Revisiting prediction of collapse in hip osteonecrosis with artificial intelligence and machine learning: a new approach for quantifying and ranking the contribution and association of factors for collapse

Research progress of femoral head necrosis in HIV-infected patients

As life expectancy among patients infected with the human immunodeficiency virus (HIV) increases, a growing number of complications have been observed. This population displays an elevated risk of ischemic necrosis of the femoral head in comparison to the general population, which may be attributed to HIV infection, antiretroviral medication use, and hormone application. Patients infected with the human immunodeficiency virus (HIV) who also have necrosis of the femoral head tend to present at an earlier age, with a rapid disease progression and a high incidence of bilateral onset. Magnetic resonance imaging (MRI) facilitates the early diagnosis of the condition, and the recommended treatment is total hip arthroplasty. Currently, the most prevalent treatment modality is total hip arthroplasty. This can effectively prevent occupational exposure when the surgery is performed in accordance with the HIV infection control guidelines. Furthermore, the surgical procedure is more time-consuming than that performed on patients without HIV. The postoperative quality of life of patients is markedly enhanced, and there are no reports of surgical complications. The precise mechanism underlying femoral head necrosis in HIV-infected patients remains unclear. Potential contributing factors have been identified, including systemic immune response, inflammatory response, local microenvironmental changes in the femoral head, and the unique anatomical structure of the femoral head, among others. A more profound comprehension of the disease's pathogenesis may facilitate the implementation of early prevention and treatment strategies, as well as the development of alternative conservative therapeutic options. This represents a promising avenue for future research. The present article reviews the epidemiological study of HIV-infected patients with osteonecrosis of the femoral head, local changes of the femoral head, possible mechanisms of osteonecrosis of the femoral head, occupational exposure during treatment, and surgical efficacy. The aim is to provide insights that can inform the diagnosis, prevention, treatment, and mechanism of this condition.

Tourniquet use in patients with sickle cell trait (SCT): Mediterranean or African ancestry influences complications, demonstrating a higher prevalence than control patients: matched study of nine hundred and forty SCT versus one thousand, two hundred and sixty three non-SCT patients

PURPOSE

One ongoing debate in orthopaedic surgery concerns using tourniquets in sickle cell trait (SCT) patients. SCT, a heterozygous carrier state of sickle cell disease (SCD), affects an estimated 300 million individuals globally with various genetic ancestries.

METHODS

A retrospective cohort study was conducted on 940 SCT patients and 1263 matched non-SCT controls who underwent limb surgeries using tourniquets between 1978 and 2018. Patient data were gathered from hospital records, blood bank information, and postoperative haemoglobin electrophoresis. Outcomes assessed included the incidence of venous thromboembolism (VTE), pulmonary embolism (PE), phlebitis, peripheral nerve impairment, and superficial infection. Covariates included age, sex, ethnicity, and diabetes status.

RESULTS

Among SCT patients, 75% were unaware of their carrier status at surgery. VTE incidence was significantly higher in SCT patients (10%) than non-SCT controls (2%), especially after prolonged tourniquet use and in upper limb procedures without anticoagulation. PE occurred in 3% of SCT patients, versus 1% in controls. Diabetes, more prevalent in SCT individuals (9% vs. 6%), further increased the complication risk. Tourniquet inflation time did not differ significantly between groups; however, neurologic complications and phlebitis were more common in SCT patients, particularly those with Mediterranean ancestry. Notably, SCT carriers of African and Mediterranean descent experienced higher complication rates than non-carriers, though complications were not confined to any single ancestry.

CONCLUSION

SCT is associated with an increased risk of thrombotic and neurologic complications during tourniquet, particularly in patients with undiagnosed SCT or diabetes and in procedures with extended ischaemia time. These results confirm the perception of SCT as a benign condition, but call for further clinical guidelines regarding tourniquet use in SCT carriers.

