Digital quantification of mutant DNA in cancer patients
`Frank Diehla,b and Luis A. Diaz Jra
`
`Purpose of review
`The accumulation of somatic mutations is the major driving
`force for tumorigenesis. These mutations uniquely
`differentiate tumor cells from their normal counterparts.
`Mutations within tumor cells and mutant DNA released by
`tumor cells into blood, lymph, stool, tissues and other bodily
`compartments can thereby be used for cancer detection.
`Here we discuss technologies available for the detection
`and quantification of mutant DNA in clinical samples and the
`value of such measurements for patient management.
`Recent findings
`Conventional mutation detection technologies are either
`qualitative or only roughly estimate the abundance of mutant
`DNA molecules. Recently-developed approaches,
`however, use single molecule counting to determine the
`genotype of each individual member of a DNA population,
`providing a more accurate and precise digital output.
`Summary
`In this review, we discuss the clinical utility of mutant DNA
`quantification in cancer patients in the context of recent
`technical advances made in digital mutation detection.
`
`Keywords
`biomarker, molecular diagnostics, mutation, quantification,
`single molecule detection
`
`Curr Opin Oncol 19:36–42. ß 2007 Lippincott Williams & Wilkins.
`
`aThe Ludwig Center for Cancer Genetics and Therapeutics, Baltimore and bThe
`Howard Hughes Medical Institute, The Johns Hopkins Kimmel Cancer Center,
`Baltimore, Maryland, USA
`
`Correspondence to Frank Diehl, PhD, The Ludwig Center for Cancer Genetics and
`Therapeutics, The Johns Hopkins Kimmel Cancer Center, 1650 Orleans Street,
`Room 590, Baltimore, MD 21231, USA
`Tel: +1 410 955 8878; fax: +1 410 955 0548;
`e-mails: fdiehl2@jhmi.edu, frank_diehl@yahoo.com
`
`Conflict of Interest and Sponsorship Statements: No conflicts declared.
`
`Current Opinion in Oncology 2007, 19:36–42
`
`Abbreviations
`
`APC
`ATP
`BCR-ABL
`EBV
`dHPLC
`EGFR
`FRET
`PCR
`PSA
`RCA
`SBE
`SMD
`
`adenomatosis polyposis coli
`adenosine triphosphate
`breakpoint cluster region-Abelson
`Epstein-Barr virus
`denaturing high performance liquid chromatography
`epidermal growth factor receptor
`fluorescence resonance energy transfer
`polymerase chain reaction
`prostate specific antigen
`rolling-circle amplification
`single base extension
`single molecule detection
`
`ß 2007 Lippincott Williams & Wilkins
`1040-8746
`
`36
`
`Introduction
`The molecular basis of cancer is rapidly being deciphered
`and recent studies suggest that a complex array of genetic
`alterations exist in most human cancers [1]. Genetic
`alterations including gene deletions, gene amplifications,
`point mutations, and chromosomal rearrangements play a
`major role in the development and progression of cancers
`and are therefore unique identifiers that distinguish
`tumor cells from their normal counterparts.
`
`In order to best use gene alterations as a biomarker in
`clinical oncology, it must be possible to detect the tumor-
`specific genetic changes in a background of DNA from
`normal cells. Gene deletions and amplifications consist of
`copy number alterations,
`instead of changes in the
`primary DNA sequence and thus are not easily dis-
`tinguishable from DNA of normal cells. Detection and
`enumeration of gene copy number changes therefore
`require tumor cell isolation or direct cellular visualization.
`For example, in-situ hybridization or laser capture micro-
`dissection followed by real-time polymerase chain reac-
`tion (PCR) is used for the quantification of v-erb-b2
`erythroblastic leukemia viral oncogene homolog 2 ampli-
`fications in breast cancer tissue [2].
`
`Unlike deletions and amplifications, point mutations and
`chromosomal rearrangements represent changes of the
`primary DNA sequence that are substantially different
`from normal DNA and thus can be detected within
`clinical samples without the need for prior tumor cell
`isolation or visualization. Mutation detection has often
`been performed qualitatively without the ability to
`quantify the amount of mutant and wild-type DNA
`present in the sample. Qualitative mutation detection
`assays can have two potential problems. First, an assay
`could yield a false-negative result because the amount of
`starting DNA is too low to detect minority mutations.
