| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
AstraZeneca, Alderley Park, Macclesfield, Cheshire SK10 4TG, United Kingdom.
2
Roy Castle International Centre for Lung Cancer
Research, 200 London Rd., Liverpool L3 9TA, United Kingdom.
3
Department of Biological Sciences, Manchester
Metropolitan University, Manchester M1 5GD, United Kingdom.
4
RSOM, AstraZeneca, Mereside, Alderley Park,
Macclesfield, Cheshire SK10 4TG, United Kingdom.
5
Molecular Genetics and Oncology Group, Clinical Dental
Sciences, University of Liverpool, Liverpool L69 3BX, United Kingdom.
a Address correspondence to this author at: AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK. Fax 44-1625-514463; e-mail simon.clayton{at}astrazeneca.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: ARMS reactions were used to detect seven commonly occurring mutations in the K-ras oncogene. These mutations produce amino acid changes in codon 12 (Gly to Ala, Arg, Asp, Cys, Ser, or Val) and codon 13 (Gly to Asp). A control reaction was used to measure the total amount of amplifiable K-ras sequence in a sample so that the ratio of mutant to wild-type sequence could be measured. Quantitative data were confirmed for a selection of samples by an independent cloning and sequencing method. The assay was used to analyze 82 lung tumor DNA samples.
Results: The assay detected K-ras mutations in 44% of adenocarcinomas, which is equivalent to frequencies reported in the literature using ultrasensitive techniques. Forty-six percent of squamous carcinomas were also positive. The ratio of mutant sequence in the tumor DNA samples was 0.04–100%.
Conclusions: The assay is homogeneous, with addition of tumor DNA sample being the only step before results are generated. The quantitative nature of the assay can potentially be used to define the analytical sensitivity necessary for any specified diagnostic application of K-ras (or other) point mutation detection.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A vast increase in knowledge of the molecular processes underlying malignant change and progression has occurred, which will be useful in the development of new therapies and in areas such as diagnosis of occult disease (4)(5)(6), monitoring of disease progress (7)(8), presymptomatic disease detection (screening), and therapy selection/optimization (9)(10)(11). In addition, there is increasing excitement about the future potential of individualized therapy (12).
The use of molecular markers in the early detection and management of lung cancer patients has been reviewed (13)(14). One such marker, mutations of the K-ras oncogene, has been reported in lung tumors (15) and a variety of other cancers (16). Clinical applications of K-ras mutation analysis in lung disease include its use as a diagnostic marker of malignancy (of greater sensitivity than cytology alone) in sputum and bronchoalveolar lavage (BAL)1 samples (17) and its use as a marker of prognosis in lung cancer patients (15)(18)(19). The presence of K-ras mutations in lung tumors has been linked with shorter disease-free and overall survival (20)(21)(22)(23). Some studies have also specified which codon 12 and 13 changes have prognostic value and which have less clinical relevance (24)(25). Improved stratification of lung tumors using molecular markers such as K-ras may lead to the identification of a patient subgroup who require a more aggressive therapy.
Techniques for the detection of K-ras mutations in clinical samples have been reviewed (26). Because the detection limits of currently used methods vary, different assays used for the same clinical application lead to different results. It is therefore vital for any novel technique to be fully validated so that the sensitivity and specificity of mutation detection in the relevant clinical sample can be determined. The limit of detection then needs to be taken into account when comparing data obtained by use of different molecular assays to answer the same clinical question (27). We investigated two approaches to K-ras mutation detection in samples from lung cancer patients. Both methods used allele-specific amplification (ARMSTM) for mutant sequence discrimination. The first method, the Elucigene K-RAS7 method (AstraZeneca Diagnostics, Abingdon, UK), is similar to a previously reported assay (28). The second, a real-time ARMS assay, provides quantitative data for samples positive for K-ras mutations. The real-time quantitative approach, with its sensitive detection and mutant sequence quantification, allows the determination of the true detection limit in any clinical application.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
allele-specific amplification
Two systems that detect the seven most commonly reported
K-ras mutations in codons 12 and 13 of exon 1 were used in
this study. The prototype of a commercially available method (Elucigene
K-RAS7) has been described previously (29).
