Keywords:
Cancer, Staging, CT, Oncology
Authors:
L. Molenaar, M. A. van der Lubbe, F. van der Krieke, M. J. Lahaye, M. Maas, F. Lalezari, R. G. H. Beets-Tan, D. M. J. Lambregts; Amsterdam/NL
DOI:
10.1594/ecr2018/C-1537
Results
In total 89 lesions were analyzed.
The distribution of response categories according to RECIST 1.1 for the two FU moments determined by radiologists is given in Table 1. There were no patients with a complete response after two FU moments.
Table 1.
Response catergories according to RECIST 1.1 by radiologists |
|
Partial response |
Stable disease |
Progressive disease |
FU1 |
19% |
75% |
6% |
FU2 |
28% |
63% |
9%
|
Kappa values to compare response between radiologists and radiographers are provided in Table 2.
For the first FU-scan,
interobserver agreement was excellent for arm-2 (FU by radiographer,
baseline measurements by radiologist) and good for arm-3 (full reading by radiographer),
with kappa values of 0.88 and 0.80,
respectively.
For the second FU-scan,
IOA was somewhat lower,
but still good for both arms (kappa 0.65 and 0.67,
respectively)
Table 2.
Interobserver agreement between the different reading arms |
|
Arm-1 vs Arm2 |
Arm-1 vs Arm-3 |
FU1 |
κ0.88 |
κ0.80 |
FU2 |
κ0.65 |
κ0.67 |
N.B.
Arm 1 = baseline + FU measurements by radiologist
Arm 2 = baseline measurments by radiologists,
FU by radiographer
Arm 3= baseline + FU measurements by radiographer
|
15% of all target lesions,
which were identified by radiographers (arm 3) at baseline,
were not deemed suitable as target lesions when checked by an independent radiologist.
An additional 31% of lesions needed minor corrections,
which were mainly related to windowing or measuring the right diameter in case of spiculation.