Advertisement

Challenges and issues in bivariate endpoints in study designs of translational medicine

  • Qian Zhou
    Correspondence
    Corresponding author. Address: Department of Medical Statistics, Clinical Trials Unit, The First Affiliated Hospital of Sun Yat-sen University, 58 Zhong Shan Er Road, Guangzhou, China, 510080. Tel./fax: +86-20-87755766-8649,.
    Affiliations
    Department of Medical Statistics, Clinical Trials Unit, The First Affiliated Hospital of Sun Yat-sen University, 58 Zhong Shan Er Road, Guangzhou, China, 510080
    Search for articles by this author

      Linked Article

      To the Editor:
      We read with great interest the recent study by Chuah et al.
      • Chuah S.
      • Lee J.
      • Song Y.
      • Kim H.D.
      • Wasser M.
      • Kaya N.A.
      • et al.
      Uncoupling immune trajectories of response and adverse events from anti-PD-1 immunotherapy in hepatocellular carcinoma.
      aiming to find immune predictors of treatment response and mechanisms of response versus immune-related adverse events (irAE) in hepatocellular carcinoma (HCC) patients treated with anti-programmed cell death 1 (PD-1). The authors noticed the recent successful combinations of immunotherapies achieved increased response rates compared to monotherapy, but with an accompanying increase in irAEs. The authors were inspired by this and conducted an innovative study to investigate how to increase response rates and reduce irAEs simultaneously. In this study, the peripheral blood samples of patients with HCC who had received anti-PD-1 monotherapy were collected for cytometry by time of flight and single-cell RNA sequencing to find the immune components related to response and toxicity, respectively. Further, on the basis of cell-cell interaction analysis, in vivo mouse models were established to verify that the immunotherapy targets can improve response rates without increasing toxicity. Defining optimal treatment combinations which balance activity and toxicity is an important issue in oncology clinical trials.
      • Ahern E.
      • Solomon B.J.
      • Hui R.
      • Pavlakis N.
      • O'Byrne K.
      • Hughes B.G.M.
      Neoadjuvant immunotherapy for non-small cell lung cancer: right drugs, right patient, right time?.
      ,
      • Agostinetto E.
      • Losurdo A.
      • Nader-Marta G.
      • Santoro A.
      • Punie K.
      • Barroso R.
      • et al.
      Progress and pitfalls in the use of immunotherapy for patients with triple negative breast cancer.
      This study provides novel insights into the preclinical study of drug combinations. However, there are a number of concerns with the study design which should be pointed out.
      In this study, when investigating the enriched immune components and cell-cell interactions for response and toxicity, the two binary endpoints of efficacy and safety were analyzed separately. The interaction or the bivariate nature of the two endpoints were ignored. Bivariate endpoints are not uncommon in clinical studies. For example, in early drug toxicity trials, there were statistical models that considered drug activity and toxicity at the same time.
      • Colin P.
      • Delattre M.
      • Minini P.
      • Micallef S.
      An escalation for bivariate binary endpoints controlling the risk of overtoxicity (EBE-CRO): managing efficacy and toxicity in early oncology clinical trials.
      But in basic and translational research, experiments are often based on a binary endpoint to explore a mechanism. The study follows the traditional design, showing the differences of targets between response and non-response and between toxicity and non-toxicity independently. This separate analysis, no matter from the perspective of bioinformatics or basic research, cannot accurately capture the combined active mechanism of response and toxicity. We believe that the study population should be reorganized into four groups in order to evaluate efficacy and safety simultaneously. As Fig. 1 shows, patients from each study cohort could be divided into four quadrants according to their two binary outcomes (i.e. response and toxicity) as a previous study on a bivariate endpoint trial design presented.
      • Kittelson J.M.
      • Spyropoulos A.C.
      • Halperin J.L.
      • Kessler C.M.
      • Schulman S.
      • Steg G.
      • et al.
      Balancing risk and benefit in venous thromboembolism trials: concept for a bivariate endpoint trial design and analytic approach.
      Patients from the upper right quadrant I achieved the most desirable result from anti-PD-1 treatment with response and non-toxicity, while the lower left quadrant reflected a negative result with no response but toxicity. In quadrant II and IV, patients faced equivocal benefit from the treatment. In terms of the purpose of this study, the authors could find suitable targets for combined immunotherapy by analyzing the difference between patients in quadrant I and other quadrants. Through this classification, we can not only compare the targets between response and no response and between toxicity and no toxicity groups, but also analyze the characteristics of patients in quadrant I, so as to find specific targets for both response and non-toxicity. Therefore, it is necessary to make such a distinction among the population first through this approach. However, one challenge of this classification is the limited sample size for each group. Although the number of each quadrant is relatively small, all four types of patients exist as expected.
      Figure thumbnail gr1
      Fig. 1Bivariate endpoints analysis maps.
      Four quadrants were presented based on the two binary endpoints of response and toxicity.
      In addition, only patients receiving monotherapy were included to evaluate the immune targets for combined therapies. To highlight the results of the combined anti-PD-1 and anti-TNFR2 treatment, a control group receiving combined immunotherapies which have been proven to be positive in randomized clinical trials should be added to the in vivo mouse models. Other issues include the limited sample size of each group for in vivo mouse models. In addition, the reporting of how to find the two targets of TNFR1 and TNFR2 from a cluster of ligand-receptor pairs is unclear.
      Besides, we re-analyzed the characteristic data of the two cohorts from Table S1 of the paper (see our Table S1) and found that the age of the Korea cohort using for validation was 10 years younger than that of the Singapore cohort. Some other variables, such as BCLC stage, viral status, extrahepatic spread, were significantly different between the two cohorts as well.
      Finally, we noticed that as shown in Fig. 5A, there were only two time points of measurements after treatment, reflecting the changes of cell proportions. The authors should provide the rationale for using the word ‘trajectory’ in the title and conclusion of the study.
      In summary, the findings of the study need to be confirmed by further well-designed large-sample animal experiments before they can be applied to clinical translational studies. Preclinical design is very important for drug development and good study design is necessary to reduce potential bias in the process of mechanism research.

