A study was presented in which sarcomas were microinjected simultaneously with several drugs to study the pharmacodynamic response after resection. This platform may represent a future way of probing efficacy of anticancer agents in the relevant model system: human tumors.
See related article by Gundle et al., p. 3958
In this issue of Clinical Cancer Research, Gundle and colleagues (1), in the ongoing battle against cancer, microinjected sarcomas simultaneously with several drugs to study the pharmacodynamic response after resection. In the Netherlands, water has traditionally been our adversary. In this battle against water, already ongoing for many centuries, many small man-made streams have been made after land was gained from areas previously covered in water. In Friesland, a region in the northern part of the Netherlands, people developed a way to bridge these streams, which is called “fierljeppen”. By using a long wooden stick, placing it in the midst of the muddy water, people try to jump over the stream getting from one piece of land to the other site, while climbing the stick as it sinks in the mud.
In many ways, our attempts in oncology to translate preclinical findings to true advances in cancer care resemble this approach. At the one piece of land, we have preclinical models for cancer and discover many potential mechanisms we may target with novel drugs. Subsequently, we try to reach the other shore: the drugs with promising results in preclinical models are tested in patients where we aim to determine the toxicity, optimal dose, potential efficacy and the optimal target population in one or a few clinical trials. Disappointingly, this approach is failure prone: traditionally only 5%–10% of the drugs we test in phase I trials receive marketing approval. This implies that the preclinical models we currently have, often fail to adequately reflect the situation we face in cancer patients underlining the need for better methods to identify the novel anticancer drugs we should pursue in clinical studies.
Many new approaches have been suggested to close the gap between preclinical and clinical realities. Part of the solution could be to develop biomarkers to select those tumors that resemble the preclinical models, as for example was done for BRAF-mutant cancers and BRAF inhibitors (2). And indeed, this approach sometimes works, but not always, while for many other agents including chemotherapeutic agents the exact targets are not known rendering selection based on molecular characteristics of tumor cells impossible. Also new methods, such as tumoroids, have their promises but for example lack the microenvironment of cancer, which appears to be so important in drug sensitivity, and represent only a fraction of the cancer cells the tumoroids are derived from, leaving us with a remaining translational divide.
Gundle and colleagues present a paper on a novel method called Comparative In Vivo Oncology (CIVO) to bridge this gap by using actual human tumors in the context of actual human patients (1). They inject microdoses of drugs into tumors of patients who are scheduled to undergo a resection or an incisional biopsy and deliver a technological framework to study the pharmacodynamic (PD) effects of these drugs in situ after resection of the injected tumor. Using their technology, they can test up to 8 drugs and/or drug combinations simultaneously for different types of effects.
Obviously, if we can measure pharmacodynamics responses in vivo in human tumors we may have a better estimation of what we may expect from novel treatments (Fig. 1). And maybe even at an individual level, this method could serve as a tool for true personalized cancer treatment. In addition, if scaled up enough, we could envision a much more efficient testing of drug combinations, enabling a deciphering of complementary pathways as different combinations can be tested in a matrix format in each tumor to uncover unique vulnerabilities. Given the large numbers of drugs we are currently testing, this would be a true improvement.
The presented technology of CIVO, however, needs some improvements as injecting therapeutic agents into a tumor presents some important conceptual issues. We need to account for a relation between dose and effect (3). This so called pharmacokinetic (PK) dimension is difficult to capture in this CIVO platform. A single injection will generate a range of doses being delivered to cells at different distances from the injection site. Thus, more precise determination of the dose–area relationship is important for further development. Moreover, many anti-cancer drugs undergo extensive and fast metabolism into active metabolites rendering these drugs less suitable for testing in this setting. Direct injection of drugs also bypasses some of the barriers certain drugs experience before being delivered into the tumor microenvironment and this could lead to overestimation of drug effects. In addition, CIVO in the study presented here was applied in patients with sarcoma, which are often located superficially, while most tumors are in general more difficult to reach for such invasive procedures. Finally, spatial heterogeneity is a given fact in cancer, especially in the larger tumors that are used for these types of research. Consequently, the effects in other parts of the tumor or in other lesions in the same patient can substantially differ from effects seen in the injected part of tumor. So all together, the study presented here is of high interest and should be considered a vignette to show where we could go, but several issues need to be clarified before we may truly rely on CIVO to guide future drug development.
Could we envision a future where CIVO could be used for more personalized treatment approaches, to start with in patients with advanced sarcoma? Personalized phenotyping is the holy grail many of us are looking for. Establishing individualized xenografts or organoid cultures to predict tumor response in an individual is such an approach, but still seems to be too cumbersome for broad application. One of the main disadvantages of both approaches is the high rate of failed cultures: if not most tumors can be tested, these platforms may proof to be less relevant (4). Using actual human tumors now provides a major advantage: in principle, data may be obtained for all tumors injected. However, to start using CIVO as predictor, we need much more data on how the responses seen in the injected tumor tissue correspond to later therapeutic efficacy observed in the same patient.
Whereas “fierljeppen” results in wet pants for those of us who are less experienced, current oncology drug development is using precious resources for ineffective trials. For the “wet pants problem,” we have found useful solutions such as bridges. Likewise, now it is time to improve the use of our precious resources in clinical trial execution in oncology. CIVO together with other methods of using relevant human biomarkers could be such a bridge and could become the real frontline for oncology drug development. And most importantly, biomarkers such as CIVO could allow us to center the drug development paradigm to where it should focus: on the patient. For the past decade, we have focused on finding patients for our trials, using advances like CIVO we should try to turn back to trying to find a treatment for our patients, also in trials.
Disclosure of Potential Conflicts of Interest
M. Lolkema reports grants from MSD, grants and personal fees from Johnson & Johnson, Sanofi, and Astellas and personal fees from Bayer, Pfizer, Amgen, Roche, Novartis, and Julius Clinical. No potential conflicts of interest were disclosed by the other author.