What’s new in MSK Imaging – September 2020

1 year ago
Incidence and Predictors of Subsequent Surgery After Anterior Cruciate Ligament Reconstruction: A 6-Year Follow-up Study

MOON Knee Group Investigation performed at Vanderbilt University Medical Center, Nashville, Tennessee, USA

American Journal of Sports Medicine

Background

Causes of repeat surgery after anterior cruciate ligament (ACL) reconstruction vary, however risk factors for these surgical procedures are not well described. If surgical or clinical factors are associated with need for repeat surgery, patients and clinicians can have a more accurate understanding of prognosis and modifiable factors to influence risk.

Question

Do clinical or surgical factors impact the rate of repeat surgery after primary ACL reconstruction? What is the incidence and types of subsequent surgery that occurred in a cohort of patients after their index ACL reconstruction? What variables were associated with the incidence of patients undergoing subsequent surgery after their index ACL reconstruction?

Study Design:

Prospective cohort study.

Participants

Patients at 7 sites, undergoing unilateral primary or revision ACL reconstruction with one of 17 surgeons between 2002 and 2008. Multi-ligamentous injuries were included.

Exclusion Criteria

Simultaneous bilateral ACL reconstruction.

Methods

Patients completed a questionnaire before their index ACL surgery and were followed up at 2 and 6 years. Patients were contacted to determine whether any of them underwent additional surgery since baseline. Operative reports were obtained, and all surgical procedures were categorized and recorded. Logistic regression models were constructed to predict which patient demographic and surgical variables were associated with the incidence of undergoing subsequent surgery after their index ACL reconstruction.

Results

The cohort consisted of 3276 patients (56.3% male) with a median age of 23 years. A 6-year follow-up was obtained on 91.5% (2999/3276) with regard to information on the incidence and frequency of subsequent surgery. Overall, 20.4% (612/2999) of the cohort was documented to have undergone at least 1 subsequent surgery on the ipsilateral knee 6 years after their index procedure.

ACL reconstruction: The most common subsequent surgical procedures were related to the meniscus (11.9%), revision ACL reconstruction (7.5%), loss of motion (7.8%), and articular cartilage (6.7%). Significant risk factors for incurring subsequent meniscus- related surgery were having a medial meniscal repair at the time of index surgery, reconstruction with a hamstring autograft or allograft, higher baseline Marx activity level, younger age, and cessation of smoking. Significant predictors of undergoing subsequent surgery involving articular cartilage were higher body mass index, higher Marx activity level, reconstruction with a hamstring autograft or allograft, meniscal repair at the time of index surgery, or a grade 3/4 articular cartilage abnormality classified at the time of index ACL reconstruction. Risk factors for incurring subsequent surgery for loss of motion were younger age, female sex, low baseline Knee injury and Osteoarthritis Outcome Score symptom subscore, and reconstruction with a soft tissue allograft.

Conclusion

Younger patients, female patients, low baseline knee injury, and osteoarthritis outcome score symptoms, and reconstruction with a soft tissue allograft may be risk factors that predispose patients to require repeat surgery after ACL reconstruction.

Link

https://doi.org/10.1177/0363546520935867

Senior editorial comment

The study is an important work with a large patient number and fills some gaps in knowledge with respect to ACL revisions. However, it has limitations of any such study with a retrospective recall bias. Many other factors, such as re-injuries, patient activity or knee instability status, patient needs or dissatisfaction with the primary surgery, and private insurance status are other obvious confounding factors.

Radiology trainees’ take-home messages:

When you review postoperative revision ACL MRI, pay closer attention to the following, as they might impact future ACL graft revisions.

  • Medial meniscal repair
  • Cartilage injury ( particularly grade 3/4)
  • cartilage repair sites.

 

Allograft Donor Characteristics Significantly Influence Graft Rupture After Anterior Cruciate Ligament Reconstructionbin a Young Active Population

Sarah Shumborski, BSc, MD, Lucy J. Salmony BAppSci (Physio), PhD, Claire Monk, BAppSci (ExPhys), Emma Heath, MPhty, Justin P. Roe MBBS, and Leo A. Pinczewski, AM, MBBS

American Journal of Sports Medicine

Background:

Graft selection in anterior cruciate ligament (ACL) surgery can be difficult in a young active population given their high rates of reinjury. Allografts allow for control over graft size and reduce morbidity of autograft harvest. There are mixed results about the use of allograft in the literature; however, the influence of the properties of the allograft on outcomes has not been considered.