`Second, an assay could yield a stochastic false-positive
`result because rare random mutations are present in a
`sample. Quantitative technologies could overcome these
`problems. They have the ability to directly measure the
`number of DNA molecules tested per assay and therefore
`ensure that the amount of starting material is sufficient to
`detect minority mutations in the predicted frequency
`range. Quantification also allows one to distinguish
`between random and pathogenic mutations by establish-
`ing a baseline mutant-to-wild-type ratio (e.g. the back-
`ground mutation frequency of human cells), where
`mutations found at a ratio below the baseline are con-
`sidered random. Quantitative assays also allow standard
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 1 of 7
`
`

`

`quality control monitoring which is a necessary require-
`ment for routine diagnostic testing [3].
`
`Several quantitative mutation detection technologies are
`currently in development, and translational studies are
`needed to demonstrate whether quantification of mutant
`DNA is beneficial for early detection and as a biomarker
`of malignancy. Technologies attempting to quantify can-
`cer mutations and their potential clinical application are
`the focus of this review.
`
`Mutation detection and quantification
`principles
`Over the past decade, many techniques have been devel-
`oped for the analysis of mutated DNA. The impetus for
`this effort has been the desire to increase sensitivity,
`specificity and efficiency while decreasing cost. Analyti-
`cal sensitivity and specificity are parameters that refer to
`the lowest amount of mutant DNA that can be detected
`in a high background of normal DNA and the ability to
`exclusively detect actual mutations, respectively. Assay
`efficiency and cost are linked and refer to the time, labor
`and cost required for each analysis. The performance of
`the currently available technologies varies considerably.
`To better understand the variability, we grouped them
`based on (i) how many mutations can be analyzed per
`assay, (ii) the order of amplification and allele discrimi-
`nation, and (iii) the quantitative nature of the assay
`signals.
`
`Screening and locus specific mutation detection
`Screening or scanning methods can detect a range of
`possible mutations in a target region. They are particu-
`larly important for the discovery of new cancer genes [4].
`For this application, the sensitivity of standard Sanger
`sequencing is sufficient. The identification of minority
`mutations, however, requires higher sensitivities and
`specificities. Current scanning technologies are in general
`not sensitive enough, due to larger numbers of positions
`analyzed. Once a new mutation is associated with a
`certain cancer, it is usually sufficient to detect it by
`locus-specific methods. Locus-specific assays have the
`advantage of being more sensitive, cheaper and easier to
`perform. Several assays that can be used for the quanti-
`fication of known and unknown mutations are described
`below.
`
`Direct and indirect allele discrimination
`Most techniques rely on the following four components:
`(i) amplification of the target sequence, (ii) discrimination
`of the mutant and wild-type DNA sequences, (iii) sep-
`aration, and (iv) detection. Approaches that amplify the
`DNA before genotype discrimination are named indirect
`detection methods. The amplification is most commonly
`accomplished by the use of PCR and therefore suffers
`from typical PCR-associated complications, such as the
`
`Quantification of mutant DNA in cancer Diehl and Diaz 37
`
`generation of random errors by the DNA polymerase and
`allele bias that results in poor quantitative accuracy.
`Direct methods, on the other hand, discriminate the
`alleles prior to the amplification or detection step. The
`presence or absence of the mutation is determined by
`allele-specific methods. DNA ligases, DNA polymerases
`and DNA nucleases, as well as thermodynamic differ-
`ences between matched and mismatched DNA duplexes
`are used to discriminate between alleles. The initial
`allele discrimination step greatly influences the speci-
`ficity and sensitivity of any direct mutation detection test.
`For some assay platforms, the DNA molecules need to be
`separated before detection by binding to a solid phase or
`by gel electrophoresis. Unfortunately, the separation step
`introduces the risk of cross-contamination and requires
`additional handling of the samples. An assay format that
`avoids these problems is a homogeneous test that does
`not require separation and can combine the amplification,
`discrimination and detection step essentially in a single
`tube [5].
