Skip Navigation LinksHome > January 2014 - Volume 22 - Issue 1 > Practicing Pathology in the Era of Big Data and Personalized...
Text sizing:
A
A
A
Applied Immunohistochemistry & Molecular Morphology:
doi: 10.1097/PAI.0000000000000022
Review Article

Practicing Pathology in the Era of Big Data and Personalized Medicine

Gu, Jiang MD, PhD*; Taylor, Clive R. MD, D Phil

Free Access
Article Outline
Collapse Box

Author Information

*Department of Pathology, Shantou University Medical College, Shantou, Guangdong, China

Department of Pathology, Keck School of Medicine, University of Southern California, Los Angeles, CA

The authors declare no conflict of interest.

Reprints: Clive R. Taylor, MD, D Phil, Department of Pathology, HMR 311, Keck School of Medicine, University of Southern California, 2011, Zonal Avenue, Los Angeles, CA 90033 (e-mail: clive.taylor@med.usc.edu).

Received October 31, 2013

Accepted October 31, 2013

Collapse Box

Abstract

The traditional task of the pathologist is to assist physicians in making the correct diagnosis of diseases at the earliest possible stage to effectuate the optimal treatment strategy for each individual patient. In this respect surgical pathology (the traditional tissue diagnosis) is but a tool. It is not, of itself, the purpose of pathology practice; and change is in the air. This January 2014 issue of Applied Immunohistochemistry and Molecular Morphology (AIMM) embraces that change by the incorporation of the agenda and content of the journal Diagnostic Molecular Morphology (DMP). Over a decade ago AIMM introduced and promoted the concept of “molecular morphology,” and has sought to publish molecular studies that correlate with the morphologic features that continue to define cancer and many diseases. That intent is now reinforced and extended by the merger with DMP, as a logical and timely response to the growing impact of a wide range of genetic and molecular technologies that are beginning to reshape the way in which pathology is practiced. The use of molecular and genomic techniques already demonstrates clear value in the diagnosis of disease, with treatment tailored specifically to individual patients. Personalized medicine is the future, and personalized medicine demands personalized pathology. The need for integration of the flood of new molecular data, with surgical pathology, digital pathology, and the full range of pathology data in the electronic medical record has never been greater. This review describes the possible impact of these pressures upon the discipline of pathology, and examines possible outcomes. There is a sense of excitement and adventure. Active adaption and innovation are required. The new AIMM, incorporating DMP, seeks to position itself for a central role in this process.
Back to Top | Article Outline

THE COMING ERA OF “BIG DATA” AND PERSONALIZED MEDICINE

The last decade has witnessed remarkable progress in the area of genomic research that will, in turn, have a direct impact on pathology and pathologic practice. The progress of pathology as a discipline has always been propelled by inventions of technologies. Indeed, modern pathology started with the introduction of the light microscope into medical practice. Subsequently the techniques of electron microscopy, histochemistry, immunohistochemistry, monoclonal antibody production, antigen retrieval, and image analysis, to mention but a few, have all played their roles in elevating pathology to a higher level.1–3 The current waves of advancement in molecular technology and data processing have made it possible to detect and analyze large amounts of pathology-related data, once more presenting pathologists with opportunities and challenges that call for action.

Application of these new techniques has shifted the understanding of disease, particularly cancer, to the molecular level, challenging pathologists to correlate these findings with existing morphologic methods, that in many instances still provide the “gold standard” for cancer diagnosis. The definitions and classifications of many diseases, particularly cancer, are being redefined, almost on a daily basis (Fig. 1). Interestingly, the data derived from these new techniques are both “large” and “small.” “Large,” in the context of the sheer volume of data generated by these new techniques, amounts that are often in the gigabyte and terabyte range, and are getting still bigger, lending the name “big data” to capture this phenomenon. “Small,” in context of the actual pathologic changes that are detected, which are at the nucleotide or protein or peptide and single-molecule level, and getting smaller, coining the name “precision medicine.” Furthermore, the myriad of changes that are detected, increasingly are being assigned to individual cancers in individual patients, a capability giving rise to the term “personalized medicine.” For the pathologist to remain relevant requires, in turn, a personalized approach to pathology; tacit recognition of the fact that each cancer in each patient has a unique molecular (genetic) signature, and that previous morphologic subtypes, for example adenocarcinoma of lung, or breast, or colon, alone no longer are sufficient for selection of therapy (Fig. 1).