elucigene K-RAS7 method
The Elucigene K-RAS7 method was used in exact compliance with the
manufacturer’s instructions. Briefly, 5 µL of each DNA sample was
added to each reaction mixture and thermocycled in a 9600 machine (PE
Applied Biosystems). AmpliTaq Gold DNA polymerase activation was
carried out at 94 °C for 20 min, followed by 36 cycles of 94 °C
for 45 s (denaturation), 63 °C for 45 s (annealing); and
72 °C for 45 s (extension), with a final extension step of
72 °C for 10 min. Amplification products were analyzed by gel
electrophoresis in accordance with the manufacturer’s recommendations.
real-time arms K-ras mutation analysis
The second system combined the approach described above with
real-time fluorescent detection of amplification products, thus
introducing quantification into the basic ARMS format (Fig. 1
). Amplification products were detected by molecular beacons
(30). Each reaction contained 50 µmol/L ARMS and common
primers, 100 µmol/L beacon, and 60 nmol/L ROX internal reference (PE
Applied Biosystems) together with the standard reaction components
described previously (29). 2'-O-Methyl RNA
beacons complementary to the sense and antisense strands of
K-ras exon 1 were designed:
5'-FAM-CGCGGUCCCUUGACGAUACAGCUAAUUCAGAACGCG-MR-3' and
5'-FAM-CGCGGUGCUGAAAAUGACUGAAUAUAAACUUGUGGACGCG-MR-3', respectively.
Both probes were labeled with a 5' 6-carboxyfluorescein (FAM)
fluorophore and a 3' methyl red (MR) monomer, which acted as a
nonfluorogenic (dark) quencher (Oswel DNA Service).
2'-O-methyl RNA and not DNA was used here to avoid
fluorescent signal generation from the 5'-3' exonuclease activity of
Taq polymerase. The cycling conditions were as follows:
94 °C for 20 min to activate the AmpliTaq Gold (PE Applied
Biosystems), followed by 50 cycles of a 94 °C denaturation step and
a 60 °C combined annealing and extension step. An additional
reaction, designed to amplify all K-ras exon 1 sequences,
was carried out under the same conditions as for the mutant-specific
reactions (Fig. 2
). When amplification occurred, relative fluorescence
increased exponentially in relation to cycle number, and the point at
which it passed a fixed threshold above baseline was defined as the
threshold cycle number, or Ct (in accordance with
PE Applied Biosystems).
|
|
With real-time ARMS K-ras mutation detection as described above, a reference data set was generated to establish background signals in the absence of mutations, using wild-type genomic DNA extracted from the peripheral blood of 12 healthy volunteers. This consisted of more than 2000 individual amplifications using DNA of a predetermined concentration. DNA was extracted using a DNA extraction method (Qiagen). DNA was quantified with care (A260 readings were taken from multiple dilutions of each stock DNA to ensure maximum accuracy of quantification). Initial template concentrations of 1 x 103, 1 x 104, and 1 x 105 diploid genome-equivalents (defined as 6–600 ng) of human DNA were selected as the dynamic range over which control data would be generated. Results of the eight real-time ARMS reactions (Ct values) were analyzed by plotting K-ras mutant ARMS reaction Ct against control Ct.
Twelve peripheral blood DNA samples were analyzed for each ARMS reaction in batches of eight replicates at three concentrations, by three independent operators. This gave a total of 288 results for each ARMS test and 2016 reactions in total. Inauthentic priming of mutant-specific ARMS primers occurs at extremely low efficiency on wild-type DNA. This gives rise to a yield of inauthentic (non-mutant-derived) product (Yi) during a 50-cycle amplification. Statistical analysis of product appearance (the Ct value at which amplification products are first detected when fluorescent signal increases above the threshold value) attributable to rare mispriming on wild-type DNA alone, from the 2016 separate reactions, was therefore performed. One-sided prediction intervals were calculated on linear regression analyses of the Yi data for each ARMS reaction, and 98% confidence limits were then set for subsequent analysis of samples of unknown mutant status. Ct values from clinical samples of unknown K-ras mutant status containing the wild-type K-ras exon 1 sequence have only a 1% probability of falling below the 98% confidence limit.
