Elevated Serum Creatinine and Low Albumin are Associated With Poor Outcomes in Patients With Liposarcoma

Transcription

1 Elevated Serum Creatinine and Low Albumin are Associated With Poor Outcomes in Patients With Liposarcoma Joannis Panotopoulos, 1 Florian Posch, 2 Philipp T. Funovics, 1 Madeleine Willegger, 1 Anke Scharrer, 3 Wolfgang Lamm, 4 Thomas Brodowicz, 4 Reinhard Windhager, 1 Cihan Ay 2 1 Department of Orthopaedics, Medical University of Vienna, Vienna, Austria, 2 Department of Medicine I, Clinical Division of Haematology and Haemostaseology, Medical University of Vienna, Vienna, Austria, 3 Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria, 4 Department of Medicine I, Clinical Division of Oncology, Medical University of Vienna, Vienna, Austria Received 23 March 2015; accepted 20 July 2015 Published online 8 September 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI /jor ABSTRACT: Low serum albumin levels and impaired kidney function have been associated with decreased survival in patients with a variety of cancer types. In a retrospective cohort study, we analyzed 84 patients with liposarcoma treated at from May 1994 to October Uni- and multivariable Cox proportional hazard models and competing risk analyses were performed to evaluate the association between putative biomarkers with disease-specific and overall survival. The median age of the study population was 51.7 (range ) years. In multivariable analysis adjusted for AJCC tumor stage, serum creatinine was highly associated with diseasespecific survival (Subdistribution Hazard ratio (SHR) per 1 mg/dl increase ¼ 2.94; 95%CI ; p ¼ 0.005). High albumin was associated with improved overall and disease-specific survival (Hazard Ratio (HR) per 10 units increase ¼ 0.50; 95%CI ; p ¼ and SHR ¼ 0.64; 95%CI ; p ¼ 0.049). The serum albumin-creatinine-ratio emerged to be associated with both overall and disease-specific survival after adjusting for AJCC tumor stage (HR ¼ 0.95; 95%CI ; p ¼ and SHR ¼ 0.96; 95%CI ; p ¼ 0.08). Our study provides evidence for a tumor-stage-independent association between higher creatinine and lower albumin with worse disease-specific survival. Low albumin and a high albumin-creatinine-ratio independently predict poor overall survival. Our work identified novel prognostic biomarkers for prognosis of patients with liposarcoma. ß 2015 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 34: , Keywords: prognostic biomarkers; liposarcoma; creatinine; albumin Conflicts of interest: None. Correspondence to: Joannis Panotopoulos (T: þ ; F: þ ; joannis.panotopoulos@meduniwien.ac.at) # 2015 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. Liposarcomas represent a spectrum of malignant tumors with adipocytic differentiation and are the most common soft tissue sarcoma (STS). 1,2 The 5-year survival rates are between 70% and 90%. 1 The clinical presentation is versatile and the disease course is difficult to predict. 1 The main histological subtypes are: Atypical lipomatous tumor (ALT)/well differentiated liposarcoma, myxoid/round cell liposarcoma, dedifferentiated liposarcoma and pleomorph liposarcoma. 2,3 The pathological subclassification and histologic grade are key prognostic factors for survival. 1,4 While ALT only carry a risk of local recurrence, the pleomorph and dedifferentiated liposarcomas are high grade malignancies with a substantial risk of metastatic disease. 