Secondary Logo

Journal Logo

Musculoskeletal Imaging

Diagnostic Value of Advanced Metal Artifact Reduction Magnetic Resonance Imaging for Periprosthetic Joint Infection

Inaoka, Tsutomu MD, PhD; Kitamura, Noriko MD, PhD; Sugeta, Masayuki MD; Nakatsuka, Tomoya MD, PhD; Ishikawa, Rumiko MD, PhD; Kasuya, Shusuke MD; Sugiura, Yoshiya MD; Nakajima, Arata MD, PhD; Nakagawa, Koichi MD, PhD; Terada, Hitoshi MD, PhD

Author Information
Journal of Computer Assisted Tomography: 5/6 2022 - Volume 46 - Issue 3 - p 455-463
doi: 10.1097/RCT.0000000000001297
  • Open

Abstract

The implantation of joint prostheses is becoming increasingly common worldwide.1–3 Although there have been marked improvements in the component design and surgical techniques used for the implantation of joint prostheses, patients still suffer from a variety of postoperative complications. Periprosthetic infection is a serious postsurgical complication in the acute and chronic phases of arthroplasty.2 In general, the diagnosis of periprosthetic joint infection (PJI) is made based on a combination of clinical findings, laboratory evaluations of blood and synovial fluid, and intraoperative findings.2,3 Although delay in the diagnosis is a significant factor in morbidity and mortality, the diagnosis can be challenging. Imaging has not been part of the diagnostic criteria of PJI3; however, there has been a demand for imaging modalities that can play a primary role in the efficiency of the work flow because accurate evaluation of the periprosthetic tissue can potentially distinguish periprosthetic infection from other causes of pain after joint replacement.

Plain radiography is the first-line imaging tool for the evaluation of metal implants in the joints.2 Plain radiography provides very limited value in soft-tissue and capsular information. Thus, radiographic findings of infection are often normal. Multidetector computed tomography (CT) and magnetic resonance imaging (MRI) are better in the visualization of the soft tissues and contrast of bone and soft tissues; however, CT and MRI are still limited because of metal susceptibility artifacts around the joints with a metal implant. Thus, CT and MRI are not recommended as a diagnostic test for PJI on clinical practice guideline of the American Academy of Orthopaedic Society.3 Despite the improvement in image quality attained with various modifications of imaging parameters of conventional pulse sequences for the reduction of susceptibility artifact in MRI, the susceptibility artifacts persist and limit the assessment of periprosthetic bone and soft tissues.4,5 Thus, advanced metal artifact reduction (MAR) sequences such as slice encoding for metal artifact correction (SEMAC) and multiacquisition variable–resonance image combination (MAVRIC) have been developed and examined in efforts to improve the image quality by reducing susceptibility artifacts in the presence of metal orthopedic implants.4,5 Furthermore, a hybrid technique merging the advantages of SEMAC and MAVRIC, that is, MAVRIC—selective (MAVRIC-SL), is developed and has been used in clinical practice.6–9 To the best of our knowledge, there have been few evaluations of PJI using the advanced MAR sequences such as SEMAC and MAVRIC.10,11 We hypothesized that the use of the advanced MAR sequences such as MAVRIC-SL would be valuable for the detection of abnormal findings and the diagnosis of PJI. Therefore, we conducted the present study to evaluate the diagnostic value of MRI with MAVRIC-SL for PJI.

MATERIALS AND METHODS

This study was conducted with the approval of the ethics committee of Toho University Sakura Medical Center (S19068). Because this study was a retrospective design and anonymous clinical data were used for research purposes, informed consent was obtained using an opt-out.

Sample Selection

Consecutive patients with joint orthopedic prostheses who underwent MRI including MAVRIC-SL sequences at our institution between April 2014 and August 2020 were included in this study. All patients were older than 18 years. We used the picture archiving and communication system to retrieve the imaging protocols and images and the patients' electronic medical records to retrieve the clinical data and clinical course.

Image Acquisition

All MRI examinations were performed with a 1.5-T system (Optima MR450W; GE Healthcare, Waukesha, WI). Hips were imaged with a 48-channel body surface coil, knees were imaged with an 8-channel knee coil, and elbows were imaged with a 16-channel surface coil. Before MRI examinations, it was explained to the patients that unusual heating sensation or any other sensations around metal joint prostheses might be felt, and if so, MRI examinations would be stopped immediately.

Two-dimensional fast spin echo (2D-FSE) and MAVRIC-SL sequences were obtained for the joints with metal implants. Initially, the 2D-FSE sequence was obtained in the transverse plane. In our experience, the transverse plane perpendicular to the long axis of a metal implant was the strongest against susceptibility artifacts due to metal implants (eg, image distortion, signal loss, and signal pileup around metal implants) and could best visualize the anatomical structures among the 3 orthogonal planes. The transverse plane images were obtained as covering the entire metal implant in the joint.

From the viewpoint of simplifying MRI for the joints with metal implants in routine practice, when severe susceptibility artifacts were observed in 2D-FSE, we decided to skip the other 2D-FSE sequences and obtain MAVRIC-SL sequences sequentially. The 2D-FSE sequences were obtained primarily with a T2-weighted imaging, followed by fluid-sensitive imaging including short tau inversion recovery (STIR) or a fat-suppressed T2-weighted imaging, and, lastly, a T1-weighted imaging. The MAVRIC-SL sequences were obtained with STIR and proton density (PD)–weighted imaging principally in the sagittal and/or coronal plane, because the transverse plane required much longer acquisition times to cover entire metal implant compared with the sagittal or coronal plane. Detailed imaging parameters are listed in Table 1. The MRI examinations were performed with different fields of view (FOVs), slice thicknesses, and contrast settings depending on the specific clinical questions asked in different joints and body sizes.