FIGURE 1
FIGURE 1
Image Tools
Back to Top | Article Outline

TOWARD PERSONALIZED PATHOLOGY

This change of paradigm in medicine results from the convergence of 2 initially independent areas of technology that have been gathering momentum following 2 of the most impressive technical advances in recent history. One is the remarkable development of DNA and protein detecting technologies, with high-throughput machines that can detect minute and precise changes at the molecular level in enormous number and variety, and at ever increasing speed. These methods include DNA sequencing, microarray technologies, comparative genomic hybridization, digital PCR, mass spectrometry, etc., each of which generates data in enormous quantities, in amounts never seen before, justifying the term “big data.” The other is the tremendous increase in capability for managing these data, quantum leaps in the capacity for storage, processing and transmission, afforded by exponential growth in computer power, internet bandwidth, and so called “cloud computing.” The former generates “big data” and the latter enables the processing of “big data.” Even more remarkably these increases in capability have been accompanied by marked reduction in cost, leading to widespread application and an entirely new concept of medical practice, characterized by Prediction, Personalization, Prevention, and Patient participation, the so-called 4P medicine.6

The human genome project consumed 3 billion dollars and 13 years, and was completed in 2003. It launched the era of genomic medicine, but it illustrated only the protein coding portion of the DNA, representing just 1% to 2% of the human genome. The subsequent ENCODE (The Encyclopedia of DNA Elements) project took 9 years with the involvement of over 400 scientists. It was completed in 2012, serving to identify many of the remaining sequences in the genome as regulatory factors, instead of “garbage genes.”7 Following these enormous efforts, it became possible to begin to see the complete picture of the human genome that governs all of human anatomy and function, and is disrupted by disease. The extraordinary complexity of gene expression, regulation, and interaction has been clearly demonstrated. Medicine and pathology are just at the beginning of understanding these intricate mechanisms and harnessing the knowledge for diagnosis, and for therapy.

Personalized medicine encompasses the concept and the intent of finding the uniqueness in each patient’s pathology, physiology, and “pathophysiology,” to tailor treatment and prevention plans to suit each individual’s condition for optimal results. This concept was made possible by dramatic increases in the capability for collecting and analyzing the tremendous amounts of medical data that modern medicine almost routinely accumulates about each patient. In addition to genetic and genomic data, there may be information on proteomics, epigenomics, lipidomics, metabolomics, transcriptomics, etc. (collectively called “omics”), and electronic records of pathology reports and laboratory tests, not to mention the 2 to 5 GB megafiles of whole slide images (WSIs), plus physical examinations and treatment regimens. Together these records form a comprehensive, but huge, database that, given appropriate methods for access and correlation, allows for detailed and continuous analysis of the patient’s physical condition, from which personalized strategies can be devised to achieve a better outcome. One example of the application of “personalized medicine” is to be found in a pilot study performed by Dr Michael Snyder and colleagues of Stanford University Medical School, in which they followed an individual for 18 months with extensive testing of genomic, proteomic, and metabolomic parameters, accompanied by extensive computer analysis. Onset of an episode of diabetes and 2 viral infections were identified, providing new insights that would not otherwise have been appreciated by the physicians caring for the patient. The result was early intervention, reversing the course of diabetes with change of life style and specifically directed treatment.8

Back to Top | Article Outline

THE IMPACT OF “BIG DATA” ON PATHOLOGY PRACTICE

The ability to generate, analyze, interpret, and store huge amounts of data inevitably is changing the platform on which pathologists perform their duties and deliver their services, with “informatics” now a recognizable subdiscipline, having a flourishing association (http://www.pathologyinformatics.org) and recognized training fellowships. However, for pathologists the challenge extends beyond assimilation of the data generated by genomics or proteomics testing into the area of anatomic pathology and into the entire electronic medical record. Already an excess of 10% of the medical record is comprised of pathology and laboratory results, with the volume of data about to increase dramatically upon introduction of digital slides, image analysis, and telepathology.3,9 With a single digital WSI requiring a file size in the 2 to 5 GB range (about 10 times the size of a radiology whole body CT file) the era of “big data” in pathology truly has arrived, and pathology will not be the same again.10–13

