Table 2.
Cross-validation and clinical validation results—before and after conversion using recalculating formulas (based on: individual hematocrit level and validated Passing-Bablok regression model)
| Statistic | Before conversion (cross-validation) | After conversion (clinical validation) | |||||
|---|---|---|---|---|---|---|---|
| MPAVAMS-LC–MS/MS vs MPAWB-LC–MS/MS |
MPAVAMS-LC–MS/MS vs MPAPL-LC–MS/MS |
MPAPL-HPLC–DAD vs MPAPL-LC–MS/MS |
MPAVAMS-LC–MS/MS vs MPAPL-HPLC–DAD |
MPAVAMSHt-LC–MS/MS vs MPAPL-LC–MS/MS |
MPAVAMSc-LC–MS/MS vs MPAPL-LC–MS/MS |
MPAVAMSc_DAD-LC–MS/MS vs MPAPL-HPLC–DAD |
|
| Passing–Bablok analysis (regression formula) | |||||||
| Passing-Bablok-regression formula | CVAMS = 0.90·CWB + 0.17 | CPL = 1.60·CVAMS + 0.08 | CHPLC = 1.01·CPL + 0.11 | CHPLC = 1.56·CPL + 0.17 | CPL = CVAMSHt — 0.13 | CPL = CVAMSc — 0.01 | CVAMSc_DAD = 1.01·CHPLC —0.02 |
|
Intercept (A) |
0.17 (− 0.06 to 0.29) |
0.08 (− 0.14 to 0.29) |
0.11 (− 0.24 to 0.30) |
0.17 (− 0.22 to 0.38) |
− 0.13 (− 0.35 to 0.03) |
− 0.01 (− 0.27 to 0.19) |
− 0.03 (− 0.35 to 0.31) |
|
Slope (B) |
0.90 (0.79–1.00) |
1.60 (1.44–1.80) |
1.01 (0.91–1.13) |
1.56 (1.38–1.78) |
1.00 (0.92–1.09) |
1.00 (0.90–1.11) |
1.01 (0.89–1.15) |
| Bland–Altman bias analysis (based on plot and %-differences bias) | |||||||
|
%-mean difference bias (− 1.96SD; + 1.96SD) |
− 6.80% (− 61.50 to 47.90%) |
54.07% (48.62–59.53%) |
+ 4.08% (− 1.21 to 9.37%) |
50.00% (1.60–98.50%) |
+ 10.90% (5.24–16.56%) |
4.76% (− 1.11 to 10.64%) |
− 1.01% (− 8.26 to 6.24%) |
|
% of paired samples fulfilled the EMA criteria (< 20%) |
70% | 0% | 12% | 12% | 88% | 86% | 72% |
| % of paired samples fulfilled the IATDMCT criteria (< 15%) | 60% | 0% | 6% | 4% | 78% | 72% | 67% |
| Predictive analysis performance factors | |||||||
| MPE | − 0.05 | − 0.92 | − 0.12 | − 0.81 | − 0.12 | − 0.01 | 0.00 |
| MPPE [%] | − 5.50 | 85.00 | 1.90 | 67.54 | 7.03 | 9.20 | 2.60 |
| RMSE [%] | 15.00 | 92.00 | 12.00 | 81.54 | 12.14 | 12.21 | 13.00 |
| MAPE [%] | 13.69 | 66.65 | 8.25 | 62.42 | 14.02 | 13.38 | 20.24 |
| Correlation coefficients | |||||||
|
Pearson’s (R2) |
0.92 (0.87–0.96; 95% CI) |
0.92 (0.86–0.95; 95% CI) |
0.93 (0.88–0.96; 95% CI) |
0.90 (0.84–0.95; 95% CI) |
0.95 (0.85–0.97; 95% CI) |
0.92 (0.87–0.96; 95% CI) |
0.91 (0.85–0.95; 95% CI) |
|
Spearman (SRCC) |
0.81 (0.68–0.89; 95% CI) |
0.88 (0.80–0.93; 95% CI) |
0.92 (0.86–0.95; 95% CI) |
0.83 (0.73–0.90; 95% CI) |
0.91 (0.84–0.95; 95% CI) |
0.88 (0.80–0.93; 95% CI) |
0.84 (0.73–0.90; 95% CI) |
|
Intraclass (ICC) |
0.91 (0.85–0.95; 95% CI) |
0.84 (0.73–0.90; 95% CI) |
0.96 (0.93–0.98; 95% CI) |
0.83 (0.74–0.91; 95% CI) |
0.94 (0.90–0.97; 95% CI) |
0.92 (0.87–0.95; 95% CI) |
0.91 (0.85–0.95; 95% CI) |
Data are presented as mean with range and confidence interval (95% CI)
VAMS volumetric absorptive microsampling, LC–MS/MS liquid chromatography-tandem mass spectrometry, HPLC–DAD high-performance liquid chromatography with diode array detection, MPAVAMS-LC–MS/MS mycophenolic acid concentration in VAMS sample determined LC–MS/MS, MPAWB-LC–MS/MS mycophenolic acid concentration in whole blood sample determined LC–MS/MS, MPAPL-LC–MS/MS mycophenolic acid concentration in plasma sample determined LC–MS/MS, MPAPL-HPLC–DAD mycophenolic acid concentration in plasma sample determined HPLC–DAD, MPAVAMSHt-LC–MS/MS mycophenolic acid concentration corrected with hematocrit, MPAVAMSc-LC–MS/MS mycophenolic acid concentration corrected with regression, MPAVAMSc_DAD-LC–MS/MS mycophenolic acid concentration corrected with regression, IATDMCT International Association of Therapeutic Drug Monitoring and Clinical Toxicology, EMA European Medicines Agency, MPE median prediction error, MPPE median percentage prediction error, RMSE root mean squared prediction error, MAPE median absolute percentage prediction error, R2 Pearson’s correlation, SRCC Spearman rank correlation coefficient, ICC Intraclass correlation coefficient