TABLE 1 - MRI Protocol for the Joints After Metal Orthopedic Implantation
Parameters 2D-FSE MAVRIC-SL
T2 STIR FS T2 T1 STIR PD
TR, ms 2300–6700 3000–9400 5400–7400 410–670 4500 3700
TE, ms 30–100 40–63 85–94 6–13 6.22–7.98 6.26–7.92
Echo train length 12 8–13 8–13 2 20–23 20–23
No. excitations 0.5–2 0.5–2 2 0.5–3 0.5 0.5
Slice thickness (range), mm 5 (2–8) 5 (3–8) 4.6 (4–6) 5 (2–8) 3–5 3–5
Spacing, % 10–25 10–25 10–25 10–25
FOV, mm2 Various Various Various Various 180 × 420–192 × 449 180 × 420–192 × 449
Bandwidth, Hz/pixel 434.0–578.7 434.0–578.7 434.0–578.7 434.0–578.7 781.2 781.2
Inversion time, ms 150 160
Acquisition time, min 3–5 3–5 3–5 3–5 4–6 5–6
T2 indicates T2-weighted image; FS, fat-suppressed; T1, T1-weighted image; PD, proton density–weighted image.

Image Analysis

All MRI data sets were assessed independently by 2 board-certificated radiologists (T.I., with 22 years' experience in musculoskeletal radiology, and M.S., with 12 years' experience in general radiology). All the MRI data sets were anonymized and randomized. One reader was completely blinded to the clinical data and final diagnosis, and the other might interpret the images clinically more than 1 month before the readout but did not know the final diagnosis. The final decisions were made by consensus.

Abnormal Findings and Overall Image Impression on MRI with MAVRIC-SL

We determined whether the abnormal findings of joint effusion, capsular thickening, pericapsular edema, soft-tissue edema, soft-tissue fluid collection, bone marrow edema pattern around the implant (BME pattern), and lymphadenopathy (LN) were present on MRI with MAVRIC-SL. Other abnormal findings were also assessed. Joint effusion was defined as intraarticular fluid collection. Capsular thickening was defined as thickened joint capsule. Pericapsular edema was defined as edema in and around the capsule. Lamellar appearance of the capsule could not always be evaluated in the hip and elbow joints. Soft-tissue edema was defined as reticular-patterned fluid signal intensity in periosteal soft tissue, in muscle, between muscle, and/or in subcutaneous soft tissue. Soft-tissue fluid collection was defined as localized fluid collection in periarticular soft tissue, in muscle, between muscle, and/or in subcutaneous soft tissue. Regarding the soft-tissue fluid collection, presence of communication to the joint and surrounding capsular-like structure was also investigated. Capsular-like structure was defined as low signal intensity band around more than half of the fluid collection. BME pattern was defined as high signal intensity similar to fluid around the metal implants within bone. Lymphadenopathy was defined as swelling of locoregional lymph nodes within the FOV. Fatty infiltration and atrophy of the muscle were not included as abnormal findings. Synovitis was not evaluated because it was difficult to be differentiated from debris in the joints without contrast-enhanced images. Finally, overall image impression for the presence or absence of PJI by the 2 readers was recorded using 6-grading system (1, definitive absence; 2, probable absence; 3, possible absence; 4, possible presence; 5, probable presence; 6, definitive presence).

Clinical Assessment

Based on the clinical history and symptoms (ie, swelling, pain, feverish, reddish, and discharge around the joint, duration of implantation), laboratory values (ie, the serum C-reactive protein level, erythrocyte sedimentation rate, and white blood cell count), available findings from additional diagnostic tests that were performed (ie, joint aspiration, surgery, and pathologic analysis), treatment, and clinical course, whether a PJI was present or not was determined. We referred to several clinical guidelines of the diagnosis of PJI.12,13 Other articular disorders including ligament tear, intraarticular free bodies, and osteoarthritis were diagnosed based on the MRI findings and patient's clinical symptoms such as joint instability and explainable pain. The final clinical diagnosis was recorded for each patient. Whether any adverse events from MRI were present was also assessed.

Statistical Analysis

The sensitivity, specificity, positive predictive values (PPVs), negative predictive values (NPVs), and odds ratios (ORs) for PJI were calculated for the abnormal findings. These values were also calculated for the combinations of abnormal findings. For demographic data, the t test, χ2 test, and Fisher exact test, and Mann-Whitney U test were used. Receiver operating characteristic (ROC) analysis was used for overall image impression for the presence or absence of PJI by the 2 readers. Probability (P) values <0.05 were considered significant. Interobserver agreements for the abnormal findings were calculated using κ statistics: a κ value has a maximum of 1.0 when agreement is perfect, whereas a κ value of 0 indicates agreement that is no better than chance agreement. The κ values of this study were interpreted according to the guideline of Landis and Koch.14 The strength of agreement quantified by a κ statistic was graded as follows: <0, poor; 0.01 to 0.20, slight; 0.21 to 0.40, fair; 0.41 to 0.60, moderate; 0.61 to 0.80, substantial; and 0.81 to 0.99, almost perfect.

RESULTS

A total of 48 joints with metal prostheses in 41 patients underwent MRI including MAVRIC-SL. We excluded 5 joints in 5 patients from the analyses: 1 involved femoral neck pinning, 1 had an intramedullary tibia fixation, 1 had small metal clips around the knee joint, 1 had acute ligament rupture after arthroplasty, and the other patient underwent MRI within 1 week after implantation of joint prosthesis. Finally, a total of 43 joints in 36 patients who underwent the MRI were included. The study population was 30 women and 6 men. Four patients underwent MRI for the same joints twice, and 7 patients underwent MRI for bilateral hip joints. At the time of the MRI examinations, the patients' mean ± SD age was 75.4 ± 8.8 years (range, 55–92 years). Seven patients had rheumatoid arthritis. Sixteen joints underwent a total hip arthroplasty; 11, bipolar hip arthroplasty; 12, total knee arthroplasty; 2, total elbow arthroplasty; 1, unilateral knee arthroplasty; and 1, total hip fusion. Because patients after arthroplasty referred from outside institutions were included, detailed information about the type of metal implants and metal composition was not always available. Fourteen joints had cobalt-chrome alloys and titanium alloys; 12 had titanium alloys; 8 had cobalt-chrome alloys; 2 had stainless steels; and 7 were unidentified. The median (interquartile range) of interval between the implantations and MRI examinations was 695 days (116–3394 days). The indications for MRI included a suspicion of PJI after arthroplasty in 23 patients, persistent unexplained pain in 10, and pain on motion in 3. A suspicion of PJI was mainly based on fever, local symptoms (pain, swelling, redness, pus leakage), and laboratory tests (C-reactive protein level, erythrocyte sedimentation rate, white blood cell count). Seven contralateral hip joints after arthroplasty within the FOV of hip MRI were included. Detailed demographic data of the patients included in this study are shown in Table 2.