Question:

Will ACL reconstruction with allografts from older donors will have a higher rate of graft rupture when compared with allograft from young donors?

Study Design:

Prospective cohort study

Participants

Patients (N = 211) aged 13 to 25 years underwent primary ACL reconstruction with fresh-frozen nonirradiated allograft over 3.5 years.

Exclusion Criteria

Patients older than 25 years of age, with additional significant ligamentous injury to the index knee, with history of prior anterior cruciate ligament injury to either knee, or seeking compensation for their injury were excluded.

Methods

Four graft types were used: patellar tendon, Achilles tendon, tibialis anterior, and tibialis posterior. Details were collected on allograft donor age and sex. At a minimum of 24 months, patients were evaluated for any further injuries and subjective analysis by International Knee Documentation Committee (IKDC) questionnaire. Comparison of variables among groups was analyzed with x2 tests for categorical data, and comparison of continuous variables was determined by Student t-test.

Results:

ACL graft rupture occurred in 23.5%. When grafts were separated into single strand (patellar and Achilles tendon) and multistrand (tibialis anterior and posterior), there was a significantly higher rate of reinjury in the single-strand grafts (29.9% vs 11%; P = .014). Grafts from female donors aged 50 years had significantly higher rates of ACL graft rupture (52.6%; P = .003) with increased odds by 6.7 times when compared with grafts from male donors aged <50 years. There was no significant difference in mean IKDC scores among the groups based on the age and sex of the allograft donor.

Conclusion

The age and sex of the allograft donor and the morphology of the graft significantly influenced the rate of ACL graft rupture in young active patients. Tendons from female donors aged ≥50 years have a higher re-rupture rates as compared with male donors of any age and younger females.

Link

https://doi.org/10.1177/0363546520938777

Senior editorial comment

We congratulate the authors for this important work. Although, there are four categories of grafts, which renders the numbers small in different groups, the study brings out few important points. It would have been even better if the failure rates were also controlled for other confounding factors, such as level of patient activity, causes of re-injuries, BMI, Diabetes, smoking, etc. among other factors.

Radiology trainees’ take-home messages:

When you review postoperative revision ACL MRI of the knee

  • Graft harvest site from the patella = Bone tendon-bone graft. This is a single strand graft.
  • Scarring in the quadriceps and hamstring muscles = think of graft harvest from these sites = multi-strand graft.
  • Higher chance of rupture for single-strand grafts

 

Assessing Femoral Trochlear Morphologic Features on Cross-Sectional Imaging Before Trochleoplasty: Dejour Classification Versus Quantitative Measurement 

Nicholas C. Nacey, Michael G. Fox, Barrett N. Luce, Dustin M. Boatman, and David R. Diduch

American Journal of Roentgenology

Background

The Dejour classification for trochlear dysplasia is a qualitative assessment of trochlear dysplasia. This study assessed the use of cross-sectional imaging to relate the quantitative and qualitative assessment of trochlear dysplasia.

Question

Can cross sectional imaging reliably, qualitatively and quantitatively categorize trochlear dysplasia as low grade (type A) or high grade (types B–D) according to the Dejour classification with good inter-rater reliability?

Study Design

Retrospective study

Participants

Patients who underwent CT or MRI prior to deepening trochleoplasty within a 9-year time period.

Exclusion Criteria

No patients excluded.

Methods

Retrospective review of CT and MRI knee examinations performed before patients undergoing deepening trochleoplasty and were independently conducted by two musculoskeletal radiologists. Each case of trochlear dysplasia was qualitatively assigned a Dejour type. Subsequently, quantitative measurements of the sulcus angle, distance from the tibial tubercle to the trochlear groove, trochlear depth, lateral trochlear inclination, trochlear facet asymmetry, and degree of patellar lateralization were performed. Quantitative data were analyzed using the mean measurements of both radiologists for all parameters except trochlear facet asymmetry, in which the presence of asymmetry was noted in binary fashion as present or absent, and the percentage of knees with asymmetry was recorded. Univariate analysis was used to compare the mean values of the different Dejour categories.

