Daria Pašalić
Department of Medical Chemistry, Biochemistry and Clinical Chemistry
Zagreb University School of Medicine
Šalata ul 2.
10 000 Zagreb, Croatia
Phone +385 (1) 4590 205; +385 (1) 4566 940
E-mail: dariapasalic [at] gmail [dot] com

Useful links

Evaluation of tumor marker HE4 assay on the Elecsys 2010 analyzer
Ćorić J
Clinical Center of Sarajevo University, Department of Clinical Chemisty, Sarajevo, Bosnia and Herzegovina
Corresponding author: coricjozo [at] hotmail [dot] com
Background: Whey-acidic protein human epididymis protein4 (HE4), a new promising biomarker for epithelial ovarian cancer (EOC). The measured HE4 value of patients sample can depending on the testing procedure use.
Methods: We evaluated a HE4 method on Elecsys 2010 analyzer. The method for quantitative determination of HE4 is direct, competitive electrochemiluminescence immunoassay. For quality control we use Elecsys PreciControl HE4 1 and 2. HE 4 was measure on sera obtained from 96 women (40 healthy and 56 with epithelial ovarian cancer).
Results: The Roche HE 4 assays showed a good linearity (r = 0.99) and precision (intrassay and total CV < 5%). The median HE4 serum concentrations was significantly higher among EOC patients than healthy females (P < 0.05). As a single marker, HE4 had a sensitivity of 78.4 % with a specificity of 95 %.
Conclusions: The presented results of the analytical evaluation methods for the determination of HE 4 on the Elecsys 2010 analyzer showed an acceptable accuracy and precision.
CYFRA 21-1 new tumor marker in Abbott iArchitect family
Županić D, CrnokrakS, Mujagić R, Honović L
Pula General Hospital, Department of Laboratory Diagnostics, Pula, Croatia
Corresponding author: danijela [dot] zupanic [at] pu [dot] t-com [dot] hr
Introduction: CYFRA 21-1 is a tumor marker that measures fragments of Cytokeratins 19 using two monoclonal antibodies KS 19-1 and BM 19-21, but its main role is to monitor the course and success of the therapy of NSCLC.
Objective: Our goal was to investigate the acceptability of Abbott reagents for use in daily routine on immunochemical analyzer Architect i2000SR to ensure traceability of our results. Validation of immunoassays was performed using CLSI / NCCLS procedure EP 15-A2.
Materials and methods: The following parameters were assessed: Precision (random error) in which we determined repeatability, interprecision and overall laboratory precision by determination of a small group (N = 30) and systematic error (deviation from expected values).
Control Level 1: 5 ng / mL (3.5 to 6.5) (Repeatability x = 5.23, Sr = 0.22, KV% = 4.2%)
(Interprecision x = 5.23, Sb = 0.123, KV% = 2.4%)
(Overall laboratory precision SL = 0.218, KV% = 4.2%)
(Systematic error is 0.23 (4.6%))
Control Level 2: 35 ng / mL (24.5 to 45.5) (Repeatability x = 34.92 ng / ml, Sr = 0.44, KV% = 1.3%)
(Interprecision x = 34.92, Sb = 0.524, KV% = 1.5%)
(Overall laboratory precision SL = 0.635, KV% = 1.8%)
(Bias -0.08 (-0.23%))
Conclusion: Based on the results of this study, we can conclude that the tested reagent CYFRA 21-1 Abbott is acceptable and we recommend its use on automated immunochemical analyzer iArchitect Abbott.
Analytical performance of tumor markers on UniCel DxI 600 using Sigma metrics
Unic A, Derek L, Marijancevic D, Serdar T, Romic Z
University Hospital Dubrava, Clinical Department for laboratory diagnostics, Zagreb, Croatia
Corresponding author: aunic [at] kbd [dot] hr
Background: The Sigma concept of a tolerance limit provides guidance for defining intended medical use in the form of an allowable total error. Performance is characterized on a sigma scale. In terms of Sigma, if a method has a value less than three is considered to be unreliable and should not be used in routine laboratory practice. The aim of this study was to assess the performance of automated immunodiagnostic chemiluminescence system UniCel DxI 600 (Beckman-Coulter, Tokyo, Japan) for qualitative detection of tumor markers.
Materials and methods: For CA19-9, CA15-3, CA125, and PSA method validation Lypochek Tumor Marker (Bio-Rad Laboratories, Marnes-la-Coquette, France; Control lot 54530- Level 1 and 3) control samples were obtained. All immunoassays were preformed on the UniCel DxI 600 analyzer according to the manufacturer instructions. Control samples were tested in triplicate every day during the 5 days according CLSI/NCLLS protocol EP15-A2. Bias was calculated using method comparison. Sigma was calculated using Ricos quality requirements.