Mutant-only reactions were set up using cloned DNA sequences corresponding to each of the seven most prevalent K-ras mutations, as described previously (29). As a control, the wild-type cloned K-ras sequence was amplified in the real-time assay to ensure that there were no significant differences in the background signals generated when either genomic DNA or amplicon was used as template. Mutant and wild-type cloned DNA templates were amplified over a concentration range similar to that described above for the genomic template. All of the cloned and amplified K-ras sequences were verified by standard sequencing methodology (data not shown).
DNA derived from lung tumor samples of unknown K-ras mutant status was analyzed in relation to the data described above. A 5-µL aliquot of each sample was added to the seven ARMS reactions and a control reaction. The reactions were thermocycled and analyzed in an ABI 7700 as described above. The ratio of mutant to wild-type K-ras exon 1 in the mutant-positive tumor samples was calculated from the data obtained using the cloned mutant K-ras sequence. This was done by interpolation from the equation of the regression analysis of results obtained using 100% mutant K-ras sequence. For example, in the codon 12 Gly-to-Arg reaction, the equation of the 100% mutant regression line was: y = 0.9955x - 0.5808, where y is the ARMS Ct, and x is the control Ct. Therefore, a sample with a control Ct of 32 cycles would have an ARMS Ct of 31.3 cycles if it contained 100% mutant (Gly-to-Arg) K-ras alone. However, if the actual ARMS Ct value was 32.3 cycles or 33.3 cycles, then the sample contained 50% or 25% mutant DNA, respectively. This calculation assumes a doubling in amplification product amount after each cycle (31). To control for this mathematical calculation, samples with established mutant-to-wild-type ratios, from cloning and sequencing experiments (29), were analyzed empirically, and the ratio of mutant to wild-type sequence was compared.
mutation detection by pcr-allele-specific oligonucleotide
hybridization
Lung tumor DNA samples, analyzed in this study by the real-time
assay, were independently analyzed by PCR-allele-specific
oligonucleotide (ASO) hybridization. Briefly, samples were amplified in
two rounds of PCR, each consisting of 35 cycles of 94 °C for 1 min,
60 °C for 1 min, and 72 °C for 1 min. PCR products were purified
using the Wizard reagent set (Promega), and their
concentrations were estimated by agarose gel electrophoresis using
molecular weight markers (Life Technologies). Approximately 40
ng of PCR product was transferred to Hybond N+ membrane (Amersham
Pharmacia Biotech). Labeled ASO probe (40 ng) was hybridized to the
membrane at 40 °C, filters were washed using 3 mol/L
tetramethylammonium chloride (Sigma) at 56 °C, and probe was
detected as described in the ECL method (Amersham Pharmacia Biotech).
The sensitivity of the ASO approach was established using carefully
prepared mixtures of mutant and wild-type genomic DNA (50%, 20%,
10%, 5%, and 1% mutant) as described previously (29). A
selection of lung tumor DNA samples, positive for a mutation by
real-time analysis, were analyzed by ASO hybridization. Tumor DNA
samples that were found to contain only wild-type K-ras by
real-time analysis were analyzed by ASO as a control.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), was confirmed by two alternative methods. In the first
method, an ASO-based assay was used to verify the results for a
selection of mutation-positive lung tumor samples. Fig. 3
shows the ASO hybridization results for the codon 12 GGT-to-CGT
(Gly-to-Arg) change. The signals for 50%, 20%, 10%, 5%, and 1%
mutant controls (mutation-positive cell line/wild-type DNA admixtures)
were higher than signals observed for controls representing the six
other K-ras mutations or tumor samples known to contain
mutations other than Gly-to-Arg. Sample 42, however, was positive for
the Gly-to-Arg mutation by real-time analysis and had a signal above
background by ASO. Signals from lung tumor samples containing high
concentrations of codon 12 Gly-to-Cys mutant DNA were higher than the
background wild-type signal (Fig. 3
). For example, sample 38 contained
100% mutant Cys DNA by real-time analysis, and this cross-reacted with
the ASO probe for the codon 12 Gly-to-Arg mutation. The probe, intended
to be specific for Arg, contains a guanosine at the first position of
codon 12, complementary to the cytosine in mutant (Gly-to-Arg)
K-ras exon 1. This interacts very weakly with the guanosine
at the same position in wild-type DNA. However, the codon 12 Gly-to-Cys
mutation, resulting from a guanosine-to-thymidine change at this
position, means that the ASO probe binds with a higher affinity and so
gives a higher signal than other codons (G-T mismatches are weakly
destabilizing compared with G-G mismatches) (32).
|
|
A second approach was used to confirm the mutation status derived from
the ARMS assays. Colorectal tumor DNA samples, in which
K-ras mutant sequences were quantified previously by cloning
and sequencing (29), were tested in the new ARMS assay.
