From Molecule to Patient and Ways to Get the Dose Precisely Right

Abstract

This theme issue “From molecule to patient” brings to the foreground the many amazing technological advances of the past decade at the molecular level, including teaching old modalities new tricks, increasing experience in quantitative pharmacology and modeling, and simulation approaches, pharmacogenomics, and a dramatic shift to include special populations, such as pediatrics and geriatrics and, importantly, also orphan indications. It is clear that the “molecule” no longer is the quintessential small molecular active moiety. It now includes monoclonal antibodies and other related biologics but also vaccines that could have a major impact in reducing the global burden of tuberculosis and HIV infection. Most recently, cell therapies have become available that offer the promise of treating and altering the course of diseases that could not be effectively treated before. These cell-based therapies include manipulated cells for transplantation or transfusions and innovative therapies referred to as advanced therapy medicinal products, such as gene therapy and engineered tissue products.1 From a clinical pharmacology perspective, these novel therapies provide interesting new challenges once a promising biologic or cell therapy can transition to first-in-human clinical testing. For instance, genetically engineered T cells represent a powerful new class of immuno-oncology cell-based therapies that offer new avenues for therapeutic success in patients with cancer.2 How do we best define the exposure-response relationship and the optimal dose, a paradigm that has been well worked out for small molecules and biologics but not so much for cell-based therapies?3 Some of these challenges may be facilitated by the mandates and incentives provided by the US Food and Drug Administration's (FDA's) 21st Century Cures Act and the sixth iteration of the Prescription Drug User Fee Act (PDUFA VI) encouraging the advancement of model-informed drug development (MIDD).4-6 With the growing experience in quantitative modeling and simulation, MIDD now seems poised to further advance the development of novel therapeutics with the goal to optimize drug-development paradigms, including reducing the need to generate clinical trial data where possible.7 Yet to date, most MIDD case studies provide mostly confirmatory evidence of efficacy and safety and have not truly been transformative in obviating actual clinical data gathering. In this respect, the recently announced Pilot Meetings Program for Model-Informed Drug Development may be the transformative initiative the modeling and simulation community has been waiting for.8 For this pilot program, MIDD projects are defined as the application of exposure-based, biological, and pharmacostatistical models derived from preclinical and clinical data sources to address drug development and regulatory issues. The envisioned MIDD approaches should improve not only clinical trial efficiency and increase the probability of regulatory success but optimize drug dosing and therapeutic individualization in the absence of dedicated trials. In this issue, Serrano and colleagues9 provide a comprehensive overview of the development history of the novel class of immune checkpoint inhibitors, all therapeutic antibodies. One of the hallmarks of cancer is the ability of cancer cells to evade the immune system. The introduction of immune checkpoint inhibitors, which are able to block negative regulators of T cells and restore immune system function, has been a transformational breakthrough in cancer therapy. This article provides a comprehensive overview of the pivotal trials leading to initial approval of the first six novel immune checkpoint inhibitors in the United States, European Union, and Japan and outlines how novel regulatory pathways, such as the Breakthrough Therapy Designation, Priority Review, Accelerated Approval and Assessment, Orphan Drug status, and use of Companion Diagnostics were granted to bring these new drugs to market more quickly. The commentary by Breckenridge10 provides a clarifying historical perspective and outlines the pros and cons of accelerated drug approval at the expense of greater uncertainty at market authorization while providing some suggestions to meeting possible public health concerns by encouraging the use of real-world data (RWD) after approval. Given the recent developments in MIDD, the recommendations for therapeutic monitoring of 5-fluorouracil in oncology by Beumer and colleagues11 provide an example of what seems an ongoing paradox in gathering and disseminating drug dosing information to better tailor doses to individual patient needs. These authors did an excellent job of systematically evaluating all available evidence to compile a comprehensive overview of existing data while providing a compelling framework to evaluate published evidence in support of therapeutic drug monitoring recommendations for any drug in oncology. In addition, as the authors note, the proposed framework would allow the identification of knowledge gaps that subsequently can be targeted for further study. This is an approach that in recent years, has been coined as RWD gathering and analysis. Real-world evidence is an umbrella term for evidence not generated in highly controlled research environments.12 The framework for the FDA's Real-World Evidence Program was published in December of 2018.13 In the document, RWD are defined as data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources, such as collections of electronic health records. Real-world evidence is the clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of RWD. RWD can also be used to improve the efficiency of clinical trials and help with informing prior probability distributions of exposure-response that subsequently can be incorporated in Bayesian pharmacostatistical models. This is where the apparent disconnect between clinical practitioners and the data and model-informed approaches applied by pharmacometricians and modelers in drug development should be bridged. In a recent contribution to this journal under the title “Why Has Model-Informed Precision