`
`Analog and digital quantification of mutations
`As discussed above, the assay strategy determines the
`specificity and sensitivity as well as the precision and
`accuracy of allele quantification. These parameters also
`depend on the nature of the assay signal generated.
`Traditional genotyping assays determine the identity
`of a particular base as an average contribution to a
`heterogeneous population of DNA molecules. Thus,
`such methods only convey an ‘analog’ signal for the
`individual members of the DNA pool. If more than
`one genotype is queried at a time, a ratio can be calcu-
`lated between the different alleles present in the reaction
`(Fig. 1a). Most assays available today are analog and have
`been reviewed elsewhere [5–7]. Here, we present
`examples of assays that have recently been used for
`quantification of mutations in clinical samples. Pyrose-
`quencing, for example, is a nucleotide extension sequen-
`cing approach where pyrophosphate is generated when a
`particular nucleotide anneals to the template and is
`incorporated by DNA polymerase. Subsequently, pyro-
`phosphate is converted to adenosine triphosphate (ATP)
`by ATP sulfurylase, which provides the energy for luci-
`ferase to oxidize luciferin and generate light. The inten-
`sity of the light is proportional to the amounts of annealed
`and extended nucleotide molecules. The peak sizes of
`the pyrogram are used to quantify the relative amount of
`each allele down to a mutant to wild-type ratio of 5% [8].
`Another example of analog mutation detection is a DNA
`endonuclease-based assay followed by denaturing high
`performance liquid chromatography (dHPLC). The
`quantitative output signals are the peak sizes on the
`dHPLC chromatograms. The assay can be used to
`quantify unknown mutations if present above 1% [9].
`Two final examples are the LigAmp assay and the Scor-
`pion primer-based quantitative PCR assay which can
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 2 of 7
`
`

`

`38 Cancer biology
`
`Figure 1 Principle of analog and digital mutation analysis
`
`(a)
`
`(b)
`
`Analog
`
`Digital
`
`100%
`90%
`80%
`70%
`60%
`50%
`40%
`30%
`20%
`10%
`0%
`
`Wild-type
`
`Mutant
`
`100%
`90%
`80%
`70%
`60%
`50%
`40%
`30%
`20%
`10%
`0%
`
`1
`
`3
`
`5
`
`7
`
`9
`11
`13
`Wild-type
`
`17
`15
`Mutant
`
`19
`
`21
`
`23
`
`25
`
`(a) In analog assays, an average signal is acquired from the mutant and wild-type DNA molecules present in the sample. The ratio between the mutant
`and wild-type signal is an estimate of the mutation frequency. (b) In digital assays, the genotype of the individual DNA molecules is determined
`separately. Counting is used to quantify the mutant and wild-type DNA molecules present in the sample.
`
`quantify mutations down to 0.01% [10,11]. As homo-
`geneous assays, these measure in real-time the accumu-
`lation of mutant and wild-type PCR product during each
`cycle. In the log-linear phase of the amplification the
`amount of the target DNA correlates with the initial
`copy numbers.
`
`A more precise and accurate approach to mutation quanti-
`fication is based on discrete counting of the mutant and
`wild-type alleles present in a sample. Techniques incor-
`porating this methodology are termed ‘digital’ as they are
`able to generate binary results (mutation present or absent)
`on each individual member of a DNA pool (Fig. 1b)
`[12,13]. The quantitative precision and accuracy of
`digital assays is limited only by the number of molecules
`being analyzed. Statistics become important for counting
`rare events, which is the case when rare mutant alleles are
`present in a high background of wild-type molecules.
`Based on the Poisson distribution, the standard deviation
`of the number of rare events equals the square root of the
`
`number of detected events. Only the measured back-
`ground events need to be subtracted from the positives
`to get the net counts. This phenomenon has been called
`Poisson noise, and it limits the precision of single molecule
`detection methods [14]. The precision can be improved
`only by increasing the number of molecules analyzed. For
`example, a precision of 10% would require the detection of
`100 mutant molecules per measurement. Considering a
`sensitivity limit of 0.01% for the detection of mutant DNA
`in plasma, this would require the analysis of a total of
`1 106 DNA molecules. This translates into 3 mg of
`human genomic DNA, which exceeds the amount of
`DNA present in most plasma samples [15].
