Table 1.

Per-reader distribution of HRCT findings (n = 77). The “difference in prevalence” columns report the p values expressing whether the prevalence of the detected HRCT features was significantly different among the three readers and on a pairwise basis (R1 versus R2, R1 versus R3, R2 versus R3)

HRCT feature	Prevalence of detection n (%, 95%CI)			Difference in prevalence
	R1	R2	R3	Among the three readers (p)a	Pairwise (p)b
					R	1	2
TLS > 6	50 (65, 48–86)	52 (68, 50–89)	53 (69, 52–90)	0.678	2	0.774
					3	0.549	1.000
GGO > 2	57 (74, 56–96)	63 (82, 63–100)	64 (83, 64–100)	0.057	2	0.146
					3	0.092	1.000
Consolidation > 2	25 (32, 21–48)	16 (21, 12–34)	22 (29, 18–43)	0.065	2	0.012
					3	0.648	0.210
Crazy-paving pattern > 2	15 (19, 11–32)	13 (17, 9–29)	8 (10, 4–20)	0.142	2	0.791
					3	0.119	0.227
OP pattern	52 (68, 50–89)	43 (56, 40–75)	29 (38, 25–54)	<0.001	2	0.035
					3	<0.001	0.014

^aCochran’s Q test

^bMcNemar test; HRCT, high-resolution computed tomography; 95%CI, 95% confidence interval; R1, reader 1; R2, reader 2; R3, reader 3; R, reader; TLS, total lung score; GGO, ground-glass opacity; OP, organizing pneumonia

Numbers in bold refer to R1, R2, and R3. Numbers in italic refer to p values when statistically significant