TABLE 2 - Patient Characteristics Included in This Study
Characteristics All PJI Non-PJI P
No. patients 36 18* 22*
Age, mean ± SD, y 74.5 ± 8.8 76.7 ± 7.4 73.9 ± 9.2 0.28
Sex 0.69
 Male 6 2* 3*
 Female 30 16* 22*
Rheumatoid arthritis 7 4 3 0.77
No. joints 43 18 25
Location of arthroplasty 0.15
 Hip joint 29 10 19
 Knee joint 12 6 6
 Elbow joint 2 2 0
Type of arthroplasty 0.19
 Total hip arthroplasty 16 4 12
 Bipolar hip arthroplasty 11 5 6
 Total hip fusion 1 0 1
 Total knee arthroplasty 12 7 5
 Unilateral knee arthroplasty 1 0 1
 Total elbow arthroplasty 2 2 0
Interval between arthroplasty and MRI, d
 Median (interquartile range) 695 (116–3394) 238 (30–1896) 1796 (609–5667) <0.05
*Number of patients: 4 patients who underwent MRI for the same joints twice and 7 patients who underwent MRI for bilateral hip joints were included.

All the 36 MRI examinations were safely performed without heating sensation or any other unusual sensations around metal joint implants. No adverse events from the MRI examinations were found. In 44% (16 of 36), only the initial sequence was obtained in 2D-FSE because of susceptibility artifacts, but multiple MAVRIC-SLs were obtained instead of the other 2D-FSE. As a result of the abbreviation of 2D-FSE, no clinical problems were found. Multiacquisition variable–resonance image combination selective including STIR and PD-weighted imaging in 1 or 2 planes were obtained in 86% (31 of 36). In 14% (5 of 36), 1 MAVRIC-SL was obtained because of a difficulty of postural control during the examinations.

Abnormal Findings and Overall Image Impression on MRI with MAVRIC-SL

The results of the detection of abnormal findings and overall image impression on MRI with MAVRIC-SL by the 2 readers are expressed in Table 3.

TABLE 3 - The Abnormal Findings Around Metal Joint Prosthesis and Overall Image Impression by 2 Readers
Reader 1 Reader 2 κ value
Abnormal findings
 Joint effusion 33 38 0.60
 Capsular thickening 23 18 0.77
 Pericapsular edema 17 12 0.74
 Soft-tissue edema 24 29 0.76
 Soft-tissue fluid collection 15 18 0.66
 With communication to the joint 12 12 0.77
 With capsule-like structure 4 6 0.71
 BME pattern 22 16 0.63
 LN 20 20 0.63
 Other abnormal findings 2 4
Overall image impression of PJI
 Presence of PJI 19 20 0.97*
 AUROC (95% CI) 0.949 (0.893–1.005) 0.926 (0.860–0.991) P = 0.534
Other abnormal findings: ligament disruption, free bodies, and patellofemoral osteoarthritis. κ value is the interobserver agreement between the 2 readers for abnormal findings.
*Weighted κ value.
AUROC, area under the receiver operating characteristic curve.

Regarding the clinical assessment, 18 (42%) of the 43 joints were diagnosed as having a PJI. Of the 18 joints with PJI, 7 (39%) showed a positive culture of joint effusion; coagulase-negative Staphylococci bacteria were detected in three joints, Gram-negative bacteria were detected in 2 joints, methicillin-sensitive Staphylococcus aureus was detected in 1 joint, and methicillin-resistance S. aureus was detected in 1 joint. The diagnosis of pyogenic arthritis was established based on the pathological specimen in 1 joint. The remaining 10 joints were diagnosed based on the patients' clinical symptoms, laboratory values, treatment, and clinical course.

Of the 25 joints diagnosed as having no PJI, 2 had ligament disruption, 1 had greater trochanteric bursitis, 1 had intra-articular free bodies, 1 had severe patellofemoral osteoarthritis, 1 had muscle denervation edema, and the remaining 19 joints had no definite abnormality on MRI. The diagnoses of ligament disruption, intraarticular free bodies, and patellofemoral osteoarthritis were made based on clinical symptoms and MRI findings. The diagnosis of greater trochanteric bursitis was established based on the surgical and pathological findings and negative culture and characteristics of the fluid collection.

