Use of Discriminant Analysis to Assess Disease Activity in Pulmonary Tuberculosis with a Panel of Specific and Nonspecific Serum Markers

Abstract

Several cell activation markers, acute phase reactants, enzymes, and antituberculous antibody serum levels have been proposed as possible markers to monitor disease activity in patients with tuberculosis. They have all shown limited sensitivity or specificity. The authors therefore attempted to generate a canonical variable using discriminant analysis, including sensitive and specific parameters, to be a reliable marker in classifying patients correctly during the course of pulmonary tuberculosis. The following parameters were selected: two soluble cell activation markers (soluble interleukin-2 receptor and sCD8); the levels of immunoglobulin (Ig) G and IgM antibodies against the A60 antigen complex; and the presence of specific antibodies directed to eight different A60 components, revealed by Western blot analysis. The tests were performed on sera from three groups of patients with pulmonary tuberculosis. The first group comprised 25 patients with onset tuberculosis, clinically active (OTCA), evaluated at the time of admission. The second group included 28 patients with chemotherapy-treated tuberculosis, clinically active (CTCA), 2 months after therapy had begun. The third group included 20 patients with tuberculosis, nonclinically active (TNCA), who had had at least 1 year of effective therapy. The authors obtained an 80.9% rate of correct classification for the three groups and a rate of 100% when OTCA and TNCA were compared. The patients with CTCA were scarcely differentiated and tended to be distributed into the two other groups. To improve the separation between patients with CTCA and those with OTCA, a second canonical variable was generated with a 91.7% rate of correct classification, as compared with 71% obtained using the sCD8 as the best single variable. The mean values of the last canonical variable were statistically different (Mann-Whitney test, P = .049) when stratified for acid fast bacilli-positive or negative CTCA patients (microscopic detection). Three patients, followed during the entire course their disease, were, as expected, correctly positioned with respect to the subsequent disease phases.

Validation of an algorithm able to differentiate small-cell lung cancer (SCLC) from non-small-cell lung cancer (NSCLC) patients by means of a tumour marker panel: analysis of the errors

“…However, this does not use information available in the quantitative data. This can be ameliorated by using the discriminant analysis, which represents one of the best methods of combining the discriminant power of more variables to obtain the most accurate classification between two or more groups (Ameglio et al, 1991(Ameglio et al, , 1994.…”

Discriminant analysis on small cell lung cancer and non-small cell lung cancer by means of NSE and CYFRA-21.1

“…This methodology is currently applied in taxonomy. Several examples of the use of this method are also reported in the literature concerning different fields [14][15][16][17]20]. Recently, serum tumour marker levels were reliably used to distinguish between primary and metastatic malignant bone tumours [16].…”

Evaluation of a novel tuberculosis complex-specific 34 kDa protein in the serological diagnosis of tuberculosis

“…In this regard, BOTHAMLEY et al [29] have found that the combined use of tests with antigens, such as p38, p19, p14, and LAM, and with monoclonal antibodies against them may considerably improve diagnostic sensitivity (up to 72%) compared to the single tests with LAM (32%) and with p19 (58%). VERBON et al [26] have obtained similar results with antigens of 10,16,24,30,38 and 70 kDa and the monoclonal antibody TB72 (specific for the p38 protein [15]). They have shown that test combinations, such as with the 10, 16, 24 kDa antigens and the antibody TB72, could improve the diagnostic sensitivity of each test.…”

“…Furthermore, the test distinguished PPD-positive patients with bacterial pneumonia presenting with a clinical picture raising the suspicion of TB. The observation that MS2-34 is specific for active TB, and it is not recognized by IgG antibodies from patients with cured or inactive disease is quite interesting, since discrimination of active from inactive patients requires many immunological markers [30]. However, the sensitivity of the MS2-34/p38 IgM test is as high as that of the MS2-34/p38 IgG test.…”