The Role of Pathologists in the Era of Personalized Medicine

Abstract

Although the field of oncology has made significant progress in improving the duration and quality of life for patients with cancer, response rates to traditional chemotherapeutic agents remain unacceptably low when compared with response rates in other disease areas.1 This situation, combined with the disturbing trend of decreasing annual new drug applications despite increasing drug development spending,2 has accelerated the adoption of the personalized medicine concept in several areas of medicine but most significantly in oncology. In this field, in the past 5 to 10 years, the terms personalized medicine, biomarkers, and translational medicine have gone from being hypothetical concepts to being widespread actualities in drug development, basic and clinical research, and clinical practice. Once obscure points of discussion at cancer conferences, these topics now dominate many congresses and have spawned dozens of biomarker meetings around the world each year.In its simplest form, personalized medicine seeks to do away with the cancer treatment paradigms of “one drug fits all” and “trial and error,” replacing them with a model centered on the pairing of molecularly classified cancer subgroups with new generations of chemotherapeutic molecules targeting the specific pathogenic mechanisms that drive the respective neoplastic processes in each subgroup. The objective is to administer the right drug to the right patient at the right time and at the right dose. This model is entirely dependent, however, on a new generation of cancer tests that accurately and comprehensively characterize patients' tumors at the DNA, RNA, and protein levels, thereby allowing physicians to identify likely responders before beginning treatment.These tests have been designated companion diagnostics to acknowledge that, in the personalized medicine model, the pharmaceutical compound and the diagnostic assay become inexorably linked. That is, without the drug, there is no need for the companion diagnostic; and perhaps more conceptually provocative, without the companion diagnostic, there may be no utility for the drug. Thus, while targeted therapies often demonstrate dramatic responses in diagnostic-driven patient subsets, the therapy's efficacy in unselected patient cohorts is typically too low for regulatory approval. The US Food and Drug Administration has published a guidance document in draft form on the codevelopment of drugs and diagnostics.3 A major point of emphasis in that document is the need to begin development of the diagnostic assay early in the drug development process, ideally in the preclinical stage, to best ensure that a robust assay can be approved by the US Food and Drug Administration and be commercialized in conjunction with the pharmaceutical compound.If one considers the large and growing number of molecularly targeted therapies that are either available commercially or in development,4 as well as the enormous amount of ongoing translational research in oncology, it becomes clear that the personalized medicine concept is not going to be a short-lived trend but rather a true paradigm shift and a new way of diagnosing and treating cancer as a disease.At the same time, the personalized medicine concept faces a number of significant challenges before its potential can be fully realized. At the scientific level, although our knowledge surrounding the molecular underpinnings of cancer continues to expand at a rapid pace, we are a long way from having a complete understanding of how cancer cells are driven, how they evolve over time, how they interact with the host microenvironment and how they react to therapeutic intervention. In drug development, we face the challenges of developing therapies against targets we do not completely understand and of selecting optimal trial designs, clinical endpoints, and biomarker strategies. In the clinic, we need to deal with the reality of acquired drug resistance, even in patients who have had dramatic initial responses, and with a previously underestimated degree of intrapatient and intratumoral heterogeneity. Although the combination of targeted therapies appears to be one solution to the latter challenges, clinical trials that include adaptive designs and combinatorial approaches have only recently been initiated, and it remains to be seen whether toxicity will be a limiting factor. At the logistic level, sample collection, especially tissue collection, continues to be challenging in the context of clinical trials. This is due to cost, patient enrollment impact, preanalytic variables, and lack of standardization. Given that the entire concept of personalized medicine is predicated on the interdependence of drugs and diagnostics, how will each of these critical components be valued, priced, and reimbursed, so they are widely available and best serve the needs of patients?The solutions to these issues will depend on a coordinated approach from a variety of disciplines, including oncology, molecular biology, and certainly, pathology.As personalized medicine takes hold and revolutionizes the way we diagnose, characterize, and treat cancer, there is a tremendous opportunity for pathologists to make major contributions to this emerging new paradigm and to dramatically improve the care of patients with cancer. Some significant changes, however, need to occur within the field of pathology for this to become a reality. Pathology, as a field, has been occupied primarily with the diagnosis, classification, and subclassification of disease. Certainly, pathologists