Results: Calculated sigma metrics for CA15-3 were 2.5 and 1.4; for CA125 6.5 and 5.9; for CA19-9 6.4 and 9.2; and for PSA-1.0 and 2.2.
Conclusion: Analytical performance for CA125 and CA19-9 is world class. On the other hand, CA15-3 and PSA analytical performance is poor or even unacceptable. Using graphic tool it becomes apparent that bias is the main problem with CA15-3 and PSA performance. With lower bias CA15-3 assay performance might even reach Six Sigma zone. Therefore, it is necessary to use the same method for individual patient. In that case, even these two methods became of excellent quality.
Prognostic value of Cyfra 21-1, CEA and NSE in patients with NSCLC
Musliu A
Faculty of Natural Science, Department of Biology, Tirana, Albania
Corresponding author: milazimshabani [at] hotmail [dot] com
Background: The aims of this study were to evaluate prognostic effects, sensitivity and specificity of Cyfra 21-1 in detecting non-small cell lung cancer.
Subjects and methods: The study included 118 randomly selected patients with NSCLC compared with control group of 30 patients with nonmalignant pulmonary disease. Histology tumor diagnosis was based on biopsy specimens obtained at bronchoscopy, lymphode biopsy or thoracotomy.Tumor markers (Cyfra 21-1,CEA and NSE) were assayed in Immunological Abbot’s Axsym System analyzers. Chemotherapy included cisplatin and carbocisplatin were aplicated in doses (80-100 mg/m2 day 1.3 wk cycle and 200-300 mg/m2 day 1.3 wk cycle for six cycles).
Results: The level of tumor markers was determined in both population in different stage of age before and after chemotherapy. Median level of Cyfra 21-1 at diagnosis was 36.2 ng/mL with range 2.8-215.0 and decreased significantly after second cycles of chemotherapy 24.3 ng/mL (range 3.4-145.0; P < 0.01). Same results we obtained in the NSE concentration (23.4 μg/mL before and 14.2 μg/mL after therapy, but CEA shows not significantly changes. Cyfra 21-1 were elevated in 22.3%, NSE in 10.8% and CEA in 16.5% of patients respectively.
Conclusion: Cyfra 21-1 reflects the extent of the disease and has an independent prognostic role along with performance status and disease stage in NSCLC.Combining Cyfra 21-1, NSE and CEA correlated with prognosis in a significant and independent manner.
Computational screening for non-small cell lung carcinoma biomarker candidates
Dundovic S (1), Debeljak Z (2), Seric V (2)
(1) General Hospital "Nasice", Department of Medical Biochemistry, Nasice, Croatia
(2) Clinical Hospital Center "Osijek", Department of Clinical Laboratory Diagnostics, Osijek, Croatia
Corresponding author: zeljko [dot] debeljak [at] gmail [dot] com
Introduction: Comparison of gene expressions obtained on control and patient samples based on computational methods allows selection of relevant genes and their protein products which may serve as biomarker candidates. For the purpose of non-small cell lung carcinoma biomarker screening a multivariate classification method known as k-nearest neighbors has been selected in this study.
Materials and methods: The study was conducted on a publicly available set of gene expressions (Showe MK et al. Cancer Res 2009;69:9202-10) which consists of 291 gene expression profiles (137 non-small cell lung carcinoma patient samples and 91 control samples), each containing 15 227 gene expressions. k-nearest neighbors method was applied to all pairs of gene expressions. t-test with Benjamini-Hochberg correction for multiple testing has been used as a univariate filter.
Results: Application of k-nearest neighbors resulted in the test set accuracy of 81.6-86.8% which corresponds to 1000 top ranking gene pairs. Less than a half of highly ranked genes were statistically significant according to the corrected t-test. The majority of the top ranking genes are responsible for the processing of gene information and for the regulation of signaling pathways.
Conclusion: k-nearest neighbors method provides insight into the relevance of single genes and gene interactions for differentiation of patient and control samples. Low correspondence between univariate results and the results of k-nearest neighbors underlines importance of the latter. This might influence the quality of candidates for non-small cell lung carcinoma biomarkers and corresponding diagnostic approaches which laboratory evaluation is underway.
Prognostic and predictive significance of ERβ1 in primary breast cancer
Jelisavac Ćosić S (1), Jakić-Razumović J (2), Sirotković-SkerlevM (3)
(1) University Hospital Centre Zagreb, Department of Oncology and Radiotherapy, Zagreb, Croatia
(2) University Hospital Centre Zagreb, Zagreb University Medical School, Department of Pathology and Cytology, Zagreb, Croatia
(3) Zagreb University Medical School, Department of Pathophysiology, Zagreb, Croatia
Corresponding author: jelisavaccosic [dot] sanda [at] gmail [dot] com
Background: Adjuvant endocrine therapy is effective treatment in breast cancer. Patients receive endocrine treatment according to steroid receptor positivity (ERa and PgR). Estrogen receptor β (ERβ) is a second estrogen receptor and its role in endocrine treatment is not fully elucidated. The aim of this research was to determine prognostic and predictive value of isoform EBβ1 in breast cancer.
