|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 3 (February 1), 2004: pp. 499-506 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.07.022 Urinary VEGF and MMP Levels As Predictive Markers of 1-Year Progression-Free Survival in Cancer Patients Treated With Radiation Therapy: A Longitudinal Study of Protein Kinetics Throughout Tumor Progression and TherapyFrom the Radiation Oncology Branch, Radiation Oncology Sciences Program, and Biometric Research Branch, Division of Cancer Treatment and Diagnosis, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD; Department of Surgery, Harvard Medical School; and Vascular Biology Program and the Department of Surgery, Children's Hospital, Boston, MA. Address reprint requests to Kevin Camphausen, MD, Radiation Oncology Branch, Radiation Oncology Sciences Program, Center for Cancer Research, National Cancer Institute, National Institutes of Health, 10 Center Dr, Building 10, Room B3/B69, Bethesda, MD, 20892-1002; e-mail: Camphauk{at}mail.nih.gov
PURPOSE: To determine the predictive value of urinary levels of two angiogenic factors, vascular endothelial growth factor (VEGF) and matrix metalloproteinase (MMPs), in a longitudinal study to determine their correlation with 1-year progression-free survival in patients with cancer. PATIENTS AND METHODS: VEGF and MMP levels were measured in the urine of 65 cancer patients at first evaluation, during therapy, and at follow-up (n = 242); normalized by creatinine levels; and compared with 16 healthy controls. The correlation of initial levels and trends of VEGF and MMPs with 1-year progression-free survival was assessed using two-sample tests and stepwise logistic regression. RESULTS: Urinary VEGF levels at presentation were different between patients with local-regional cancer and normal controls, and between patients with metastatic prostate cancer and local-regional disease (P = .04 and .01, respectively). Similar results were found with MMP measurement (P = .03 and .0001, respectively). Of those patients subsequently treated with radiation, VEGF levels at presentation between patients with no evidence of disease (NED) after radiation and those who had persistent or recurrent disease after radiotherapy were also different (P = .039). The comparison between angiogenic factor levels taken at least 1 month postradiotherapy and the last level taken during treatment was the strongest predictor of patient 1-year progression-free survival (P = .004). Similarly, the overall MMP trend was also significantly associated with 1-year progression-free survival, as was the individual MMP-2 trend (P = .004 and .001, respectively). Stepwise logistic regression revealed that the VEGF trend comparing postradiation levels with last level taken during treatment was an independent predictor of progression-free survival (P = .02). CONCLUSION: This small exploratory study suggests that the angiogenic urinary trends of VEGF and MMPs may be useful predictive markers for progression-free survival in cancer patients after the completion of radiotherapy.
Tumor growth is dependent on angiogenesis (the development and recruitment of new blood vessels). The therapeutic opportunity to target the endothelial cells that support tumor growth rather than the cancer cells themselves was first recognized by Judah Folkman, MD [1]. Despite a vast array of established tumor markers, none have been shown to have general predictive value across all tumor subtypes. In theory, circulating angiogenic factors could identify patients at risk for persistent or recurrent disease, regardless of tumor type, because the process of angiogenesis is ubiquitous to cancer. Indeed, multiple investigators have explored the use of angiogenic factors as possible general tumor markers [2-15]. Although angiogenic proteins, such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor, and matrix metalloproteinases (MMPs), may have been shown to have prognostic value in specific tumor types, few studies have individually explored the utility of angiogenic proteins as general tumor markers across multiple tumor types, including solid tumors and hematologic malignancies [3]. A tumor marker that could consistently identify patients at risk for failure would allow selection of these patients for more aggressive or alternate treatment. Furthermore, to our knowledge, each of the aforementioned studies focused on the magnitude of either the VEGF or MMP level at presentation. No study has explored the dynamic trend of these protein levels through therapy and its possible predictive significance. VEGF, originally known as vascular permeability factor, is one of the most potent and well-characterized proangiogenic proteins. It is expressed by a variety of human solid tumors, as well as neoplastic myeloid and leukemic cells [16]. VEGF has specific mitogenic activity on endothelial cells, and promotes extravasation of proteins from tumor vessels, which creates a fibrin matrix infrastructure allowing for stromal cell invasion and tumor development [17]. MMPs are a family of zinc-dependent enzymes that enable new vessels to invade the surrounding extracellular matrix. MMPs function during the normal physiologic processes of tissue repair and morphogenesis. Pathologically, MMPs have been implicated in diseases associated with excess degradation of extracellular matrix, such as tumor invasion and metastasis. Increased MMP expression has been linked to more aggressive metastatic behavior, and increased expression of MMPs has been documented in numerous tumor types [18]. In this study, we sought to determine the prognostic value of urinary VEGF and MMP measurement at presentation across multiple cancer types. We also examined the trend of VEGF and MMP levels through radiotherapy in relation to clinical outcome. Although prognostic factors are measurements available at the time of diagnosis that are associated with disease-free or overall survival, predictive factors are defined as any measurement associated with response or lack thereof to a particular therapy [19]. Thus, we also sought to identify the predictive value of VEGF and MMP measurement in relation to radiotherapy.