These samples were analyzed using the real-time assay, and
mutant/wild-type quantities were compared. The results of these
experiments are summarized in Table 1
. The reported frequencies obtained by the two approaches were
largely in agreement, and discrepancies may have been attributable to
sampling issues because different aliquots of the tumor DNA samples
were analyzed at two different geographic sites with a long time
interval between measurements (3 years). These two approaches confirmed
both the qualitative and quantitative aspects of the real-time ARMS
assay, within the sensitivity range of the ASO procedure (
5% mutant
target).
|
An example of the Elucigene K-RAS7 assay results are shown in Fig. 4
. Tumor DNA was extracted from both paraffin-embedded and frozen
tissue samples. The paraffin-embedded tissue samples in general showed
either partial or complete loss of the upper control product (360 bp).
This is characteristic of amplified DNA obtained from this sample
source in which genomic DNA is generally degraded. The lower control
product (111 bp) usually is present even when the upper control is
missing. The ARMS products were 158–190 bp in length, and were
therefore largely unaffected by the DNA degradation observed for
paraffin-embedded samples. However, in samples where the upper control
product was missing and the lower control was present (although absence
of an ARMS product is probably attributable to the fact that the sample
is wild type), DNA fragmentation may have occurred, i.e., fragmentation
of target mutant genomic DNA above 111 bp (lower control product size)
in length. These events were recorded as wild type in the results
summarized in Table 2
.
|
|
Tumor samples 45, 75, 84, and 120 were positive for codon 12 GGT-to-GAT
(Gly-to-Val) mutations, as indicated by the presence of ARMS products
on the gel (Fig. 4
). DNA concentrations after extraction from the lung
tissue were not standardized before amplification, and this most likely
accounted for the visible differences in overall product yields among
the samples. One advantage of the real-time assay is that initial DNA
concentrations do not require standardization.
K-ras mutations were detected in 15% (9 of 62) lung tumor
DNA samples by Elucigene KRAS-7 ARMS analysis (Table 2
.). Nineteen
percent of adenocarcinomas and 10% of squamous cell carcinomas were
found to have mutations. No K-ras mutations were detected
for the other lung tumors analyzed. The frequencies of the mutant
sequences in K-ras-positive lung tumor samples were
calculated using real-time analysis, by interpolation from the equation
of the regression analysis for 100% mutant DNA for each ARMS
reaction, and were expressed as percentage of total
K-ras exon 1 in the samples (Table 3
). All mutations detected by the Elucigene K-RAS7 method were
also detected by real-time analysis. Twenty additional squamous cell
carcinomas were analyzed by the real-time assay to further characterize
K-ras mutations occurring in this tumor type. The results of
the real-time analyses are summarized in Table 4
. The ratios of mutant to wild-type sequences in the positive
samples were expressed as percentages of the total number of
K-ras exon 1 sequences (Table 5
).
|
|
The nature of ARMS mispriming on non-mutant DNA is determined by the 3'
nucleotide species that mispairs with the wild-type nucleotide in the
target sequence. These mismatches are not absolute and produce rare
priming events, occurring at extremely low efficiencies, that depend on
the strength of the mismatched base pair interaction (32).
These rare mispriming events were controlled for in our assay, thus
allowing the reliable analysis of clinical samples of unknown
K-ras status. The strength of the mismatched base pair
interaction is reflected in the Yi data for each
test (see Fig. 5
). For example, in the codon 12 Gly-to-Ala test (Fig. 5B
), the
ARMS primer specific for the GGT-to-GCT change has a 3' terminal
guanosine that is strongly destabilized from the guanosine at position
2 in codon 12 in wild-type K-ras. This means that
Yi resulting from inauthentic priming events in
this assay is rarely seen. In the codon 12 Gly-to-Asp test (Fig. 5A
),
the ARMS primer has a 3' terminal thymidine mismatched with the
guanosine in wild-type K-ras exon 1. This interaction is
less destabilizing, and so Yi appears earlier in
the amplification process.
|
Negligible variations in Yi data were observed
between operators or between DNA samples of equivalent DNA
concentration (results not shown). Fig. 5
shows the real-time data for
all lung tumor samples in all seven ARMS tests. As described in
Materials and Methods, tumor samples were analyzed in
relation to Yi data to determine mutational
status. There was a 1% chance of a sample appearing below the lower
limit of the 98% confidence interval if it contained wild-type
sequence alone, and if such a case was observed, it was considered
positive for the mutation interrogated by that specific ARMS primer.