Materials and methods: In the study were included 150 consecutive primary breast cancer cases operated at University hospital Zagreb (year 2002-2003). Overall and disease free survival were recorded until January 1st, 2011. Immunohistochemistry was used for EBβ1 determination (anti-ERβ1 clone PPG5/10, Dako, Denmark).
Results: ERβ1 and ERα expressions were not correlated (P = 0.178, r = 0.123). There was no association of ERβ1 with age, menopausal status, lymph node status, tumor size, histological grade, histological type, nuclear grade, Nottingham prognostic index, proliferation index Ki67 expression, steroid receptor status and HER-2/neu status. In univariate Cox-regression analysis, ERβ1 was associated with overall survival (P = 0.026; HR = 0.46; 95%CI 0.24-0.92) but there was no association with disease free survival (P = 0.054; HR = 0.51; 95%CI 0.25-1.04). Kaplan-Meyer survival curve analysis confirmed prognostic significance of ERβ1 for overall survival. In the subgroup of patients that received endocrine treatment (N = 93), there was no statistical difference in survival between patients with high and those with low ERβ1 [overall survival (P = 0.27; HR = 1.73; 95%CI 0.65-4.7), disease free survival (P = 0.08; HR = 2.25; 95%CI 0.91-4.94)].
Conclusions: Cancer tissue expression of ERβ1 is a prognostic, but not a predictive marker in primary breast cancer.
Comparative study of two chemiluminescent methods for determining free PSA
Fernandez Leivas A, Garcia Moreira V, Llorente Torres A, Beridze N, Garcia Alonso S, Fernandez Rodriguez E
Hospital de Cabueñes, Servicio de Analisis Clinicos, Gijon, Spain
Corresponding author: anafleivas [at] hotmail [dot] com
Background: The % free PSA (fPSA) is, on average, lower in prostate cancer than the benign prostatic hyperplasia and has been commonly used as an aid in the diagnosis of prostate cancer when PSA is between 4-10 ng/mL.
Materials and methods: In our study, fPSA were analyzed in 50 serum samples of patients with total PSA between 4-10 ng/mL. The results obtained using the methods ADVIA Centaur® fPSA test (Siemens) and fPSA Immulite2000 (Siemens), were compared in order to ensure transferability of results between both methods. Total PSA was measured by ADVIA Centaur® method. Analysis of data was performed using the MedCalc®, using Passing-Bablok nonparametric regression test, Pearson´s correlation and agreement between methods was evaluated using the Kappa statistics, according to the following groups: positive > 15% and negative < 15% fPSA.
Results: The following values of fPSA were obtained (range, median) 0.39-2.72, 1.191 (95% CI: 1.037 to 1.346) and 0.27-2.46, 1.135 (95% CI: 0.985 to 1.285) for Immulite2000 and ADVIA Centaur® respectively, yielding a correlation coefficient of 0.9519 between both methods (P < 0.001). The Passing-Bablok regression equation obtained had a Slope = 0.9583 (95% CI 0.8729 to 1.0596) and Intercept = 0.0196 (95% CI: -0.0837 to 0.1084). The concordance of % freePSA is good, yielding a Kappa index of 0.797 (95%CI: 0.628-0.966).
Conclusions: Both methods of fPSA show a good correlation (0.9 <= R <1) and concordance. Results between methods are transferable.
Soluble urokinase plasminogen activator receptor (suPAR) in breast cancer
TarapaczJ, RychlikU, WójcikE, StasikZ, KulpaJ
Center of Oncology – Maria Sklodowska-Curie Memorial Institute, Cracow Division, Poland, Clinical Biochemistry, Kraków, Poland
Corresponding author: z5tarapa [at] cyfronet [dot] pl
Introduction: The urokinase plasminogen activator system plays an important role in many processes involved with cancer invasion and metastasis. The elevated levels of uPAR as well as uPA and PAI-1 are associated with poor prognosis in cancer patients. Recently, similar suggestions are brought forward to soluble form of this receptor – suPAR. The aim of study was the evaluation of relationship between suPAR and selected biochemical and clinical parameters in breast cancer patients.
Material and methods: The study of suPAR, CA 15-3, CEA, CRP was performed in the group of 146 breast cancer patients before surgical treatment and in the reference group of 43 healthy women.
