DeVita. Cancer

189

Chapter 15  Precision Medicine in Oncology

helps to define the drug-induced pharmacologic effect associated with the therapeutic activity of the agent; however, a PD effect does not, by itself, ensure patient benefit because biochemical events downstream of the drug target may impair tumor cell kill- ing. Thus, PD biomarkers may or may not be predictive of patient benefit (or toxicity) from a specific intervention. Although the de- velopment of predictive biomarkers is a critical goal of precision oncology, the first step in the process is establishing molecular proof-of-mechanism (PoM) in human tumors during early-phase clinical trials. 58 It is in later stage trials that definitive relationships between biomarker modulation and clinical response are estab- lished, so-called proof-of-concept (PoC) studies. Although the need to clarify the dose range and schedule that correlates with a molecular response in a tumor (rather than simply with pharmacokinetics in the circulation or normal tissue toxicity) seems clear, the development of robust assays for PD assessment has lagged behind the development of clinical DNA sequencing for cancer. 59 This may be due to the difficulties inher- ent in generating a variety of assays to examine a broad range of biochemical targets using many different analytical platforms with- out the availability of high-quality reagents. 60 Developing novel, mechanistically based combination therapies targeting multiple signaling defects has been enormously challenging in the absence of establishing the doses of the agents to be used in combination that are needed to produce the degree of target inhibition required for optimal tumor cell killing. 61 Furthermore, failure to confirm PoM using appropriate PD assays in early-phase clinical studies has resulted in costly failures of phase III clinical trials. 62 If validated PD assays are available for characterization of spe- cific molecular species, important insights into drug action can be detailed from clinical trials that involve small numbers of pa- tients; such studies can lead to the confirmation (or rejection) of prevailing hypotheses regarding the PoM of specific drug classes. 63 Recent technologic improvements in microscopy, dye technology, and image processing now permit the quantitative assessment of multiple aspects of pathway engagement following drug treatment across a spatial context using formalin-fixed, paraffin-embedded biopsy tissues. Such quantitative immunofluorescence techniques are capable of monitoring heterogeneous, individual cell-to-cell changes in tumoral DNA damage response following exposure to DNA-damaging agents, for example. 64 Improvements in proteom- ics now also support high-throughput evaluation of reverse-phase protein arrays, permitting the development of proteomic response signatures associated with drug treatment in vivo. 65 It is likely that these new approaches will expand the use of PD to a wider range of signaling pathways and will broadly enhance understanding of the molecular mechanisms of drug action in patients.

approach to cancer therapeutics. The lack of human immune components in the PDX tumor microenvironment not only limits the evaluation of immunotherapeutic agents in these models but also prohibits examination of the role human microenvironmen- tal elements play in the control of tumor growth by molecularly targeted drugs. It is also clear that as the PDX passage number in- creases in vivo, clonal populations not predominant in the original tumor may be selected for outgrowth, affecting gene copy number, mutational spectrum, and the antiproliferative activity of therapy. 50 In addition to the advent of PDX models to guide the use of molecular characterization of human tumors, new methods for three-dimensional in vitro culture of malignant cells (from surgi- cal or biopsy samples) have been developed, facilitating tumor cell proliferation with a rich matrix of essential growth factors; such complex organotypic (organoid) cultures retain many phenotypic and genotypic characteristics of cancer stem cells from primary or metastatic tumors and thus may provide a closer approximation of human tumor biology than prior two-dimensional cell culture approaches (see Fig. 15.2). 51,52 Organoid cultures can also be used for drug screening and for the propagation of PDX models, provid- ing a means to rapidly move back and forth between low-passage in vitro and in vivo models originating from the same clinically and molecularly annotated surgical or biopsy specimen. 53 As the range of tumors amenable to organoid culture expands (from colon, pancreatic, prostate, and breast cancers) to other common malignancies, it is likely that these model systems will accelerate understanding of the molecular interactions that underlie the spa- tial context of therapeutic response.

ROLE OF MOLECULAR PHARMACODYNAMICS AND DIAGNOSTICS IN PRECISION ONCOLOGY Human Biospecimens for Molecular Characterization

CANCERTHERAPEUTICS

Molecular characterization of human tumors, whether it is for DNA or RNA sequencing, epigenetic studies, proteomics, or im- munohistochemistry, starts with the collection of biospecimens of high quality. 54 Although research biopsy techniques are improving, national standards for specimen acquisition, processing, fixation, and storage have not been universally implemented, which hin- ders the comparability of results obtained across multiple research sites. 55 It is not infrequent, particularly for radiologically guided needle biopsies, that the percentage of tumor in the specimen ob- tained may be  5%, possibly sufficient for pathologic diagnosis but often inadequate for any protein-based molecular assay. 36 Fur- thermore, the utility of the assay to be performed may be dimin- ished by insufficient attention to stabilizing the analyte of interest immediately after acquiring the specimen. Unfortunately, tech- niques to stabilize biomarkers are infrequently examined before biomarker studies are initiated. 56,57 Attention to the details of tissue acquisition, either tumor or normal tissues, is fundamental to the success of the molecular characterization efforts that are at the heart of a precision approach to oncology.

Molecular Pharmacodynamics in Precision Oncology Evaluating therapeutic regimens for cancer based on the degree of their target engagement and the downstream molecular events initiated by a drug-target interaction requires pharmacodynamic (PD) biomarkers: molecular species that are altered in response to an oncologic drug or treatment and that, if measured, can be used to establish the mechanism of action of the drug. A PD biomarker Predictive Diagnostic Assays Although PoM studies are critical first steps in the precision ap- proach to cancer therapeutics, target engagement does not neces- sarily confer therapeutic efficacy, in part because of the remarkable redundancy of growth-controlling signal transduction pathways in human tumors. 66 Furthermore, other than the predictive power of mutational profiles, only a modest number of predictive biomark- ers for cancer have been deployed successfully in the clinic. Such biomarkers need to provide accurate correlations with a clinical outcome following specific treatment interventions, unlike prog- nostic biomarkers that convey associations with clinical benefit, appropriate for populations that are not unequivocally related to the treatment’s mechanism of action. Well-known predictive bio- markers include the use of HER2 expression to select breast can- cer patients who may benefit from trastuzumab; the Oncotype DX and MammaPrint assays to establish the relative risk/benefit of ad- juvant chemotherapy in the setting of early-stage breast cancer 67,68 ; and the detection of mutant KRAS , which predicts for the lack of efficacy of cetuximab in colon cancer. 69 Copyright © 2018 Wolters Kluwer H alth, Inc. Unauthorized reproduction of the article is prohibited.