Overall, of the 43 joints, joint effusion, capsular thickening, pericapsular edema, soft-tissue edema, soft-tissue fluid collection, BME pattern, and LN were found in 33 (77%), 23 (53%), 17 (40%), 24 (56%), 16 (37%), 19 (44%), and 16 (37%), respectively. Regarding the soft-tissue fluid collection, the presence of communication to the joint and capsular-like structure was 12 and 4, respectively. Of the 18 joints with PJI, joint effusion, capsular thickening, pericapsular edema, soft-tissue edema, soft-tissue fluid collection, BME pattern, and LN were found in 18 (100%), 16 (89%), 13 (72%), 16 (89%), 13 (72%), 9 (50%), and 7 (39%), respectively. On the other hand, of the 25 joints without PJI, these findings were found in 15 (60%), 7 (28%), 4 (16%), 8 (32%), 3 (12%), 10 (40%), and 9 (36%), respectively. Table 4 provides the sensitivity, specificity, PPV, NPV, and OR for each of the abnormal findings suggesting PJI. Regarding the abnormal findings to suggest PJI, there were differences for capsular thickening (PPV, 70%; NPV, 90%; OR, 20.6; P < 0.001), soft-tissue fluid collection (PPV, 81%; NPV, 81%; OR, 19.1; P < 0.001), soft-tissue edema (PPV, 67%; NPV, 89%; OR, 17; P < 0.001), pericapsular edema (PPV, 76%; NPV, 81%; OR, 13.7; P < 0.001), and joint effusion (PPV, 55%; NPV, 100%; OR, 12; P < 0.05), but not for BME pattern (P = 0.734) and LN (P = 0.963) (Figs. 1–4). In addition, there were differences for soft-tissue fluid collection with no capsular-like structure (PPV, 83%; NPV, 74%; OR, 14.4; P < 0.01) and with communication to the joint (PPV, 75%; NPV, 71%; OR, 7.3; P < 0.05). In particular, the combinations of joint effusion and capsular thickening (PPV, 74%; NPV, 95%; OR, 53.8; P < 0.001), capsular thickening and soft-tissue fluid collection (PPV, 92%; NPV, 77%; OR, 37.7; P < 0.001), joint effusion and soft-tissue fluid collection (PPV, 87%; NPV, 82%; OR, 29.9; P < 0.001), soft-tissue fluid collection and soft-tissue edema (PPV, 87%; NPV, 82%; OR, 29.9; P < 0.001), and capsular thickening and soft-tissue edema (PPV, 82%; NPV, 85%; OR, 25.6; P < 0.001) highly suggested PJI. Table 5 provides the sensitivity, specificity, PPV, NPV, and OR for the combinations of abnormal findings suggesting PJI.

TABLE 4 - Abnormal Findings to Suggest PJI
Sensitivity Specificity PPV NPV OR (95% CI) P
Joint effusion 100% (18/18) 40% (10/25) 55% (18/33) 100% (10/10) 12 (1.374–104.8) <0.05
Capsular thickening 89% (16/18) 72% (18/25) 70% (16/23) 90% (18/20) 20.6 (3.721–113.7) <0.001
Pericapsular edema 72% (13/18) 84% (21/25) 76% (13/17) 81% (21/26) 13.7 (3.090–60.30) <0.001
Soft-tissue edema 89% (16/18) 68% (17/25) 67% (16/24) 89% (17/19) 17 (3.127–92.43) <0.001
Soft-tissue fluid collection 72% (13/18) 88% (22/25) 81% (13/16) 81% (22/27) 19.1 (3.899–93.23) <0.001
 With communication to the joint 50% (9/18) 88% (22/25) 75% (9/12) 71% (22/31) 7.3 (1.605–33.51) <0.05
 With no communication to the joint 22% (4/18) 100% (25/25) 100% (4/4) 64% (25/39) 7.1 (0.726–70.32) 0.160
 With capsular-like structure 17% (3/18) 96% (24/25) 75% (3/4) 62% (24/39) 4.8 (0.456–50.50) 0.380
 With no capsular-like structure 56% (10/18) 92% (23/25) 83% (10/12) 74% (23/31) 14.4 (2.579–80.13) <0.01
BME pattern 50% (9/18) 60% (15/25) 47% (9/19) 63% (15/24) 1.5 (0.442–5.092) 0.734
LN 39% (7/18) 64% (16/25) 44% (7/16) 59% (16/27) 1.3 (0.357–4.542) 0.963