`
`Digital mutation quantification methods
`Various approaches have been described that allow the
`confined digital analysis of single molecules. We will
`review direct and indirect methods that can be used
`for mutation quantification in DNA fragments at known
`or unknown base positions (Table 1).
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 3 of 7
`
`

`

`Quantification of mutant DNA in cancer Diehl and Diaz 39
`
`[40]
`[33,34]
`
`[38]
`[15]
`[36]
`
`[19]
`[23]
`[16]
`
`[22]
`
`References
`
`[29–32]
`
`K-ras
`
`1/100–1000
`
`p53
`
`1/100000a
`
`CFTR
`
`ND
`
`CFTR
`
`1/500–1500
`
`Flowcytometry
`platereader
`Fluorescence
`
`machine
`
`Real-timePCR
`microscope
`
`Scanningelectron
`
`Microscope
`Fluorescence
`
`EGFR
`
`1/1000
`
`GS20,454Corp.
`
`NA
`
`APC
`
`1/500
`
`Gelscanner
`
`PIK3CA,p53
`
`K-ras,
`
`1/10000
`
`Flowcytometry
`
`APC
`
`K-ras
`
`1/1000
`
`1/10000
`
`Flowcytometry
`
`PAGE
`
`separation
`
`Magnetic
`
`separation
`
`Magnetic
`
`separation
`
`Magnetic
`
`NA
`
`separation
`
`Magnetic
`
`NA
`
`NA
`
`NA
`
`K-ras
`
`Gene
`
`1/1000
`
`Microfluidicdevice
`
`Sensitivityb
`
`Detection
`
`Separation
`
`Direct digital mutation detection
`Direct digital quantification methods are based on the
`discrimination of the genotype of single DNA molecules
`followed by single molecule amplification and/or detec-
`tion. As described above, the enzymatic or physical
`methods used to distinguish mutant from wild-type
`sequences limit the sensitivity and specificity of these
`assays. For example, the initial allele discrimination of
`DNA molecules immobilized on glass slides can be done
`using a ligation assay [16] or a surface-invasion cleavage
`assay [17]. These strategies can at best detect one mutant
`in 1500 wild-type molecules. After the allele discrimi-
`nation reaction,
`allele-specific
`circularized single-
`stranded DNA molecules are hybridized to the ligation
`or cleavage products and used as a template for sub-
`sequent rolling-circle amplification (RCA). RCA pro-
`duces a long single-stranded concatemerized molecule
`containing multiple copies of the complementary circular
`starting sequence. The amplified DNA sequences are
`then labeled by hybridizing allele-specific fluorescent
`oligonucleotides to the RCA products. The image of
`the slide taken after the labeling is used to count the
`individual molecules. Recently, a similar solution-based
`approach was used for the digital analysis of DNA from
`pathogens [18]. This homogeneous test has the poten-
`tial to be adapted to mutation detection and quantifi-
`cation. Briefly, the assay uses ligation for the circulariza-
`tion of padlock probes,
`subsequent RCA, and a
`microfluidic device attached to a fluorescence microscope
`to count the molecules. Another type of assay allows the
`absolute quantification of unknown mutations in DNA
`fragments, but not the number of wild-type DNA mol-
`ecules [19]. The test is based on capturing the target
`sequences on beads coated with complementary probes,
`selective digestion of wild-type molecules, and the sub-
`sequent quantification of intact mutant molecules by
`digital PCR (see below). The high sensitivity of this test
`(1/100 000) can only be accomplished by performing
`several rounds of mutant enrichment, which unfortu-
`nately makes the assay difficult to use for clinical appli-
`cations.