Results: In breast cancer patients in comparison with the reference group there were found significantly higher levels of suPAR, CA 15-3, CEA and CRP. Significantly higher levels of CA 15-3 had patients in more advanced stages of disease, with tumor greater than 2 cm, and CRP levels > 3 mg/L. There were no significant differences between analyzed biochemical factors between groups selected in respect to: lymph node status, histological grade, percentage of S-phase cells, steroids receptor status. In the group of breast cancer patients with HER2 overexpression significantly lower suPAR and significantly higher CA 15-3 levels and the percentage of S-phase cells there were observed.
Conclusions: In breast cancer patients suPAR level seems to be associated with tendency to inflammation. The reciprocal relationship between HER2 status and suPAR levels may indicate potential role of this receptor in prognosis of some breast cancer patients.
BRAF: potential prognostic marker in colorectal cancer
RakoI (1), Jakić-Razumović J (2), SabolI (3), Sertić J (1), PleštinaS (4)
(1) University Hospital Center Zagreb, Department of Laboratory Diagnostics, Zagreb, Croatia
(2) University Hospital Center Zagreb , Department of Pathology and Cytology, Zagreb, Croatia
(3) Institute Rudjer Boskovic Zagreb, Laboratory of Molecular Virology and Bacteriology, Zagreb, Croatia
(4) University Hospital Center Zagreb, Department of Oncology, Zagreb, Croatia
Corresponding author: irako [at] kbc-zagreb [dot] hr
Introduction: Incidence and mortality of colorectal cancer (CRC) are on constant increase and represent one of the major health problems in Croatia. Activation of BRAF oncogene is implicated in colorectal carcinogenesis. BRAF protein is a serin/threonine kinase, component of a conserved signaling pathway that regulates cellular responses to exstracellular signals. In human cancer, it is commonly activated by hotspot mutation of the BRAF gene which leads to a single aminoacid substitution p.V600E. The aim was to determine the incidence of BRAF gene mutations in CRC patients in Croatia and to assess whether they are linked with clinicopathological features of poor prognosis.
Materials and methods: Sections from 113 formalin-fixed paraffin-embedded tumor samples were evaluated for the BRAF mutation using LightCycler PCR with allele-specific fluorescent probe melting curve analysis. Obtained results were confirmed by sequencing method.
Results: Our results show that BRAF gene mutation p.V600E was detected in 8.8% (10/113) CRC samples. Statistical analysis revealed a significant association between the BRAF mutation and Dukes’stage (P = 0.04) where all mutations were found in tumors classified as Dukes’C. Incidence of mutation was higher in males, patients older than 60 years, tumors bigger than 5 cm, tumors with angioinvasion and poor differentiated tumors, but no significant association was found. All BRAF gene mutations were detected in colon cancers.
Conclusions: The incidence of BRAF gene mutation in CRC patients in Croatia is within commonly accepted limits. Higher incidence in tumors with poor prognostic markers shows that BRAF gene mutation play a role in the progression of CRC.
Assessment of serum CA 15-3 and CEA concentrations in patients with breast cancer
Tuteska J (1), Stojkovski V (2), Arapceska M (3), Lozanovski Z (1), Risteska K (1)
(1) University "St. Kliment Ohridski", Medical Nursing College, Depatment of Biochemistry, Bitola, Macedonia, TFYR
(2) University "Ss. Cyril and Methodius", Faculty of Veterinary Medicine, Depatment of Biochemistry, Skopje, Macedonia, TFYR
(3) University "St. Kliment Ohridski", Faculty of Biotechnical Sciences, Department of Natural Sciences, Bitola, Macedonia, TFYR
Corresponding author: jtuteska [at] yahoo [dot] com
Background: Breast cancer is the most common malignancy in women. The aim of this study was assessment of serum CA 15-3 and CEA concentrations in patients with breast cancer and monitoring of their status by these tumor markers.
Material and methods: Serum samples were acquired from female patients with breast cancer (N = 224) grouped according pTNM staging. Control samples (N = 44) were collected from healthy females. Concentrations of tumor markers were measured by the immunochemical analyzer Vitros ECI with enhanced chemiluminiscence.
Results: According obtained results concentrations of CA 15-3 in patients in grade IIIb (102.50 ± 10.61 U/mL) and grade IV (134.50 ± 179.74 U/mL) were 86.67% and 89.84% significantly higher (P < 0.01, P < 0.05 respectively) than in healthy women (13.66 ± 8.55 U/mL). Concentration of CEA in patients in grade IV (19.3 ± 17.4 ng/mL) was 92.64% higher (P < 0.01) compared with healthy women (1.42 ± 1.02 ng/mL). Three months after surgery in patients was noted increase of CA 15-3 values, which after six months were 71.14% higher, despite CEA values that were in referent range.

Conclusions: Because of low sensitivity tumor markers CA 15-3 and CEA are more suitable for monitoring of state of patients after surgery than for early screening or diagnosis of breast cancer.