By linear regression analysis,
correlations between age and sex and ROI values and pathologic,
operative and imaging grades were all insignificant (p > 0.05).
Furthermore,
the difference between 1st and 2nd ROI measurements was insignificant (AM bundle; p = 0.236,
PL bundle; p = 0.824).
The mean ROI values of the AM and PL bundles of the ten normal volunteer on T2* mapping images were 17.9 (range,
16.8 – 20.2) and 17.9 (range,
16.2 – 20),
respectively.
The mean ROI values of AMB and PLB on T2* mapping images and corresponding pathologic,
operative and T2-weighted imaging grades are presented in table 1.
Histologic results showed that 58 % (31 of 53) of the AM bundles and 62 % (33 of 53) of the PL bundles exhibited moderate to severe degenerative changes (grades 3-4).
In operative fields,
the number of patients with grade 3 (moderate partial tear) and grade 4 (complete tear) was four and seven for AM bundles,
and five and five for PL bundles.
On oblique coronal T2-weighted images,
13 and 10 patients in AM bundle,
and 9 and 12 patients in PL bundle were interpreted as grade 3 and 4,
respectively.
According to table 1,
higher pathologic grades had higher mean ROI values for both AM and PL bundles and this result was statistically significant (AM bundle; Pearson’s coefficient = 0.557 (p = 0.00),
PL bundle; Pearson’s coefficient = 0.384 (p = 0.005)).
The correlation between ROI values and operative grades was prominently linear for AM bundles (Pearson’s coefficient = 0.513 (p = 0.000)),
but poor for PL bundles (Pearson’s coefficient = 0.254) and statistically insignificant (p = 0.093).
The correlation between ROI values and imaging grades on T2-weighted MR images was linear for AM bundles (Pearson’s coefficient = 0.362 (p = 0.008),
but negligible for PL bundles (Pearson’s coefficient = 0.040 (p = 0.774).
Table 1.
Mean ROI values of AM and PL bundles of anterior cruciate ligaments on T2* mapping images and corresponding pathologic,
operative and imaging grades.
Grade
|
AM bundle
|
PL bundle
|
P
|
O
|
T2
|
P
|
O
|
T2
|
0
|
*19.9
(n = 1)
|
25.1 (19.5– 36.5)
(n = 23)
|
25.6 (23.3– 29.9)
(n = 4)
|
21.2
(n = 1)
|
30.7
(22-43.1)
(n = 20)
|
31.2 (29.3– 33.7)
(n = 3)
|
1
|
23
(19.8– 27.4)
(n = 8)
|
28.6
(19.9– 36.7)
(n = 8)
|
25.8
(19.5– 38.9)
(n = 17)
|
27
(22–33.7)
(n = 8)
|
29.8 (18.2– 48.6)
(n = 12)
|
32.2 (22.9– 45.3)
(n = 17)
|
2
|
25.7 (19.5– 35)
(n = 13)
|
28.2
(19.5– 38.9)
(n = 11)
|
28.1
(21.4– 36.7)
(n = 9)
|
31.2 (23.5– 45.3)
(n = 11)
|
31.6
(22–38.5)
(n = 11)
|
28.7 (21.2– 38.1)
(n = 12)
|
3
|
29.4 (21.7– 40)
(n = 14)
|
31.8
(27.2– 39.7)
(n = 4)
|
27.4
(19.9– 36.5)
(n = 13)
|
31.5 (18.1– 43.1)
(n = 16)
|
32.7 (28.4– 37.4)
(n = 5)
|
31.8 (18.2– 48.6)
(n = 9)
|
4
|
30.5
(20 – 39)
(n =17)
|
32.2
(29.1– 35)
(n = 7)
|
31.8
(19.8– 39.7)
(n = 10)
|
33.9
(22– 48.6)
(n =17)
|
35.8 (28.2– 46.1)
(n = 5)
|
32.4 (25.8– 37.6)
(n = 12)
|
Note.
– AM = anteromedial,
PL = posterolateral,
P = pathologic,
O = operative,
T2 = oblique coronal T2-weighted image,
* ROI value
We also examined correlations between pathologic grades and imaging grades on oblique coronal T2-weighted images,
T2* mapping images and operative grades for AM and PL bundles.
Results are represented in table 2.
Table 2.
The correlation between pathologic and imaging grades for oblique coronal T2-weighted images,
T2* mapping images,
and operative grades.
(n = Pearson’s correlation coefficient)
|
AM bundle
|
PL bundle
|
Pathology & T2WI
|
0.482 (p = 0.000)
|
0.315 (p = 0.022)
|
Pathology & †T2* map
|
0.445 (p = 0.001)
|
0.247 (p = 0.074)
|
Pathology & operative findings
|
0.331 (p = 0.027)
|
0.311 (p =0.036)
|
Operative findings & T2WI
|
0.562 (P = 0.000)
|
0.634 (P = 0.000)
|
Operative findings & †T2* map
|
0.725 (p = 0.000)
|
0.514 (p = 0.000)
|
Note.- †; imaging grades
Except the correlation between pathologic grades and imaging grades on T2* mapping images in PL bundle,
imaging grades on MR images and visual grades on operative field had significant correlation with pathologic grades about each AM and PL bundle.
However,
all kappa values were < 0.3,
which indicated only fair to slight agreement (see Table 3).
Table 3.
Agreements between imaging and pathologic and macroscopic grades in the operative field.
|
AM bundle
|
PL bundle
|
Operative findings & T2WI
|
0.168 (p = 0.003)
|
0.094 (p = 0.128)
|
Operative findings & T2* map
|
0.3 (p = 0.000)
|
0.211 (p = 0.002)
|
Pathology & T2* map
|
-0.005 (p =0.940)
|
0.039 (p = 0.568)
|
Pathology & T2WI
|
0.082 (P = 0.335)
|
0.163 (P = 0.062)
|
Operative findings & pathology
|
0.071 (p = 0.201)
|
0.016 (p = 0.776)
|
Using operative results as standard references,
for complete tears,
the sensitivity,
specificity and accuracy of oblique T2-weighted MR imaging were 100,
94,
and 96 %,
respectively for AM bundles and 100,
90,
and 92 %,
respectively for PL bundles,
and for incomplete tears,
these were 71,
59,
and 63 %,
respectively for AM bundles and 80,
62,
and 67 %,
respectively for PL bundles.
The sensitivity,
specificity and accuracy of T2* mapping for complete tears were 86,
93,
and 92 %,
respectively for AM bundles and 80,
79,
and 79 %,
respectively for PL bundles and for incomplete tears,
corresponding values were 53,
74,
and 68 % for AM bundles and 56,
76,
and 70 % for PL bundles.