1,2 For individual risk estimation several prognostic factors (size, depth, site, grade, age at dignosis, and resection margins) have been associated with overall survival (OS) in STS. 1,2,5 In recent studies the concept of the involvement of sytemic inflammation, acute phase proteins in cancer progression has been postulated. Specifically, elevated preopeartive CRP and neutrophil/lymphocyte (N/L) ratio as markers of systemic inflammatory response have been found to be associated with decreased overall survival in STS Growing evidence emerged that low serum albumin levels and impaired kidney function as reflected by elevated serum creatinine are associated with decreased survival in cancer patients, though the exact mechanism remains still unknown The impact of serum creatinine and albumin on risk of mortality and survival in patients with STS has not yet been investigated. We set up a retrospective cohort study to analyze the association of serum creatinine and albumin with overall and disease-specific survival in patients with liposarcoma. Serum albumin is generally applied to assess the nutritional status, severity of disease and prognosis of cancer, but can be interfered by many factors. Due to the volatility of albumin its application in survival prediction is limited. 15 Therefore, we hypothesized that taking serum albumin and creatinine together to create a albumin-creatinineratio (ACR), the limitations of albumin could be avoided. PATIENTS AND METHODS The detailed inclusion and exclusion criteria for our study have been described previously. 10 In brief, we followed 148 patients with histologically confirmed liposarcoma who underwent surgery at the Department of Orthopaedic Orthopaedics, Medical Unviversity of Vienna, Austria from 1994 to Fifty four patients had incomplete clinical and laboratory data and were excluded from the present study. Twenty six patients have been operated due to their sarcoma at another institution prior referral to our department because of incomplete excision or non-radical resection margins (defined as prevalent cases). A cut-off of 12 weeks was set and all patients with a longer duration between index surgery and referral to our hospital (n ¼ 10) were excluded. Complete clinical pathological data were available for 84 patients (i.e., the final study population) and were collected from medical reports retrospectively (sex, age, tumor site, histology, tumor size, tumor depth, resection margins, tumor stage, laboratory 533

2 534 PANOTOPOULOS ET AL. parameters, adjuvant radiotherapy, adjuvant chemotherapy, local recurrence). Laboratory data from routine investigations were obtained 1 14 d before first surgical treatment. Albumin was measured using a bromocresol green colorimetric assay, 23 and creatinine with the Jaffe method. 24 Histopathological diagnoses were done according to the current WHO classification for soft tissue and bone tumors, 3 and reconfirmed by an experienced pathologist specialized in STS at our hospital. Resection margins were classified according to Enneking et al. 25 Tumor-stage was generated according to AJCC criteria. 26 Grading (i.e., G1 G3 according to the French Federation of Cancer Centers Sarcoma Group (FNCLCC) grading system). 27 The study was approved by the ethics commitee of the Medical University of Vienna, Austria (EC number 1833/2013). LEVEL OF EVIDENCE III (RETROSPECTIVE COHORT STUDY) Statistical Analysis All statistical analyses were performed using Stata (Windows version 13.0, Stata Corp., College Station, TX). Continuous data were summarized using medians and ranges, and categorical data by absolute frequencies and percentages. Correlations were evaluated with Spearman s correlation coefficient. Average levels of creatinine, albumin and the ACR were compared between patients with early (AJCC stages I þ II) versus advanced (AJCC stages III þ IV) disease using Wilcoxon s rank-sum test. Follow-up time was defined as the time from index surgery to death or last known alive. The median followup time was estimated using the method of Schemper and Smith. 