Back to Top | Article Outline

THE NEW MOLECULAR TECHNOLOGY IS A POWERFUL ADDITION TO HISTOPATHOLOGY

In 2014 obtaining a “tissue diagnosis” still remains the cornerstone of diagnosis for cancer. However, while the microscopic examination of a tissue section by a skilled pathologist has survived the test of time, more than 150 years, a histopathologic diagnosis has inherent limitations of subjectivity and precision. On the one hand, these limitations have been exposed by newly emerged genetic, molecular, and digital technologies, but on the other they may also be minimized or repaired by the supplementary information that these methods bring. The histopathologic changes that characterize cancer are a result of long-term gradual alterations of gene expression and protein structure that accumulate over months or years. When the structure and appearances of a cell or groups of cells have changed to the extent that can be visualized with a light microscope, disruption of the cellular pathways and control mechanisms may be extensive, and to a large degree irreversible. In this context the abnormalities identified by light microscopy can no longer be regarded as “early.” In theory in situ hybridization and immunohistochemistry can detect cumulative changes of nucleotide sequences (including mutations) or proteins indicative of changes in cellular pathways that antedate typical morphologic changes. However, the analytical sensitivity of these techniques precludes identification of molecular changes smaller than about 200 nucleotides or a few dozen antigenic epitopes. By these methods cellular localization is precise, but measurement of the real amount of such changes is semiquantitative at best, pending the arrival of true quantification,14 and single-molecule detection that already is on the near horizon. Compared with changes at the nucleotide or protein levels, the current evaluation of morphologic changes by pathologists with a microscope seems rough and imprecise. On the one hand, many clinically significant molecular abnormalities do not lead to detectable morphologic changes. On the other, morphologic changes that appear identical to the surgical pathologist may harbor more than one, and often many, changes at the molecular level. This new discovery is well exemplified by a single-tumor type, adenocarcinoma of lung, that has been found to manifest numerous different mutations, many with different therapeutic import, but with no morphologic clue as to which is which, or what the treatment should be (Fig. 1). There is a dawning realization that morphologically visible changes, the basis of surgical pathology for more than a century, are predictive only of a general need for treatment, but give little guidance as to which of the new targeted therapies applies. Similar conclusions may be drawn for other (almost all) major cancer types (Fig. 1).

Molecular techniques offer some advantages, but are still evolving. Sequencing and array technologies can generate data in minute detail and large quantity, and may reveal the reasons underlying changes in morphologic appearance. New techniques can detect mutations, insertions, deletions, and switches of single or segments of nucleotide sequences,15 working at the level of specific genes, exons, or the entire genome.15–17 It is also possible to detect alterations of mRNA, small RNA, long noncoding DNA, and DNA methylation patterns that may eventually lead to morphologic changes that are not identifiable with current histopathologic means,16 thus offering possibilities of earlier diagnosis. Last, and not least, changes in DNA, RNA. or protein may also identify molecules that have the potential to serve as targets for new drugs. This possibility of targeted therapies and “companion diagnostics” is currently a powerful driving force in the pharmaceutical industry.

However, there are attendant disadvantages. The various machines tend to be expensive, with rapid obsolescence, plus performing these tests without having in place the necessary data processing and informatics expertise is pointless. Also molecular methods (sequencing, PCR, mass spectrometry) typically are performed using tissue extracts, in which the exact nature and number of contributing (cancer) cells is only inferred from examination of adjacent tissue section. Because of this dissociation from morphologic features, it may be difficult to separate the signatures of abnormal cells from admixed normal elements, and the practical detection sensitivity for any mutation is limited to samples containing at least 10% of (tumor) cells having that particular mutation.

With current approaches, it is hard to capture the initial earliest events in the development of neoplasm, not least because these changes frequently occur sometime before the lesion becomes clinically evident, as discussed below. Thus, the dynamics and continuous nature of disease progression are difficult to appreciate. Although examination of a formal biopsy of tissue continues to be an important part of conventional pathology, other less invasive approaches offer the potential to evaluate pathologic changes earlier, and also continuously during the clinical course. Repeated targeted needle biopsies meet this latter need in part, but are often not acceptable to patients. A new and promising technique, termed “liquid biopsy,” samples peripheral blood and detects either circulating tumor cells or free DNA in circulation, DNA shed by the diseased cells, including tumor cells.18 Peripheral blood samples are easily obtainable and well tolerated by most patients, whereas the freely circulating DNA can be detected and analyzed by deep sequencing or PCR. No microscope is needed. The information obtained has the potential to be as valuable as a histopathology. Most importantly, the DNA “signature” of cancer may be detected in this way much earlier in the course of development, even before occurrence of a clinically detectable tumor allows the possibility of biopsy.19 These types of approaches, “beyond the tissue biopsy,” have also extended to analyses of relevant body fluids, such as urine for genitourinary cancer, sputum for lung cancers, and stool for gastrointestinal cancers.20