`
`Several years ago, fluorescence-based single molecule
`detection (SMD) approaches were introduced. These
`are based on the direct visualization of individual fluor-
`escently labeled DNA molecules without the need for
`enzymatic amplification [20,21] So far only one of these
`approaches has been used for the enumeration of point
`mutations in solution [22]. The assay is based on the
`allele-specific ligation of fluorescence resonance energy
`transfer (FRET) probes generating molecular beacons
`upon successful ligation. A laser-based fluorescence sys-
`tem attached to a heatable microfluidic device is used to
`detect photon bursts generated by the FRET reaction.
`SMD technologies have also been developed for the
`analysis of single DNA molecules immobilized on solid
`
`Pyrosequencing
`
`nonmagneticbeads
`
`Emulsion,PCRon
`
`Proteintruncationtest
`
`Singlebaseextension
`
`Singlebaseextension
`
`Molecularbeacons
`
`PCRinsolution
`
`magneticbeads
`Emulsion,PCRon
`magneticbeads
`Emulsion,PCRon
`magneticbeads
`Emulsion,PCRon
`
`bMutant/wild-type.
`aQuantificationofmutantpopulationonly;
`NA,notapplicable;ND,notdetermined.
`
`Indirect,scanning
`
`stopcodons
`
`Indirect,scanningfor
`
`Indirect,locus-specific
`
`Indirect,locus-specific
`
`pyrosequencing
`
`DigitalPCRand
`truncationtest
`
`DigitalPCRandprotein
`singlebaseextension
`
`DigitalPCR,RCAand
`
`baseextension
`
`DigitalPCRandsingle
`beaconhybridization
`
`Indirect,locus-specific
`
`DigitalPCRandmolecular
`
`beaconhybridization
`
`Molecularbeacons
`
`PCRinsolution
`
`Indirect,locus-specific
`
`DigitalPCRandmolecular
`
`Restrictiondigest
`
`PCRinsolution
`
`Direct,locus-specific
`
`Invasivecleave
`
`oligonucleotides
`
`Ligationof
`
`molecularbeacons
`
`Ligationof
`
`NA
`
`Direct,locus-specific
`
`RCA
`
`Direct,locus-specific
`
`LigationandRCA
`
`NA
`
`Direct,locus-specific
`
`LigationandFRET
`
`digitalPCR
`
`Restrictiondigestand
`
`goldnanoparticles
`Invasivecleavageand
`
`Discrimination
`
`Amplification
`
`Assayprinciple
`
`Technique
`
`Table1Digitalassaysappliedtomutationdetection
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 4 of 7
`
`

`

`40 Cancer biology
`
`there is only one base-specific
`supports. Currently,
`approach described in the literature that could readily
`be applied to the enumeration of mutations. The test is
`based on a surface-invasion cleavage assay followed by
`the detection of mutant alleles using gold nanoparticles
`[23]. Recently, technologies for the sequencing of single
`DNA molecules on solid support have also been intro-
`duced. These are intended for the low-cost resequencing
`of complete genomes, but could eventually be adapted
`for the quantification of genetic variations in clinical
`samples [24–26]. Most notable are two approaches that
`use cyclic array sequencing. These rely on the extension
`of a DNA template hybridized to immobilized primers by
`a polymerase in the presence of fluorescently-labeled
`nucleotides [25,27]. Unfortunately, the sensitivity of
`these methods is limited by the error rates of the
`DNA polymerases currently available for sequencing.
`These error rates are at least one log higher compared
`with DNA polymerases conventionally used for PCR.
`
`Indirect digital mutation detection
`Indirect digital mutation quantification involves an
`initial compartmentalized amplification of single DNA
`molecules followed by allele discrimination and detec-
`tion. Cloning of DNA fragments followed by the sequen-
`cing of individual bacterial colonies is the most basic
`form of digital analysis. Unfortunately, this approach is
`time-consuming and labor-intensive. In 1988, it was
`demonstrated that PCR can be efficiently performed
`on single DNA templates [28]. This opened the way
`for several applications, reviewed elsewhere [12,13].