Results

A total of 35 patients (29 female patients and six male patients; mean age, 21.1 years) with 39 affected knees (17 right knees and 22 left knees) were included. Readers had exact qualitative agreement using Dejour classification for 30 of 39 knees (77% [κ = 0.77; 95% CI, 0.62–0.91]) and agreement on classification of low-grade versus high-grade dysplasia for 36 of 39 knees (92%). For these 36 knees, the mean differences in measurements of low- versus high-grade dysplasia, respectively, were as follows: for sulcus angle, 153° versus 168° (p < 0.001); for trochlear depth, 4 versus 1 mm (p < 0.001); for lateral trochlear inclination, 12 versus 7 mm (p < 0.02); and for decreased trochlear facet asymmetry, 13% versus 92% (p < 0.001). Trochlear depth, lateral trochlear inclination, and trochlear facet asymmetry were also different in comparisons of knees with Dejour type B and C trochlear dysplasia versus those with Dejour types B and D (all p < 0.05). No quantitative measurement differentiated between trochlear dysplasia of Dejour types C and D. The distance from the tibial tubercle to the trochlear groove and the degree of patellar lateralization were not statistically different between low- and high-grade dysplasia.

Conclusion

Qualitative use of the Dejour classification accurately categorizes trochlear dysplasia as low grade or high grade in 92% of cases, with exact agreement reached in 77% of cases. Furthermore, the trochlear depth, lateral trochlear inclination, trochlear facet asymmetry, and sulcus angle can differentiate between low-grade and high-grade dysplasia, with trochlear depth, lateral trochlear inclination, and trochlear facet asymmetry useful for differentiating between Dejour types B and C and Dejour types B and D.

Link

https://doi.org/10.2214/ajr.19.22400

Senior editorial comment

This is an interesting study with nice images and measurement illustrations. The low-grade vs high-grade trochlear classification though seems artificial with not much clinical relevance and accuracy and has been described without a reference standard or true negative cases in the cohort. Demonstration of inter-reader reliability contributes to body of knowledge.

Radiology trainees’ take-home messages:

When you review CT or MRI of the knee for dysplasia be familiar with

Dejour classification of dysplasia.  Higher grade = likely candidates for trochleoplasty

Type A : shallow trochlea (sulcus angle > 145°)

Type B : flat or shallow trochlea and a prominent supratrochlear spur or bump projecting anterior to the femoral shaft.

Type C : flattened trochlear facet with lateral facet convexity and medial facet hypoplasia.

Type D : flattened trochlea with an abrupt cliff on the medial side adjacent to a pronounced supratrochlear spur.

  • Make measurements 3 cm proximal to the medial tibiofemoral joint line, which will correspond to the medial femoral epiphyseal scar.

 

Spectrum of magnetic resonance imaging findings in transplanted multiple myeloma patients with hip/pelvic pain (according to MY-RADS): A single center experience

Federica Rosi, Lorenzo Torri, Alida Dominietto, Alberto Stefano Tagliafico

European Journal of Radiology

Background

Standardized multiple myeloma reporting has been developed with MY-RADS. This study applies this reporting criteria to post-transplant patients in the setting of pain, both to identify patterns of involvement and correlate with causes of pain/clinical parameters.

Question

What is  the expected spectrum of MRI findings in transplanted Multiple Myeloma (MM) patients with hip/pelvic pain?  Do specific MR findings correlate with standard prognostic clinical outcomes?

Study Design

Retrospective study

Participants

A total of 54 consecutive MM patients who were status post bone marrow transplant and underwent MRI for hip/pelvic pain from January 2017 to December 2019 were included.

Exclusion Criteria

Studies with inaccurate images due to artifacts, patients affected by other known malignancies, patients with pain coming from other regions were excluded.

Methods

A retrospective interpretation of the MRI findings of 54 MM patients with hip/pelvic pain were done according to MY-RADS guidelines. MRI findings included: type of bone marrow involvement and incidental findings (osteonecrosis and fractures). Inter- and intra-reader agreement were calculated using Cohen’s kappa test. Survival and relapse rates, type of transplantation and days of hospitalization were correlated with MRI findings.