F1
FIGURE 1:
A 77-year-old woman with rheumatoid arthritis after right total elbow arthroplasty 12 years ago. She suddenly presented with pain, swelling, and fever in her right elbow joint. Methicillin-sensitive S. aureus was detected. A, Sagittal conventional 2D-FSE STIR image of the right elbow joint. Image distortion and signal loss around the metal implant can be seen. B, Sagittal MAVRIC-SL PD-weighted image and (C) sagittal MAVRIC-SL STIR image of the right elbow joint. Image distortion and signal loss around the metal implant are significantly reduced. Bone marrow edema pattern in the humerus can be seen (thick arrow). Soft-tissue fluid collections with no capsular-like structure can clearly be seen (arrow heads). Severe soft-tissue edema in periarticular, periosteal, and subcutaneous areas can be seen (thin arrows). Both readers recorded definitive presence of PJI.
F2
FIGURE 2:
A 68-year-old woman 37 days after right total knee arthroplasty. She had a mild pain in the right knee joint and mildly elevated inflammatory reactions. From joint effusion, Gram-positive bacteria were detected. A, Sagittal conventional 2D-FSE STIR image of the right knee joint. Severe image distortion and signal loss can be seen predominantly at the joint area. Although subcutaneous edema can be observed, the information at the joint area is quite limited. B, Sagittal MAVRIC-SL PD-weighted image and (C) sagittal MAVRIC-SL STIR image of the right knee joint. Image distortion and signal loss around the metal implants are clearly diminished. A small amount of joint effusion with capsular thickening of the suprapatellar bursa can be seen (arrow heads). Short tau inversion recovery image shows that pericapsular edema and extensive soft-tissue edema can be seen (thin arrows). The patella and femur are well depicted and shows bone marrow edema patterns (thick arrows). Both readers recorded definitive presence of PJI.
F3
FIGURE 3:
An 81-year-old woman 695 days after bipolar hip arthroplasty. She felt a mild but persistent pain at the right hip joint. Mild elevation of inflammatory reactions was found. Gram-positive bacteria were detected. A, Coronal MAVRIC-SL PD-weighted image of the right hip joint. Image distortion and signal loss around the metal implants are minimized. B, Coronal MAVRIC-SL STIR of the right hip joint. Bone marrow edema patterns around the metal implants can be seen (thick arrows). Associated pericapsular edema (arrow head) and soft-tissue edema particularly in periosteal areas (thin arrows) can be seen. Joint effusion is almost absent. One reader recorded possible presence of PJI, and the other recorded probable presence of PJI.
F4
FIGURE 4:
A 71-year-old man 8 months after left bipolar hip arthroplasty. He had sustained pain in the left hip joint and mildly elevated inflammatory reactions, which gradually resolved. Therefore, this case was clinically diagnosed as non-PJI. A, Coronal MAVRIC-SL PD-weighted image of the left hip joint. Image distortion and signal loss around the metal implants are minimized. B, Coronal and (C) sagittal MAVRIC-SL STIR images of the left hip joint. Bone marrow edema patterns around the metal stem within the left femur can be seen (arrows). However, there are no other associated abnormal findings such as soft-tissue edema. One reader recorded possible absence of PJI, and the other recorded probable absence of periprosthetic joint infection.
TABLE 5 - Combinations of Abnormal Findings to Suggest PJI
Sensitivity Specificity PPV NPV OR (95% CI) P
Joint effusion and capsular thickening 94% (17/18) 76% (19/25) 74% (17/23) 95% (19/20) 53.8 (5.871–493.6) <0.001
Joint effusion and pericapsular edema 72% (13/18) 84% (21/25) 76% (13/17) 81% (21/26) 13.7 (3.090–60.30) <0.001
Joint effusion and soft-tissue edema 89% (16/18) 72% (18/25) 70% (16/23) 90% (18/20) 20.6 (3.721–113.7) <0.001
Joint effusion and soft-tissue fluid collection 72% (13/18) 92% (23/25) 87% (13/15) 82% (23/28) 29.9 (5.066–176.5) <0.001
Joint effusion and BME pattern 50% (9/18) 80% (20/25) 64% (9/14) 69% (20/29) 4 (1.040–15.38) 0.082
Joint effusion and LN 39% (7/18) 76% (19/25) 54% (7/13) 63% (19/30) 2 (0.539–7.538) 0.476
Capsular thickening and soft-tissue edema 78% (14/18) 88% (22/25) 82% (14/17) 85% (22/26) 25.6 (4.978–132.3) <0.001
Capsular thickening and soft-tissue fluid collection 61% (11/18) 96% (24/25) 92% (11/12) 77% (24/31) 37.7 (4.123–345.0) <0.001
Capsular thickening and pericapsular edema 72% (13/18) 84% (21/25) 76% (13/17) 81% (21/26) 13.7 (3.090–60.30) <0.001
Capsular thickening and BME pattern 44% (8/18) 80% (20/25) 62% (8/13) 67% (20/30) 3.2 (0.829–12.35) 0.166
Capsular thickening and LN 39% (7/18) 84% (21/25) 64% (7/11) 66% (21/32) 3.3 (0.801–13.94) 0.179
Pericapsular edema and soft-tissue edema 61% (11/18) 88% (22/25) 79% (11/14) 76% (22/29) 11.5 (2.485–53.43) <0.01
Pericapsular edema and soft-tissue fluid collection 50% (9/18) 96% (24/25) 90% (9/10) 73% (24/33) 24 (2.650–217.4) <0.01
Pericapsular edema and BME pattern 39% (7/18) 88% (22/25) 70% (7/10) 67% (22/33) 4.7 (1.006–21.64) 0.090
Pericapsular edema and LN 33% (6/18) 88% (22/25) 67% (6/9) 65% (22/34) 4 (0.838–19.10) 0.155
Soft-tissue edema and soft-tissue fluid collection 72% (13/18) 92% (23/25) 87% (13/15) 82% (23/28) 29.9 (5.066–176.5) <0.001
Soft-tissue edema and BME pattern 44% (8/18) 88% (22/25) 73% (8/11) 69% (22/32) 5.9 (1.279–26.90) <0.05
Soft-tissue edema and LN 39% (7/18) 92% (23/25) 78% (7/9) 68% (23/34) 7.3 (1.300–41.20) <0.05
Soft-tissue fluid collection and BME pattern 33% (6/18) 92% (23/25) 75% (6/8) 66% (23/35) 5.8 (1.003–32.95) 0.088
Soft-tissue fluid collection and LN 33% (6/18) 96% (24/25) 86% (6/7) 67% (24/36) 12 (1.293–111.3) <0.05
BME pattern and LN 22% (4/18) 84% (21/25) 50% (4/8) 60% (21/35) 1.5 (0.321–7.012) 0.904

Regarding the overall image impression for PJI of the 2 readers, areas under the ROC curve (95% confidence interval [CI]) were 0.949 (0.893–1.005) and 0.926 (0.860–0.991) with no significant difference between the 2 readers (P = 0.534). Interobserver agreement (weighted κ value) for presence or absence (6-grading system) of PJI was 0.97.

DISCUSSION

The advanced MAR sequences such as MAVRIC-SL can allow detailed assessments of periprosthetic process after arthroplasty.6–9 To our knowledge, however, there have been few evaluations of PJI using the advanced MAR sequences such as SEMAC and MAVRIC.10,11 The expectation of MRI with the advanced MAR sequences for the evaluation of PJI has been increasing.15,16 This study examined the diagnostic value of MRI with MAVRIC-SL for PJI at 1.5 T. Overall, of the 43 joints, joint effusion was detected in 33 (77%), capsular thickening in 23 (53%), pericapsular edema in 17 (40%), soft-tissue edema in 24 (56%), soft-tissue fluid collection in 16 (37%), BME pattern in 19 (44%), and LN in 16 (37%). Of the 18 joints with PJI, joint effusion was detected in 18 (100%), capsular thickening in 16 (89%), soft-tissue edema in 16 (89%), pericapsular edema in 13 (72%), soft-tissue fluid collection in 13 (72%), BME pattern in 9 (50%), and LN in 7 (39%). Regarding the interobserver agreements of the abnormal findings between the 2 readers, κ values were 0.60 to 0.77. Capsular thickening (PPV, 70%, NPV, 90%; OR, 20.6), soft-tissue fluid collection (PPV, 81%; NPV, 81%; OR, 19.1), soft-tissue edema (PPV, 67%; NPV, 89%; OR, 17), pericapsular edema (PPV, 76%; NPV, 81%; OR, 13.7), and joint effusion (PPV, 55%; NPV, 100%; OR, 12) suggested PJI. Soft-tissue fluid collection without capsular-like structure (PPV, 83%; NPV, 74%; OR, 14.4) or with communication to the joint (PPV, 75%; NPV, 71%; OR, 7.3) suggested PJI. In particular, the combinations of joint effusion and capsular thickening (PPV, 74%; NPV, 95%; OR, 53.8; P < 0.001), capsular thickening and soft-tissue fluid collection (PPV, 92%; NPV, 77%; OR, 37.7; P < 0.001), joint effusion and soft-tissue fluid collection (PPV, 87%; NPV, 82%; OR, 29.9; P < 0.001), soft-tissue fluid collection and soft-tissue edema (PPV, 87%; NPV, 82%; OR, 29.9; P < 0.001), and capsular thickening and soft-tissue edema (PPV, 82%; NPV, 85%; OR, 25.6; P < 0.001) highly suggested PJI. Regarding the BME pattern, the combination with soft-tissue edema raised the possibility of PJI (PPV, 73%; NPV, 69%; OR, 5.9; P < 0.05). Regarding the overall image impression of the 2 readers, weighted κ value was 0.97 and areas under the ROC curve were 0.949 (95% CI, 0.893–1.005) and 0.926 (95% CI, 0.860–0.991) with no significant difference (P = 0.534).

