Advertisement

AJNR Awards, New Junior Editors, and more. Read the latest AJNR updates

Research ArticleBRAIN

Brain Tumor Classification by Proton MR Spectroscopy: Comparison of Diagnostic Accuracy at Short and Long TE

Carles Majós, Margarida Julià-Sapé, Juli Alonso, Marta Serrallonga, Carles Aguilera, Juan J. Acebes, Carles Arús and Jaume Gili
American Journal of Neuroradiology November 2004, 25 (10) 1696-1704;

Article Figures & Data

Figures

  • Fig 1.
    Fig 1.

    Mean spectra at short TE (TE = 30) of the five tumor types included in the study.

  • Fig 2.
    Fig 2.

    Mean spectra at long TE (TE = 136) of the five tumor types included in the study.

  • Fig 3.
    Fig 3.

    Box plots show the distribution of some relevant resonances. The horizontal line is the median, the ends of the boxes are the upper and lower quartiles, and the vertical lines show the full range of values in the data. The extreme points (*) are outliers.

    A, LIP13 at short TE.

    B, LIP13 at long TE.

    C, Ala at short TE.

    D, Ala at long TE.

    E, CR at short TE.

    F, CR at long TE.

    G, CHO at short TE.

    H, CHO at long TE.

    I, Gly/MI at short TE.

    J, Gly/MI at long TE.

  • Fig 4.
    Fig 4.

    Scatter plots show the distribution of the complete set of cases at short TE (A) and long TE (B) by using the two first discriminant functions obtained for linear discriminant analysis (x axis, value obtained with the first discriminant function; y axis, value obtained with the second discriminant function). Open circle indicates meningiomas; open square, low-grade astrocytomas; open diamond, anaplastic astrocytomas; open triangle, glioblastomas-metastases. The large black symbols show the centroid for every tumor group. With this method, new cases with unknown diagnosis are classified according to their distance to the centroids.

Tables

  • TABLE 1:

    Analysis of significant differences between tumor groups at TE 30

    Pairwise ComparisonsResonances Showing Significant Differences†
    P* < .001P* < .01P* < .05
    MEN vs LGAGLXAlaGly/MI
    MEN vs AAAlaCR
    MEN vs GBM-METLIP09, LIP13, Ala, GLX, CR, CHO, Gly/MI
    LGA vs AA
    LGA vs GBM-METLIP09, LIP13, CR, Gly/MI
    AA vs GBM-METLIP13, CR, CHOLIP09, Gly/MI

    • Note.—Differences of P corrected values (P*) <.05 were considered to be statistically significant. MEN indicates meningioma; LGA, low grade astrocytoma; AA, anaplastic astrocytoma; GBM-MET, glioblastoma-metastasis group; GLX, glutamate and glutamine; Ala, alanine; LIP, lipid; CR, creatine plus phosphocreatine; CHO, choline and other thmethylamine-containing compounds; Gly/MI, glycine and/or myo-inositol.

    • † Only the seven resonances showing significant differences in the previously applied Kruskal-Wallis test for multiple samples were considered (LIP09, LIP13, Ala, GLX, CR, CHO, and Gly/MI).

  • TABLE 2:

    Analysis of significant differences between tumor groups at TE 136

    Pairwise ComparisonsResonances Showing Significant Differences†
    P* < .001P* < .01P* < .05
    MEN vs LGAAla, GLXCHO
    MEN vs AAAlaGLX
    MEN vs GBM-METLIP09, LIP13, LACT, Ala, GLX, CHONACC
    LGA vs AA
    LGA vs GBM-METLIP09, LIP13LACT, CR
    AA vs GBM-METLIP09, LIP13, CRNACC, CHO

    • Note.—Differences of P corrected values (P*) <.05 were considered to be statistically significant. MEN indicates mengingioma; LGA, low grade astrocytoma; AA, anaplastic astrocytoma; GBM-MET, glioblastoma-metastasis group; Ala, alanine; GLX, glutamate and glutamine; LIP, lipid; CHO, choline and other trimethylamine-containing compounds; CR, creatine plus phosphocreatine; NACC, N-acetyl-containing compounds; LACT, lactate.

    • † Only the eight resonances showing significant differences in the previously applied Kruskal-Wallis test for multiple samples were considered (LIP09, LIP13, LACT, Ala, NACC, GLX, CR, and CHO).

  • TABLE 3:

    Results of brain tumor classification by two TE (30 and 136) into four groups with linear discriminant analysis, evaluated with the leave-one-out method (in parenthesis are shown the tumors of each group according to the definitive diagnosis)

    Short TE ClassificationLong TE ClassificationTotal
    MENLGAAAGBM-MET
    MEN28 (28 MEN)2 (1 AA, 1 GBM-MET)1 (1 MEN)6 (3 GBM-MET, 2 MEN, 1 LGA)37 (31 MEN, 4 GBM-MET, 1 AA, 1LGA)
    LGA1 (1 MEN)8 (6 LGA, 2 AA)2 (1 AA, 1 GBM-MET)5 (2 AA, 1 LGA, 1 GBM-MET)16 (8 LGA, 5 AA, 2 GBM-MET, 1 MEN)
    AA06 (2 LGA, 2 AA, 2 GBM-MET)8 (6 AA, 1 GBM-MET, 1 LGA)3 (2 GBM-MET, 1 AA)17 (9 AA, 5 GBM, 3 LGA)
    GBM-MET6 (3 GBM-MET, 3 MEN)6 (6 GBM-MET)1 (1 GBM-MET)68 (65 GBM-MET, 2 MEN, 1 AA)81 (75 GBM-MET, 5 MEN, 1 AA)
    Total35 (32 MEN, 3 GBM-MET)22 (9 GBM-MET, 8 LBA, 5 AA)12 (7 AA, 3 GBM-MET, 1 LGA, 1 MEN)82 (71 GBM-MET, 4 MEN, 3 LGA, 4 AA)151 (86 GBM-MET, 37 MEN, 16 AA, 12 LGA)

    • Note.—Classification accuracy of the whole tumor set into four groups was higher at short TE (123/151; .81 [.75–.88]) (tumors correctly classified/total of tumors; accuracy [95% confidence intervals]) than at long TE (118/151; .78 [.71–.85]). MEN indicates meningioma; LGA, low grade astrocytoma; AA, anaplastic astrocytoma; GBM-MET, glioblastoma-metastasis group.

Back to top

In this issue

Advertisement
Print
Download PDF
Email Article
Cite this article
0 Responses
Respond to this article
Bookmark this article
Brain Tumor Classification by Proton MR Spectroscopy: Comparison of Diagnostic Accuracy at Short and Long TE
Carles Majós, Margarida Julià-Sapé, Juli Alonso, Marta Serrallonga, Carles Aguilera, Juan J. Acebes, Carles Arús, Jaume Gili
American Journal of Neuroradiology Nov 2004, 25 (10) 1696-1704;
del.icio.us logo Twitter logo Facebook logo Mendeley logo

Jump to section

Advertisement