Duplicate analysis of each sample markedly reduced the chance of a
sample appearing as mutation-positive twice when it contained wild-type
sequence alone. The quantity of mutant sequence in these samples was
calculated by interpolation using the 100% mutant data (not shown).
Fifty cycles is the endpoint of the assay, and samples were ascribed a
Ct of 50 cycles if they had not yielded a
detectable ARMS product by this point in the amplification process. The
majority of tumor samples were negative for K-ras mutations
in each test. This provided the opportunity to observe the nature of
mispriming events for individual ARMS primers on the wild-type
K-ras exon 1 sequence. At high concentrations of starting
genomic target DNA (>1 x 104 copies),
mispriming events did occur, and Yi appeared
relatively early and with limited variability. At lower starting
template concentrations, however, mispriming events were rarer and
amplification varied stochastically, with detection of
Yi dramatically reduced to the extent that
samples often gave no detectable product up to 50 cycles.
Addition experiments conducted with low ratios of mutant to
wild-type DNA added to a K-ras mutant-negative tumor sample
showed that this type of sample still yielded ARMS products below the
background cutoff for wild-type DNA if mutant sequence was present
(data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There are two major reasons for the increased sensitivity of the real-time assay over the Elucigene method and other assay systems. The assay was designed to identify mutation-positive samples as "not wild type" by comparison to a data set obtained using wild-type DNA alone. This maximizes sensitivity compared with alternative methods that specify a cutoff at which the "signal" is considered positive in relation to a positive control containing a low mutant fraction. In addition, the behavior of the ARMS primers on wild-type DNA was established over a dynamic concentration range (100-fold) in a 50-cycle amplification. This range was chosen by reference to DNA yields routinely obtained from various clinical samples entering our laboratory (tissue and cell suspensions in body fluids). Therefore, unknown clinical samples containing different DNA amounts in the fixed assay volume could all be analyzed in the same 50-cycle amplification. Measurement and standardization of the DNA concentration before any molecular assay can be unrealistic in the clinical setting when overall yields are low and the sample is limited. If an assay involves PCR with an early fixed endpoint (to maintain specificity), then unknown samples with low DNA concentrations and K-ras mutations will be reported as false negatives.
It is worth pointing out that a high percentage of squamous cancers were positive for K-ras mutations in this study. Mutations in the K-ras oncogene are not restricted to adenocarcinomas, and other investigators have demonstrated this (24)(34) or referred to it (28). There may be several reasons for differences in perception concerning the occurrence of K-ras mutations and whether they are restricted to the adenocarcinoma tumor type. These may include differences in the molecular assays used and heterogeneity of the tumor tissue. Multiple sections from a tumor sample may indicate either adenocarcinoma or squamous histological types, and this gives rise to the possibility that a histological definition of tumor type may not reflect the true nature of the cell population analyzed for K-ras mutations. There were no differences in the amounts of K-ras mutant sequence between the adenocarcinoma and squamous carcinoma types in this study.
The clinical significance of the quantity of K-ras mutation in a positive clinical sample is yet to be established. One study has separated K-ras results into low-fraction and high-fraction positives. The authors reported an increased stage of disease and decreased survival linked to the high-fraction positives, although statistical significance was not reached with the limited number of samples investigated (34). Mutation of the K-ras oncogene is considered an early event in lung cancer development (35), although not necessarily the first (36), and measurement of the mutation fraction may simply reflect disease stage. Larger studies using quantitative assays may establish the clinical relevance of molecular markers in relation to treatment or survival endpoints. Quantitative data may also be important when looking for malignant cells distant to the site of a tumor, such as in the screening or early diagnosis setting, where less-invasive clinical procedures are used. Quantitative data could be used in establishing the required assay sensitivity for a particular clinical application. For example, in lung cancer this would involve the analysis of sputum and BAL samples to potentially detect clinically occult tumors. Matched K-ras mutations have been demonstrated in adenocarcinomas of the lung and corresponding sputum (37) and BAL (38) samples in lung cancer patients. To date, we have used our assay on 108 bronchial lavage samples (58 from patients with a confirmed malignancy and 50 from non-malignant cases). In five mutation-positive BAL samples, cytology reported no malignant cells present despite malignancy being confirmed (data not shown). Three of eight BALs from confirmed adenocarcinomas were positive in the assay. One of 50 non-malignant cases was positive, with the patient reported as having heart disease.