X-ray based radiomics machine learning models for predicting collapse of early-stage osteonecrosis of femoral head

This study aimed to develop an X-ray radiomics model for predicting collapse of early-stage osteonecrosis of the femoral head (ONFH). A total of 87 patients (111 hips; training set: n = 67, test set: n = 44) with non-traumatic ONFH at Association Research Circulation Osseous (ARCO) stage II were retrospectively enrolled. Following data dimensionality reduction and feature selection, radiomics models were constructed based on anteroposterior (AP), frog-lateral (FL), and AP + FL combined view using random forest (RF), support vector machine (SVM), and stochastic gradient descent (SGD). After the optimal radiomics model was selected based on areas under the curve (AUC), its performance on the test set was compared with that of orthopaedists using receiver operating characteristic (ROC) curves and confusion matrices. Among all radiomics models, the SVM-based AP + FL combined view model (AP + FL-Rad_SVM) achieved the highest individual performance demonstrating an AUC of 0.904 (95% CI 0.829 -0.978) in the test set, which was significantly better than that of three attending surgeons (p = 0.014, 0.004, and 0.045, respectively). The SVM model based on AP + FL views of hip X-ray exhibited excellent ability in predicting the collapse of ONFH and showed superior performance compared with less experienced orthopaedic surgeons. This model may inform clinical decision-making for early-stage ONFH.

Acetabular coverage exerts minimal influence on femoral head collapse and the necessity for surgical intervention in patients with osteonecrosis of femoral head

PURPOSE

The acetabular coverage in osteonecrosis of the femoral head (ONFH) affects the need for surgical intervention, and the collapse of the femoral head remains unclear. This study aimed to evaluate the relation between the acetabular coverage and the need for surgical treatment and femoral head collapse.

METHODS

The study included 158 patients with 252 hips with glucocorticoid administration and idiopathic ONHF without osteoarthritis changes. The mean age at the first visit was 45.2 years, and the mean follow-up period was 92.2 months. All ONFH hips were subsequently divided into two groups: those needing surgical intervention and those without surgery. Additionally, it divided 167 initially non-collapsed hips into those that either later collapsed or not. Radiographic parameters with the centre-edge angle, acetabular roof obliquity, sharp angle, and necrotic location, following the guidelines of the Japanese Investigation Committee, were evaluated.

RESULTS

There were no significant differences in radiographic parameters between the 106 hips that underwent surgery and the 146 hips without surgery. Among the 167 hips without initial collapse, 91 eventually collapsed while 76 did not; their radiographic findings have no significant differences. The necrotic locations were significantly larger in hips requiring surgical intervention or femoral head collapse. Furthermore, 21.8% (55 out of 252 hips) had acetabular dysplasia, which did not significantly correlate with the necessity for surgical treatment or the incidence of femoral head collapse.

CONCLUSIONS

Acetabular coverage has little effect on the necessity for surgical treatment and femoral head collapse in ONFH patients over a long-term follow-up.

Factors related to collapse progression in Japanese Investigation Committee classification type B osteonecrosis of the femoral head

PURPOSE

This study aimed to identify factors related to collapse progression in Japanese Investigation Committee classification type B osteonecrosis of the femoral head (ONFH) and to identify patients who would benefit from surgical treatment.

METHODS

This study included 41 patients (56 hips) with type B ONFH with a minimum follow-up of three years. Based on a ≥ 3 mm collapse progression in ONFH, we categorised patients into two groups: collapse progression and no collapse progression. Sagittal and coronal computed tomography images were used to measure the necrotic region relative to the intact femoral head diameter. The ratios of the necrotic regions of transverse and vertical diameter in coronal and sagittal images are defined as mediolateral transverse and mediolateral vertical, anteroposterior transverse and anteroposterior vertical, respectively. Demographic data and these imaging findings were compared between the two groups. We established a cut-off value for predicting collapse progression through receiver operating characteristic analysis and determined survival rates.

