This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Allegra, C. J. Articles by Benedetti, J. K. Search for Related Content PubMed PubMed Citation Articles by Allegra, C. J. Articles by Benedetti, J. K. Related Articles Related Article Social Bookmarking                            What's this?

Don Quixote and the Quest for Personalized Medicine

Division of Hematology/Oncology, University of Florida Shands Cancer Center, Gainesville, FL

Jacqueline K. Benedetti

University of Washington and Southwest Oncology Group Statistical Center (and Fred Hutchinson Cancer Research Center), Seattle, WA

Is it time to dream the impossible dream? The promise of personalized medicine has gained much publicity in the popular press, but the realization of this goal remains in its infancy. Our expanding understanding of cancer biology and developments in technology have been translated into assays capable of simultaneously investigating tens of thousands of genes and gene products for mutational and expression status.¹ The availability of these multiplex technologies has tremendous potential to radically alter management strategies through the identification of precise molecular diagnostics, to tailor treatment to specific characteristics of both patient and tumor.

While current technology affords the ability to measure many markers, the level of evidence required to identify and validate the importance of an individual marker must fulfill many requirements. Among these are: presence of a reliable and reproducible assay; adequate patient samples to identify an initial signal; existence of an independent set of patient samples to validate the signal detected during the initial marker exploration; sufficient magnitude of signal to indicate meaningful clinical usefulness; and appropriate interpretation of the signal. This latter point is both important and dependent on the design of the study. Most initial marker exploration is generally conducted in groups of similarly treated patients. Thus, any link to outcome merely suggests the marker's potential as a prognostic factor, one that identifies which patients have poorer or worse outcome. These results do not assess whether or not the marker will differentiate outcome for patients given an alternate treatment. To interpret a marker as selectively predictive of which patients should receive specific treatment, studies must be performed within the confines of a randomized trial. In this case, it would be expected that a comparison between the experimental arm and standard would yield a positive effect only for one level of the marker, but not for the other (or if the magnitude of the effect is vastly different between the two marker levels). Such a marker would be declared a predictive factor. Given all of these requirements, it is not surprising that most recent studies have failed to identify clinically meaningful markers.

In the past few decades, the expression level or mutations in individual genes or proteins, often related to the known mechanism(s) of action of the therapeutic agents, were extensively explored for utility as both prognosticators and predictors of benefit associated with specific therapeutic interventions. These efforts have been further augmented by the advent of newer cytotoxic agents and targeted therapies. While many of these investigations resulted in the identification of genes or proteins that suggested potential use as markers, most of the differences in outcome were of limited clinical value and were often based on small exploratory trials. This observation was reinforced by the Colon Cancer Working Group of the Program for the Assessment of Clinical Cancer Tests which reviewed more than 100 research articles considered to be the most promising reports on potential prognostic and predictive factors for colorectal cancer.² In addition to finding that most studies were based on small sample sizes, the group concluded that the studies used inconsistent assays even for common markers and yielded conflicting results; thus, few markers emerged with recommendations for further study. A general lack of validation studies further dampened enthusiasm. Presently, only a handful of markers have gained widespread recognition as clinically useful predictive markers in patients with nonhematologic malignancies. These include estrogen/progesterone receptors, human epidermal growth factor receptor 2 status, and cKit. More recently, the mutational status of ras has been identified as a potentially powerful marker of benefit associated with the use of anti-epidermal growth factor receptor antibodies.³

The report in this issue of the Journal by Braun et al⁴ explores 11 potential predictive markers of benefit associated with the firstline use of combination therapy with fluorouracil plus either oxaliplatin or irinotecan compared to fluorouracil alone, followed sequentially by the addition of oxaliplatin or irinotecan in patients with metastatic colorectal cancer treated in the Fluorouracil, Oxaliplatin, CPT-11: Use and Sequencing (FOCUS) trial.^4,5 The central question addressed by this report is how the use of combination chemotherapy compared with single-agent therapy alters progression-free survival in patient subgroups defined by the 11 proposed predictive markers. The markers were identified a priori, and a substantial number of annotated tissue samples (approximately 1,300) were successfully analyzed. The authors developed a two-stage strategy for screening the markers, and ultimately were only able to demonstrate a significant treatment interaction for topoisomerase (Topo) -1. Notable among the negatives were immunohistochemical analysis of the mismatch repair enzymes MLH1 and MSH2 and single nucleotide polymorphisms in ERCC1 and ERCC2. The authors observed an improved progression-free survival in patients with moderate and high Topo1 levels for the patient groups treated initially with combination chemotherapy compared with those treated with single-agent fluorouracil (FU), treated on progression with either oxaliplatin or irinotecan. In addition to its predictive value, the authors identified elevated levels of Topo1 to be a poor prognostic factor. Topo1 is the primary intracellular target of irinotecan and other camptothecin analogs and is responsible for relaxing supercoiled DNA through the production of readily repairable single strand DNA nicks.⁶ However, if Topo1 is bound by irinotecan, the DNA damage is not reparable and ultimately results in the formation of lethal DNA fragmentation. In keeping with the findings of the present clinical report, preclinical investigations have consistently demonstrated that irinotecan sensitivity is directly related to the cellular level of Topo1.^7,8 Given this concordance, it is tempting to conclude that the predictive value of Topo1 levels is directly linked with the mechanism of action of irinotecan; however, there are features that suggest at least one alternative hypothesis. In addition to the predictive value of Topo1 for patients treated with the FU and irinotecan combination, this study noted a similar predictive value for patients treated with FU and oxaliplatin, in which the relationship between Topo1 levels and the drug's mechanism of action is far less clear. This suggests that the predictive value of Topo1 may not be mechanistically based. Rather, it may be a reflection of more global differences in tumor biology, such as a lowered apoptotic threshold, that perhaps determine sensitivity in favor of more intense combinations of various chemotherapeutic agents rather than specific combinations. As such, the noted predictive interaction may not be specific for the agents investigated, but this would need to be further explored.