This prospective clinical study (02C0064) was approved by the Institutional Review Board of the Clinical Center, National Cancer Institute, National Institutes of Health (Bethesda, MD). Informed consent was obtained from each patient. Urine samples were obtained from a cohort of 65 patients, comprised primarily of patients with prostate, breast, brain, and hematologic malignancies, who were evaluated for external-beam radiation therapy at the National Institutes of Health (Table 1). No patient underwent chemotherapy concurrent with the course of radiation. Samples were also obtained from 16 normal controls. Samples were taken at presentation, semiweekly during the course of treatment, and at the first follow-up visit 1 month after the completion of radiotherapy. Each urine sample was aliquoted, frozen immediately after collection, and stored frozen (-20°C) until assayed. Aliquots were thawed for duplicate or triplicate VEGF assays. VEGF levels were determined using a commercial chemiluminescent immunoassay (QuantiGlo system; R&D Systems, Minneapolis, MN). Standardization by urinary creatinine concentration was obtained by dividing the VEGF concentration (in picograms per milliliter) of a particular sample by its corresponding creatinine level (in milligrams per deciliter), such that normalized VEGF levels are reported in units of picograms per milligram creatinine.
MMP levels were determined using zymography. Samples were run on commercial 10% gelatin zymogram kits (Novex system, Carlsbad, CA). Gels were stained and photographed for permanent record [3]. Both proenzyme and activated proteinases appeared as zones of substrate clearing. Different MMPs were distinguished from one another on the basis of their molecular weights. High molecular weight MMP, MMP-9/neutrophil gelatinase-associated lipocalin (NGAL), MMP-9, MMP-2, and low molecular weight MMP (LMW-MMP) were determined using this methodology [20]. As positive controls, purified human MMPs and commercially available MMPs were subjected to gel electrophoresis as described above. Two investigators, who were each blinded to the clinical status of the individuals from whom the urine samples were obtained, then scored zymograms independently. Although concordance was quite high between investigators, discrepancies in scoring were evaluated by a third impartial investigator. A patient's MMP levels at presentation were determined by scoring the presence or absence of each type of MMP examined on the zymogram, on a scale of 0 to 5, with 0 meaning that the patient expressed no detectable MMPs, and 5 meaning that all five types of MMPs were demonstrated by gel zymogram. A patient's 1-year progression-free survival was defined as the presence or absence of detectable disease at last follow-up visit, approximately 1 year from initial presentation for radiotherapy (mean, 12.7 months; range, 10.5 to 14.9 months). All patients were evaluated at 1 month postradiotherapy with a complete history and physical examination, directed radiographic examination, and serum and urine collection. Patients were then evaluated using routine standard-of-care guidelines for the patient's disease site including complete history and physical examination, directed radiographic examination, and laboratory work-up. A patient is defined as having active disease if persistent or recurrent bulk disease is detected. A patient is defined as having no evidence of disease (NED) if no active disease is detected.
