Development of a clinical diagnostic tool to differentiate multiple myeloma from bone metastasis in patients with destructive bone lesions (MM-BM DDx)

Table 2 Multivariable fractional polynomial logistic regression model for diagnostic prediction of multiple myeloma. (imputed dataset with a total n = 586)

Predictor	Covariate transformation			ß	95% CI	P-value
Predictor	Terms	df	Formula	ß	95% CI	P-value
Intercept				−2.28	−2.63, −1.93	< 0.001
Hemoglobin	Out	0	–	–	–	–
Log serum creatinine	Linear	1	Log creatinine-0.0237	1.28	0.80, 1.75	< 0.001
Log serum globulin	Linear	4	Log globulin^-0.5-0.8714	− 92.64	− 114.80, −70.49	< 0.001
	FP2		Log globulin^-0.5*^{Log (Log globulin)}-0.2400	−48.14	−60.13, −36.15	< 0.001
Log alkaline phosphatase	Linear	1	Log ALP-4.9318	−0.97	−1.38, −0.56	< 0.001
Serum calcium	Out	0	–	–	–	–

Abbreviations: df Degrees of freedom, CI Confidence interval, Log Natural logarithm function, FP2 Second-degree fractional polynomial, ALP Alkaline phosphates

ISSN: 2731-4553