28 Death-due-to-liposarcoma (disease-specific survival) and death-from-any-cause (overall survival) were considered as co-primary endpoints, with deathdue-to-liposarcoma being adjudicated by a panel of two authors (JP and MW). Cumulative risks of deathfrom-any-cause were calculated with the 1-Kaplan-Meier estimator, whereas cumulative risks of death-fromliposarcoma were calculated using a competing risk approach. Here, we used cumulative incidence estimators according to Marubini & Valsecchi (Stata program) 29 and considered death from causes other than liposarcoma as the competing event of interest. To compare the survival experience between two or more subgroups, we used logrank-tests (overall survival endpoint) and Gray s tests (disease-specific survival endpoint). For these subgroup comparisons, we dichotomized patients at the following cut-offs: (1) Elevated creatinine was defined according to our in-house laboratory cut-off as > 1.2 mg/dl for males and > 1.1 mg/dl for females; (2) Decreased albumin was defined according to our inhouse laboratory cut-off as < 35g/L; (3) Decreased ACR was defined as an ACR below the 25th percentile (i.e., Q3) of its distribution (cut-off: < units). This ACR cut-off was selected arbitrarily, because validated cutoffs for the ACR presently do not exist. Uni- and multivariable modeling of time-to-death was performed with Cox Proportional Hazards models (overall survival endpoint) and Fine & Gray Proportional Subdistribution Hazards models (disease-specific survival endpoint). The proportional (sub-) hazards assumption was assessed by fitting interactions between the predictor variables and the natural logarithm of follow-up time, and did not reveal time-dependent effects (data not shown). In multivariable analysis, we adjusted for a binary AJCC tumor stage variable (stages I þ II vs. stages III þ IV), because this variable allowed us to include information on both tumor grade and tumor TNM stage into a single parsimonious variable. RESULTS Baseline characteristics of the study cohort are shown in Table 1. The median age of the cohort was 51.7 years (range ), and male sex was slightly predominant (n ¼ 50 (59.5%)). The vast majority of tumors were located subfascially on the extremities. Twenty-three patients (27.4%) suffered from AJCC stage III or IV disease at baseline. The median levels of creatinine, albumin, and the ACR were comparable between patients with early (i.e., AJCC stages I þ II) and advanced (i.e., AJCC stages III þ IV) disease (all ranksum p > 0.66, Table 2). We did not observe a correlation between albumin and creatinine (r ¼ 0.00, p ¼ 0.99). Albumin was positively correlated with hemoglobin (r ¼ 0.36, p ¼ 0.001), and negatively correlated with C-reactive protein (CRP, r ¼ 0.31; p ¼ 0.005). The albumin-creatinine- Ratio (ACR) strongly correlated with both albumin and creatinine (r ¼ 0.6 and 0.7; both p < 0.001). There was weak evidence for inverse correlations between the ACR and fibrinogen (r ¼ 0.20; p ¼ 0.06), and also between the ACR and CRP (r ¼ 0.28; p ¼ 0.009). During a median follow-up period of 5.6 years (95% CI: ) 16 patients (19.1%) died. Nine of these deaths (56.3%) were adjudicated to be disease-related (i.e., deaths due to liposarcoma). The median age was 51.7 (range ) years. The 1-, 5- and 10-year cumulative risks (95%CI) of death-from-any-cause (i.e., 1-Overall Survival) were 1.2% ( ), 12.2% ( ), and 25.9% ( ), respectively. The 1-, 5-, and 10-year cumulative risks of death-due-toliposarcoma (i.e., disease-specific probability of death) were 1.2% (95%CI: ), 7.2% ( ), and 15.7% ( ), respectively. In univariable analysis, elevated levels of serum creatinine were associated with a worse overall survival experience (Hazard ratio (HR) per 1 mg/dl increase: 3.27; 95%CI ; p ¼ 0.04), and were highly associated with worse disease-specific survival (Subdistribution Hazard Ratio (SHR) per 1 mg/dl increase ¼ 4.72: 95%CI ; p < 0.001, Table 3). Whereas the 5-, 10-, and 15-year risks of death-dueto-liposarcoma were 11.1%, 46.7%, and 46.7% in patients with creatinine above the cut-off, the corresponding risks in patients with creatinine levels below or equal to the cut-off were only 6.7%, 11.4%, and 16.3%, respectively (Gray s test p ¼ 0.02, Fig. 1). We also observed a univariable association between higher serum albumin levels and a more favorable