Back to Top | Article Outline

NEXT-GENERATION SEQUENCING

Among the recently emerged techniques, next-generation sequencing has perhaps had the greatest impact. Within a few years of its appearance, the method has led to dramatic advances in understanding the structure and function of the genome. The increased speed, efficiency, and resolution of next-generation sequencing has greatly facilitated the detection of genetic, genomic, and epigenomic alterations, including single-nucleotide mutations, small insertions and deletions, chromosomal rearrangements, copy number variations, and DNA methylation changes. Comprehensive analysis of cancer genomes including the whole genome, exome, and transcriptome, has provided a new basis for understanding the pathology of cancer, particularly growth, differentiation, and metastasis, in some instances leading to better diagnosis and therapy.

The dramatic increase in the efficiency of sequencing, coupled with rapid reductions in cost, and the pace of ongoing technical improvements has created both excitement and expectation. In <10 years, the time and cost of DNA sequencing of a human genome has been reduced by a factor in excess of 1 million. A personal genome has about 100 GB of data, or the equivalent of about 100,000 digital photographs. Data from a million genomes would constitute hundreds of petabytes of data, the generation and storage of which would have been almost inconceivable just a decade ago. The advent of cloud computing and cloud storage provides the means and capacity for handling these huge amounts of data. A few years ago, the cost of sequencing a whole human genome was more than 1 million dollars, and it took more than 1 month. Today the cost of this sequencing has been reduced by more than 100-fold and it takes a few hours.21 Some have argued that a revolution in medical service would occur when the cost of whole genome sequencing falls to <$1000 per case and takes only a few hours to complete. Today, both goals are imminent and the trends show no sign of slowing. Many believe that by 2020, most major hospitals in the developed world will sequence the whole genome as a “routine” test.21,22

Back to Top | Article Outline

THE STUDY OF CANCER EXEMPLIFIES THE PROGRESS IN MOLECULAR PATHOLOGY

The diagnosis of cancer is central to the work of the pathologist. Molecular and genetic techniques have made impressive headway in this field. The accumulated findings to date have extended our knowledge well beyond the current morphologic classifications of cancer (Fig. 1). Literally thousands of alterations in the so called “cancer genome” have been detected, revealing about 140 genes that when altered by intragenic mutations can “drive” tumorigenesis. Of these, 71 are tumor suppressor genes and 54 are oncogenes. A typical tumor contains a few to more than a dozen of these “driver gene” mutations.21 About 95% of these mutations are single-base substitutions, and those remaining are deletions or insertions of one or a few bases. In one study, 90.7% of the base substitutions resulted in missense changes, 7.6% in nonsense changes, and 1.7% in alterations of splice sites or untranslated regions immediately adjacent to the start and stop codons.21,23–25 Some cancer types generally carry relatively few mutations—for example, medulloblastomas, testicular germ cell tumors, acute leukemias, and carcinoids,26,27 whereas others, such as lung cancer and melanomas, have many more mutations, occasionally as many as 100,000.27–29 Even within a particular morphologic cancer type, individual tumors may display wide variation in the prevalence and specific types of base substitutions. Most mutations are “passengers” or “bystanders” that confer no selective growth advantage for the tumors; these represent “noise” that the pathologist must address and set aside, part of the data processing challenge. The driver genes that do offer growth advantages can be classified into a dozen major signaling pathways that regulate 3 core cellular processes: cell fate, cell survival, and maintenance of the genome.30–33

In a recent report of The Cancer Genome Atlas project, somatic variants across 3281 specimens of 12 tumor types were sequenced and analyzed. The study identified 127 significantly mutated genes (SMGs).34 The average number of SMGs varied in different tumor types. Most tumors had 2 to 6 SMGs indicating that only a few genes are critical in carcinogenesis. Mutations in transcriptional factors and regulators may show tissue specificity. The effects on cancer behavior and patient survival of the SMGs were identified with clinical association analysis. Most occurrences of SMGs appeared to be related to clonal and subclonal lineages of cancer cells in temporal order, reflecting the heterogeneity of expression that is demonstrated by methods that conserve morphology. An understanding of the molecular landscape of tumors provides the foundation for pathologists to develop methods of diagnosis that integrate molecular and morphologic methods for individualized treatment.35

Equipped with this insight, the evaluation of cancer behavior by gene sequencing or PCR has become more precise and meaningful to a degree that cannot be matched by conventional pathology alone. The major driver gene mutations identified in common cancers are presented in Table 1.21 It seems certain that detection and evaluation of the expression of these driver genes by the appropriate best method, whether immunohistochemistry, in situ hybridization, sequencing, or other molecular method, will be demanded of pathologists in the near future.