How did the authors succeed with respect to the five criteria listed for marker evaluation? Topo1 was evaluated using immunohistochemistry; before recommending general use of such a marker, one would want some indication of the reliability of these assays in different laboratories. Moreover, the cutoffs for identification of the medium and high levels were not discussed, and would require justification and validation. With respect to study design and interpretation, 1,300 specimens provided sufficient materials for testing. The focus was on treatment/marker interactions and the appropriate measure for the interpretation of predictive factors within a clinical trial. Importantly, the results of this trial include the appropriate cautions regarding the need for validation studies, which are apparently already being conducted.⁹

One may conclude that these data support the use of more aggressive chemotherapy in patients with a worse prognosis (eg, high Topo1 levels) regardless of the cellular mechanism that accounts for the benefit associated with combination chemotherapy as first-line of therapy in this patient population. Conversely, it is reasonable to suggest that patients with low Topo1 levels may be safely managed with sequential agents. However, as cautioned by the authors, these data—despite the large sample size of the present study—require independent validation before Topo1 levels should be considered a marker for clinical decision making.

In this era of ever-increasing investigations of markers for individual components of a treatment regimen, it is also worth considering whether one can translate information from separate marker studies into a recipe for personalized treatment. Does it necessarily follow that a patient who has kRAS wild type and has high Topo1 levels should automatically receive both irinotecan and cetuximab? Is it possible that these two mechanisms do not act independently and that perhaps we also need to assess whether the combined knowledge of the two may dictate a different treatment strategy? Given the requirements for validation of a single marker, and the increased availability of new generations of treatments, such questions are indeed sobering in the search for personalized treatment strategies. This should not discourage our quest, but remind us that we need to be cognizant of the challenges ahead. Otherwise, we risk jousting with windmills.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: Carmen J. Allegra, Almac Diagnostics (C) Stock Ownership: None Honoraria: None Research Funding: None Expert Testimony: None Other Remuneration: None

REFERENCES

1. Lockhart DJ, Winzeler FA: Genomics, gene expression and DNA arrays. Nature 405:827-836, 2000[CrossRef][Medline]

2. Taube SE, Jacobson JW, Lively TG: Cancer diagnostics: Decision criteria for marker utilization in the clinic. Am J Pharmacogenomics 5:357-364, 2005[CrossRef][Medline]

3. Amado RG, Wolf M, Freeman D, et al: Panitumumab (pmab) efficacy and patient-reported outcomes (PRO) in metastatic colorectal cancer (mCRC) patients (pts) with wild-type (WT) KRAS tumor status. Proc Am Soc Clin Oncol GI Cancer Symposium; (abstr 278), 2008

4. Braun MS, Richman SD, Quirke P, et al: Predictive biomarkers of chemotherapy efficacy in colorectal cancer: Results from a large randomized trial. J Clin Oncol 26:2690-2698, 2008[Abstract/Free Full Text]

5. Seymour MT, Maughan TS, Ledermann JA, et al: For the FOCUS Trial Investigators and the National Cancer Res Institute Colorectal Clinical Studies Group: Different strategies of sequential and combination chemotherapy for patients with poor-prognosis advanced colorectal cancer—A randomized controlled trial. Lancet 370:143-152, 2007[CrossRef][Medline]

6. Tsao YP, Russo A, Nyamuswa G, et al: Interaction between replication forks and topoisomerase 1-DNA cleavage complexes: Studies in a cell-free SV-40 DNA replication system. Cancer Res 53:5908, 1993[Abstract/Free Full Text]

7. Pommier V, Pourquier P, Urasaki Y, et al: Topoisomerase 1 inhibitors: Selectivity and cellular resistance. Drug Resistance Update 2:307-318, 1999[CrossRef][Medline]

8. Woessner RD, Eng WK, Hofmann GZ, et al: Camptothecin hyperresistant P388 cells: Drug-dependent reduction in topoisomerase 1 content. Oncol Res 4:481, 1992[Medline]

9. Koopman M, Antonini NF, Douma J, et al: Sequential verses combination chemotherapy with capecitabine, irinotecan and oxaliplatin in advanced colorectal cancer (CAIRO): A phase III randomized controlled trial. Lancet 370:135-142, 2007[CrossRef][Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Article

• Predictive Biomarkers of Chemotherapy Efficacy in Colorectal Cancer: Results From the UK MRC FOCUS Trial
  Michael S. Braun, Susan D. Richman, Philip Quirke, Catherine Daly, Julian W. Adlard, Faye Elliott, Jennifer H. Barrett, Peter Selby, Angela M. Meade, Richard J. Stephens, Mahesh K.B. Parmar, and Matthew T. Seymour
  JCO 2008 26: 2690-2698 [Abstract] [Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Allegra, C. J. Articles by Benedetti, J. K. Search for Related Content PubMed PubMed Citation Articles by Allegra, C. J. Articles by Benedetti, J. K. Related Articles Related Article Social Bookmarking                            What's this?