also provide prognostic information in the form of histologic grade, and in a small number of cases, they provide information, such as HER2 status, that is directly related to therapy. For the most part, however, pathologists have not played an active role in driving forward personalized medicine. Pathology, as a field, needs to move beyond the realm of diagnosis and mere classification/subclassification of diseases, to become the field responsible for providing personalized medicine information. Pathologists need to be the experts in the medical system who answer the following questions for patients and their treating physicians:If pathologists do not embrace personalized medicine, other fields and various nonpathology biotechnology companies will provide answers to these questions, but they will do so in the absence of morphologic data, ultimately generating information that will be inherently of less value. In a recent article, Thomas Giordano, MD, outlines what he calls personalized predictive pathology and appropriately asks, “Will molecular profiling become part of the routine pathologic assessment of cancers performed by surgical pathologists, or will it become the domain of subspecialty labs, either academic or commercial, with surgical pathologists observing from the sidelines?”5(p402) He adds, “Prediction of therapeutic response by molecular profiling is the logical and natural extension of the work of surgical pathologists.”5(p403)Pathologists are the only professionals in the health care/scientific system that can interpret genomic, gene expression, and proteomic data in the context of tumor morphology. Only the pathologist can comment on the presence, absence, and differential expression of biomarkers in tumor cells versus normal cells, in in situ tumors versus invasive tumors, and in different grades and patterns of a tumor present within one sample. Only the pathologist can provide a comprehensive assessment of companion diagnostics and other biomarkers on a per patient, per clone, and per cell basis. The correlation of biomarker data with traditional histomorphology lies exclusively in the realm of the anatomic surgical pathologist and holds the promise of unlocking crucial insights into the biology of tumors and, therefore, into the most rational and effective therapeutic strategies.Pathologists will certainly require novel tools and technologies to enable this promise. For decades, pathologists have had the ability to correlate the expression of proteins with histomorphologic patterns by using immunohistochemical methods. Recently, because of the advent of light microscopy-based in situ hybridization methods, gene amplification has been added to the brightfield repertoire of the surgical pathologist, allowing pathologists to visualize individual copies of genes and amplification states in the context of morphology. Unlike fluorescence in situ hybridization, these technologies use traditional counterstains and generate high-quality cellular morphology, allowing for distinction between different aspects of the specimen, such as invasive and in situ tumor components. This type of technology is a critical advance for the field of pathology because, for the first time, pathologists are able to generate genomic profiles of patients' tumors that incorporate histologic pattern information. For example, in a breast cancer case that contains both an invasive ductal and an invasive lobular component, pathologists can use these novel technologies to observe separately the HER2 gene status of the 2 components. Several studies have investigated the frequency and extent of heterogeneity in breast cancer,67 and it will be important for pathologists to continue to further this type of work by correlating intratumoral biomarker heterogeneity with differential clinical responses to targeted therapies.Although the visualization of amplification status in the context of histomorphology represents a major advance in the field of pathology, it is only the first of many types of molecular genomic technologies that pathologists will need to acquire. As we expand our knowledge of the differences between cancer cells and their normal counterparts at the DNA, RNA, and protein levels, it is becoming clear that the characterization of tumors as part of the personalized medicine paradigm will need to occur in a comprehensive manner and will not be possible with single markers or technologies. The evolving picture of personalized medicine for non–small cell lung cancer is a recent example supporting this notion.To varying degrees, numerous studies have demonstrated the usefulness of epidermal growth factor receptor (EGFR) mutations, copy number/amplification status, and to a lesser extent, protein expression by immunohistochemistry to predict responses to EGFR-specific therapies, such as erlotinib (Tarceva, Genentech, South San Francisco, California and OSI Pharmaceuticals, Melville, New York) and gefitinib (Iressa, AstraZeneca Pharmaceuticals, Wilmington, Delaware), in patients with advanced non–small cell lung cancer.8–10 Although not perfect, these markers can be used to enrich patient cohorts for responders, and in some cases, the clinical responses in patients with metastatic disease have been dramatic. One notable patient subset that is immediately resistant to this approach (ie, primary resistance) is that of patients harboring mutations in the KRAS gene. Activating mutations in this parallel pathway serves to bypass EGFR inhibition and to maintain downstream signaling through growth and proliferation pathways.Unfortunately, even the group of patients with