Results

1/52 patient presented normal bone marrow pattern, 10/52 focal pattern, 26/52 diffuse pattern and 15/52 mixed based on MY-RADS descriptions. No cases of micronodular pattern were reported. Among the incidental findings, n = 6 osteonecrosis and n = 5 pathological fractures were found, with average length of stay higher. The intra- and inter-reader agreement assessing MY-RADS, were good (k value between 0.61−0.8). Focal pattern was most represented in patients with osteonecrosis and the worst survival rate. Diffuse pattern was most represented in relapses. No statistically significant correlations were found between bone marrow infiltration patterns and the type of transplantation.

Conclusion

MRI can recognize different infiltration patterns and complications in transplanted MM patients with hip/pelvic pain, correlating with causes of pain and clinical parameters.

Link

https://doi.org/10.1016/j.ejrad.2020.109154

Senior editorial comment

Thank you for this interesting work. The study confirms that MRI has an important role in MM surveillance both pre- and post- marrow replacement. Whole body CT is becoming popular in this domain but as MRI becomes more cost-effective, it will probably be a better study to follow-up such patients.

Radiology trainees’ take home message:

Be familiar with MY RADS components. See table below.

Messiou, Christina, et al. “Guidelines for acquisition, interpretation, and reporting of whole-body MRI in myeloma: myeloma response assessment and diagnosis system (MY-RADS).” Radiology 291.1 (2019): 5-13.

 

The differentiation between aneurysmal bone cyst and telangiectatic osteosarcoma: a clinical, radiographic and MRI study

Umme Sara Zishan, Ian Pressney, Michael Khoo & Asif Saifuddin

Skeletal Radiology

Background

Aneurysmal bone cyst (ABC) and telangiectatic osteosarcoma (TOS) may be difficult to differentiate due to similar clinical and imaging features, including young presentation, long bone involvement, lytic appearance on radiography and fluid-fluid levels on MRI. However, clinical management differs significantly. Confident imaging differentiation could positively affect patient care in difficult cases.

Question

Are there reliable imaging characteristics that may aid in differentiation of ABC and TOS?

Study Design

Retrospective study

Participants

Patients with confirmed histopathological diagnosis of ABC or TOS and pre-surgical radiograpsh and/or MRI available over an 11 year period.

Exclusion Criteria

No cases were excluded.

Materials and methods

Retrospective review of all histologically confirmed ABC and TOS over an 11-year period. Data recorded include age at presentation, sex, skeletal location and various radiographic and MRI features. Statistical analysis of recorded variables and inter-observer agreement was performed.

Results

This retrospective study included 183 patients, 92 males and 91 females. Mean age at presentation of 18.4 years (range 1–70 years); 152 cases of ABC and 31 TOS. No significant difference between age and sex. TOS was significantly less likely to involve the axial skeleton; no difference related to location within the bone. Radiographic findings significantly favouring ABC included a less aggressive pattern of bone destruction, a purely lytic appearance, an expanded but intact cortex, no periosteal response and no soft tissue mass. MRI features significantly favouring ABC included smaller tumor size (maximum mean dimension 46 mm compared to 95 mm for TOS), absence of soft tissue mass, > 2/3 of the lesion filled with fluid levels and thin septal enhancement following contrast.

Conclusions

Several radiographic and MRI features aid in the differentiation between ABC and TOS., including size, associated soft tissue mass, percent of lesion with fluid-fluid levels, and  Lesions with a geographic Type 1A or IB pattern of bone destruction which are completely filled with FFLs on MRI can confidently be diagnosed as ABC.

Link

https://doi.org/10.1007/s00256-020-03432-w

Senior editorial comment

Congratulations. Excellent work and good points made confirming the aggressiveness of TOS as compared to ABC on imaging despite a relatively smaller sample size of TOS.

Radiology trainees’ take-home messages:

When you review radiographs and MRI, and you are trying to differentiate ABC or TOS, use the criteria to help

TOS :

  • less likely to involve the axial skeleton
  • larger tumor size (maximum mean dimension 95 mm)
  • aggressive bone destruction with only partial fluid levels  – Need biopsy

ABC:

  • less aggressive pattern of bone destruction
  • purely lytic appearance, an expanded but intact cortex
  • no periosteal response
  • no soft tissue mass.
  • Smaller tumor size (maximum mean dimension 46 mm)
  • > 2/3 of the lesion filled with fluid levels and thin septal enhancement following contrast.