3 PROGNOSTIC BIOMARKERS OF LIPOSARCOMA 535 Table 1. Baseline Characteristics of the Study Population Number (n) % Sex Male Female Tumour site Extremities Trunk Other Depth Epifascial Subfascial Histology Highly differentiated Myxoid Dedifferentiated Pleomorph Mixed Not other specified (NOS) Grading G G G Resection margins Wide Focal marginal Marginal Focal intralesional Intralesional Not known Tumour stage (AJCC) Stage IA Stage IB Stage IIA Stage IIB Stage III Stage IV Adjuvant Therapy Radiation Chemotherapy Both Tumour size (cm) > Age at baseline Median Range Laboratory parameters Median Range C-reactive protein (mg/dl) Hemoglobin (g/dl) Alcalic phosphatase (U/L) White Blood Count (G/L) NLR Platelet Count (G/L) Categorical variables are reported as absolute frequencies (n) and percentages (%). Continuous variables are reported as medians with ranges. NLR, Neutrophil-to-Lymphocyte-ratio. Table 2. Distribution of Albumin, Creatinine and the ACR According to AJCC Tumor Stage Laboratory Parameters AJCC I þ II AJCC III þ IV Albumin (g/l) 42.3 ( ) 42.6 ( ) Creatinine (mg/dl) 0.98 ( ) 0.97 ( ) Albumin-Creatinine ratio (ACR) (units) 43.2 ( ) 42.9 ( ) Variables are reported as medians with ranges in brackets. overall survival experience (HR per 10 g/l increase ¼ 0.50; 95%CI ; p ¼ 0.03). Whereas the 5-, 10-, and 15-year risks of death-from-any-cause were 9.3%, 22.4%, and 36.7% in patients with albumin 35 g/l, the corresponding risks in patients with albumin levels below this cut-off were 26.9%, 41.5%, and 100.0%, respectively (log-rank test p ¼ 0.02, Fig. 2). However, albumin did not appear to be significantly associated with disease-specific survival (SHR per 10 g/l increase ¼ 0.69; 95%CI ; p ¼ 0.19, Table 3). The albumin-creatinine-ratio (ACR) emerged as a favorable univariable predictor for both overall survival (HR per 10 units increase ¼ 0.59, 95%CI ; p ¼ 0.01) and disease-specific survival (SHR ¼ 0.95; 95%CI ; p ¼ 0.024). Whereas the 5-, 10-, and 15-year risks of death-fromany-cause were 19.6%, 41.4%, and 100.0% in patients with an ACR below the 25th percentile of its distribution (cut-off: units), the corresponding risks in patients with albumin levels above or equal this cut-off were 10.1%, 21.9%, and 27.5%, respectively (log-rank test p ¼ 0.01, Fig. 3). In multivariable analysis adjusted for AJCC tumor stage, there was a borderline statistically significant association between creatinine and overall survival (HR per 1 mg/dl increase ¼ 2.86; 95%CI ; p ¼ 0.06). However, the highly siginificant univariable association between elevated creatinine and worse disease-specific survival prevailed after multivariable adjustment for tumor stage (SHR per 1 mg/dl increase ¼ 2.94; 95%CI ; p ¼ 0.005, Table 4). Albumin was associated with both overall and disease-specific survival after adjusting for tumor-stage, indicating a tumor-stageindependent prognostic value of albumin for both outcomes (Table 4). The tumor-stage adjusted ACR emerged to be associated with both overall and diseasespecific survival (HR per 10 units increase ¼ 0.95; 95% CI ; p ¼ and SHR ¼ 0.96; 95%CI ; p ¼ 0.008, respectively). DISCUSSION In the present study we demonstrated that serum albumin and creatinine are significantly associated with risk of mortality and overall survival in patients with liposarcoma. In a competing risk analysis creatinine was also predictive of disease specific survival. Moreover, we identified a novel prognostic factor, the serum albumin-creatinine-ratio (ACR), for survival

