ISSN print edition: 0366-6352 ISSN electronic edition: 1336-9075 Registr. No.: MK SR 9/7 Published monthly

Evaluation of colon cancer elements contents in serum using statistical methods

D. Milde, J. Macháček, and V. Stužka

Department of Analytical Chemistry, Palacký University, Tř. Svobody 8, 771 46 Olomouc, Czech Republic

E-mail: david.milde@upol.cz

Received: 9 February 2007  Revised: 10 May 2007  Accepted: 5 June 2007

Abstract: Classification of normal and different cancer groups (TNM classification) with univariate and multivariate statistical methods according to the contents of Cu, Fe, Mn, Se, and Zn in blood serum is discussed. All serum samples were digested by acid mixture in a microwave mineralization unit prior to the analysis by atomic absorption spectrometry. Results show that univariate methods can distinguish normal and cancer groups. Level of selenium evaluated as arithmetic mean with its standard deviation in colorectal cancer patients was (42.61 ± 23.76) µg L⁻¹. Retransformed mean was used to evaluate levels of managanese (11.99 ± 1.71) µg L⁻¹, copper (1.05 ± 0.06) mg L⁻¹, zinc (2.14 ± 0.21) mg L⁻¹, and iron (1.82 ± 0.22) mg L⁻¹. Conclusions of multivariate statistical procedures (principal component analysis, hierarchical, and k-means clustering) do not correlate very well with the division of serum samples according to the TNM classification.

Keywords: colon cancer - serum - atomic absorption spectrometry - univariate analysis - multivariate analysis

Full paper is available at www.springerlink.com.

DOI: 10.2478/s11696-007-0046-z

Chemical Papers 61 (5) 348–352 (2007)