The real-time assay used here offers a unique opportunity to assess the clinical significance of K-ras mutations in lung and other cancers. The sensitive nature of the assay provides the opportunity to screen large numbers of clinical samples for malignant cells in the early detection setting. The utility of K-ras mutation detection in BAL samples as a precursor to lung cancer diagnosis has been demonstrated previously (37). At present, we are using this method to investigate the utility of K-ras mutation detection for pancreatic and colorectal disease management. The assay has been designed to facilitate high-throughput sample analysis after routine nucleic acid extraction procedures for a range of clinical materials.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  W. A. Franklin, J. Haney, M. Sugita, L. Bemis, A. Jimeno, and W. A. Messersmith KRAS Mutation: Comparison of Testing Methods and Tissue Sampling Techniques in Colon Cancer J. Mol. Diagn., January 1, 2010; 12(1): 43 - 50. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Weichert, C. Schewe, A. Lehmann, C. Sers, C. Denkert, J. Budczies, A. Stenzinger, H. Joos, O. Landt, V. Heiser, et al. KRAS Genotyping of Paraffin-Embedded Colorectal Cancer Tissue in Routine Diagnostics: Comparison of Methods and Impact of Histology J. Mol. Diagn., January 1, 2010; 12(1): 35 - 42. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Jimeno, W. A. Messersmith, F. R. Hirsch, W. A. Franklin, and S. G. Eckhardt KRAS Mutations and Sensitivity to Epidermal Growth Factor Receptor Inhibitors in Colorectal Cancer: Practical Application of Patient Selection J. Clin. Oncol., March 1, 2009; 27(7): 1130 - 1136. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Archambeault, N. J. Flores, A. Yoshimi, C. P. Kratz, M. Reising, A. Fischer, P. Noellke, F. Locatelli, P. Sedlacek, C. Flotho, et al. Development of an allele-specific minimal residual disease assay for patients with juvenile myelomonocytic leukemia Blood, February 1, 2008; 111(3): 1124 - 1127. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N T van Heek, S J Clayton, P D J Sturm, J Walker, D J Gouma, L A Noorduyn, G J A Offerhaus, and J C Fox Comparison of the novel quantitative ARMS assay and an enriched PCR-ASO assay for K-ras mutations with conventional cytology on endobiliary brush cytology from 312 consecutive extrahepatic biliary stenoses J. Clin. Pathol., December 1, 2005; 58(12): 1315 - 1320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Willis, T. Lange, S. Demehri, S. Otto, L. Crossman, D. Niederwieser, E. P. Stoffregen, S. McWeeney, I. Kovacs, B. Park, et al. High-sensitivity detection of BCR-ABL kinase domain mutations in imatinib-naive patients: correlation with clonal cytogenetic evolution but not response to therapy Blood, September 15, 2005; 106(6): 2128 - 2137. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ogino, T. Kawasaki, M. Brahmandam, L. Yan, M. Cantor, C. Namgyal, M. Mino-Kenudson, G. Y. Lauwers, M. Loda, and C. S. Fuchs Sensitive Sequencing Method for KRAS Mutation Detection by Pyrosequencing J. Mol. Diagn., August 1, 2005; 7(3): 413 - 421. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. R. Hodgson, S. J. Clayton, F. Girdler, I. Brotherick, B. Shenton, D. Browell, M. Stuart, J. C. Fox, P. Ceuppens, C. A. Foy, et al. ARMSTM Allele-specific Amplification-based Detection of Mutant p53 DNA and mRNA in Tumors of the Breast, Clin. Chem., April 1, 2001; 47(4): 774 - 778. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or -negative (
). The
two regression lines represent mean Yi (—–) and the
99% confidence limit for Yi data (· · ·
·).