4 536 PANOTOPOULOS ET AL. Table 3. Univariable Associations With Overall and Disease-Specific Survival Cause-Specific Hazards Analysis of Overall Survival Competing Risk Analysis of Disease-Specific Survival Variable HR 95%CI p-value SHR 95%CI p-value Creatinine (per 1 mg/dl increase) < Albumin (per 10 g/l increase) Albumin-creatinine-ratio (ACR, per 1 unit increase) AJCC tumor stage (I þ II vs. III þ IV) Age (per 10years increase) Female sex Overall survival was analyzed using Cox proportional hazards regression. Disease-specific survival was analyzed using Fine & Gray proportional subdistribution hazards regression. HR, hazard ratio; SHR, subdistribution hazard ratio; 95%CI, 95% confidence interval. Figure 1. Cumulative Risk of Death-due-to-liposarcoma according to baseline serum creatinine. Risks were calculated using competing risk estimators that incorporated death-from-anycause as the competing event of interest. Cut-off for creatinine was set at 1.1 mg/dl for females and 1.2 mg/dl for males. in patients with a liposarcoma, the most common entity of soft tissue sarcoma. The ACR emerged to be associated with both overall and disease-specific survival after adjusting for AJCC tumor-stage, which captures tumor specific information on grade as well as tumor size. Multivariable adjustment did not materially alter (sub-) hazard ratios and confidence intervals (and thus p-values). Synoptically, it appears that the ACR captures prognostic information from both creatinine and albumin which ultimately renders it as a tumor-stage independent predictor of overall- and disease-specific survival in patients with liposarcoma. The median levels of albumin, creatinine and ACR are highly comparable between patients with early and advanced tumor stages (I þ II vs. III þ IV) (Table 2), which implicates that the postulated biomarkers are not affected by advanced tumor stage. Previous reports investigated the influence of renal function and albumin on outcome in cancer patients. Joergensen et al. 30 associated increasing levels of ACR with increased incidence of cancer, especially bladder and renal cancer. Impaired estimated glomerular filtra- Figure 2. Cumulative Risk of Death-from-any-cause according to baseline serum albumin. Risks were calculated using the 1- Kaplan Meier product limit estimator. Cut-off for albumin was set at 35 g/l. Figure 3. Cumulative Risk of Death-from-any-cause according to baseline albumin-creatinine-ratio (ACR). Risks were calculated using the 1-Kaplan Meier product limit estimator. Cut-off for the ACR was set at the 25th percentile of the ACR distribution (cutoff: units).

5 PROGNOSTIC BIOMARKERS OF LIPOSARCOMA 537 Table 4. Multivariable Associations With Overall and Disease-Specific Survival Cause-Specific Hazards Analysis of Overall Survival Competing Risk Analysis of Disease-Specific Survival Variable HR 95%CI p-value SHR 95%CI p-value Creatinine (per 1 mg/dl increase) AJCC tumor stage (I þ II vs. III þ IV) Albumin (per 10 g/l increase) AJCC tumor stage (I þ II vs. III þ IV) Albumin-creatinine-ratio (ACR, per 1 unit increase) AJCC tumor stage (I þ II vs. III þ IV) Overall survival was analyzed using Cox proportional hazards regression. Disease-specific survival was analyzed using Fine & Gray proportional subdistribution hazards regression. HR, hazard ratio; SHR, subdistribution hazard ratio; 95%CI, 95% confidence interval. tion rate was found to be a significant predictor for bladder cancer recurrence and progression 12 and elevated preoperative serum creatinine was associated with worse cancer-specific survival. 22 Hypoalbuminaemia was foundtobeanindependentpredictorofpooroutcomein metastatic Ewing s sarcoma, in patients with malignant pleural mesothelioma and non-small lung cancer However, underlying mechanisms for this association in cancer patients have not been elucidated. Lin et al. 11 regarded in their study albuminuria as a paraneoplastic and inflammatory process. Cytokines modulate albumin synthesis, albumin catabolism and transcapillary leak. 13 Chronic inflammation triggers endothelial dysfunction leading to albumin loss. In consequence, low serum albumin may mirror a poor performance status of cancer patients who are at increased risk of death. Albumin helps to maintain intravascular oncotic pressure and acts as a radical scavenger. 16 It is not known whether the association is a general oncogenic effect or attributed to a specific cancer. The current evidence, in the authors opinion, shows an association and not a causation for increased risk of death. 