Tf. Bathen et al., Analysis and classification of proton NMR spectra of lipoprotein fractionsfrom healthy volunteers and patients with cancer or CHD, ANTICANC R, 20(4), 2000, pp. 2393-2408
Human blood plasma samples from 52 subjects were collected and the very low
density lipoprotein (VLDL) intermediate density lipoprotein (IDL) low dens
ity lipoprotein (LDL) and high density lipoprotein were isolated by serial
ultra centrifugation. 600 MHz H-1 NMR spectra of the lipoprotein fractions
were acquired. The methyl and methylene regions in the spectra of VLDL, LDL
and HDL were utilised in further analyses via Kohonen neural networks (KNN
) and generative topographic mapping (GTM) two related examples of (unsuper
vised learning) self-organising feature mapping techniques. Systematic vari
ations in lipoprotein profiles can be substantially visualised through the
use of KNN and GTM. The relationship between the sample positions in the Ko
honen plot was visualised by surface plots of the corresponding VLDL and HD
L cholesterol and VLDL triglyceride contents. The GTM maps of the VLDL and
HDL fractions were used to investigate the individual properties of selecte
d samples. A large number of the cancer patients were found clustered in th
e VLDL GTM map, and GTM map positions of samples in relation to CHD, diabet
es and renal failure could be found. Although the study group here consider
ed is heterogeneous in respect to age, sex, type of disease and medications
within each defined class, classification of VLDL and HDL data with probab
ilistic neural network (PNN) was quite successful with respect to the group
ings: cancer, CHD, volunteers and other (comprising patients with other dis
eases). Statistics based on 15 independent sets of PNN calculations gave tr
ue positive fractions usually higher than 0.83 and false positive fractions
lower than 0.088. Attempts to use the corresponding LDL data and four clas
ses were uniformly poor although some classifications (e.g, volunteer versu
s CHD) were easily performed.