non– small cell lung cancer who initially respond to EGFR inhibitors inevitably relapse after a period of 2 to 3 years because of drug resistance.11 The mechanisms underlying this secondary or acquired resistance are being elucidated and are providing insight into the future of personalized medicine and the types of diagnostic and therapeutic approaches that will be necessary to overcome them. One mechanism of acquired resistance in this setting is the emergence of second-site mutations in the EGFR gene, such as the T790M point mutation, which renders first-generation EGFR tyrosine-kinase inhibitors less effective competitors for adenosine triphosphate.1213 A second mechanism involves the amplification of the MET gene, which along with ERBB3/HER3, serves to bypass the inhibition of EGFR via the phosphatidylinositol-3-kinase pathway.1415 Finally, a third mechanism of acquired resistance to EGFR tyrosine kinase inhibitors elucidated in A431 cell lines is the activation of the IGF1R pathway via IRS-1.16This example of acquired resistance via multiple mechanisms is being replicated with other targeted therapies in other indications, and suggests that therapeutic strategies incorporating combinations of pathway-specific targeted molecules will be required to overcome the redundancy and parallel nature of aberrant growth signaling in cancer. At the same time, if pathologists are to have a significant role in personalized medicine, they will need to have access to in situ and tissue-based diagnostic tools that span the gamut of genetic derangements in cancer. In addition to protein expression and gene copy number, pathologists may require tools that allow them to visualize point mutations, insertions, deletions, translocations, mRNA expression, methylation, and other epigenetic events, to name a few. The in situ detection of micro-RNAs has already been demonstrated and represents a positive development in the expansion of the pathologist's tool kit.17In the era of personalized medicine, the practice of pathology will need to undergo some changes to remain aligned with the needs of oncologists and their patients. Traditional histomorphology and morphologic diagnosis will not disappear but, rather, will take on more importance as pathologists expand their ability to comprehensively profile patients' tumors with next-generation, in situ methods. The pathologist's role will be to integrate the data from these diverse technologies into a coherent stream of information for patients and clinicians (Figure 1). The pathology report of the future will contain not only diagnostic and prognostic information but also critical predictive information pertaining to which drugs will be effective in a particular patient. An interesting opportunity for pathologists that emerges from these changes is that of direct patient interaction. As the amount and complexity of data that directly affect the care of the patient increasingly falls under the purview of pathologists, patients are more frequently seeking the expertise of pathologists to best understand their diagnoses, tumor profiles, and resulting implications for treatment. Rather than infringing on the responsibilities of oncologists, pathologists who take on this patient-facing role can be seen as providing information that is complementary to, and synergistic with, the discussions that oncologists have with patients. Some pathology departments have begun to experiment with making pathologists available to patients a few hours a week for consultations with very positive results. Other pathologists have organized educational programs for cancer patients that focus on the interpretation of their surgical pathology reports.1819Pathology-training programs also need to change to ensure that pathology residents and fellows are aware of the developments taking place in personalized medicine. As the links between drugs and diagnostics continue to strengthen, it becomes increasingly important for pathologists to be knowledgeable about, or at least exposed to, developments in clinical oncology, molecular biology, and translational science. Developments in these areas have direct implications for the diagnosis and molecular profiling of cancer and, therefore, for pathology. According to Dr Giordano, “Most pathology training programs are not adequately preparing their trainees in molecular profiling.”5(p403) This deficiency needs to be addressed if we are to ensure that new pathologists are with the and knowledge necessary for them to be active in the era of personalized medicine. are in which pathology residents can to their to these the annual organized by the and of Pathology, of and of Pathology, pathology trainees can meetings organized by the for and the of the developments in cancer research and translational are and is this of an annual on and that are the field of pathology. trainees can also the and research to personalized medicine by such as New of of and have important contributions to make in the practice of personalized medicine, including its oncology and companion diagnostics (Figure and clinical pathologists are increasingly being by pharmaceutical and diagnostic companies because these have the importance of having in the development of oncology drugs and companion diagnostics (Figure the pharmaceutical pathologists are in of drug development, including preclinical research, translational medicine, and clinical drug The most for pathologists are in the areas of research and translational in preclinical research are best to pathologists who have either an of medicine or of degree or an in basic This work typically involves