potential bias was introduced by two key factors: the line of systemic therapy and the administration of pembrolizumab in combination with chemotherapy. An imbalance in these variables between the groups may substantially affect the study outcomes. Combination therapy has been associated with longer OS compared with monotherapy strategies.³ In table 1 of the paper, 58% of patients in the hybrid dosing group received pembrolizumab plus chemotherapy, compared with 66% in the flat dosing group. This imbalance in baseline characteristics could significantly influence the final results.

The main purpose of this study was to determine whether patients weighing less than 90 kg could receive a lower dose of pembrolizumab without compromising efficacy. To adequately address this question, detailed information and interpretation of the subgroup analyses should be provided. For example, in figure 4A, patients weighing less than 65 kg in the hybrid group exhibited significantly longer median OS compared with those in the flat dosing group. Furthermore, in online supplemental figure 8, 100 mg of pembrolizumab failed to demonstrate non-inferiority in PFS compared with 200 mg among patients weighing less than 65 kg (HR 0.90, 95% CI 0.68 to 1.19). As previously noted, had PFS been selected as the primary endpoint, the conclusion of the study might have been entirely different regarding the dosing strategy.

In summary, the study by Michiel M Smeenk and colleagues raises important questions regarding the modulation of pembrolizumab dosing based on patient weight rather than a fixed flat dose. It would be valuable for the authors to address the concerns raised, in order to more accurately assess the efficacy of pembrolizumab at an optimal dose in patients with advanced non-small cell lung cancer. Therefore, a well-designed prospective study is warranted. Given the inherent limitations of non-inferiority studies, head-to-head comparative studies may offer a more appropriate approach for evaluating dosing strategies.⁴