TABLE 1
TABLE 1
Image Tools

Genomic sequencing has also given insight into the nature of cancer growth and metastasis. The number of mutations occurring in tumors at different stages of progression of colorectal and pancreatic cancers has been determined,36,37 resulting in 2 unambiguous understandings.

First, usually it takes a very long time to develop metastatic cancer. Thousands of different “cancer genomes” have been sequenced at different stages of carcinogenesis. Many cancers take decades to become clinically apparent, and the “incurable” stage, metastasis, occurs only a few years or months before death. It follows that the majority of patients dying from cancer do so only because the cancer was not detected during the first 90% of its development, when it may have been amenable to surgery or other therapy.21 It is likely that successful detection at a much earlier phase in the evolution of the cancer may lead to as much as a 75% reduction in cancer deaths.38

Second, virtually all of the mutations found in metastatic lesions are already present in subsets of cells in the primary tumors. Despite intensive effort, consistent genetic alterations that distinguish primary cancer from metastasis have yet to be identified. Also, while the rate of occurrence of point mutations in cancer cells is similar to that of normal cells, the rate of occurrence of significant chromosomal changes in cancer is elevated,39 and most solid tumors display widespread changes in chromosome number, as well as overt deletions, inversions, translocations, and other abnormalities. The technique of comparative genomic hybridization, which detects copy number changes at the chromosomal level, is valuable in this type of analysis.

Many of these changes in tissue can be detected with current pathologic tools such as immunohistochemistry and in situ hybridization. Correlation of molecular changes with histopathologic findings is an intriguing exercise for pathologists. It is reasonable to believe that many significant molecular alterations may have morphologic and structural correlates, that may lead to recognizable histopathologic changes, some of which may be best appreciated by image analysis and “machine learning.” However, it should be noted that not all genomic alterations lead to changes of protein expression, and certainly not to morphologic change. Also regulatory elements, DNA methylation, epigenetic modification, and the like affect expression of other proteins, are not likely to be directly detectable morphologically.

Back to Top | Article Outline

PATHOLOGIC DIAGNOSIS FOR NEW THERAPY: THE END OF CLASSIFICATION AS WE KNOW IT?

In recent years, the proteins altered by driver mutations have become targets for successful anticancer drug development.40 The recognition that certain tumors contain activating mutations in driver genes that encode protein kinases has led to the development of small-molecule inhibitor drugs targeting those kinases.21 In addition, monoclonal antibodies have been developed to target cell surface receptors of driver genes to treat derivative cancers. At the time of writing, at least 39 targeted therapy drugs for cancer have been approved by the FDA and at least 45 targeted cancer therapies are in clinical trials (http://www.cancer.gov/cancertopics/factsheet/Therapy/targeted) with more in the pipeline. A list of currently available targeted therapeutic drugs and their corresponding gene and protein targets is presented in Table 2. This list is certain to grow, along with the demand for pathologists to develop and use suitable detection techniques, variously termed as “predictive markers,” “companion diagnostics,” or “advanced personalized diagnostics.” Driver genes exert their influence through protein expression, and mutated genes may produce proteins that are sufficiently abnormal in structure so as to appear foreign to the immune system. These altered proteins are, in a sense, true tumor-specific antigens and provide ideal targets for tumor immunology.41 Immunohistochemistry, in situ hybridization and other molecular methods are applicable to evaluating expression and distribution of these antigens and associated mRNAs as companion diagnostics.

TABLE 2
TABLE 2
Image Tools

In the face of this flood of new data it is abundantly clear that the histopathologic classification of cancer no longer is sufficient and must be supplemented by molecular methods to reach the optimal therapeutic approach. A single example suffices, although new instances occur almost daily. A decade ago human epidermal growth factor receptor-2 (HER2) emerged as the target for therapy with Herceptin (anti-HER2 human monoclonal antibody), delivering effective treatment of HER2-positive cancers.42 With HER2-negative tumors, Herceptin has no beneficial effect, but only causes toxicity. Initially discovered in about a quarter of breast cancers, HER2 overexpression is now found in some cases of stomach cancer, lung cancer and a few other cancer types. HER2 detection by immunohistochemistry with specific antibodies, supplemented as needed by FISH for HER2 gene amplification, has become a routine in even small hospitals. It is now more meaningful for pathologists to inform the oncologists (and patients) if the tumor is HER2 positive or negative than to subclassify the various recognized histologic subtypes, such as papillary adenocarcinoma or squamous-cell carcinoma. Lung cancer provides another example (Fig. 1). Only a small fraction of lung cancer patients have EGFR gene mutations or ALK gene translocations, but only these patients will respond to the appropriate drug,43 for all others the treatment would be detrimental. Morphology alone is unable to identify these “responders.”