Among the advanced MAR sequences, MAVRIC-SL, a hybrid technique merging the advantages of SEMAC and MAVRIC, can expectedly reduce both in-plane and through-plane susceptibility artifacts and improve the visualization of bone-implant interface and surrounding soft tissues at arthroplasty.17,18 Although the overall image quality of MAVRIC-SL is slightly limited with respect to spatial resolution, noise and contrast, and fat suppression, the significant reduction of artifact size and image distortion can improve the visualization of joint anatomy and diagnostic confidence regarding implant-associated abnormalities.19 Hayter et al20 previously reported that the MAVRIC, a prototype of MAVRIC-SL, better depicted periprosthetic findings at the joints, particularly at larger elliptical-shaped metal components. Thus, the advanced MAR sequences of MAVRIC-SL can effectively detect abnormal findings around metal joint prostheses, particularly at the joint area, and lead to the diagnosis of PJI.

Magnetic resonance imaging findings of infection in native joints have previously been reported: joint effusion, synovial proliferation, synovial edema, soft-tissue edema, soft-tissue fluid collection, soft-tissue fistula, bone marrow edema, periosteal reaction, bone destruction, LN, and contrast-enhancement.4,5,21–23 Among them, joint effusion is reported to be quite common in infection in native joints on conventional MRI.21 Several investigators reported that simple joint effusion was frequently detected in patients after arthroplasty when evaluated by the advanced MAR sequences.10,11,24,25 Our results also showed that the finding of joint effusion was always found in the PJIs. It had a high NPV (100%) but a relatively low PPV (55%) for PJI. Therefore, the presence of joint effusion is not always suggestive of PJI, but the absence of joint effusion may be a good sign for no PJI. However, the combinations with capsular thickening (PPV, 74%; NPV, 95%; OR, 53.8), soft-tissue fluid collection (PPV, 87%; NPV, 82%; OR, 29.9), soft-tissue edema (PPV, 70%; NPV, 90%; OR, 20.6), or pericapsular edema (PPV, 76%; NPV, 81%; OR, 13.7) highly suggested PJI.

Regarding the soft-tissue abnormalities such as soft-tissue edema and soft-tissue fluid collection, in the CT study by Cyteval et al,23 abnormal findings in the soft tissues, particularly intramuscular and perimuscular fluid collections, were accurate for the diagnosis of PJI. These findings had high PPV (100%) and NPV (88%) for PJI. They addressed that PJI induced an increase in joint effusion and periprosthetic fluid collections. The advanced MAR sequences revealed more detailed findings around the joints with a metal implant, and synovial edema, capsule edema, and muscle edema showed a high diagnostic accuracy for PJI.10,11 Galley et al11 described that intramuscular fluid collections were specific for PJI, but they were not always present. Our results showed that the soft-tissue fluid collection suggested PJI (PPV, 81%; NPV, 81%; OR, 19.1), and the soft-tissue edema suggested PJI (PPV, 67%; NPV, 89%; OR, 17). In addition, the combination with soft-tissue fluid collection or soft-tissue edema almost always raised the possibility of PJI. In particular, regarding the BME pattern, the combination with soft-tissue edema definitively suggested PJI.

As mentioned previously, the combination of joint effusion and soft-tissue fluid collection raised the possibility of PJI. The presence of communication to the joint or the absence of capsular-like structure of the soft-tissue fluid collection was suggestive of PJI. Berquist et al26 reported that bursa-like cavities were frequently seen after hip arthroplasty on arthrography. Although the noninfected cavities generally showed smooth-walled capsular extensions, the infected bursa-like cavities were more irregular. According to the sonographic study of hip arthroplasty patients by van Holsbeeck et al,27 joint effusion and extraarticular fluid extension superficial to and/or between muscles were strongly suggestive of PJI. These studies using arthrography and sonography strongly supported our results.

Unexpectedly, the finding of BME pattern was not always suggestive of the presence of PJI in our study. According to the previous studies using the advanced MAR MRI,25,28 periprosthetic bone edema was common in symptomatic and asymptomatic patients after hip arthroplasty regardless of age, sex, acetabular cup inclination, stem alignment, femoral torsion, or subsidence of the femoral stem. Therefore, the periprosthetic findings after arthroplasty were considered not only by infection but also by reactive changes for metal implantation. Galley et al11 disclosed that periosteal reaction, capsular edema, and intramuscular edema highly suggested PJI. In our study, the combination of BME pattern with soft-tissue edema raised the PPV (73%) of PJI. We consider that the finding of BME pattern alone may not always be suggestive of PJI. However, BME pattern associated with soft-tissue edema definitively raised the possibility of PJI.

