Bone marrow lesions can be subtyped into groups with different clinical outcomes using two magnetic resonance imaging (MRI) sequences

Table 3 Relationship between the presence of BML according to sequence at baseline and tibial cartilage volume¹

	No BML	BML in T2 only	BML in T1 and T2	p value⁴
Baseline cartilage volume (mm³)²
Medial compartment
	n = 273	n = 7	n = 16
Medial	1705 (32)	1552 (200)	1743 (133)	0.72
Medial, adjusted²	1698 (22)	1722 (139)	1809 (95)	0.52
Lateral compartment
	n = 278	n = 2	n = 16
Lateral	2014 (39)	2462 (456)	2176 (161)	0.33⁵
Lateral, adjusted²	2020 (28)	2356 (323)	2137 (117)	0.34⁵
Annual change in cartilage volume (mm³)
Medial compartment
	n = 248	n = 7	n = 15
Medial	22 (3)	−11 (19)	42 (13)	0.07
Medial, adjusted³	22 (3)	−13 (19)	45 (14)	0.04
Lateral compartment
	n = 254	n = 2	n = 14
Lateral	22 (3)	70 (36)	55 (13)	0.02⁵
Lateral, adjusted³	22 (3)	69 (36)	55 (14)	0.02⁵

BML bone marrow lesion
¹One participant had a BML present only on T1-weighted sequences, and was not included
²Estimated marginal means (standard error of the mean, SEM) used to compare groups adjusted for baseline age, gender, body mass index (BMI)
³Estimated marginal means (SEM) used to compare groups adjusted for baseline age, gender and BMI. Annual change is expressed such that a negative value represents a gain in cartilage volume
⁴ p value for F test, pairwise comparisons among the estimated marginal means across all groups, except where indicated
⁵For lateral compartment, the comparison was between those with no BML at baseline and those with BML visible using both T1- and T2-weighted images, as n = 2 for BML visible using T2-weighted images only

ISSN: 1478-6362