The testing was performed with ELIS

The testing was performed with ELISA technology. All procedures
were performed according to the manufacturer's instructions making use of
an ELISA plate washer PW40 (Sanofi Pasteur). Read-outs of the
20 Microtiterplate were digitally saved and further used for the data reduction.
All data analysis has been done by making use of standard curves of OD
values obtained by the Microtiterplate reader (Multiscan EF type 35,
ThermoScientific) against concentrations as provided by the individual
manufacturers of the kits. Individual measured patient sample values were 25 obtained by interpolation of the sample OD value and the OD values of the
standard curve obtained in each run.
The results of the individual markers in the individual samples are presented in Table 2 for the serum samples and in Table 3 for the urine samples. The urine samples were corrected for creatinin.
The data from the patients and control persons were collected and for every parameter two values were determined, if possible. To obtain the two values for every parameter a Receiver Operating Characteristic (ROC)
was determined using the statistical program Medcalc (version 11.4.0.0,
5 http://www.medcalc.org ). A first value was determined which would
discriminate between the two groups being the criterion for the Cut-off for
having the disease, set at > 95% of the calculated Positive Predictive Value
(+PV). Any sample value measured exceeding this first discriminative value
(Cut-off-Dis, for diseased) is considered discriminative for the diseased
10 state. The second value was the reciprocal situation: being the criterion for
the Cut-off for not having the disease, set at > 95% of the calculated
Negative Predictive Value (-PV). Any sample value measured exceeding this
second discriminative value (Cut-off-Excl., excluding the diseased state) is
considered discriminative for the healthy state. Any value found in between 15 the Cut-off-Dis and Cut-off-Excl is considered non-discriminative. The
parameters where the measurements enabled the determination of one or
both of these cut-off values were regarded as being sufficient discriminatory
to be used as a marker for the present การประดิษฐ์.


20 On basis of results of the Cut-off-Dis and Cut-off-Excl. for each
parameter the results were further analyzed by assigning the samples being Positive, Negative and indeterminate. A positive sample was assigned the value +1, a Negative sample the value '-1', and the indeterminate the value
`0'. Next, a set of markers was constructed. Per sample, in each set
25 parameters, the sum of all Pos, Neg and indeterminate results (expressed in
+1, -1, 0) was determined, which resultant values were then tested for
sensitivity and specificity by ROC analysis: optimal cut-off results obtained with each set of markers were used to assign patients being diseased or to exclude controls from being diseased from the resultant +PV and —PV for
30 each set of markers.

Three sets of data were analysed separately to make choices for combinations of biomarkers in each set. The sets were based on the body-
fluid in which the biomarkers are detected: serum/plasma, urine, or the
5 combination serum/plasma and urine. In all three sets, based on increasing
contribution, a reduction in the numbers of biomarkers can be reached,
starting with the 4th order (i.e. all markers) down to the 3rd, the 2nd and
ultimately to the 1st order. Significance of a contribution of biomarkers in
the respective order is based on the position of the Area under the Curve 10 (AUC) of the measured combination in the phenotype randomisation
histogram of the AUC as defined for ROC analysis. Significance is any
position within the right 0% to 5% region. The significances shown in Table
4 are given upon phenotype randomisation against the total number of
biomarkers included in the 4th - 1st order biomarkers. The AUC varies from 15 0.500 to 1.000. The higher the value of the AUC, the better its sensitivity
and specificity. The AUC is therefore chosen to evaluate the performance of
each combination of biomarkers and to make choices for optimization
purposes.
Within each 'order' three or four analyses are done, based on
20 different inclusion criteria at 90% and 95% respectively. The more stringent
the inclusion criterion, the less the number of biomarkers that are included fulfil the inclusion criteria (participating biomarkers'). Upon increasing the stringency starting in the 4th order, the amount of participating biomarkers
decreases, allowing to make an optimisation choice. Those biomarkers under 25 a certain condition that are found not to participate under higher
stringencies are excluded in the next lower order, stepwise.