Trends of VEGF levels were evaluated using three methods: (1) patients were placed into a binary up-down categorization, according to whether their VEGF levels increased or decreased from previous levels; (2) the quantitative difference between sample measurements was calculated; and (3) the VEGF slope (or rate of change of VEGF level) was calculated. t tests and MMP trend was evaluated similarly using method 1. MMP trend, however, was also subdivided by specific MMP types. Thus, each MMP type (high molecular weight MMP, MMP-9/NGAL, MMP-9, MMP-2, and low molecular weight MMP) was independently evaluated. Trend methods 2 and 3 were not applied to MMP calculations because the nature of the data suited categoric data analysis, rather than analysis of continuous variables. These three types of trend methods were evaluated for trend during treatment, comparing the last sample drawn during therapy with the initial pretreatment sample; posttreatment trend, comparing the sample drawn 1 month after the completion of radiotherapy to the initial sample at presentation; and a second posttreatment trend, comparing the sample drawn after the completion of radiotherapy to the last sample drawn during radiotherapy.
Two hundred forty-two urine samples were analyzed for VEGF, creatinine, and MMP levels. Thirty-five patients who received radiation therapy were included, as well as 30 patients with metastatic prostate cancer who had experienced treatment failure after primary therapy, and 16 normal controls (Table 1).
VEGF Alone: VEGF Levels at Presentation
Those patients subsequently treated with radiotherapy (n = 35) were divided into two groups according to 1-year progression-free survival. Patients with detectable disease included patients with prostate, brain, hematologic, and renal cell cancers. Prostate cancer patients with possible biochemical failure (ie, an elevated prostate-specific antigen level) were not considered patients with detectable disease. The initial VEGF levels of patients with detectable disease after therapy (n = 9; mean, 471.6 pg/mg creatinine) were higher than in patients with NED after therapy (n = 25; mean, 314.0 pg/mg creatinine; P = .039; Fig 1B). One patient was lost to follow-up 2 weeks after therapy and is not included in this analysis. Although the average values of NED patients were lower than in patients with detectable disease after the completion of radiotherapy, the degree of overlap between these two groups is large. This overlap of measurements, at presentation, limits the value of a single urinary VEGF measurement to predict response to therapy (Fig 1B). Furthermore, examining the initial pretreatment VEGF level represents only one point in time during the patient's treatment course. Tracking the VEGF levels throughout treatment may give a more effective measure of disease status.
MMP Levels at Presentation
VEGF Binary Trend Evaluation (Method 1) Using the binary evaluation method described, VEGF trends during and after radiotherapy were significantly correlated to disease detectability after therapy. When the 1-month postradiotherapy VEGF level was compared with the last level taken during treatment, the trend was correlated with disease status after therapy, with an increasing VEGF trend indicating detectable disease (relative risk [RR], 6.23; 95% CI, 1.21 to 31.9; P = .01; Fig 3). VEGF trends increased in patients with detectable disease, or decreased or remained stable in patients with that had NED, in 75% of the patients. Sensitivity and specificity for 1-year progression-free survival were calculated to be 86% (95% CI, 59% to 100%) and 72% (95% CI, 54% to 90%), respectively, with six of seven active patients demonstrating an increasing VEGF trend, and 18 of 25 NED patients demonstrating a stable or decreasing VEGF trend.
Calculation of likelihood ratios (LR) indicated that this trend had significant discriminatory power as well (LR 3.1 ± 0.2; Table 2). Neither of the other two VEGF trends calculated with this binary method was predictive of disease detectability after therapy.
VEGF Quantitative and Rate Trend (Methods 2 and 3) None of the quantitative trends calculated using method 2 were significantly correlated to 1-year progression-free survival (Table 3). Of the method 3 trends that calculate the rate of change in VEGF level, only the comparison of the posttreatment measurement with the last measurement taken during treatment was correlated to 1-year progression-free survival after therapy (P = .032; Fig 3). This is in keeping with the binary trend evaluation (method 1).
