Table 5 Comparison of ChatGPT, radiologist A + B and radiologist E’s diagnosis with the final pathological diagnosis
No. (Percentage) | |||
|---|---|---|---|
ChatGPT | Radiologist A + B | Radiologist D | |
The first diagnosis is consistent with the pathological diagnosis | 21(38.9%) | 26(48.1%) | 17(31.5%) |
One of diagnoses matches pathological diagnosis | 17(31.5%) | 20(37.0%) | 18(33.3%) |
None of the diagnosis is consistent with the pathological diagnosis | 16(29.6%) | 8(14.8%) | 19(35.2%) |