RESULTS

Type B ONFH had a 17.8% collapse progression rate. The mediolateral transverse, mediolateral vertical, anteroposterior transverse, and anteroposterior vertical were significantly higher in the collapse progression group (P < 0.01). Mediolateral transverse was an independent risk factor of collapse progression (hazard ratio, 1.27; 95% confidence interval, 1.03-1.57; P = 0.03), with an optimal cut-off of 45.6%. The 5-year survival rates with collapse progression as the endpoints were 57.0 and 94.9% in the mediolateral transverse of ≥ 45.6 and < 45.6%, respectively.

CONCLUSION

A mediolateral transverse of ≥ 45.6% predicts collapse progression in patients with type B ONFH.

Evaluation of the long-term patient-reported outcomes after hip arthroplasty or joint preserving with Sugioka femoral osteotomy in patients with femoral head osteonecrosis

PURPOSE

This study retrospectively evaluated long-term clinical outcomes and patient-reported outcome measures (PROMs) in patients with osteonecrosis of the femoral head (ONFH) who underwent transtrochanteric rotational osteotomy (TRO), curved varus osteotomy (CVO), and total hip arthroplasty (THA).

METHODS

We retrospectively reviewed the 109 hips in 96 patients (46 men, 50 women) who underwent CVO, TRO, or THA for ONFH treatment. The mean follow-up period for the TRO, CVO, and THA groups was 14.8, 11.5, and 13.3 years, respectively.

RESULTS

The THA conversion rate of the TRO patients was significantly higher than that of the patients with CVO, and the final clinical scores in the patients with TRO did not improve compared with preoperative scores. Postoperative PROMs showed that the total and pain scores of the patients with THA were significantly higher than those of patients with TRO and CVO, while the PROM score did not change between patients with TRO and CVO. The analysis further showed that the preoperative type C2, stage 3A, or postoperative type C1 and C2 were significant predictors of decreased final PROM scores.

CONCLUSION

This study found that CVO and THA are clinically effective treatments for ONFH, with significant improvements compared with preoperative scores. However, THA was associated with significantly higher PROMs and pain scores than those of CVO and TRO in long-term follow-up. Furthermore, our results suggest that postoperative PROMs depend mainly on the preoperative level of collapse and postoperative transposed intact ratio of the articular surface of the femoral head.

Mesenchymal Stem Cell Therapy for Bone Repair of Human Hip Osteonecrosis with Bilateral Match-Control Evaluation: Impact of Tissue Source, Cell Count, Disease Stage, and Volume Size on 908 Hips

We investigated the impact of mesenchymal stem cell (MSC) therapy on treating bilateral human hip osteonecrosis, analyzing 908 cases. This study assesses factors such as tissue source and cell count, comparing core decompression with various cell therapies. This research emphasizes bone repair according to pre-treatment conditions and the specificities of cell therapy in osteonecrosis repair, indicating a potential for improved bone repair strategies in hips without femoral head collapse. This study utilized a single-center retrospective analysis to investigate the efficacy of cellular approaches in the bone repair of osteonecrosis. It examined the impact on bone repair of tissue source (autologous bone marrow concentrate, allogeneic expanded, autologous expanded), cell quantity (from none in core decompression alone to millions in cell therapy), and osteonecrosis stage and volume. Excluding hips with femoral head collapse, it focused on patients who had bilateral hip osteonecrosis, both pre-operative and post-operative MRIs, and a follow-up of over five years. The analysis divided these patients into seven groups based on match control treatment variations in bilateral hip osteonecrosis, primarily investigating the outcomes between core decompression, washing effect, and different tissue sources of MSCs. Younger patients (<30 years) demonstrated significantly better repair volumes, particularly in stage II lesions, than older counterparts. Additionally, bone repair volume increased with the number of implanted MSCs up to 1,000,000, beyond which no additional benefits were observed. No significant difference was observed in repair outcomes between different sources of MSCs (BMAC, allogenic, or expanded cells). The study also highlighted that a 'washing effect' was beneficial, particularly for larger-volume osteonecrosis when combined with core decompression. Partial bone repair was the more frequent event observed, while total bone repair of osteonecrosis was rare. The volume and stage of osteonecrosis, alongside the number of injected cells, significantly affected treatment outcomes. In summary, this study provides comprehensive insights into the effectiveness and variables influencing the use of mesenchymal stem cells in treating human hip osteonecrosis. It emphasizes the potential of cell therapy while acknowledging the complexity and variability of results based on factors such as age, cell count, and disease stage.