MMPs Binary Trend (Method 1) Binary overall MMP trends during and after radiotherapy were significantly correlated to disease detectability after therapy. In keeping with our VEGF results, only the trend comparing the postradiation measurement to the treatment measurement taken during treatment was significantly correlated to 1-year progression-free survival (RR, 7.72; 95% CI of RR, 1.51 to 39.0; P = .004). MMP trends increased in patients with detectable disease or decreased or remained stable in patients that had NED in 81% of the patients. Sensitivity and specificity for disease detectability after therapy were estimated to be 86% and 79%, respectively, with six of seven patients with active disease demonstrating concurrent increasing MMP trends, and 19 of 24 patients with NED demonstrating stable or decreasing MMP trends (Table 2). MMP trends at other time points measuring disease status after therapy were not statistically significant. Focusing on individual MMP trends, only MMP-2 binary trend comparing the postradiotherapy MMP-2 level to the last MMP-2 level taken during treatment was found to significantly correlate with disease detectability after therapy (RR, 18.17; 95% CI of RR, 1.12 to 293.9; P = .001). High molecular weight MMP, MMP-9/NGAL, MMP-9, and LMW-MMP trends were not found to be significantly associated with 1-year progression-free survival.
Correlation Between VEGF and MMPs
In this study, we evaluated the urinary level of VEGF and MMPs in patients with cancer evaluated for external-beam radiotherapy. Previous studies have examined the significance of the level of an angiogenic factor measured at presentation. Although our study found significance of VEGF and MMP levels for 1-year progression-free survival at presentation, we hypothesized that the dynamic change in VEGF or MMP level through a course of therapy and follow-up might provide more predictive clinical information than the static measurement of the initial level. We analyzed patient urine samples, rather than serum samples, to avoid confounding from the potential sequestration of VEGF in blood and the presence of MMPs also found in normal blood [22]. A direct correlation between measured VEGF levels and tumor VEGF production is further complicated by the sequestration of VEGF by platelets, and by the binding of VEGF to plasma alpha-2 macroglobulin, which makes VEGF unavailable to some antibodies. Furthermore, clot formation during the process of collecting blood samples induces platelet activation and subsequent release of multiple cytokines, including VEGF [22]. Some authors argue that VEGF levels should be measured in plasma to be clinically useful. However, this is complicated by the equilibrium between free VEGF and that sequestered within platelets. Hematopoietic cells in patients with cancer may also produce VEGF, unrelated to that generated by tumors [23]. Thus, rigorous definitions for VEGF measurement are needed to provide the best predictive information [23]. VEGF and MMP levels at presentation were higher in patients with cancer compared with those in normal controls. Patients with metastatic disease also had higher VEGF and MMP levels compared with patients with local-regional disease who were being evaluated for external-beam radiation. One weakness of our study was that our metastatic patient cohort comprised only one tumor type—prostate cancer. These patients, aside from being all male, also tended to be older than our average cancer patient population (Table 1). Furthermore, our normal controls were younger than average. Additional studies using matched controls and a more diverse metastatic patient cohort would strengthen our findings. Previous studies [2-15] have demonstrated prognostic utility for initial angiogenic factor levels in serum or urine. In our study, we found that urinary VEGF or MMP levels at presentation were good markers for the presence of disease. Although there was a significant difference between normal controls, patients with local-regional cancer, and patients with metastatic disease at presentation, the degree of overlap in the range of values for each category renders the initial VEGF level, for prediction of the tumors response to therapy, less than ideal (Fig 1A and 1B). Monitoring a patient's VEGF and MMP trends during and after radiotherapy may give a better sense of the 1-year progression-free survival after the completion of treatment. We used multiple approaches to determine whether trends could be significantly associated with 1-year progression-free survival after radiotherapy. For both VEGF and MMPs, the progression-free survival after radiotherapy was associated with the dynamic trend of either the VEGF or MMP levels from the last measurement taken during treatment to the first follow-up measurement. The VEGF rate of change (or slope) as well as the binary classification for this two–time point comparison was associated with 1-year progression-free survival (Fig 3). The positive correlation between VEGF slope and disease detectability probably reflects tumor activity after the completion of radiation. Tumors that recurred or persisted after radiation therapy produced relatively greater amounts of proangiogenic factors than tumors that were successfully eradicated. Patients who were classified as having NED at the completion of this study predominantly had flat or decreasing VEGF slopes, possibly indicating a decrease in tumor growth and angiogenesis. Therefore, a patient's VEGF level trend may serve as a predictive marker for a tumor's response to radiotherapy. Collapsing