the use of cell or to the of novel compound the and of biomarkers, as well as to understand the molecular of therapeutic targets and, therefore, potential clinical typically are and include of a pathologists will also in translational within the pharmaceutical the traditional research and development that have at pharmaceutical having a group or that work at the of research and development has now become practice. are various such as and and have a pathologist as either a or the group The translational medicine is to the of novel basic and scientific knowledge into clinical going from to companies now that it is no or for basic to work in from their clinical development a compound in the and it over the to clinical for clinical trials and development is not an approach that will work in the era of targeted and personalized medicine. research and development work from the very beginning of the drug process, and in many pharmaceutical translational medicine are this Pathologists who work in these types of will typically work with both research and clinical development in developing and biomarker strategies for several or a of For example, with early (ie, or this may include an immunohistochemical assay to a protein and to For a the objective may be to an assay that is predictive of drug response (ie, a companion may include with as well as work to and diagnostic may also in the clinical trial or the especially for that deal with sample and biomarker and The of tissue as a necessary of biomarker development, is becoming in both early and clinical trials. In various clinical trial the pathologist critical expertise the subclassification of tumors, with histologic and the of different types of or sample can have a significant on the quality and of clinical trial The of high-quality tissue has to be very challenging in large clinical trials. For example, in the 3 trial for the drug erlotinib and OSI only of patients had for and only of patients had tissue available for or fluorescence in situ hybridization in the clinical development of drug development are not typically available to pathologists without a pathologist has in one of the if he or may become a clinical the development of early or therapeutic the diagnostics pathologists can to the entire life of a diagnostic including development, clinical regulatory and The specific available to a pathologist will depend on the focus of the diagnostic with the number of being at companies that have tissue-based diagnostic include pathologist, medical and medical the a pathologist may be responsible for and the medical and scientific of diagnostic The pathologist may also or in to various novel or technologies or the of a novel assay Pathologists can provide critical this to ensure that these 2 processes occur in the That is, new diagnostic begin with the of a medical or an of medical which in the for technologies. this can to of and regulatory the development pathologists can to the and of and diagnostic this are to ensure that the diagnostic the required and needs and an of and The typically studies which the pathologist may interact with pathologists or other physicians who are as the primary for these can also an important role in the regulatory of a diagnostic to or the medical of and clinical diagnostic trials that will serve as the for by the US Food and Drug Administration and other regulatory In pathologists can a role and after the of the diagnostic by to the as a and to the and and with pathologists and other physicians to to a Pathologists may also be responsible for and with to understand the potential medical of that are as only or to understand the medical need for new This type of role can be to pathologists who to maintain academic including and data from their with traditional in a or the role of a pathologist in either the pharmaceutical or diagnostic is different in and time Rather than and of individual patient cases, a pathologist in to the development of a therapeutic a diagnostic or Rather than as an individual in a a pathologist in has a that with in a and often in a Rather than being the in a pathologists in need to become to in a and in which they will have a direct of in or perhaps pathologists in work on that typically take or years to and in some cases, they will work on that ultimately do not make it to in frequently require and to of the time or more in for pathologists in is typically in the form of a an annual with both and individual and an that may include or typically an initial in when they a however, the potential in can that of clinical practice Although some pathologists may these differences as may them as differences certainly be before a pathologist the of a in Pathologists who have made the from clinical practice to a in a in related to the absence of of medical light of an ongoing in the understanding of cancer and the approaches necessary to treat molecularly patient subsets, the field of pathology at a For the field of pathology to remain in an era of personalized medicine, pathologists need to become the of information, critical answers to questions not only to diagnosis but also to and response driving the clinical practice of molecular predictive pathology and taking on new in pharmaceutical and diagnostics development, pathologists have the opportunity to and tools that allow the in situ characterization of patient tumors in a comprehensive way that directly specific treatment these concepts and molecular genomic data in the context of traditional pathologists will a role in the era of personalized medicine and will directly improve the of patients with cancer.