An Evaluation of the use of reformatted coronal oblique proton density fat suppressed fast spin echo images in ACL tears

1
An Evaluation of the use of reformatted coronal
oblique proton density fat suppressed fast spin
echo images in ACL tears

• Joseph Castillo B.Sc M.Sc

2
Introduction

• MR evaluation of the ACL is not done in isolation
  but is part of a routine knee protocol.
• MR diagnoses of the ACL are usually based on
  oblique sagittal and coronal images using T1, T2,
  GRE and STIR sequences.

3
Limitations

• The unique diagonal course and its size of 3.5cm
  pose limitations to standard imaging.
• Partial volume effects of the ACL with adjacent
  soft tissues is the commonest limitation
• Another limitation with sagittal view is the
  pulsating artifact from the popliteal artery.

4
Strategies

• Knee flexion.
• Oblique sagittal images prescribed parallel to
  lateral femoral condyle using an axial image.
• Oblique sagittal images prescribed parallel to
  ACL using a coronal image
• Use of 3D GRE.
• Supplementing with axial imaging.
• Use of secondary signs.

5
Thin slice direct coronal obliques

• Two studies used thin slice direct coronal
  obliques prescribed parallel to ACL using a
  sagittal image. (Katahira et al, 2001 and Hong et
  al 2003)
• Increased Diagnostic Accuracy
• Increased Diagnostic Efficacy
• Increased Scanning time by 4min.

6
Grading the Severity of ACL injury

• Grading of the ACL injury on MRI can be of help
  during the treatment decision making process.
  Not every one who has an ACL partial tear
  requires ACL injury requires surgical
  reconstruction. In this regard it would be
  beneficial to be able to determine whether the
  ACL injury is classified as low or high grade
  partial tear.

7
Purpose of the Study

• indirect coronal oblique image batch reformatted
  from an optimized coronal high resolution (2mm)
  slices along the course of the ACL would provide
  additional information on the status of the
  anterior cruciate ligament without prolonging
  scanning time.

8
Scanning Parameters

• TR 4200ms
• TE 20
• ETL 3 (Kojima et al, 1999)
• RBW 20.83
• Matrix 256 x 256
• FOV 24
• Slice 2mm gap 0
• Fat suppression

9
Resultant Image
10
Equipment

• 1.5T whole-body scanner (GE Medical Systems,
  Milwauke, Wisconsin) operating with Twin Speed
  gradients.
• A dedicated transmit-receive coil.
• Patient positioned with knee slightly flexed and
  with a slight external rotation.

11
Materials and Methods

• Patients could be of either gender.
• Had a history of trauma
• Had a request to assess ACL and/or menisci
  pathology by MRI
• had never undergone ACL repair
• This criteria yielded 87 patients.
• A review of MRI reports disclosed 10 partial
  tears.

12
Results

• Images were reviewed by 2 radiologists
  independently
• Initially, each radiologist evaluated the status
  of ACL by routine knee protocol (A)
• A Coronal PD, Sagittal PD, Coronal STIR,
  Sagittal T2

13
Results

• Then later (after 10days) in combination with
  reformatted coronal oblique images (Protocol B)
• And later on (10days), using the reformatted
  coronal oblique images on their own (Protocol C)

14
Diagnostic Accuracy

• Interreader agreement for the presence or absence
  of ACL Tear was measured using Confidence levels
  (5 point scale)
• 0 intact, 1 probably intact, 2 possible
  tear, 3 probably tear, 4 definite tear

15
MR Grading Criteria Grade 0

• Intact ligament. A and B, sagittal PD-fs (TR/TE
  4200/20) show normal ACL and sagittal T2
  (TR/TE/FA 600/15/20) show almost normal ACL but
  with some high signal at the tibial attachment.
  C. RCO shows normal contour with distinct fibers
  at the tibial attachment

16
Grade 1 low grade partial tear

• A, sagittal PD-fs and B, sagittal T2 shows more
  than half of the ligament substance disrupted at
  the midportion and tibial attachment of the ACL
  which suggests a high grade tear. C, RCO shows
  oedematous ACL with diffusely increased signal
  intensity, but ligament continuity at both
  femoral and tibial sites are preserved. Low
  grade partial tear is more likely.

17
Grade 2 High Grade partial tear

• A, Sag PD-fs and B Sag T2 show ACL disruption at
  the Femoral attachment. C, RCO shows marked
  thinning (more than half of the ACL disrupted)
  but with preserved continuity. High grade
  partial tear was suggested by both readers.

18
Grade 3 Complete tear

• A PD-fs, B T2 GRE and C (RCO) show a complete
  tear of the ACL.

19
Results

• Interreader agreement was calculated using
  weighted ? statistics.
• 0.0 lt ? lt 0.2 Poor correlation
• 0.2 lt ? lt 0.4 Fair correlation
• 0.4 lt ? lt 0.6 Moderate correlation
• 0.6 lt ? lt 0.8 Good correlation
• 0.8 lt ? lt 1.0 Very Good correlation

20
Confidence Protocol A

• Interreader agreement for the presence or absence
  of ACL tear

21
Confidence Protocol B

• Interreader agreement for the presence or absence
  of ACL tear

22
Confidence Protocol C

• Interreader agreement for the presence or absence
  of ACL tear

23
Confidence all protocols

• Protocol A Moderate correlation
• Protocol B Good correlation
• Protocol C Poor correlation

24
Grading Protocol A

• Interreader agreement for grading the ACL without
  the reformatted coronal oblique images.

25
Grading Protocol B

• Interreader for grading ACL using protocol A and
  the indirect coronal oblique

26
Grading Protocol C

• Interreader agreement using the reformatted
  coronal obliques only

27
Limitations of the study

• Indirect signs were not removed
• Due to small number of cases this study was
  exploratory.
• Radiologists viewed printed images rather than
  using the advantage windows system.
• A comparison of sensitivity and specificity using
  arthroscopic findings could have added value to
  the study.

28
Conclusion

• I found that the use of additional reformatted
  coronal oblique images improved the accuracy of
  ACL tear detection.
• Reformatted oblique coronal images also improved
  the accuracy of grading ACL injury and can help
  in the decision making process for treatment.

29
Thank You