Deep learning enables the differentiation between early and late stages of hip avascular necrosis

OBJECTIVES

To develop a deep learning methodology that distinguishes early from late stages of avascular necrosis of the hip (AVN) to determine treatment decisions.

METHODS

Three convolutional neural networks (CNNs) VGG-16, Inception ResnetV2, InceptionV3 were trained with transfer learning (ImageNet) and finetuned with a retrospectively collected cohort of (n = 104) MRI examinations of AVN patients, to differentiate between early (ARCO 1-2) and late (ARCO 3-4) stages. A consensus CNN ensemble decision was recorded as the agreement of at least two CNNs. CNN and ensemble performance was benchmarked on an independent cohort of 49 patients from another country and was compared to the performance of two MSK radiologists. CNN performance was expressed with areas under the curve (AUC), the respective 95% confidence intervals (CIs) and precision, and recall and f1-scores. AUCs were compared with DeLong's test.

RESULTS

On internal testing, Inception-ResnetV2 achieved the highest individual performance with an AUC of 99.7% (95%CI 99-100%), followed by InceptionV3 and VGG-16 with AUCs of 99.3% (95%CI 98.4-100%) and 97.3% (95%CI 95.5-99.2%) respectively. The CNN ensemble the same AUCs Inception ResnetV2. On external validation, model performance dropped with VGG-16 achieving the highest individual AUC of 78.9% (95%CI 51.6-79.6%) The best external performance was achieved by the model ensemble with an AUC of 85.5% (95%CI 72.2-93.9%). No significant difference was found between the CNN ensemble and expert MSK radiologists (p = 0.22 and 0.092 respectively).

CONCLUSION

An externally validated CNN ensemble accurately distinguishes between the early and late stages of AVN and has comparable performance to expert MSK radiologists.

CLINICAL RELEVANCE STATEMENT

This paper introduces the use of deep learning for the differentiation between early and late avascular necrosis of the hip, assisting in a complex clinical decision that can determine the choice between conservative and surgical treatment.

KEY POINTS

• A convolutional neural network ensemble achieved excellent performance in distinguishing between early and late avascular necrosis. • The performance of the deep learning method was similar to the performance of expert readers.

Revision hip arthroplasty dislocation risk calculator: when to select dual mobility, large heads, constrained liners, or a standard head size? Testing one hundred thousand hip revisions with artificial intelligence

PURPOSE

Preventing dislocation with large head (≥ 36 mm), dual mobility, or constrained acetabular liner is another option than a standard (≤ 32 mm). Many other dislocations risk factors than size of the femoral head exist after hip arthroplasty revision. Predicting dislocation with a calculator according to the implant, to the indication of revision, and to patient's risks could allow a better surgery decision.

METHODS

Our search method covers the period from 2000 to 2022. A total of 470 relevant citations on hip major revision (cup or stem or both revisions) were identified with artificial intelligence comprising 235 publications of 54,742 standard heads comprising 142 publications of 35,270 large heads, comprising 41 publications of 3945 constrained acetabular components, and 52 publications of 10,424 dual mobility implants. We considered four implant types (standard, large head, dual mobility, or constrained acetabular liner) as the entry layer of the artificial neural network (ANN). Indication for revision THA was the second hidden layer. Demographics, spine surgery, and neurologic disease were the third layer. Implant revision, reconstruction process as next input (hidden layer). Surgery-related factors, and so on. The output was a postoperative dislocation or not.