Back to Top | Article Outline

PATHOLOGISTS SHOULD TAKE THE LEAD IN PERSONALIZED MEDICINE

Many pathologists had been reluctant to depart from the relative security of morphology into the unknown territory of the genome and proteome. Consequently, oncologists have attempted to set up these tests within their own departments, often lacking adequate internal experience and quality control, or they have simply utilized small or large reference laboratories that have proliferated to meet the perceived need. As many more targeted therapies are developed (Table 2), pathologists face a choice, either of taking on the responsibility for detecting these genomic and molecular “lesions,” thereby claiming “ownership” of these new diagnostic and predictive tests, or becoming less important in the coming era of personalized medicine.10,11

To incorporate these new techniques into the pathology laboratory, requires training in the use of new equipment and protocols, and awareness of all aspects of testing starting from tissue collection and processing. Demands are similar to those of high quality immunohistochemistry.44–46 The samples must be collected freshly and processed as soon as possible. Extraction of DNA and RNA from formalin-fixed and paraffin-embedded tissue samples, while less ideal than from fresh or frozen samples, is possible and is in reality increasingly common, response to the fact that formalin-fixed and paraffin-embedded tissue is often the only option.47,48

Although the changes are initiated at the DNA and RNA levels, the effects and actions are exerted by the encoded proteins, lipid, and carbohydrates, and these changes often are within the power of detection by current pathologic technology, by immunohistochemistry or FISH.49 There is another major, even essential reason for involvement of pathologists. Driver gene mutations are closely related to cancer clonality, with marked heterogeneity of expression of various subclones in many tumors, which can only be evaluated by histopathologists. Tissue extracts, for PCR or next-generation sequencing, lose all of the morphologic correlation, and are unable to assess the possible critical significance of heterogeneity, a reason why there are few studies of this problem.

Molecular pathology thus has the potential to form the foundation of personalized medicine. The integration of Diagnostic Molecular Pathology into Applied Immunohistochemistry and Molecular Morphology recognizes this potential and this reality. Furthermore, the pathology community has been alerted to the potential and problems of molecular pathology and several initiatives are being pursued. At a recent conference held in October 2010 at Cold Spring Harbor Laboratory, NY, named “Genome-Era Diagnostics and Preemptive Care: A stakeholder summit,” 7 projects were proposed (Table 3).11 These recommendations provide as good a place to start as any.

TABLE 3
TABLE 3
Image Tools
Back to Top | Article Outline

CONCLUSIONS

The traditional task of the pathologist is to assist physicians in making the correct diagnosis of diseases at the earliest possible stage and to assist them in developing and evaluating the most effective treatment for each individual patient. In this respect morphologic analysis (the traditional tissue diagnosis) is a tool, but of itself is not the purpose of pathology practice.

Molecular and genomic techniques are rapidly expanding. Their use demonstrates clear value in the diagnosis and treatment of diseases tailored specifically to individual patients. Pathologists and pathology laboratories have the experience and infrastructure to accurately perform these assays, and to collect, analyze, interpret, distribute, and store the huge volumes of associated data. Already the mechanisms are in place for managing the gigabyte data sets generated by digital WSIs, in laboratory PACS systems3,9 and Pathology Informatics is a thriving subdiscipline. There is no profession better suited than Pathology to take the lead in establishing these assays, correlating the findings with existing modes of cancer diagnosis including morphology, and providing integrated reports tailored to individual patients and to the clinical need.

Active adaption and innovation is required, driven by a sense of excitement and adventure. A passive or negative attitude is likely to marginalize pathologists with dire consequences for the discipline in the long term.

Back to Top | Article Outline

REFERENCES

1. Gu J .Computerized three- and four-dimensional modeling of biomedical images for morphologists and pathologists.Cell Vision. 1997; 4:6–15.

2. Taylor CR, Cote RJ .Immunomicroscopy: A diagnostic tool for the surgical pathologist. 2005; :3rd ed.Philadelphia:Elsevier Saunders.