 

Deep learning evaluation of pelvic radiographs for position, hardware presence, and fracture detection

Gene Kitamura

European Journal of Radiology

Background

Deep learning models have been shown to be useful in augmenting both the efficiency and accuracy of reporting radiologists. Recent papers have shown the utility of deep learning in detecting hip fractures with pelvic radiographs, but there is a paucity of research utilizing deep learning to detect pelvic and acetabular fractures.

Question

Can deep learning models be developed in order to detect pelvic radiograph position, hardware presence, and pelvic/acetabular fractures?

Study Design

Retrospective review

Participants

Patients >18 years of old, over an approximately 10 year time period with pelvic radiographs obtained in an emergency or inpatient setting.

Exclusion Criteria

Suboptimal quality due to technical factors, as determined by a board-certified musculoskeletal radiologist.

Methods

Data was retrospectively acquired between 8/2009–6/2019. Pelvic radiographs including “fracture” in the impression, hardware, and randomly acquired pelvic radiographs were included. A subset of the data was split into 4 position labels and 2 hardware labels to create position labeling and hardware detecting models. The remaining data was parsed with these trained models, labeled based on 6 “separate” fracture patterns, and various fracture detecting models were created. A receiver operator characteristic (ROC) curve, area under the curve (AUC), and other output metrics were evaluated.

Results

The position and hardware models performed well with AUC of 0.99−1.00. The AUC for proximal femoral fracture detection was as high as 0.95, which was in line with previously published research. Pelvic and acetabular fracture detection performance was as low as 0.70 for the posterior pelvis category and as high as 0.85 for the acetabular category with the “separate” fracture model.

Conclusion

Deep learning models that can successfully detect pelvic imaging position, hardware presence, and pelvic and acetabular fractures with AUC loss of only 0.03 for proximal femoral fracture. This may serve as an important aid to interpreting radiologists.

Link

https://doi.org/10.1016/j.ejrad.2020.109139

Senior editorial comment

AI related research is advancing at an exponential pace. Nice work! We expect AI to do more in future and perform relevant angulation and displacement measurements of such fractures.

Radiology trainees’ take-home messages:

AI is here to stay and we need to adapt to it.

Detection of femoral and pelvic fractures will help in reducing morbidity and mortality of patients.

Using AI software to move patients with fractures to the top of the lists in the ER would be particularly useful.

 

CT-guided discitis-osteomyelitis biopsies: needle gauge  and microbiology results

Jad S. Husseini & F. Joseph Simeone & Sandra B. Nelson & Connie Y. Chang

Skeletal Radiology

Background

Rate of positive culture on needle biopsy varies throughout the literature, with limited data on factors that determine positivity rate. There are negative impacts on patient care and antibiotic stewardship in the setting of negative cultures on biopsy with imaging findings of osteomyelitis.

Question

Do microbiology results have a relationship with needle gauge used for CT-guided biopsies of suspected discitis-osteomyelitis?

Study Design

Retrospective study

Participants

All patients who underwent CT guided biopsy for suspected discitis-osteomyelitis at one institution between 2002 and 2019.

Exclusion Criteria

Patients with previously biopsy or clinically proven diagnosis of discitis-osteomyelitis at the level of the procedure and patients who did not receive sufficient follow up to establish a clinical diagnosis of osteomyelitis.

Methods

Biopsy location, needle type and gauge, microbiology, pathology, and clinical and imaging follow-up were obtained through chart review. Yield, sensitivity, specificity, and accuracy were calculated. A pairwise analysis of different needle gauges was also performed with calculations of odds ratios. Naïve Bayes predictive modeling was performed.

