Clinical Benefit Scales and Trial Design: Some Statistical Issues

First-line immune checkpoint inhibitors in low programmed death-ligand 1-expressing population

Introduction: Inhibitors of programmed cell death 1 (PD1) and its ligand (PDL1) have exhibited favorable long-term survival in many types of advanced-stage cancer and current approvals have to date been granted in certain tumour types irrespective of PD-L1 status. Methods: We extracted the following information: study sample size, trial period, cancer types, intervention of treatment, type of PD-L1 antibody, immunohistochemistry (IHC) scoring method, number and percentage of PD-L1 < 1% population, and median follow- up time. PD-L1 expression was defined as percentage of number of PD-L1-stained tumor cells (TPS), area of tumor infiltrated by PD-L1-stained immune cells (IPS), number of PD-L1-stained cells (tumor cells, lymphocytes and macrophages; CPS). Different trials used distinct method to define low PD-L1 expression. The risk of bias of the included trials was assessed by using the Cochrane risk of bias tool for RCTs. Results: Here, a total of 34 trials were included to extract individual patient data (IPD) to evaluate the survival benefit of first line PD1/PDL1 inhibitors vs. standard-of-care (SOC) in patients with PDL1 < 1%. In term of anti-PD-1/PD-L1 monotherapy, OS (HR = 0.90, 0.81-1.01) and PFS (HR = 1.11, 0.97-1.27) between PD-1/PD-L1 inhibitor group and SOC group were comparable. In term of anti-PD-1/PD-L1 combination therapy, PD-1/PD-L1 inhibitor group exhibited longer OS (median 19.5 months vs. 16.3 months; HR = 0.83, 0.79-0.88, p < 0.001) and PFS than those of SOC group (median 8.11 months vs. 6.96 months; HR = 0.82, 0.77-0.87, p < 0.001).Subgroup analysis showed that survival benefit was mainly observed in non-small cell lung cancer (NSCLC) (HROS = 0.74; HRPFS = 0.69; p < 0.001), small-cell lung cancer (SCLC) (HROS = 0.58, p < 0.001; HRPFS = 0.55, p = 0.030), esophageal squamous cell carcinoma (ESCC) (HROS = 0.62, p = 0.005; HRPFS = 0.79, p < 0.001), melanoma (HROS = 0.53, p < 0.001) and nasopharyngeal carcinoma (NPC) (HRPFS = 0.35, p = 0.013). Conclusion: Anti-PD-1/PD-L1 combinational therapy rather than monotherapy exhibit survival benefit in the low PD-L1 population in the first-line setting, and the survival benefit was mainly observed in specific tumor types.

Clinical data mining: challenges, opportunities, and recommendations for translational applications

Clinical data mining of predictive models offers significant advantages for re-evaluating and leveraging large amounts of complex clinical real-world data and experimental comparison data for tasks such as risk stratification, diagnosis, classification, and survival prediction. However, its translational application is still limited. One challenge is that the proposed clinical requirements and data mining are not synchronized. Additionally, the exotic predictions of data mining are difficult to apply directly in local medical institutions. Hence, it is necessary to incisively review the translational application of clinical data mining, providing an analytical workflow for developing and validating prediction models to ensure the scientific validity of analytic workflows in response to clinical questions. This review systematically revisits the purpose, process, and principles of clinical data mining and discusses the key causes contributing to the detachment from practice and the misuse of model verification in developing predictive models for research. Based on this, we propose a niche-targeting framework of four principles: Clinical Contextual, Subgroup-Oriented, Confounder- and False Positive-Controlled (CSCF), to provide guidance for clinical data mining prior to the model's development in clinical settings. Eventually, it is hoped that this review can help guide future research and develop personalized predictive models to achieve the goal of discovering subgroups with varied remedial benefits or risks and ensuring that precision medicine can deliver its full potential.