Subjective score of chest x-ray difficulty in detecting lung nodule before and after AI data is provided in whole responders (N=63)
| Level | Resident year | Before AI-based reading | After AI-based reading | t | P-value | |||
|---|---|---|---|---|---|---|---|---|
| M |
SD | M | SD | |||||
| 1 |
Total | 4.10 | 0.82 | 4.79 | 0.48 | -6.349 | <0.001 | |
| 1st year | 4.04 | 0.95 | 4.83 | 0.48 | -3.800 | 0.001 | ||
| 2nd year | 4.12 | 0.70 | 4.76 | 0.44 | -5.416 | <0.001 | ||
| 3rd year | 4.14 | 0.77 | 4.77 | 0.53 | -3.130 | 0.005 | ||
| 2 | Total | 3.76 | 0.84 | 4.32 | 0.80 | -4.577 | <0.001 | |
| 1st year | 3.79 | 1.02 | 4.46 | 0.59 | -3.391 | 0.003 | ||
| 2nd year | 3.59 | 0.62 | 4.06 | 1.09 | -1.725 | 0.104 |
||
| 3rd year | 3.86 | 0.77 | 4.36 | 0.73 | -2.730 | 0.013 | ||
| 3 | Total | 3.35 | 1.18 | 4.11 | 1.00 | -5.035 | <0.001 | |
| 1st year | 3.38 | 1.41 | 4.25 | 0.99 | -3.715 | 0.001 | ||
| 2nd year | 3.24 | 1.30 | 4.00 | 0.87 | -2.626 | 0.018 | ||
| 3rd year | 3.41 | 0.80 | 4.05 | 1.13 | -2.309 | 0.031 | ||
| 4 | Total | 2.57 | 1.10 | 3.46 | 1.13 | -5.663 | <0.001 | |
| 1st year | 2.63 | 1.24 | 3.92 | 1.02 | -4.630 | <0.001 | ||
| 2nd year | 2.47 | 1.12 | 2.82 | 1.19 | -1.461 | 0.163 |
||
| 3rd year | 2.59 | 0.96 | 3.45 | 1.01 | -3.472 | 0.002 | ||
| 5 | Total | 2.98 | 1.20 | 3.54 | 1.06 | -3.798 | <0.001 | |
| 1st year | 3.29 | 1.23 | 3.67 | 1.13 | -1.519 | 0.142 |
||
| 2nd year | 2.71 | 1.10 | 3.12 | 1.17 | -1.514 | 0.150 |
||
| 3rd year | 2.86 | 1.21 | 3.73 | 0.83 | -3.600 | 0.002 | ||
M, mean; SD, standard deviation; JSRT, Japanese Society of Radiological Technology.
aLevel of chest x-ray is equal to the JSRT score of each chest x-ray. For example, JSRT score of level 1 chest x-ray is 1. bThe score ranges from 1 to 5, and higher score means low difficulty in detecting nodule on chest x-ray. cDenote statistical significance at P<0.05.