The finding of LN around the joints after arthroplasty is well recognized, particularly in the inguinal region of patients after hip arthroplasty.29 Recently, Albano et al30 reported that lymph nodal indices on conventional MRI, particularly those related to the nodal number, represented biomarkers of infection after hip arthroplasty. Interestingly, they described that the combinations of the lymph nodal indices with other features such as synovitis, extracapsular edema, bone edema, and extracapsular fluid collection slightly improved the accuracy of PJI. In the recent CT study by Isern-Kebschull et al,31 enlarged iliac lymph nodes and soft-tissue accumulation after hip arthroplasty were suggestive of PJI. In our series, there was no significance of the LN alone for PJI; however, the combinations with soft-tissue fluid collection or soft-tissue edema actually increased the PPVs for PJI. Therefore, it was quite valuable to evaluate LN in combination with soft-tissue fluid collection or soft-tissue edema.

There are the disadvantages of longer acquisition time and lower image quality in the advanced MAR sequences compared with 2D-FSE. Gutierrez et al17 described that the addition of the advanced MAR sequences might contribute directly to determine the treatment for symptomatic patients with metal joint implants and at least 1 or 2planes of advanced MAR images with reduced artifact size and image distortion might be valuable to answer the clinical questions. Filli et al28 also reported that the advanced MAR sequences altered the patient management in approximately 30% of symptomatic patients after arthroplasty. In our series, at least 1 2D-FSE was needed for the anatomical orientation of MRI for arthroplasty patients because the localizer images were of poor quality. Only 1 2D-FSE was obtained, and the other 2D-FSE was skipped in 44% of the MRI examinations because of susceptibility artifacts. Multiacquisition variable–resonance image combination selective provided minimized susceptibility artifacts and sufficient image quality for the evaluation of bone and soft tissues around the metal implants. Therefore, there were no clinical problems in the abbreviation of 2D-FSE. In 5 cases (14%), 1 MAVRIC-SL was obtained because of a difficulty in postural control. Postural control during the MRI examination in patients with PJI was a major challenge. However, we believe that the use of MRI with the advanced MAR sequences may be valuable for the diagnosis of PJI and may be recommended for PJI. Although applications and methods have been developed to decrease the acquisition times,32,33 further progress should be expected in the future.

Our study has some limitations, with the relatively small number of cases including being the first and most important. The small number of cases might weaken the statistical power of the results, especially for the evaluation of abnormal findings. The study included patients referred from outside institutions, and thus, the type of metal implants, metal composition, and the joints with a metal implant varied. Therefore, the significance of abnormal findings depending on the type of joint, method of surgery, and duration after surgery could not be assessed. Further studies with a large population would be needed. Second, all the MRI examinations were ordered based on the decisions of clinicians, mainly orthopedic surgeons. Therefore, there might be patient selection bias. In fact, a relatively high frequency (23 of 36 [64%]) of the MRI examinations was indicated because of a suspicion of PJI. Lastly, we did not use contrast materials for the assessment of infection of bone and soft tissues. However, earlier studies indicated that contrast materials were typically not necessary for the primary diagnosis of infection but might be useful to define the fluid collections and sinus tracts.

CONCLUSIONS

Magnetic resonance imaging with MAVRIC-SL was safely performed in patients after arthroplasty at 1.5 T. Capsular thickening, soft-tissue fluid collection, soft-tissue edema, pericapsular edema, and joint effusion were significant for PJI. Regarding the soft-tissue fluid collection, the presence of communication to the joint or the absence of capsular-like structure likely suggested PJI. The combinations of joint effusion, capsular thickening, pericapsular edema, soft-tissue fluid collection, and soft-tissue edema definitively raised the possibility of PJI. Regarding the overall image impression on MAVRIC-SL MRI, high interobserver agreement for PJI was found. The demands for the early diagnosis and arcuate evaluation of prosthesis-related complications may continue to expand. Using the advanced MAR sequences such as MAVRIC-SL, the clinical value of MRI would increase to evaluate the presence and severity of infection in the joints after arthroplasty and to determine the management plan for patients with PJI.