Results

241 (age: 59 ± 18 years; 88 [35%] F, 162 [65%] M) biopsies were performed. There were 3 (1%) 11 gauge (G), and 13 (5%) 12-G biopsies; 23 (10%) 13-G biopsies; 75 (31%) 14-G biopsies; and 90 (37%) 16-G, 33 (14%) 18-G, and 4 (2%) 20 G biopsies. True disease status (presence of infection) was determined via either pathology findings (205, 86%) or clinical and imaging follow-up (36, 14%). The most common true positive pathogen was Staphylococcus aureus (31, 33%). Overall biopsy yield, sensitivity, specificity, and accuracy were 39%, 56%, 89%, and 66%, respectively. Pooled biopsy yield, sensitivity, specificity, and accuracy was 56%, 69%, 71%, and 69% for 11–13-G needles and 36%, 53%, 91%, and 65% for 14–20-G needles, respectively, with an odds ratio between the two groups of 2.29 (P = 0.021). Pooled biopsy yield, sensitivity, specificity, and accuracy was 48%, 63%, 85%, and 68% for 11–14-G needles and 32%, 49%, 91%, and 64% for 16–20-G needles, respectively, with an odds ratio between the two groups of 2.02 (P = 0.0086).

Conclusion

The use of a larger inner bore diameter/lower gauge biopsy needle may increase the likelihood of culturing the causative microorganism for CT-guided biopsies of discitis-osteomyelitis, and particularly appears associated with an increase in sensitivity for osteomyelitis. This may be useful for interventional radiologists and result in greater, positive impact on patient care in patients undergoing biopsy for suspected osteomyelitis.

Link

https://doi.org/10.1007/s00256-020-03439-3

Senior editorial comment

Thank you for interesting work! Biopsy yield for bone infections remains abysmal, barely at 50% or less. Pathology expertise and pathologist’s willingness to render a hard diagnosis of osteomyelitis also varies based on their skill and confidence level. Given such confounders, this is a useful study. With widespread use of power drill related larger core biopsies, the results are likely to be better than historical studies.

Radiology trainees’ take-home messages:

Consider all aspects of a procedure and its impact on the patient.
Higher gauge of needle (when safe to use) may increase yield on bone biopsy for infection.

 

Modernization of bone age assessment: comparing the accuracy and reliability of an artificial intelligence algorithm and shorthand bone age to Greulich and Pyle

Mina Gerges & Hayley Eng & Harpreet Chhina & Anthony Cooper

Skeletal Radiology

Background

Greulich and Pyle (GP) is one of the most common methods to determine bone age from hand radiographs. In recent years, new methods have been developed to increase the efficiency in bone age analysis, including the shorthand bone age (SBA) and automated artificial intelligence algorithms.

Question

Does the use of newer methods (SBA or automated artificial intelligence algorithms) affect the accuracy/reliability of bone age assessment?

Study Design

Retrospective review

Participants

Patients in one hospital’s database, aged 12.5-16 and 10-14 years of age for males/females respectively.

Exclusion Criteria

Diagnosis or evidence of ipsilateral hand/wrist fracture within 2 years prior to the exam and diagnosis of a skeletal dysplasia or endocrine disease that may affect stature.

Methods

426 patients (equal males and females) had their bone age determined by two separate raters using the SBA and GP methods. Three weeks later, the two raters repeated the analysis of the radiographs. The raters timed themselves using an online stopwatch. De-identified radiographs were securely uploaded to an automated algorithm developed by a group of radiologists in Toronto. The gold standard was determined to be the radiology report attached to each radiograph, written by experienced radiologists using GP.

Results

Intraclass correlation between each method and the gold standard fell within the range of 0.8–0.9, highlighting significant agreement. Most of the comparisons showed a statistically significant difference between the new methods and the gold standard; however, it may not be clinically significant as it ranges between 0.25 and 0.5 years. A bone age is considered clinically abnormal if it falls outside 2 standard deviations of the chronological age; standard deviations are calculated and provided in GP atlas.

Conclusion

The shorthand bone age method and the automated algorithm produced values that are statistically different but clinically in agreement with the gold standard while reducing analysis time.

Link

https://doi.org/10.1007/s00256-020-03429-5

Senior editorial comment

A good use of AI with repetitive and tedious task. Thank you for the work!

Radiology trainees’ take-home messages:

Using AI software in order to help in performing a repetitive task may be a good use of this technology, especially when shown to perform well against the gold standard.

References
  • Share