      Financial support

      The authors received no financial support to produce this manuscript.

      Conflict of interest

      The author declares no conflicts of interest that pertain to this work.
      Please refer to the accompanying ICMJE disclosure form for further details.

      Supplementary data

      The following are the supplementary data to this article:

      References

        • Chuah S.
        • Lee J.
        • Song Y.
        • Kim H.D.
        • Wasser M.
        • Kaya N.A.
        • et al.
        Uncoupling immune trajectories of response and adverse events from anti-PD-1 immunotherapy in hepatocellular carcinoma.
        J Hepatol. 2022 Sep; 77: 683-694
        • Ahern E.
        • Solomon B.J.
        • Hui R.
        • Pavlakis N.
        • O'Byrne K.
        • Hughes B.G.M.
        Neoadjuvant immunotherapy for non-small cell lung cancer: right drugs, right patient, right time?.
        J ImmunoTherapy Cancer. 2021; 9e002248
        • Agostinetto E.
        • Losurdo A.
        • Nader-Marta G.
        • Santoro A.
        • Punie K.
        • Barroso R.
        • et al.
        Progress and pitfalls in the use of immunotherapy for patients with triple negative breast cancer.
        Expert Opin Investig Drugs. 2022; 9: 1-25
        • Colin P.
        • Delattre M.
        • Minini P.
        • Micallef S.
        An escalation for bivariate binary endpoints controlling the risk of overtoxicity (EBE-CRO): managing efficacy and toxicity in early oncology clinical trials.
        J Biopharm Stat. 2017; 27: 1054-1072
        • Kittelson J.M.
        • Spyropoulos A.C.
        • Halperin J.L.
        • Kessler C.M.
        • Schulman S.
        • Steg G.
        • et al.
        Balancing risk and benefit in venous thromboembolism trials: concept for a bivariate endpoint trial design and analytic approach.
        J Thromb Haemost. 2013; 11: 1443-1448