Machine learning-based Diagnostic model for determining the etiology of pleural effusion using Age, ADA and LDH

Respiratory Research

Table 1 Clinical characteristic of the cohorts

	Beijing Chao-Yang Cohort	Training Set	Validation Set	p
n	742	522	220
Diagnosis (%)				0.998
Malignant pleural effusion	397 (53.5)	278 (53.3)	119 (54.1)
Tuberculous pleural effusion	253 (34.1)	178 (34.1)	75 (34.1)
Parapneumonic pleural effusion	31 (4.2)	22 (4.2)	9 (4.1)
Transudative pleural effusion	33 (4.4)	24 (4.6)	9 (4.1)
Others	28 (3.8)	20 (3.8)	8 (3.6)
Age (median [IQR]), years	64 [51, 74]	64 [50, 73]	65 [52, 75]	0.368
Sex (%)				0.806
Female	280 (37.7)	195 (37.4)	85 (38.6)
Male	462 (62.3)	327 (62.6)	135 (61.4)
Pleural effusion
ADA (median [IQR]), U/L	15 [9, 36]	15.00 [9, 36]	14 [9, 36]	0.911
LDH (median [IQR]), U/L	302 [186, 488]	285 [184, 479]	330 [191, 515]	0.104
Total protein (median [IQR]), g/L	47.4 [41.5, 51.8]	47.2 [41.5, 51.9]	47.5 [41.5, 51.3]	0.931
Glucose (median [IQR]), mmol/L	5.94 [4.63, 7.29]	6.00 [4.63, 7.29]	5.74 [4.64, 7.32]	0.578
Chloride (median [IQR]), mmol/L	106.2 [103.4, 108.8]	106.2 [103.2, 109.0]	106.5 [103.7, 108.6]	0.988
Total cell counts (median [IQR]), cells/μl	6972 [3105, 22154]	6903 [3037,22406]	7087 [3313, 20708]	0.931
Mononuclear cell percentage (median [IQR]), %	91[78, 96]	90 [78, 96]	91 [78, 96]	0.839

Data are presented as the median [IQR] for Non-normally distributed continuous variables, number (%) for categorical variables. For non-normally distributed continuous variables, Mann–Whitney U test was used for comparison; for categorical variables, Chi-square test was used for comparison. ADA Adenosine Deaminase, LDH lactate dehydrogenase

ISSN: 1465-993X