the VEGF or MMP trends into binary categories was also significant when posttreatment levels were compared with the last levels taken during treatment. For example, when VEGF levels trended upward, they were highly sensitive for 1-year progression-free survival after radiotherapy (Table 2). The magnitude of these changes was not a significant predictor for 1-year progression-free survival. Rather, it was the increase or decrease in these VEGF levels after the completion of treatment that was significant for response to therapy. The rationale for examining VEGF and MMP trends after completion of therapy is that angiogenic protein levels during radiotherapy are confounded by the therapeutic intervention. Tumors upregulate their expression of angiogenic factors such as VEGF in response to the cytotoxic insult of radiation [24]. During the course of treatment, the VEGF level represents the total of basal tumor production and radiation-induced upregulation. Therefore, trends that encompass time points within the patient's radiation course might be confounded by this radiation effect. By using the slope methodology or binary calculation, the contribution of the absolute magnitude of the last VEGF or MMP level taken during treatment is reduced because this magnitude may be confounded by a radiation-induced upregulation. Increasing levels (or a positive trend) indicated that there was continued growth and neovascularization after the completion of radiotherapy. The measurements in this group of patients, as hypothesized, were correlated with 1-year progression-free survival, whereas those patients in whom the trend was negative tended to have controlled disease (Fig 3). To evaluate further the predictive value of VEGF and MMP trend measurements, we examined their sensitivity, specificity, and positive and negative LRs for 1-year progression-free survival (Table 2). Sensitivity and specificity values are not dependent on prevalence, so they serve as more objective indicators of the quality of VEGF or MMP trend as tests for 1-year progression-free survival [3]. The sensitivity and specificity observed for VEGF and MMP trends as clinical markers of 1-year progression-free survival were indeed similar to, if not better than, those observed for serum tumor markers currently in use to monitor disease activity [25-30]. One other commonly used measure to assess predictive tests is LR, which is the ratio of the probabilities of diagnosis positive and negative, such that the greater the LR, the more likely the diagnosis will be positive. LR is also independent of prevalence, but can be used to convert pretest probability to posttest probability. Currently, some investigators believe LR is the best statistical method to evaluate a predictive test [31]. The strong likelihood ratios for both VEGF and MMP trends postradiotherapy suggest that these two trends might serve as robust predictive tests if put to general clinical use. Interestingly, within the MMP group examined, the MMP-2 enzyme demonstrated significant correlation with patients' 1-year progression-free survival (RR, 18.17; 95% CI, 1.12 to 293.9; P = .001). Sensitivity and specificity were 100% and 71%, respectively (Table 2). Positive LR was also strong at 3.43. This finding was similar to previously published results, which found that MMP-2 correlated the most strongly with 1-year progression-free survival in comparison with some other disease types [3]. Again, we found that postradiation MMP-2 trend correlated the most strongly with 1-year progression-free survival. Finally, this study was unique for measuring both VEGF and MMP levels in patients. Previous studies that have examined multiple angiogenic factors have not studied these two proteins in the same urine samples. We examined VEGF and MMPs to focus on key molecules required for tumor angiogenesis. By studying these two proteins together, we hoped to determine whether one or both would serve as clinically useful predictors of 1-year progression-free survival after radiotherapy. Therefore, our results are consistent with previous studies indicating that VEGF and MMP levels may be elevated in cancer patients [2-15]. However, we found that VEGF and MMP trends provide the clinician significant predictive information on the 1-year progression-free survival of cancer patients and their tumors' response or lack of response to radiation therapy treatment. These data support the utility of tracking VEGF and MMP trends as a tool for tracking the clinical course of cancer patients. Given the high LRs of these trends, the data also suggest that analysis of VEGF and MMP trends provides insight into tumor behavior before recurrence is detected by other diagnostic methods. We recognize that the sample size of this study is small and the number of statistical comparisons is large. Thus, larger, longer prospective studies are needed to confirm these preliminary conclusions and to obtain more precise estimates of diagnostic accuracy (ie, sensitivity and specificity) for these markers. Additional studies are also needed to explore urinary VEGF and MMP trends and their relationship to prognosis after other therapeutic modalities, such as surgery or chemotherapy. In conclusion, the evaluation of urinary levels and trends of angiogenic factors such as VEGF and MMPs may play an important role in monitoring disease progression after radiotherapy. Urinary VEGF and MMP trends could emerge as important tools in establishing disease recurrence and selecting patients that would benefit most from alternate or additional therapy.