`One such technique, which came to be known as digital
`PCR, proved to be a powerful tool for single molecule
`counting and quantification of somatic mutations in
`clinical samples [29]. In digital PCR assays, multiple
`PCR reactions are performed in parallel at DNA con-
`centrations so low that most reactions contain zero or one
`template molecule and thus can be amplified clonally.
`Each resulting DNA pool is then analyzed individually
`for the presence of mutant and wild-type sequences by
`using fluorescent allele-specific molecular beacons. The
`digital PCR approach has been used for detection of
`K-ras mutations in various clinical samples [29 –32].
`Protein truncation tests which can be used to scan for
`stop codons within a target sequence have also been
`combined with digital PCR [33,34]. In contrast to more
`conventional mutation detection methods, the sensi-
`tivity of digital PCR approaches is not limited by the
`DNA polymerase error rate. In the worst-case scenario,
`an error occurs in the first cycle of a double-stranded
`single molecule PCR reaction which will result in a
`mutant to wild-type ratio of one to four. This would
`translate to a mixed mutant/wild-type signal
`in the
`subsequent genotyping assay and thus be excluded from
`the calculation. The sensitivity is instead limited by the
`number of molecules that can be analyzed and the false-
`
`positive rate of the mutation detection assay. In prin-
`ciple, the latter has the least influence on the detection
`limit as the analysis only needs to distinguish analytes
`that are exclusively wild-type or exclusively mutant.
`The main limitation of the originally described digital
`PCR techniques was the cost and labor involved in
`performing a large number of individual PCR reactions.
`To address these issues, several methods allowing
`millions of single-molecule PCR reactions to be per-
`formed in a single assay have been developed. One way
`to achieve this is by performing single-molecule PCRs in
`a thin polyacrylamide film poured on a glass microscope
`slide [35]. The amplification results in discrete DNA
`colonies in the polymer matrix (polonies). Another
`approach involves BEAMing (beads, emulsions, ampli-
`fication and magnetics) which allows single-molecule
`PCR reactions to be performed on magnetic beads in
`water-in-oil emulsions [36]. An alternative way to gen-
`erate beads coated with clonally amplified PCR products
`is by adding single beads into wells of a picotiter plate
`that can be used for digital PCR [37]. The bead suspen-
`sion obtained after solid-phase PCR accurately reflects
`the DNA diversity present in the template population
`and can therefore be used for mutation quantification.
`Recently, BEAMing followed by single base extension
`(SBE) and flow cytometry was used to quantify the level
`of mutated DNA circulating in the plasma of colorectal
`]. The detection limit of this assay
`tumor patients [15
`was one mutant DNA molecule in a background of 10 000
`wild-type DNA molecules. This threshold was deter-
`mined by the error rate of the DNA polymerase used for
`the preamplification of the limited amounts of plasma
`DNA. The sensitivity of BEAMing-based assays can also
`be constrained by the signal-to-noise ratio of the bead-
`based SBE assay. Therefore, RCA was used to increase
`the DNA copy number on the beads, making the SBE
`assay more specific [38]. Besides analyzing a mutation at
`a specific location, the DNA on beads can also be used as
`a template for sequencing. 454 Life Sciences developed
`an approach to sequence DNA on individual beads in
`parallel [39]. This system is commercially available and
`has been used for the identification and quantification of
`epidermal growth factor receptor (EGFR) gene mutations in
`lung cancer biopsies [40]. Another strategy used DNA
`coated beads immobilized in a polyacrylamide gel as a
`template for cycle sequencing by ligation [41].
`
`The clinical application of counting mutations
`Cancers are currently managed by a variety of clinical
`markers, which generally include patient symptoms,
`radiographic evaluation,
`routine laboratory tests and
`pathologic evaluation. These markers are used not only
`for diagnosis, but also as prognostic and predictive mar-
`kers, for tumor staging, and as markers for tumor response
`and detection of residual disease. More sensitive and
`specific biomarkers could aid cancer diagnosis and man-
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 5 of 7
`
`

`

`Quantification of mutant DNA in cancer Diehl and Diaz 41
`
`Conclusions
`The development of digital technologies has resulted in
`the ability to precisely and accurately count mutant and
`normal DNA molecules present in a sample. This is an
`improvement over the analog approaches, which can only
`estimate relative amounts. Indirect digital PCR-based
`assays are also more sensitive. When digital assays are
`performed in a highly parallel setup that only requires low
`amounts of enzymes, digital and analog approaches run at
`a similar cost.
