Page 136 - 89Zr-Immuno-PET:Towards a Clinical Tool to Guide Antibody-based Therapy in Cancer
P. 136

                                Chapter 6
Reliability. ICC data are presented in Table 4. For eligible VOI, ICC values for SUVmax, SUVpeak and SUVmean were ≥ 0.90 for 89Zr-rituximab (D3, D6) and for 89Zr- cetuximab (D6). For 89Zr-trastuzumab, ICC values ≥ 0.82 were obtained. For volume and TLU ICC values were > 0.66 for all mAbs. In addition, ICC values for all combinations of 2 observers were calculated (Supplemental Table 2).
Table 4 Reliability of
tumor uptake quantification for 89Zr-immuno-PET
   SUVmax
SUVpeak
SUVmean
Volume (mL)
TLU (%ID)
all eligible
all eligible
all eligible
all eligible
all eligible
89Zr-rituximab
D3
1.00; n=26 (1.00-1.00)
1.00; n=23 (1.00-1.00)
1.00; n=26 (1.00-1.00)
1.00; n=23 (1.00-1.00)
0.92; n=26 (0.84-0.96)
0.90; n=23 (0.80-0.96)
0.85; n=26 (0.74-0.92)
0.87; n=23 (0.77-0.94)
0.90; n=26 (0.83-0.95)
0.90; n=23 (0.82-0.96)
D6
1.00; n=32 (1.00-1.00)
1.00; n=28 (1.00-1.00)
1.00; n=32 (1.00-1.00)
1.00; n=28 (1.00-1.00)
0.95; n=32 (0.91-0.97)
0.94; n=28 (0.90-0.97)
0.12; n=32 (-0.07-0.36)
0.72; n=28 (0.55-0.85)
0.65; n=32 (0.47-0.79)
0.83; n=28 (0.72-0.91)
89Zr-cetuximab
D3
0.93; n=7 (0.77-0.99)
NA*
0.94; n=7 (0.82-0.99)
NA*
0.93; n=7 (0.79-0.99)
NA*
0.80; n=7 (0.46-0.96)
NA*
0.86; n=7 (0.61-0.97)
NA*
D6
0.72; n=10 (0.41-0.91)
1.00; n=8 (NA-NA)
0.75; n=10 (0.46-0.92)
1.00; n=8 (1.00-1.00)
0.79; n=10 (0.56-0.96)
0.92; n=8 (0.73-0.98)
0.83; n=10 (0.60-0.95)
0.83; n=8 (0.56-0.96)
0.89; n=10 (0.72-0.97)
0.89; n=8 (0.70-0.98)
89Zr- trastuzumab
D4
0.97; n=28 (0.95-0.99)
0.97; n= 23 (0.95-0.99)
0.83; n=28 (0.69-0.92)
0.82; n= 23 (0.64-0.91)
0.94; n=28 (0.89-0.97)
0.93; n= 23 (0.86-0.97)
0.67; n=28 (0.48-0.82)
0.66; n= 23 (0.45-0.82)
0.71; n=28 (0.54-0.84)
0.70; n= 23 (0.51-0.85)
  Data presented as ICC (95% confidence interval). * NA = ICC not available, 2 eligible VOI
DISCUSSION
Interobserver reproducibility for tumor uptake measures was investigated, as knowledge of measurement error is required for future clinical application of 89Zr- immuno-PET. Interobserver reproducibility was excellent for SUVmax and SUVpeak (variability of 0%) and very reasonable for SUVmean (variability of 7%), especially
134























   134   135   136   137   138