The authors indicated no potential conflicts of interest.
M.A.M. was supported in part by Global Medical Products Inc. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Folkman J: Tumor angiogenesis: Therapeutic implications. N Engl J Med 285:1182–1186, 1971
2. Lindeholm B, Tavelin B, Grankvist K, et al: Vascular endothelial growth factor is of high prognostic value in node-negative breast carcinoma. J Clin Oncol 16:3121–3128, 1998
3. Moses MA, Wiederschain D, Loughlin KR, et al: Increased incidence of matrix metalloproteinases in urine of cancer patients. Cancer Res 58:1395–1399, 1998 4. Kraft A, Weindel K, Ochs A, et al: Vascular endothelial growth factor in the sera and effusions of patients with malignant and nonmalignant disease. Cancer 85:178–187, 1999[CrossRef][Medline]
5. Smith BD, Smith GL, Carter D, et al: Prognostic significance of vascular endothelial growth factor protein levels in oral and oropharyngeal squamous cell carcinoma. J Clin Oncol 18:2046–2052, 2000
6. Verheul HM, Hoekman K, Jorna AS, et al: Targeting vascular endothelial growth factor blockade: Ascites and pleural effusion formation. Oncologist 5:45–50, 2000 7. Gerhards S, Jung K, Koenig F, et al: Excretion of matrix metalloproteinases 2 and 9 in urine is associated with a high stage and grade of bladder carcinoma. Urology 57:675–679, 2001[CrossRef][Medline] 8. Lengyel E, Schmalfeldt B, Konik E, et al: Expression of latent matrix metalloproteinase 9 predicts survival in advanced ovarian cancer. Gynecol Oncol 82:291–298, 2001[CrossRef][Medline]
9. Poon RT, Fan ST, Wong J: Clinical implications of circulating angiogenic factors in cancer patients. J Clin Oncol 19:1207–1225, 2001 10. Cai W, Song JD: Expression of matrix metalloproteinase and tissue inhibitors in ovarian tumors. Ai Zheng 21:91–94, 2002[Medline] 11. Kaban K, Herbst RS: Angiogenesis as a target for cancer therapy. Hematol Oncol Clin North Am 16:1125–1171, 2002[CrossRef][Medline] 12. Kausch I, Bohle A: Molecular aspects of bladder cancer: Prognostic markers of bladder cancer. Eur Urol 41:15–29, 2002[CrossRef][Medline] 13. Kuittinen O, Soini Y, Turpeenniemi-Hujanen T: Diverse role of MMP-2 and MMP-9 in the clinicopathological behavior of Hodgkin's lymphoma. Eur J Haematol 69:205–212, 2002[CrossRef][Medline] 14. Ondruschka C, Buhtz P, Motsch C, et al: Prognostic value of MMP-2, -9 and TIMP-1, -2 immunoreactive protein at the invasive front in advanced head and neck squamous cell carcinomas. Pathol Res Pract 198:509–515, 2002[CrossRef][Medline] 15. Sienel W, Hellers J, Morresi-Hauf A, et al: Prognostic impact of matrix metalloproteinase-9 in operable non-small cell lung cancer. Int J Cancer 103:647–651, 2003[CrossRef][Medline]