`
`Beyond the application of early detection and mutation
`defining diagnosis, mutant DNA quantification could also
`be used to estimate tumor burden in already diagnosed
`cancer patients. The amount of mutant DNA could
`predict recurrence more specifically than currently avail-
`able protein biomarkers. We envision a simple mutation-
`specific blood test that can diagnose cancer at earlier
`stages than before and measure patients’ tumor burden.
`In addition, a blood test of this nature will have the
`potential to define therapy based on tumor-specific
`mutations and/or detect treatment-resistant mutations
`as they arise.
`
`Acknowledgements
`We thank Will Hendricks, Nickolas Papadopoulos, Ian Cheong, Kerstin
`Schmidt, and Bert Vogelstein for critical review of the manuscript.
`
`References and recommended reading
`Papers of particular interest, published within the annual period of review, have
`been highlighted as:
`
`of special interest
` of outstanding interest
`Additional references related to this topic can also be found in the Current
`World Literature section in this issue (pp. 71–74).
`
`1
`
`2
`
`Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat
`Med 2004; 10:789–799.
`
`Tokunaga E, Oki E, Nishida K, et al. Trastuzumab and breast cancer: devel-
`opments and current status. Int J Clin Oncol 2006; 11:199–208.
`
`3
`
`Strom CM. Mutation detection, interpretation, and applications in the clinical
`laboratory setting. Mutat Res 2005; 573 (1–2):160–167.
`This review highlights the role of quantitative assays in diagnostic testing.
`
`4
`
`5
`
`Sjoblom T, Jones S, Wood LD, et al. The consensus coding sequences of
`human breast and colorectal cancers. Science 2006 [Epub ahead of print].
`
`Foy CA, Parkes HC. Emerging homogeneous DNA-based technologies in the
`clinical laboratory. Clin Chem 2001; 47:990–1000.
`
`6 Whitcombe D, Newton CR, Little S. Advances in approaches to DNA-based
`diagnostics. Curr Opin Biotechnol 1998; 9:602–608.
`
`7 Mike Makrigiorgos G. PCR-based detection of minority point mutations. Hum
`Mutat 2004; 23:406–412.
`
`8 Ogino S, Kawasaki T, Brahmandam M, et al. Sensitive sequencing method for
`KRAS mutation detection by pyrosequencing. J Mol Diagn 2005; 7:413–
`421.
`
`9
`
`Janne PA, Borras AM, Kuang Y, et al. A rapid and sensitive enzymatic method
`for epidermal growth factor receptor mutation screening. Clin Cancer Res
`2006; 12 (3 Pt 1):751–758.
`
`10 Shi C, Eshleman SH, Jones D, et al. LigAmp for sensitive detection of single-
`nucleotide differences. Nat Methods 2004; 1:141–147.
`
`11
`
` Kimura H, Kasahara K, Kawaishi M, et al. Detection of epidermal growth factor
`
`receptor mutations in serum as a predictor of the response to gefitinib in
`patients with nonsmall-cell lung cancer. Clin Cancer Res 2006; 12:3915–
`3921.
`Report on a blood-based assay to detect EGFR mutations in lung cancer. They
`used a Scorpion Amplified Refractory Mutation System for mutation detection.
`
`agement by giving physicians additional information to
`make decisions. For example, PSA is a widely-used
`marker available for prostate cancer, but lacks specificity
`to be used independently for the diagnosis of malignancy
`[42]. Markers for other solid cancers currently used in the
`clinic suffer from similar problems.
`
`New markers that are currently being developed are
`either based on the in-vivo or ex-vivo detection of cancer.