Description of the choices.
1. Urine plus serum/plasma.
Testing significance at stringency to include 7.5% for the 90 %
and 95% conditions, shows a phenotype randomisation significance of
p=0.04. Under this condition, all 40 biomarkers are included making each
biomarker a potential candidate irrespective the body fluid in which it is 5 tested (randomisation run characteristics: AUC-ree=0.855, AUC.
random=0,801+/- 0.032, number of runs=3025, fraction left of 0,855=95.7%,
active biomarkers 40/40, Software Randomisation check version 1.05, see
table).
4th order: all 40 tested biomarkers appeared to be included.
10 3rd order: excluded 19 biomarkers at inclusion condition
20%/12.5%, remaining 21 biomarkers.
2nd order: excluded 3 biomarkers at inclusion condition 25%/12.5%, remaining 16 biomarkers.
1st order: excluded 5 biomarkers at inclusion condition 25%/15%,
15 remaining 11 biomarkers.
Condition of maximum performance, AUC.a,=0,879, reached with
21 biomarkers: 10 serum biomarkers plus 11 urine biomarkers. Condition of optimal performance, AUCmin=0,858, reached with 11 biomarkers:
combination of 5 biomarkers in serum and 6 biomarkers in urine.
20
2. Urine.
A stringency to include 10% at 90 % and 95%, obtains a
significance of p=0.046. Under this condition, all 19 biomarkers are included
making each urine biomarker tested a potential candidate. Randomisation 25 run characteristics: AUC-rea i= 0.781, AUC-random = 0,721+/- 0.037, number of
runs=3538, fraction left of 0,721 = 95.4%, active biomarkers 19/19, Software
Randomisation check version 1.05, data not shown in table.
4th order: all 19 biomarkers appeared to be included.
3rd order excluded 4 biomarkers at inclusion condition 14%/10%:
30 remaining 15 biomarkers.
2nd order excluded 4 biomarkers at inclusion condition 20%/12.5%: remaining 11 biomarkers.
1st order excluded 4 biomarkers at inclusion condition 25%/15%: remaining 7 biomarkers.
5 Condition of maximum performance, AUCmax=0,825, reached with
11 biomarkers. Condition of optimum performance, AUCopt=0,823, reached with 11 biomarkers.


3. Serum/plasma
10 A stringency to include 7.5% at 90% and 95% shows a significance
of p=0.030. Under this condition, all 21 biomarkers are included making
each serum/plasma biomarker tested a potential candidate. Randomisation run characteristics: AUC-rea=0.793, AUC-random = 0,727+/- 0.036, number of runs=3154, fraction left of 0,721=97.0%, active biomarkers 21/21, Software
15 Randomisation check version 1.05, data not shown in table.
4th order: 21 biomarkers included.
3rd order excluded 4 biomarkers at inclusion condition 15%/7.5%, remaining 17 biomarkers.
2nd order excluded 7 biomarkers at inclusion condition20%/12.5%,
20 remaining 10 biomarkers.
1st order excluded 4 biomarkers at inclusion condition 25%/14%, remaining 6 biomarkers.
Condition of maximum performance, AUC.=0,793, reached with
21 biomarkers. Condition of optimum performance, AUCopt=0,775, reached
25 with 10 biomarkers.

ผลลัพธ์ (ไทย) 3:

ทดสอบการปฏิบัติ + เทคโนโลยี ทุกขั้นตอนมีการปฏิบัติตามคำสั่ง

เป็นผู้ผลิตให้ใช้วิธี pw40 ( ซาโนฟี่ปาสเตอร์จานเครื่องซักผ้า ) อ่านลึกหนาบางของ
20 microtiterplate ถูกเซ็นชื่อแบบดิจิทัลบันทึกและเพิ่มเติมที่ใช้สำหรับข้อมูลทั่วไป
ทั้งหมดการวิเคราะห์ข้อมูลทำโดยการใช้เส้นโค้งมาตรฐานของ OD