11 It remains further to be elucidated whether serum creatinine is a surrogate for inflammation due to tumor cytokines or for comorbidities. Chronic low-grade inflammation may lead to intrarenal vascular dysfunction (resulting in elevated serum creatinine) and transvascular leakage for macromolecules (i.e., decreased serum albumin). 11,30 Regarding inflammation, CRP, neutrophil-lymphocyte ration, fibrinogen have been reported as negative prognostic agents in tumor survival. 6 8,31 In our analysis we found a correlation between albumin-creatinine ratio with fibrinogen and CRP. If we consider the latter variables as surrogate for inflammation, our findings appear to be biological plausible. The direction of the hazard and subdistribution hazard ratios indicate that higher creatinine is associated with a worse prognosis, whereas higher albumin is associated with a more favorable prognosis. This is clinically plausible as albumin is supposed to be a marker of performance and nutritional status. A major limitation of this retrospective study is that due to the relatively small event rate, we could not adjust for comorbidities in the multivariable analysis. However, the strength of our report is the homogeneity of our single-center study population. The inclusion of the few prevalent cases in the present study may not reflect a survivorship bias because we set the cut off of the prevalent cases of 12 weeks and excluded all patients that had a longer duration between initial surgery and re-resection. No significant difference in outcome was reported when re-resection within 12 week after initial surgery. 32 CONCLUSION Our study provides evidence for a tumor-stage-independent association between higher creatinine and lower albumin with worse disease-specific survival. Low albumin and a high Albumin-Creatinine-Ratio predict independently poor overall survival. In conclusion, this work identified novel prognostic biomarkers for prognosis of patients with liposarcoma. We further identified a novel prognostic biomarker for disease and overall prognosis in liposarcoma, the albumin-creatinine-ratio. These biomarkers could be used for individual risk estimation and integrated in existing prognostic models for soft tissue sarcoma after future studies have confirmed their prognostic value. AUTHORS CONTRIBUTIONS J. Panotopoulos, F. Posch, and C. Ay performed research design, analysis and interpretation of data, and drafted the mauscript. M. Willegger performed acquisition of data and drafted the manuscript. FT. Funovics, A. Scharrer, W. Lamm, T. Brodowicz, and R. Windhager revised critically the paper. All authors have read and approved the submitted and final versions of the paper. REFERENCES 1. Rutkowski P, Trepka S, Ptaszynski K, et al Surgery quality and tumor status impact on survival and local control of resectable liposarcomas of extremities or the trunk wall. [Internet]. Clin Orthop Relat Res 471: Available from fcgi?dbfrom¼pubmed&id¼ &

6 538 PANOTOPOULOS ET AL. 2. Haniball J, Sumathi VP, Kindblom LG, et al Prognostic factors and metastatic patterns in primary myxoid/roundcell liposarcoma. Sarcoma 2011: Fletcher CDM, Unni KK, Mertens F Pathology & Genetics. IARC 1 p. 4. Haniball J, Sumathi VP, Kindblom LG, et al Prognostic factors and metastatic patterns in primary myxoid/round-cell liposarcoma. Sarcoma. 