`In-vivo detection strategies rely on imaging technologies
`and have been reviewed elsewhere [43]. Biomarkers for
`ex-vivo detection most commonly include cancer-specific
`proteins or nucleic acids that can be assayed within cancer
`tissues, blood or other bodily fluids collected from cancer
`patients [44,45]. Of these markers, gene mutations are
`arguably the most specific. Thus far, however, mutations
`have only been routinely applied in the clinic for the
`diagnosis of hereditary cancer syndromes. To apply
`mutation detection for early detection and management
`of malignancies, sensitive and specific assays are essen-
`tial. Several investigational studies have started using
`quantitative approaches for this purpose. These studies
`include the noninvasive detection of mutant adenomatosis
`polyposis coli (APC) DNA in stool or plasma from subjects
`with colorectal
`adenomatous polyps
`and cancers
`[15,33]. Furthermore, the detection of mutations in
`the EGFR gene in biopsies and plasma has been under
`investigation partly due to a number of reports showing
`that these mutations are directly related to the efficacy of
`treatments targeting EGFR [9,11,40].
`
`Quantitative diagnostics have also become an essential
`part of the management of chronic myeloid leukemia
`patients treated with imatinib mesylate (Gleevec), a drug
`that appears to target the breakpoint cluster region Abel-
`son (BCR-ABL) fusion protein [46]. A real-time quanti-
`tative reverse transcriptase PCR (RT-PCR) assay of the
`BCR-ABL RNA transcript provides a measure of the total
`leukemia cell mass and thus can be used for the monitor-
`ing of patients. In fact, the rate and size of change in
`BCR-ABL load can predict long-term response to therapy
`and potential cure. The remaining challenge will be the
`standardization of BCR-ABL quantification across differ-
`ent institutions [47]. Another approach that may become
`standard in clinical practice is the quantification of
`Epstein-Barr virus (EBV) DNA in plasma, which is
`closely associated with nasopharyngeal carcinoma [48].
`It appears that plasma EBV DNA is useful for determin-
`ing prognosis and monitoring response to treatment. New
`concepts that could be applied to a broader range of
`tumors, in particular solid malignancies, are currently
`being investigated. These are based on the counting of
`intact tumor cells in the bone marrow or circulation
`[49,50], the quantification of DNA with tumor-specific
`methylation patterns [51], or the detection of tumor-
`specific RNA [52].
`
`Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
`
`The Johns Hopkins University Exhibit JHU2004 - Page 6 of 7
`
`

`

`42 Cancer biology
`
`12 Pohl G, Shih Ie M. Principle and applications of digital PCR. Expert Rev Mol
`Diagn 2004; 4:41–47.
`
`33 Traverso G, Shuber A, Levin B, et al. Detection of APC mutations in fecal DNA
`from patients with colorectal tumors. N Engl J Med 2002; 346:311–320.
`
`13
`
` Kraytsberg Y, Khrapko K. Single-molecule PCR: an artifact-free PCR
`
`approach for the analysis of somatic mutations. Expert Rev Mol Diagn
`2005; 5:809–815.
`Recent review about digital PCR approaches for the analysis of somatic mutations.
`
`14 Castro A, Williams JG. Single-molecule detection of specific nucleic acid
`sequences in unamplified genomic DNA. Anal Chem 1997; 69:3915–
`3920.
`
`15 Diehl F, Li M, Dressman D, et al. Detection and quantification of mutations in
`
`the plasma of patients with colorectal tumors. Proc Natl Acad Sci USA 2005;
`102:16368–16373.
`Application of BEAMing for the quantification of mutant DNA in plasma from
`cancer patients.
`
`16 Lizardi PM, Huang X, Zhu Z, et al. Mutation detection and single-molecule
`counting using isothermal
`rolling-circle amplification. Nat Genet 1998;
`19:225–232.
`
`17 Nie B, Shortreed MR, Smith LM. Scoring single-nucleotide polymorphisms at
`the single-molecule level by counting individual DNA cleavage events on
`surfaces. Anal Chem 2005; 77:6594–6600.
`
`18
`Jarvius J, Melin J, Goransson J, et al. Digital quantification using amplified
`
`single-molecule detection. Nat Methods 2006; 3:725–727.