RESULTS

Of the 104,381 hips that underwent a major revision, a second revision for dislocation was performed for 9234 hips. In each implant group, dislocation remained the first cause of revision. The rate of second revision for dislocation as a percentage of first revision procedures was significantly higher in the standard head group (11.8%) than in the constrained acetabular liner group (4.5%), the dual mobility group (4.1%), and the large head group (6.1%). Instability of a previous THA, infection, or periprosthetic fracture as the indication for revision was increased risk factors as compared with aseptic loosening. One hundred variables were used to create the calculator with the best parameter combination of data and ranking the different factors, according to the four implant types (standard, large head, dual mobility, or constrained acetabular liner).

CONCLUSION

The calculator can be used as a tool to identify patients at risk for dislocation after hip arthroplasty revision and individualize recommendations to select another option than a standard head size.

Prognosis of hip osteonecrosis after cell therapy with a calculator and artificial intelligence: ten year collapse-free survival prediction on three thousand and twenty one hips

PURPOSE

Several reports have identified prognostic factors for hip osteonecrosis treated with cell therapy, but no study investigated the accuracy of artificial intelligence method such as machine learning and artificial neural network (ANN) to predict the efficiency of the treatment. We determined the benefit of cell therapy compared with core decompression or natural evolution, and developed machine-learning algorithms for predicting ten year collapse-free survival in hip osteonecrosis treated with cell therapy. Using the best algorithm, we propose a calculator for "prognosis hip osteonecrosis cell therapy (PHOCT)" accessible for clinical use.

METHODS

A total of 3145 patients with 5261 osteonecroses without collapses were included in this study, comprising 1321 (42%) men and 1824 (58%) women, with a median age of 34 (12-62) years. Cell therapy was the treatment for 3021 hips, core decompression alone for 1374 hips, while absence of treatment was the control group of 764 hips. First, logistic regression and binary logistic regression analysis were performed to compare results of the three groups at ten years. Then an artificial neural network model was developed for ten year collapse-free survival after cell therapy. The models' performances were compared. The algorithms were assessed by calibration, and performance, and with c-statistic as measure of discrimination. It ranges from 0.5 to 1.0, with 1.0 being perfect discrimination and 0.5 poor (no better than chance at making a prediction).

RESULTS

Among the 3021 hips with cell therapy, 1964 hips (65%) were collapse-free survival at ten years, versus 453 (33%) among those 1374 treated with core decompression alone, and versus 115 (15%) among 764 hips with natural evolution. We analyzed factors influencing the prediction of collapse-free period with classical statistics and artificial intelligence among hips with cell therapy. After selecting variables, a machine learning algorithm created a prognosis osteonecrosis cell therapy calculator (POCT). This calculator proved to have good accuracy on validation in these series of 3021 hip osteonecroses treated with cell therapy. The algorithm had a c-statistic of 0.871 suggesting good-to-excellent discrimination when all the osteonecroses were mixed. The c-statistics were calculated separately for subpopulations of categorical osteonecroses. It retained good accuracy, but underestimated ten year survival in some subgroups, suggesting that specific calculators could be useful for some subgroups. This study highlights the importance of multimodal evaluation of patient parameters and shows the degree to which the outcome is modified by some decisions that are within a surgeon's control, as the number of cells to aspirate, the choice of injecting in both the osteonecrosis and the healthy bone, the choice between unilateral or bilateral injection, and the possibility to do a repeat injection.

CONCLUSION

Many disease conditions and the heterogeneities of patients are causes of variation of outcome after cell therapy for osteonecrosis. Predicting therapeutic effectiveness with a calculator allows a good discrimination to target patients who are most likely to benefit from this intervention.