REFERENCES

1. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007;89:780–785. doi:10.2106/JBJS.F.00222.
2. Kapadia BH, Berg RA, Daley JA, et al. Periprosthetic joint infection. Lancet. 2016;387:386–394. doi:10.1016/S0140-6736(14)61798-0.
3. Della Valle C, Parvizi J, Bauer TW, et al. The diagnosis of periprosthetic joint infections of the hip and knee. J Bone Joint Surg Am. 2011;93:1355–1357. doi:10.2106/JBJS.9314ebo.
4. Fritz J, Lurie B, Miller TT, et al. MR imaging of hip arthroplasty implants. Radiographics. 2014;34:E106–E132. doi:10.1148/rg.344140010.
5. Talbot BS, Weinberg EP. MR imaging with metal-suppression sequences for evaluation of total joint arthroplasty. Radiographics. 2016;36:209–225. doi:10.1148/rg.2016150075.
6. Koch KM, Brau AC, Chen W, et al. Imaging near metal with a MAVRIC-SEMAC hybrid. Magn Reson Med. 2011;65:71–82. doi:10.1002/mrm.22523.
7. Koff MF, Burge AJ, Koch KM, et al. Imaging near orthopedic hardware. J Magn Reson Imaging. 2017;46:24–39. doi:10.1002/jmri.25577.
8. Jungmann PM, Agten CA, Pfirrmann CW, et al. Advances in MRI around metal. J Magn Reson Imaging. 2017;46:972–991. doi:10.1002/jmri.25708.
9. Hargreaves BA, Worters PW, Pauly KB, et al. Metal-induced artifacts in MRI. AJR Am J Roentgenol. 2011;197:547–555. doi:10.2214/AJR.11.7364.
10. Li AE, Sneag DB, Greditzer HG, et al. Total knee arthroplasty: diagnostic accuracy of patterns of synovitis at MR imaging. Radiology. 2016;281:499–506. doi:10.1148/radiol.2016152828.
11. Galley J, Sutter R, Stern C, et al. Diagnosis of periprosthetic hip joint infection using MRI with metal artifact reduction at 1.5T. Radiology. 2020;296:98–108. doi:10.1148/radiol.2020191901.
12. Parvizi J, Tan TL, Goswami K, et al. The 2018 definition of periprosthetic hip and knee infection: an evidence-based and validated criteria. J Arthroplasty. 2018;33:1309–1314. doi:10.1016/j.arth.2018.02.078.
13. Signore A, Sconfienza LM, Borens O, et al. Consensus document for the diagnosis of prosthetic joint infections: a joint paper by the EANM, EBJIS, and ESR (with ESCMID endorsement). Eur J Nucl Med Mol Imaging. 2019;46:971–988. doi:10.1007/s00259-019-4263-9.
14. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–174.
15. Zanetti M. The expanding role of MRI in the evaluation of periprosthetic hip joint infection. Radiology. 2020;296:109–110. doi:10.1148/radiol.2020201419.
16. Hynes JP, Murphy MC, Gillespie C, et al. Periprosthetic hip joint infection and its diagnosis. Radiology. 2020;297:E240. doi:10.1148/radiol.2020202755.
17. Gutierrez LB, Do BH, Gold GE, et al. MRI near metallic implants using MAVRIC SL: initial clinical experience at 3T. Acad Radiol. 2015;22:370–379. doi:10.1016/j.acra.2014.09.010.
18. Choi SJ, Koch KM, Hargreaves BA, et al. Meta artifact reduction with MAVRIC SL at 3-T MRI in patients with hip arthroplasty. AJR Am J Roentgenol. 2015;204:140–147. doi:10.2214/AJR.13.11785.
19. Kretzschmar M, Nardo L, Han MM, et al. Metal artefact suppression at 3 T MRI: comparison of MAVRIC-SL with conventional fast spin echo sequences in patients with hip joint arthroplasty. Eur Radiol. 2015;25:2403–2411. doi:10.1007%2Fs00330-015-3628-0.
20. Hayter CL, Koff MF, Shah P, et al. MRI after arthroplasty: comparison of MAVRIC and conventional fast spin-echo techniques. AJR Am J Roentgenol. 2011;197:W405–W411. doi:10.2214/AJR.11.6659.
21. Karchevsky M, Schweitzer ME, Morrison WB, et al. MRI findings of septic arthritis and associated osteomyelitis in adults. AJR Am J Roentgenol. 2004;182:119–122. doi:10.2214/ajr.182.1.1820119.
22. Anderson H, Toms AP, Cahir JG, et al. Grading the severity of soft tissue changes associated with metal-on-metal hip replacements: reliability of an MR grading system. Skeletal Radiol. 2011;40:303–307. doi:10.1007%2Fs00256-010-1000-7.
23. Cyteval C, Hamm V, Sarrabere MP, et al. Painful infection at the site of hip prosthesis: CT imaging. Radiology. 2002;224:477–483. doi:10.1148/radiol.2242010989.
24. Nardo L, Han M, Kretschmar M, et al. Metal artifact suppression at the hip: diagnostic performance at 3.0 T versus 1.5 tesla. Skeletal Radiol. 2015;44:1609–1616. doi:10.1007%2Fs00256-015-2214-5.
25. Germann C, Filli L, Jungmann PM, et al. Prospective and longitudinal evolution of postoperative periprosthetic findings on metal artifact-reduced MR imaging in asymptomatic patients after uncemented total hip arthroplasty. Skeletal Radiol. 2021;50:1177–1188. doi:10.1007%2Fs00256-020-03666-8.
26. Berquist TH, Bender CE, Maus TP, et al. Pseudobursae: a useful finding in patients with painful hip arthroplasty. AJR Am J Roentgenol. 1987;148:103–106. doi:10.2214/ajr.148.1.103.
27. van Holsbeeck MT, Eyler WR, Sherman LS, et al. Detection of infection in loosened hip prostheses: efficacy of sonography. AJR Am J Roentgenol. 1994;163:381–384. doi:10.2214/ajr.163.2.8037036.
28. Filli L, Jungmann PM, Zingg PO, et al. MRI with state-of-the-art metal artifact reduction after total hip arthroplasty: periprosthetic findings in asymptomatic and symptomatic patients. Eur Radiol. 2020;30:2241–2252. doi:10.1007%2Fs00330-019-06554-5.
29. Yanny S, Cahir JG, Barker T, et al. MRI of aseptic lymphocytic vasculitis-associated lesions in metal-on-metal hip replacements. AJR Am J Roentgenol. 2012;198:1394–1402. doi:10.2214/AJR.11.7504.
30. Albano D, Messina C, Zagra L, et al. Failed total hip arthroplasty: diagnostic performance of conventional MRI features and locoregional lymphadenopathy to identify infected implants. J Magn Reson Imaging. 2021;53:201–210. doi:10.1002/jmri.27314.
31. Isern-Kebschull J, Tomas X, Garcia-Diez AI, et al. Value of multidetector computed tomography for the differentiation of delayed aseptic and septic complications after total hip arthroplasty. Skeletal Radiol. 2020;49:893–902. doi:10.1007%2Fs00256-019-03355-1.
32. Kaushik SS, Marszalkowski C, Koch KM. External calibration of the spectral coverage for three-dimensional multispectral MRI. Magn Reson Med. 2016;76:1494–1503. doi:10.1002/mrm.26065.
33. Zochowski KC, Miranda MA, Cheung J, et al. MRI of hip arthroplasties: comparison of isotropic multiacquisition variable–resonance image combination selective (MAVRIC SL) acquisitions with a conventional MAVRIC SL acquisition. AJR Am J Roentgenol. 2019;213:W277–W286. doi:doi.org/10.2214/AJR.19.21606.
Keywords:

MRI; metal artifact reduction; MAVRIC-SL; periprosthetic joint infection; arthroplasty

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.