Dosing Not Yet Become Common Clinical Reality?” the issues preventing widespread implementation into clinical practice of model-informed precision dosing were discussed.14 What is presented reflects nearly half a century of model-based approaches as a means of adjusting doses for individual patients. It highlights the fact that advanced technologies that used to be lacking (e.g., computational power and rapid turnaround assays) have become available. It also highlights several reasons why model-informed precision dosing for many still is more of a scientific endeavor with a paucity of evidence to definitively show improved outcomes and cost-effectiveness, which hampers broad-scale adoption. In a similar effort, colleagues at the University of North Carolina provide a strong plea for the implementation of precision dosing strategies as they would improve the efficiency of drug development, decrease healthcare costs, and address a significant public health need in terms of precision medicine delivery.15 Their proposed framework for the development of precision dosing tools starts early in the development of a new drug and continues after regulatory approval. Sponsors should be encouraged to consider whether their drug would benefit from precision dosing, and these considerations should be included as part of the investigational new drug application and approval processes. The collected exposure-response data and models that will help guide biomarker and pharmacokinetic/pharmacodynamic (PK/PD) assessments performed throughout phase I–III studies will provide valuable insights into the level of variability in drug exposure and the need to refine strategies to predict and control for this variability at the individual patient level. The information collected during the development process should be summarized as PK/PD models, which then ideally are provided in the label and made available online. In addition, to further refine and expand existing models, data from postmarketing surveillance registries could be further mined using modeling approaches to answer questions that could not be fully addressed during the regular drug-development process. So what would it take to make model-informed precision dosing common practice? It all starts with clinical champions who see the potential of bringing pharmacometrics directly to the bedside. This will require great emphasis on efforts to educate and train healthcare professionals, both at the pregraduate and graduate levels and those working in clinical practice, also to overcome some of the cultural differences between healthcare professionals and the modeling community. Several initiatives are ongoing, including efforts by the adult and pediatric National Institutes of Health (NIH) supported T32 clinical pharmacology training networks to disseminate know-how on modeling and simulation and pharmacometrics approaches through the Sumner J. Yaffe Webinar Series in Pediatric Clinical Pharmacology as well as through hands-on training. One perceived hurdle to broad dissemination of model-informed precision dosing is the relative scarcity of point-of-care assays for concentration and biomarker measurement. Here, a companion diagnostic approach, as sometimes introduced in the early stage of drug development, will help improve the feasibility of integrating precision dosing in patient care. In oncology, the use of companion diagnostics has already become a reality. Diagnostic tests are being incorporated into clinical trials as an attractive approach to better identify which patients may respond best to new treatments. Moreover, in ongoing efforts to expedite the availability of safe and effective cancer treatments for patients, the FDA is streamlining the review process of cancer diagnostics that are being developed in conjunction with clinical drug trials.16 In addition, we do not have simple tools for clinical interpretation of PK/PD data. We do have software packages that can help with interpretation of concentration time data, but they all require fairly laborious manual data input and a sophisticated level of training. What we need are electronic health record integrated decision support tools that automatically pull patient-specific data from the system and generate PK/PD model-informed recommendations via an intuitive graphical user interface—just one click away.17, 18 This would allow us to track the amount of drug and biomarkers in the body as a dynamic “molecular state,” in vivo, and in real time. In this respect, it is exciting to see the life-changing technological breakthroughs to treat type 1 diabetes and its complications. In June of this year, the FDA approved the first-ever implantable continuous glucose monitoring system.19 The device consists of a glucose biosensor with a transmitter that is implanted subcutaneously in the upper arm and a mobile app that displays concentration values and trends while issuing alerts on out of range glucose values.20, 21 Despite the advances is biosensor and nanotechnology, we do not have similar applications for the continuous tracking of drugs. However, the areas of biosensors and wearable technology are ever more converging and will likely soon deliver the first groundbreaking noninvasive system that could be used to measure and transmit real-time data in the precise management of drug exposure and response.22 So this time is different. We do have the tools to bring pharmacometrics and model-informed precision dosing directly to the bedside by integrating it into the electronic health record. The common approach, to date, has been to leave modeling tools in the hand of trained staff members who use it to give consults but reach only a minority of our patients. A more promising approach is to provide all the actionable information directly to the physician and clinical teams in a comprehensible format. To quote Daniel Kurnik, MD, Director of Clinical Pharmacology at Rambam Medical Center in Haifa Israel—“This is as revolutionary as making Wikipedia available to everybody, rather than keeping the Encyclopedia Britannica locked up in the library for the relatively few well-trained pharmacometricians.” The authors declared no conflict of interest.