3. Taylor CR .From microscopy to whole slide digital images: a century and a half of image.Appl Histochem Mol Morph. 2011; 19:491–493.

4. . Catalogue of Somatic Mutations in Cancer (COSMIC) .http://www.cancer.sanger.ac.uk/

Cell line project. Available at: http://www.cancer.sanger.ac.uk/. Accessed October 20, 2013


5. Beadling C, Heinrich MC, Warrick A, et al .Multiplex mutation screening by mass spectrometry evaluation of 820 cases from a personalized cancer medicine registry.J Mol Diagn. 2011; 13:504–513.

6. Hood L, Friend SH .Predictive, personalized, preventive, participatory (P4) cancer medicine.Nat Rev Clin Oncol. 2011; 8:184–187.

7. Gerstein MB, Kundaje A, Hariharan M, et al .Architecture of the human regulatory network derived from ENCODE data.Nature. 2012; 489:91–100.

8. Chen R, Mias GI, Li-Pook-Than J, et al .Personal omics profiling reveals dynamic molecular and medical phenotypes.Cell. 2012; 148:1293–1307.

9. Taylor CR .Whole slide imaging—issues for use in diagnostic pathology: “routine;” stains, immunohistochemistry and predictive markers.Biotehnics and Histochem. 2013;

In press


10. Walk EE .The role of pathologists in the era of personalized medicine.Arch Pathol Lab Med. 2009; 133:605–610.

11. Tonellato PJ, Crawford JM, Boguski MS, et al .A national agenda for the future of pathology in personalized medicine: report of the proceedings of a meeting at the Banbury Conference Center on genome-era pathology, precision diagnostics, and preemptive care: a stakeholder summit.Am J Clin Pathol. 2011; 135:668–672.

12. Matloff E, Caplan A .Direct to confusion: lessons learned from marketing BRCA testing.Am J Bioeth. 2008; 8:5–8.

13. Hogarth S, Javitt G, Melzer D .The current landscape for direct-to-consumer genetic testing: legal, ethical, and policy issues.Annu Rev Genomics Hum Genet. 2008; 9:161–182.

14. Taylor CR, Levenson RM .Quantification of immunohistochemistry—issues concerning methods, utility and semiquantitative assessment II.Histopathology. 2006; 49:411–424.

15. Parsons DW, Jones S, Zhang X, et al .An integrated genomic analysis of human glioblastoma multiforme.Science. 2008; 321:1807–1812.

16. . The Cancer Genome Atlas Network .Comprehensive molecular characterization of human colon and rectal cancer.Nature. 2012; 487:330–337.

17. Williams ES, Hegde M .Implementing genomic medicine in pathology.Adv Anat Pathol. 2013; 20:238–244.

18. Crowley E, Di Nicolantonio F, Loupakis F, et al .Liquid biopsy: monitoring cancer-genetics in the blood.Nat Rev Clin Oncol. 2013; 10:472–484.

19. Dawson SJ, Tsui DW, Murtaza M, et al .Analysis of circulating tumor DNA to monitor metastatic breast cancer.N Engl J Med. 2013; 368:1199–1209.

20. Dunn BK, Jegalian K, Greenwald P .Biomarkers for early detection and as surrogate endpoints in cancer prevention trials: issues and opportunities.Recent Results Cancer Res. 2011; 188:21–47.

21. Vogelstein B, Papadopoulos N, Velculescu VE, et al .Cancer genome landscapes.Science. 2013; 339:1546–1558.

22. Stratton MR .Exploring the genomes of cancer cells: progress and promise.Science. 2011; 331:1553–1558.

23. Stephens PJ, Tarpey PS, Davies H, et al .The landscape of cancer genes and mutational processes in breast cancer.Nature. 2012; 486:400–404.

24. Jones S, Zhang X, Parsons DW, et al .Core signaling pathways in human pancreatic cancers revealed by global genomic analyses.Science. 2008; 321:1801–1806.

25. Wood LD, Parsons DW, Jones S, et al .The genomic landscapes of human breast and colorectal cancers.Science. 2007; 318:1108–1113.

26. Parsons DW, Li M, Zhang X, et al .The genetic landscape of the childhood cancer medulloblastoma.Science. 2011; 331:435–439.

27. Greenman C, Stephens P, Smith R, et al .Patterns of somatic mutation in human cancer genomes.Nature. 2007; 446:153–158.

28. Pleasance ED, Stephens PJ, O’Meara S, et al .A small-cell lung cancer genome with complex signatures of tobacco exposure.Nature. 2010; 463:184–190.