2011, Article ID , 10 p. doi: /2011/ Kattan MW, Leung DHY, Brennan MF Postoperative nomogram for 12-year sarcoma-specific death. J Clin Oncol 20: Szkandera J, Absenger G, Liegl-Atzwanger B, et al Elevated preoperative neutrophil/lymphocyte ratio is associated with poor prognosis in soft-tissue sarcoma patients. Br J Cancer 108: Szkandera J, Pichler M, Liegl-Atzwanger B, et al The elevated pre-operative plasma fibrinogen level is an independent negative prognostic factor for cancer-specific, diseasefree and overall survival in soft-tissue sarcoma patients. J Surg Oncol 109: Nakamura T, Matsumine A, Matsubara T, et al Clinical significance of pretreatment serum C-reactive protein level in soft tissue sarcoma. Cancer 118: Pichler M, Hutterer GC, Stojakovic T, et al High plasma fibrinogen level represents an independent negative prognostic factor regarding cancer-specific, metastasis-free, as well as overall survival in a European cohort of nonmetastatic renal cell carcinoma patients. Br J Cancer 109: Panotopoulos J, Posch F, Alici B, et al Hemoglobin, alkalic phosphatase, and C-reactive protein predict the outcome in patients with Liposarcoma. J Orthop Res 33: Lin Y-S, Chiu F-C, Lin J-W, et al Association of albuminuria and cancer mortality. Cancer Epidemiol Biomarkers Prev 19: Rausch S, Hennenlotter J, Todenh ofer T, et al Impaired estimated glomerular filtration rate is a significant predictor for non-muscle-invasive bladder cancer recurrence and progression-introducing a novel prognostic model for bladder cancer recurrence. Urol Oncol 32: Aung A, Alqudihy S, Rybicki L, et al Does serum albumin and creatinine predict survival of inpatient palliative care patients? Am J Hosp Palliat Care 31: Biswas B, Rastogi S, Khan SA, et al Hypoalbuminaemia is an independent predictor of poor outcome in metastatic Ewing s sarcoma family of tumours: a single institutional experience of 150 cases treated with uniform chemotherapy protocol. Clin Oncol (R Coll Radiol) 26: Yao Y, Zhao M, Yuan D, et al Elevated pretreatment serum globulin albumin ratio predicts poor prognosis for advanced non-small cell lung cancer patients. J Thorac Dis 6: Yao Z-H, Tian G-Y, Yang S-X, et al Serum albumin as a significant prognostic factor in patients with malignant pleural mesothelioma. Tumour Biol 35: O~nate-Oca~na LF, Aiello-Crocifoglio V, Gallardo-Rincon D, et al Serum albumin as a significant prognostic factor for patients with gastric carcinoma. Ann Surg Oncol 14: Ruiz-Tovar J, Martn-Perez E, Fernandez-Contreras ME, et al Impact of preoperative levels of hemoglobin and albumin on the survival of pancreatic carcinoma. Rev Esp Enferm Dig 102: Wang C-Y, Hsieh M-J, Chiu Y-C, et al Higher serum C-reactive protein concentration and hypoalbuminemia are poor prognostic indicators in patients with esophageal cancer undergoing radiotherapy. Radiother Oncol 92: Lambert JW, Ingham M, Gibbs BB, et al Using preoperative albumin levels as a surrogate marker for outcomes after radical cystectomy for bladder cancer. Urology 81: Roth GA, Lebherz-Eichinger D, Ankersmit HJ, et al Increased total cytokeratin-18 serum and urine levels in chronic kidney disease. Clin Chim Acta 412: Morizane S, Iwamoto H, Masago T, et al Preoperative prognostic factors after radical nephroureterectomy in patients with upper urinary tract urothelial carcinoma. Int Urol Nephrol 45: Hill PG The measurement of albumin in serum and plasma. Ann Clin Biochem 22: Bartels H, B ohmer M, Heierli C Serum creatinine determination without protein precipitation. Clin Chim Acta 37: Enneking WF, Spanier SS, Goodman MA A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop Relat Res Sobin LH, Wittekind CH TNM Classification of Malignant Tumours. Hoboken, New Jersey: John Wiley and Sons. 27. Coindre JM Grading of Soft Tissue Sarcomas: Review and Update. Arch Pathol Lab Med 130: Schemper M, Smith TL A note on quantifying followup in studies of failure time. Control Clin 17: Coviello V, Boggess M Cumulative incidence estimation in the presence of competing risks. Stata J 4: Jørgensen L, Heuch I, Jenssen T, et al Association of albuminuria and cancer incidence. J Am Soc Nephrol 19: Ay C, Dunkler D, Pirker R, et al High D-dimer levels are associated with poor prognosis in cancer patients. Haematologica 97: Funovics PT, Vaselic S, Panotopoulos J, et al The impact of re-excision of inadequately resected soft tissue sarcomas on surgical therapy, results, and prognosis: a single institution experience with 682 patients. J Surg Oncol 102: