|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.03.7598 on February 27 2006 © 2006 American Society of Clinical Oncology. Serum Circulating Human mRNA Profiling and Its Utility for Oral Cancer Detection
From the School of Dentistry and Dental Research Institute, Division of Head & Neck Surgery/Otolaryngology, David Geffen School of Medicine; Department of Biostatistics, School of Public Health, Henry Samueli School of Engineering and Applied Science; Jonsson Comprehensive Cancer Center and Molecular Biology Institute, University of California, Los Angeles (UCLA); and the School of Medicine, University of Southern California, Los Angeles, CA. Address reprint requests to David T. Wong, DMD, DMSc, University of California Los Angeles, School of Dentistry, Dental Research Institute, 73-017 CHS, 10833 Le Conte Ave, Los Angeles, CA 90095; e-mail: dtww{at}ucla.edu
PURPOSE: The purpose of this study is to explore the presence of informative RNA biomarkers from human serum transcriptome, and evaluate the serum transcriptome diagnostics for disease detection. Oral squamous cell carcinoma (OSCC) was selected as the proof-of-concept disease. PATIENTS AND METHODS: Blood samples were collected from patients (n = 32) with primary T1/T2 OSCC and matched healthy patients (n = 35). Circulating RNA was isolated from serum and linearly amplified using T7 polymerase. Microarrays were applied for profiling transcriptome in serum from 10 cancer patients and controls. The differential gene expression was analyzed by combining the present calls, t tests, and fold-change statistics. Quantitative polymerase chain reaction (PCR) was used to validate the selected candidate RNA markers identified by microarray. Receiver operating characteristic curve and classification models were exploited to evaluate the diagnostic power of these markers for OSCC.
RESULTS: Human serum circulating mRNAs were presented by reverse transcriptase PCR. Microarray identified 2,623 ± 868 probes assigned present calls in OSCC (n = 10) versus 1,792 ± 165 in healthy patients (n = 10), indicating a higher complexity of serum transciptome in OSCC patients (P = .002, Wilcoxon test). Three hundred thirty-five serum RNAs exhibited significantly differential expression level between the two groups (P < .05, t test; fold CONCLUSION: The utility of serum transcriptome diagnostics is successfully demonstrated for OSCC detection. This novel concept could be developed as an adjunctive tool for disease diagnosis.
Accurately identified, biomarkers may provide new avenues for early cancer detection, and constitute targets for cancer risk assessment.1-3 However, the clinical utility of biomarkers to monitor tumor initiation, progression, and outcome is relatively limited. Biomarkers are compromised by their insufficient diagnostic sensitivity and specificity; and are thus not used for definitive diagnosis, but as an auxiliary approach to assist in clinical decision making. Emerging high-throughput technologies, including microarray and mass spectrometry, provide global information to observe genetic and proteomic alterations and to facilitate the discovery of new biomarkers with improved sensitivity and specificity. Moreover, the development of powerful bioinformatics methodologies has contributed to the measurement of thousands of gene expressions simultaneously.4,5 It is also essential that the detection tools are sufficiently noninvasive to allow widespread applicability. According to the Early Detection Research Network of the National Cancer Institute, the cancer-screening program is aimed at detecting tumors at a stage early enough that treatment is likely to be successful.6 However, most of the currently used markers have been identified either in cancer cell lines or in biopsy specimens from late invasive or metastatic cancers. The invasive nature of a biopsy also makes it unsuitable for cancer screening in high-risk populations. This suggests an imperative need for developing new diagnostic tools that would improve early detection.7 Recent advances in the field of biologic science have sparked new interest in the area of identifying cancer biomarkers in bodily fluids.8 It has been shown that identical mutation present in the primary tumor can be identified in the bodily fluids of the affected patients.8 Advantages of using bodily fluid as a diagnostic tool also resulted from its relatively noninvasive manner. mRNA in blood, semen, urine, and saliva has been proved as a novel resource to supplant conventional tools for disease identification.9,10 Cancer-related nucleic acids in blood, urine, and CSF have been used successfully as cancer biomarkers.11-13 In this scheme, circulating mRNA biomarkers in serum and plasma are the most widely investigated biomarkers for cancer detection. By reverse transcriptase polymerase chain reaction (RT-PCR), mRNA markers have been the targets for identifying patients with colorectal, breast, lung, and thyroid cancers, and malignant melanoma.14-18 While these studies have shown promise, they all were performed by testing a single mRNA marker. The diagnostic sensitivity and specificity were thus limited. Here, we report an effort to embark on examining and analyzing the global mRNA profiling in serum. High-density oligonucleotide microarrays were used for the first time in the global transcriptome profiling from serum. The serum transcriptome diagnostics are expected to help identify more informative biomarkers for cancer detection. The diagnostic utility of serum mRNA biomarkers was tested by using oral squamous cell carcinoma (OSCC) as the proof-of-concept disease.
Patients Thirty-two OSCC patients were recruited from medical centers at the University of California, Los Angeles, CA, and the University of Southern California (USC), Los Angeles, CA. Patients had recently been diagnosed with primary stage I (T1N0M0) or stage II (T2N0M0) OSCC, and had not received any prior treatment. Thirty-five healthy donors were recruited as controls from University of California, Los Angeles. All participants provided consent according to the institutional review boards (IRBs) of University of California, Los Angeles, and USC. Whole blood was collected according to IRB-approved procedure and centrifuged at 1,000x g for 10 minutes at 15°C. Superase-In RNase inhibitor (100 ± U/mL; Ambion Inc, Austin, TX) was promptly added to the segregated serum.
RNA Isolation and RT-PCR
Microarray Sera from 10 OSCC patients (eight male, two female; age, 51 years [standard deviation {SD}, 9.0 years]) and 10 sex and age matched healthy donors (age, 49 years [SD, 5.6 years]) were used for microarray. Serum mRNA was linearly amplified using RiboAmp RNA Amplification kit (Arcturus, Mountain View, CA). The Affymetrix Human Genome U133A Array (Santa Clara, CA),10 was applied for serum transcriptome profiling. The raw data were imported into DNA-Chip Analyzer 1.3 software (http://biosun1.harvard.edu/complab/dchip/) for normalization and model-based analysis.19 DNA-Chip Analyzer 1.3 gives the expression index, which represents the amount of mRNA/gene expression and another parameter, present call, which determines whether or not the mRNA transcript was present in the sample.20 S-plus 6.0 (version 3.3, Statistical Sciences, Seattle, WA) was used for all statistical tests. We used three criteria to determine differentially expressed genes between OSCCs and controls. First, we excluded genes that were assigned as absent call in all samples. Second, a two-tailed t test was used to compare the average gene expression levels between the two groups. The critical  level of .05 was defined for statistical significance. Third, fold changes were calculated for those genes that showed statistically significant differences (P < .05). Only those that exhibited at least two-fold change were included for additional analysis.
Quantitative PCR
Prediction Models Another model, classification and regression trees (CART), was also constructed by S-plus 6.0. CART fits the classification model by binary recursive partitioning, where each step involves searching for the predictor variable that results in the best split of the affected versus the healthy groups.24 CART used the entropy function with splitting criteria determined by default settings. With this approach, the parent group containing the entire samples (N = 67) was subsequently divided into affected and healthy. The initial tree was pruned to remove all splits that did not result in sub-branches with different classifications.
Sixty-seven participants including 32 OSCC patients and 35 control subjects, were recruited. There were no significant differences in terms of mean age: OSCC patients, 49.3 years (SD, 7.5 years); healthy patients, 47.8 years (SD, 6.4 years; t test P = .84). The sex distribution in OSCC group was 10:22 (female:male) and in control group was 14:21 (  2 test P  .99). We matched the smoking history by determining the pack per year history (t test P = .77). The presence and integrity of human serum mRNAs were evaluated by RT-PCR and electrophoresis (Fig 1). Transcripts from four housekeeping genes (ACTB, B2M, GAPDH, and RPS9) could be consistently detected. The longest PCR products amplified covered 56.8% (ACTB), 85.9% (B2M), 93.4% (GAPDH), and 88.9% (RPS9) of the full length of the respective mRNAs. This indicates intact circulating mRNAs existing in blood in a cell-free phase.
The HG U133A microarrays25 were used to profile and identify the differences in serum mRNA transcriptomes between cancer patients and healthy controls. Among the 14,268 genes included by the criteria described herein, we identified 335 genes with a P value  .05 and a fold change of at least 2. Among them, there were 223 upregulated and 112 downregulated genes in the OSCC group. The number of mRNAs that were assigned present call on each array are presented in Table 2. The OSCC group has significantly more present probes (2,623 ± 868) than controls (1,792 ± 165), indicating a higher complexity of serum transcriptome in OSCC (P = .002, Wilcoxon test). A more stringent criterion was applied to select candidate markers, which requires the present calls be consistently assigned among all cancers (n = 10) or all controls (n = 10). We identified 62 transcripts from the differently expressed 335 genes with such criterion. Of interest is that these 62 transcripts are all upregulated in OSCC serum.
qPCR was performed to validate the microarray findings on all of the enrolled sera from 32 OSCC patients and 35 controls. Ten significant upregulated candidates were selected from the list of 62 based on their reported cancer association: H3F3A, TPT1, FTH1, NCOA4, ARCR, THSMB, PRKCB1, FTL1, COX4I1, and SERP1. Table 3 presents the quantitative alterations of these 10 mRNA in serum from OSCC determined by qPCR. The corresponding microarray data for those 10 candidates are also presented in Table 3. Five transcripts (H3F3A, TPT1, FTH1, NCOA4, and ARCR) were confirmed to be significantly elevated in OSCC sera (t test P < .05). Although qPCR showed higher mean copy number of THSMB, PRKCB1, FTL1, COX4I1, and SERP1 in OSCC, controls were not statistically significant (t test P > .05).
Using the qPCR data, logistic regression models were built using six serum transcripts examined in previous step: ARHA, FTH1, H3F3A, TPT1, COX4I1, and FTL1. These six transcripts in combination provided the best prediction, which was then validated by the leave-one-out method. Of 67 leave-one-out cross-validation, the same model was consistently the best logistic model, with the same six markers as the one from the entire set of data (81%; 54 of 67). The error rate was 31.3% (21 of 67). A cutoff probability of 44%, a sensitivity of 84%, and a specificity of 83%, were obtained by ROC analysis. The final model predicts correctly for 56 of 67 (83.5%) patients, and it misclassifies six from the control group and five from the OSCC group. The area under the curve was 0.88 for the final regression model (Fig 2).
The fitted CART model used the serum mRNA concentrations of THSMB and FTH1 as predictor variables for OSCC (Fig 3). THSMB, chosen as the initial split, with a threshold of 4.59–17 M, produced two child groups from the parent group, containing a total of 67 samples. Forty-seven samples with the THSMB concentration less than 4.59E-17 M were assigned into the Healthy-1 group, while 20 samples with THSMB concentration at least 4.59–17 M were assigned into Cancer-1 group. The Healthy-1 group was partitioned further by FTH1 with a threshold of 8.44–16 M. The resulting subgroups, Healthy-2 contained 28 samples with FTH1 concentration less than 8.44–16 M, and Cancer-2 contained 19 samples with FTH1 concentration of at least 8.44–16 M. Consequently, the 67 serum samples involved were classified into the healthy group and the cancer group. The healthy group was composed of Healthy-2 including a total of 28 samples, 25 from healthy subjects and three from cancer patients. Thus, by using the combination of THSMB and FTH1 for OSCC prediction, the overall specificity is 71% (25 of 35). The cancer group was composed of 39 samples from the Cancer-1 group and 29 samples from cancer patients and 10 from healthy subjects from the Cancer-2 group. Therefore, the overall sensitivity is 91% (29 of 32).
Cancer caused approximately 600,000 deaths of Americans in 2004.26 A disappointing cancer prognosis is most probably attributed to diagnostic delay.27,28 This fact is particularly the case of OSCC, of which the overall 5-year-survival rate has remained low at approximately 30% to 50% in past decades.29,30 Recently, there has been interest in the potential use of nucleic acid makers in bodily fluid for the purpose of exploring better biomarker detection for cancers. Cancer-associated genetic changes often lead to altered gene expression patterns identified in bodily fluids, draining the affected organ long before the resulting cancer phenotypes are manifested. Tumor associated mRNA biomarkers have been identified in serum, plasma, urine, semen, CSF, and saliva,9,31 and applied for cancer diagnosis.12,16,32 Salivary transcriptome has been profiled by microarray technology and used for OSCC diagnosis with promise.31 A simple blood test for early cancer detection has been the dream of all researchers and physicians. Recovering the circulating mRNA biomarkers representing tumor genetic alterations in the serum from cancer patients should provide more insights into this possibility. This study contributes to the previous discoveries in the field of developing tumor-derived mRNA in the serum/plasma of cancer patients.18,33-35 The serum transcriptomes of OSCC patients and healthy subjects were assessed and compared. To our knowledge this is the first article where circulating mRNA in serum is globally profiled by microarray. Approximately 1,800 different mRNA species exist in the serum of healthy subjects. Enders et al36 reported the RNA species in plasma are predominantly in the particle-associated forms. The isolation method we used may have decreased the particle-associated RNAs, which may explain less mRNA in serum detected than that in saliva using the same array platform as previously reported.10 The different characteristics of serum-free RNA could be addressed by using more efficient methodology, like guanidinium-phenol precipitation, as suggested.37 It was known that the increased plasma DNA was correlated to tumor burden, such as metastasis,38 and the integrity of plasma DNA significantly increased in cancer as well.39 Additional investigation is needed into whether the increased serum mRNAs in OSCC possess the characteristics similar to those found in this study. It is reasonable to predict more human mRNAs will be identified in serum by more comprehensive coverage methodologies. It has been presumed that human mRNA is highly degraded in serum because RNase exists in blood.40 This has recently changed with several articles clearly demonstrating the presence of amplifiable RNA in plasma/serum.41 The extracellular RNA was reported stable in serum for up to 3 hours.37 Our study confirmed that the overall quality of serum RNA could meet the demand for RT-PCR, RT-qPCR, and microarray assays. However, discrepancies were found when validating serum microarray data by qPCR. Previous studies by others also suggested disagreement of RT-PCR and microarray data.42-47 Increased numbers of absent calls by the microarray software and increased separation between the location of the PCR primers and the microarray probes both led to reduced agreement.43 The different methodologies for RNA amplification (ie, exponentially by PCR or linearly by T7 transcription) yielded transcripts varying significantly in estimated length, GC content, and expression level. The correlation of expression intensities using the different amplification methods was considerably lower (R2 = 0.52).47 These results highlight the known need to carry out validation of microarray data by more accurate and complementary technologies.42 The multifactorial nature of oncogenesis and the heterogeneity in oncogenic pathways make it unlikely that a single biomarker will detect all cancers with high specificity and sensitivity. The most promising approach is to use panels of biomarkers to produce diagnostic and predictive information that is more powerful.48 This concept led us to construct prediction models using multiple statistical strategies to identify best combinations of serum biomarkers for OSCC. Though encouraging, we understand that the current sensitivity (84%) and specificity (83%) cannot meet the clinical screening tool demands. Efforts are underway to validate other candidate markers and to combine them to generate a higher power for oral cancer discrimination. The patients enrolled in this study were diagnosed with primary stage I (T1N0M0) or stage II (T2N0M0) OSCC. Though they were all classified as early clinical stage, tumors in stage II have worse prognosis.49 Also tumors originated from different sites in the mouth, like the tongue, mouth floor, and buccal mucosa, and have different cancer development characteristics.50 Limited by the small overall sample size (N = 67), we were not able to generate additional important data to see if the microarray data can segregate T1 from T2 oral cancer lesions. The next step will be to validate our results in an independent cohort with a larger sample size and more restricted inclusion criteria. Serum transcriptome diagnostics for cancer detection meet the demands for a noninvasive diagnostic tool. Additional exploration is needed before its full clinical utility can be realized. Little is known about the biologic origin of circulating RNA. Cell death or apoptosis, has been postulated as the mechanism responsible for the release of nucleic acids into plasma.51,52 There was evidence showing the cell-free nucleic acids in serum might be partially tumor derived.40 Also, this proof-of-concept study clearly needs reproduction and careful clinical validation. Of particular importance is the need to detect preneoplastic states and states of early tumor recurrence for which there is no tissue visible by imaging and for which no systemic markers are currently available.
The authors indicated no potential conflicts of interest.
Absent/present call: When using an Affymetrix (Santa Clara, CA) Microarray, the detection algorithm uses probe pair intensities to generate a detection P value and assigns a call. Each probe pair in a probe set is considered as having a potential vote in determining whether the measured transcript is detected (present) or not detected (absent). Amplicon: The DNA product of a polymerase chain reaction reaction, usually an amplified segment of a gene or DNA. Leave-one-out cross validation: Test sets of one sample are chosen and the accuracy of the model derived from the remaining (n –1) samples is scored. This is repeated for all the "n" samples so that every sample acts as a test set. The predictive error obtained can be used as a measure of internal validation of the predictive power of the classifier developed using the full data set. All aspects of the classifier development process should be repeated from scratch for each leave-one-out training set; including all aspects of gene selection or tuning parameter optimization. Logistic regression model: A multivariable prediction model in which the log of the odds of a time-fixed outcome event is related to a linear equation. ROC (receiver operating characteristic) curves: ROC curves plot the true positive rate (sensitivity) against the false-positive rate (1-specificity) for different cut-off levels of a test. The area under the curve is a measure of the accuracy of the test. An area of 1.0 represents a perfect test (all true positives), whereas an area of 0.5 represents a worthless test. T1N0M0: T1 refers to a tumor 2 cm or less in greatest dimension, N0 means no regional lymph node metastasis, and M0 means no distant metastasis. T2N0M0: T2 refers to a tumor more than 2 cm but not more than 4 cm in greatest dimension, N0 equals no regional lymph node metastasis, and M0 equals no distant metastasis. Transcriptome: The complete expressed product of the entire genome in a particular cell, tissue, or biofluid at a specific point in time.
Supported by US Public Health Service (PHS) Grant No. RO1 DE15970, a UCLA Jonsson Comprehensive Cancer Center Grant (D.T.W.), PHS Grant No. T32 DE07296-07, and a Cancer Research Foundation of American fellowship (X.Z.). Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Negm RS, Verma M, Srivastava S: The promise of biomarkers in cancer screening and detection. Trends Mol Med 8:288-293, 2002[CrossRef][Medline] 2. Sidransky D: Emerging molecular markers of cancer. Nat Rev Cancer 2:210-219, 2002[CrossRef][Medline] 3. Srinivas PR, Kramer BS, Srivastava S: Trends in biomarker research for cancer detection. Lancet Oncol 2:698-704, 2001[CrossRef][Medline] 4. Lipshutz RJ, Fodor SP, Gingeras TR, et al: High density synthetic oligonucleotide arrays. Nat Genet 21:20-24, 1999[CrossRef][Medline] 5. Diamandis EP: Mass spectrometry as a diagnostic and a cancer biomarker discovery tool: Opportunities and potential limitations. Mol Cell Proteomics 3:367-378, 2004 6. Sullivan Pepe M, Etzioni R, Feng Z, et al: Phases of biomarker development for early detection of cancer. J Natl Cancer Inst 93:1054-1061, 2001 7. Barker PE: Cancer biomarker validation: Standards and process: Roles for the National Institute of Standards and Technology (NIST). Ann N Y Acad Sci 983:142-150, 2003 8. Sidransky D: Nucleic acid-based methods for the detection of cancer. Science 278:1054-1059, 1997 9. Juusola J, Ballantyne J: Messenger RNA profiling: A prototype method to supplant conventional methods for body fluid identification. Forensic Sci Int 135:85-96, 2003[CrossRef][Medline] 10. Li Y, Zhou X, St John MA, et al: RNA profiling of cell-free saliva using microarray technology. J Dent Res 83:199-203, 2004 11. Anker P, Mulcahy H, Chen XQ, et al: Detection of circulating tumour DNA in the blood (plasma/serum) of cancer patients. Cancer Metastasis Rev 18:65-73, 1999[CrossRef][Medline] 12. Rieger-Christ KM, Mourtzinos A, Lee PJ, et al: Identification of fibroblast growth factor receptor 3 mutations in urine sediment DNA samples complements cytology in bladder tumor detection. Cancer 98:737-744, 2003[CrossRef][Medline] 13. Wong LJ, Lueth M, Li XN, et al: Detection of mitochondrial DNA mutations in the tumor and cerebrospinal fluid of medulloblastoma patients. Cancer Res 63:3866-3871, 2003 14. Kopreski MS, Benko FA, Gocke CD: Circulating RNA as a tumor marker: Detection of 5T4 mRNA in breast and lung cancer patient serum. Ann N Y Acad Sci 945:172-178, 2001 15. Bunn PJ Jr: Early detection of lung cancer using serum RNA or DNA markers: Ready for "prime time" or for validation? J Clin Oncol 21:3891-3893, 2003 16. Wong SC, Lo SF, Cheung MT, et al: Quantification of plasma beta-catenin mRNA in colorectal cancer and adenoma patients. Clin Cancer Res 10:1613-1617, 2004 17. Fugazzola L, Mihalich A, Persani L, et al: Highly sensitive serum thyroglobulin and circulating thyroglobulin mRNA evaluations in the management of patients with differentiated thyroid cancer in apparent remission. J Clin Endocrinol Metab 87:3201-3208, 2002 18. Kopreski MS, Benko FA, Kwak LW, et al: Detection of tumor messenger RNA in the serum of patients with malignant melanoma. Clin Cancer Res 5:1961-1965, 1999 19. Li C, Wong WH: Model-based analysis of oligonucleotide arrays: Expression index computation and outlier detection. Proc Natl Acad Sci U S A 98:31-36, 2001 20. Li C, Hung Wong W: Model-based analysis of oligonucleotide arrays: Model validation, design issues and standard error application. Genome Biol 2:1-11, 2001 21. Donia D, Divizia M, Pana A: Use of armored RNA as a standard to construct a calibration curve for real-time RT-PCR. J Virol Methods 126:157-163, 2005[CrossRef][Medline] 22. Ginzinger DG: Gene quantification using real-time quantitative PCR: An emerging technology hits the mainstream. Exp Hematol 30:503-512, 2002[CrossRef][Medline] 23. Renger R, Meadows LM: Use of stepwise regression in medical education research. Acad Med 69:738, 1994 24. Lemon SC, Roy J, Clark MA, et al: Classification and regression tree analysis in public health: Methodological review and comparison with logistic regression. Ann Behav Med 26:172-181, 2003[CrossRef][Medline] 25. Affymetrix: Affymetrix Technical Note: New Statistical Algorithms for Monitoring Gene Expression on GeneChip® Probe Arrays. 2001 26. American Cancer Society: Statistics for 2004: Cancer facts and figures 2004. Atlanta, GA, American Cancer Society, 2004 27. Wildt J, Bundgaard T, Bentzen SM: Delay in the diagnosis of oral squamous cell carcinoma. Clin Otolaryngol 20:21-25, 1995[Medline] 28. Fong KM, Srivastava S, Gopal-Srivastava R, et al: Molecular genetic basis for early cancer detection and cancer susceptibility, in Srvastava S, Hensen DE, Gazdar A (eds): Molecular Pathology of Early Cancer, IOS Press, 1999, pp 13-26 29. Epstein JB, Zhang L, Rosin M: Advances in the diagnosis of oral premalignant and malignant lesions. J Can Dent Assoc 68:617-621, 2002[Medline] 30. Mao L, Hong WK, Papadimitrakopoulou VA: Focus on head and neck cancer. Cancer Cell 5:311-316, 2004[CrossRef][Medline] 31. Li Y, St John MA, Zhou X, et al: Salivary transcriptome diagnostics for oral cancer detection. Clin Cancer Res 10:8442-8450, 2004 32. Anker P, Mulcahy H, Stroun M: Circulating nucleic acids in plasma and serum as a noninvasive investigation for cancer: Time for large-scale clinical studies? Int J Cancer 103:149-152, 2003[CrossRef][Medline] 33. Chen XQ, Bonnefoi H, Pelte MF, et al: Telomerase RNA as a detection marker in the serum of breast cancer patients. Clin Cancer Res 6:3823-3826, 2000 34. Silva JM, Dominguez G, Silva J, et al: Detection of epithelial messenger RNA in the plasma of breast cancer patients is associated with poor prognosis tumor characteristics. Clin Cancer Res 7:2821-2825, 2001 35. Dasi F, Lledo S, Garcia-Granero E, et al: Real-time quantification in plasma of human telomerase reverse transcriptase (hTERT) mRNA: A simple blood test to monitor disease in cancer patients. Lab Invest 81:767-769, 2001[Medline] 36. Ng EK, Tsui NB, Lam NY, et al: Presence of filterable and nonfilterable mRNA in the plasma of cancer patients and healthy individuals. Clin Chem 48:1212-1217, 2002 37. El-Hefnawy T, Raja S, Kelly L, et al: Characterization of amplifiable, circulating RNA in plasma and its potential as a tool for cancer diagnostics. Clin Chem 50:564-573, 2004 38. Jung K, Stephan C, Lewandowski M, et al: Increased cell-free DNA in plasma of patients with metastatic spread in prostate cancer. Cancer Lett 205:173-180, 2004[CrossRef][Medline] 39. Wang BG, Huang HY, Chen YC, et al: Increased plasma DNA integrity in cancer patients. Cancer Res 63:3966-3968, 2003 40. Stroun M, Anker P, Maurice P, et al: Neoplastic characteristics of the DNA found in the plasma of cancer patients. Oncology 46:318-322, 1989[Medline] 41. Johnson PJ, Lo YM: Plasma nucleic acids in the diagnosis and management of malignant disease. Clin Chem 48:1186-1193, 2002 42. Dallas PB, Gottardo NG, Firth MJ, et al: Gene expression levels assessed by oligonucleotide microarray analysis and quantitative real-time RT-PCR: How well do they correlate? BMC Genomics 6:59, 2005 43. Etienne W, Meyer MH, Peppers J, et al: Comparison of mRNA gene expression by RT-PCR and DNA microarray. Biotechniques 36:618-620, 622, 624-626, 2004 44. Jenson SD, Robetorye RS, Bohling SD, et al: Validation of cDNA microarray gene expression data obtained from linearly amplified RNA. Mol Pathol 56:307-312, 2003 45. Morgan KT, Ni H, Brown HR, et al: Application of cDNA microarray technology to in vitro toxicology and the selection of genes for a real-time RT-PCR-based screen for oxidative stress in Hep-G2 cells. Toxicol Pathol 30:435-451, 2002[Medline] 46. Puskas LG, Zvara A, Hackler L Jr, et al: RNA amplification results in reproducible microarray data with slight ratio bias. Biotechniques 32:1330-1334, 1336, 1338, 1340, 2002 47. Wadenback J, Clapham DH, Craig D, et al: Comparison of standard exponential and linear techniques to amplify small cDNA samples for microarrays. BMC Genomics 6:61, 2005 48. Bernard PS, Wittwer CT: Real-time PCR technology for cancer diagnostics. Clin Chem 48:1178-1185, 2002 49. O'Brien CJ, Lauer CS, Fredricks S, et al: Tumor thickness influences prognosis of T1 and T2 oral cavity cancer–but what thickness? Head Neck 25:937-945, 2003[CrossRef][Medline] 50. Lip and oral cavity, in American Joint Committee on Cancer: AJCC Cancer Staging Manual (ed 6), New York, NY, Springer, 2002, pp 23-32 51. Jahr S, Hentze H, Englisch S, et al: DNA fragments in the blood plasma of cancer patients: Quantitations and evidence for their origin from apoptotic and necrotic cells. Cancer Res 61:1659-1665, 2001 52. Biggiogera M, Bottone MG, Pellicciari C: Nuclear RNA is extruded from apoptotic cells. J Histochem Cytochem 46:999-1005, 1998 Submitted August 5, 2005; accepted December 13, 2005. This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2). Five cancer-related gene transcripts were consistently validated by quantitative PCR on serum from OSCC patients (n = 32) and controls (n = 35). The best combination of biomarkers yielded a receiver operating characteristic curve value of 88%, sensitivity (91%), and specificity (71%) in distinguishing OSCC. 