29. Ding L, Getz G, Wheeler DA, et al .Somatic mutations affect key pathways in lung adenocarcinoma.Nature. 2008; 455:1069–1075.

30. Kunkel TA .Evolving views of DNA replication (in)fidelity.Cold Spring Harbor Symposia on Quantitative Biology. 2009; 74:91–101.

31. Kerbel RS .Tumor angiogenesis.N Engl J Med. 2008; 358:2039–2049.

32. Perrimon N, Pitsouli C, Shilo BZ .Signaling mechanisms controlling cell fate and embryonic patterning.Cold Spring Harbor Perspectives in Biology. 2012; 4:a005975

33. Chung AS, Ferrara N .Developmental and pathological angiogenesis.Annu Rev Cell Dev Biol. 2011; 27:563–584.

34. Kandoth C, McLellan MD, Vandin F, et al .Mutational landscape and significance across 12 major cancer types.Nature. 2013; 502:333–339.

35. Taylor CR, Becker KF .Liquid morphology: immunochemical analysis of proteins extracted from formalin fixed paraffin embedded tissues: combining proteomics with immunohistochemistry.Appl Immunohistochem Mol Morphol. 2011; 19:1–9.

36. Jones S, Chen WD, Parmigiani G, et al .Comparative lesion sequencing provides insights into tumor evolution.Proc Natl Acad Sci U S A. 2008; 105:4283–4288.

37. Yachida S, Jones S, Bozic I, et al .Distant metastasis occurs late during the genetic evolution of pancreatic cancer.Nature. 2010; 467:1114–1117.

38. Colditz GA, Wolin KY, Gehlert S .Applying what we know to accelerate cancer prevention.Sci Transl Med. 2012; 4:127rv124

39. Pelizzola M, Ecker JR .The DNA methylome.FEBS Lett. 2011; 585:1994–2000.

40. Stuart D, Sellers WR .Linking somatic genetic alterations in cancer to therapeutics.Curr Opin Cell Biol. 2009; 21:304–310.

41. Kirkwood JM, Butterfield LH, Tarhini AA, et al .Immunotherapy of cancer in 2012.CA Cancer J Clin. 2012; 62:309–335.

42. Incorvati JA, Shah S, Mu Y, et al .Targeted therapy for HER2 positive breast cancer.J Hematol Oncol. 2013; 6:38

43. Janku F, Stewart DJ, Kurzrock R .Targeted therapy in non-small-cell lung cancer—is it becoming a reality? Nat Rev Clin Oncol. 2010; 7:401–414.

44. Taylor CR .Quantifiable internal reference standards for immunohistochemistry; the measurement of quantity by weight.Appl Immunohistochem Mol Morph. 2006; 14:253–259.

45. Hewitt SM, Robinowitz M, Bogen SA, et al .Quality Assurance for Design Control and Implementation of Immunohistochemistry Assays: Approved Guidelines. 2011; :2nd ed.Wayne, PA:Clinical Lab Standards Institute.

46. Taylor CR .Focus on Biospecimens: the issue is the tissue. Observations on the “NCI-NIST fitness for purpose quality assessment and standards workshop”. October 2010.Appl Histochem Mol Morph. 2011; 19:95–98.

47. Shi SR, Balgley BM, Taylor CR. Shi S-R, Taylor CR .Symbiosis of Immunohistochemistry and Proteomics: Marching to a New Era.Antigen Retrieval Immunohistochemistry Based Research and Diagnostics. 2010; .Hoboken, NJ:John Wiley and Sons, Inc.; 391–398.

48. Shi SR, Taylor CR, Fowler CB, et al .Complete solubilization of formalin-fixed paraffin embedded tissues may improve proteomic studies.Proteomics Clin Appl. 2013.;

doi.10.1002.prca.201200031


49. Rekhtman N, Tafe LJ, Chaft JE, et al .Distinct profile of driver mutations and clinical features in immunomarker-defined subsets of pulmonary large-cell carcinoma.Mod Pathol. 2013; 26:511–522.

Keywords:

immunohistochemistry; molecular pathology; next-generation sequencing; PCR; digital pathology; surgical pathology; diagnostic molecular pathology; companion diagnostics; predictive markers; advanced personalized diagnostics; personalized medicine; big data; mass spectrometry; targeted therapy; morphologic classification; genomics; proteomics; metabolomics; liquid morphology

Copyright © 2013 by Lippincott Williams & Wilkins

Login

Article Tools

Images

Share

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.