16. Mayerhofer M, Valent P, Sperr WR, et al: BCR/ABL induces expression of vascular endothelial growth factor and its transcriptional activator, hypoxia inducible factor-1a, through a pathway involving phosphoinositide 3-kinase and the mammalian target of rapamycin. Blood 100:3767–3775, 2002 17. Dvorak HF, Nagy JA, Berse B, et al: Vascular permeability factor, fibrin, and the pathogenesis of the tumor stroma formation. Ann N Y Acad Sci 667:110–111, 1992 18. Burke PA, DeNardo SA: Antiangiogenic agents and their promising potential in combined therapy. Crit Rev Oncol Hematol 39:155–171, 2001[Medline] 19. Clark GM: Prognostic and Predictive Factors, in Harris JA, Lippman ME, Morrow M, Hellman S (eds): Diseases of the Breast. Philadelphia, PA, Lippincott-Raven Publishers, 1996, pp 461–479 20. Yan L, Borregaard N, Kjeldsen L, et al: The high molecular weight urinary matrix metalloproteinase (MMP) activity is a complex of gelatinase B/MMP-9 and neutrophil gelatinase-associated lipocalin (NGAL): Modulation of MMP-9 activity by NGAL. J Biol Chem 246:37258–37265, 2001 21. Agresti, A: Categorical Data Analysis. New York, NY, Wiley Press, 1990 22. Nakayama Y, Sako T, Shibao K, et al: Prognostic value of plasma vascular endothelial growth factor in patients with colorectal cancer. Anticancer Res 22:2437–2442, 2002[Medline] 23. Hormbrey E, Gillespie P, Turner K, et al: A critical review of vascular endothelial growth factor analysis in peripheral blood: Is the current literature meaningful? Clin Exp Metastasis 19:651–663, 2002[CrossRef][Medline]
24. Gorski DH, Beckett MA, Jaskowiak NT, et al: Blockade of the vascular endothelial growth factor stress response increases the antitumor effects of ionizing radiation. Cancer Res 59:3374–3378, 1999 25. Prestigiacomo AF, Lilja H, Pettersson K, et al: A comparison of the free fraction of serum prostate specific antigen in men with benign and cancerous prostates: The best case scenario. J Urol 156:350–354, 1996[CrossRef][Medline]
26. Coley CM, Barry MJ, Fleming C, et al: Early detection of prostate cancer. Part I: Prior probability and effectiveness of tests. Ann Intern Med 126:394–406, 1997
27. Einhorn N, Sjovall K, Knapp RC, et al: Prospective evaluation of serum CA 125 levels for early detection of ovarian cancer. Obstet Gynecol 80:14–18, 1992
28. Carlson KJ, Skates SJ, Singer DE: Screening for ovarian cancer. Ann Intern Med 121:124–132, 1994 29. van der Schouw YT, Verbeek AL, Wobbes T, et al: Comparison of four serum tumour markers in the diagnosis of colorectal carcinoma. Br J Cancer 66:148–154, 1992[Medline] 30. Moertel CG, Fleming TR, Macdonald JS, et al: An evaluation of the carcinoembryonic antigen (CEA) test for monitoring patients with resected colon cancer. JAMA 270:943–947, 1993[Abstract]
31. Greenhalgh T: How to read a paper: Papers that report diagnostic or screening tests. BMJ 315:540–543, 1997 Submitted July 7, 2003; accepted November 21, 2003. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 tests were used to test for an association between marker values at presentation and 1-year progression-free survival. Fisher's exact test was used to test for an association between changes in VEGF using VEGF method 1 and 1-year progression-free survival. t tests were used to test for an association between changes in VEGF using methods 2 and 3, and 1-year progression-free survival. Sensitivity and specificity were estimated using standard formulas. Relative risk and 95% CIs were estimated with small sample corrections [
