Aneurysmal bone cyst: Is selective arterial embolization effective as curettage and bone grafting?

Aneurysmal bone cyst (ABC) is a lytic benign bone lesion representing about 1% of all primary bone tumors. Method to treat ABC's have developed over time. The standard of care cure for ABC has been curettage with or without bone grafting of the defect but is burdened by recurrence rates of approximately 25%–31%. Based on the assumption that ABCs usually supplied by one or more pathological feeding arteries, selective arterial embolization has been described as an adjuvant preoperative procedure to reduce intra‐operative hemorrhage, and as primary treatment for lesions in difficult surgical access. In the current study, we therefore asked whether (1) a single or a repeat selective arterial embolization (SAE) for treating ABCs would produce comparable healing rates compared with curettage and bone grafting; (2) evaluated the relationship of recurrence in relation to the site of the cyst, the age, and gender of the patients; and (3) the two techniques differ in term of long‐term complication.

recurred patients were treated with SAE (four patients need two while seven need three SAE to heal), and eight patients with curettage and bone grafting.Thirty-eight out of 46 (82%) patients experienced bone ossification regardless the number of SAE.The overall rate of local recurrence for all patients was 26.7%.SAE group presented a lower complication rate (6%) where two patients experienced skin necrosis, and one limb-length discrepancies (2% of all cohort).
Discussion: The use of SAE is an attractive option to treat ABC as it combines on one hand a lower complication rate than curettage and bone grafting, on the other it can be carried out in case of nonresectable ABCs, significantly reducing the size of viable ABC lesions, fostering bone remodeling and mineralization, and most importantly, significantly improving the patient's quality of life.

| INTRODUCTION
Aneurysmal bone cyst (ABC) is a lytic benign bone lesion representing about 1% of all primary bone tumors. 1,2In about 70% of cases, it occurs as a primary lesion, in other 30% of cases it is found as secondary lesions arising in other bone lesions such as giant cell tumors, chondroblastoma, osteoblastomas, and other primary malignant bone cystic tumors. 1,3imary ABC, it was first described by Virchow in 1876 and then by Jaffe et al. 4 who assumed a hematic filling of pre-existing bone lesions with high internal pressure.The most recent evidence has shown the existence of a typical translocation of these primary forms that leads to the activation of the gene USP6 placed on 17p13, but this has not been demonstrated in secondary forms. 5sed on the staging system of musculoskeletal neoplasms by Enneking et al., 6 ABC can be classified into latent (grade 1), active (grade 2), and aggressive (grade 3).Sometimes because of its radiological characteristics of aggressiveness it can be sometimes mistaken for a malignant tumor.
In some cases, ABCs can result in very rapid bone cyst expansion and can be considered locally aggressive, 2,6 leading to a local bone erosion and pathological fracture. 7It represents a problem, especially in spine lesions, where mild to severe neurological damage can occur. 2ey typically occur in the metaphysis of long bones (about 50%-60%), but any osseous segment may be involved; other frequent sites are spine and sacrum (about 20%-30%), especially located in posterior element with extension to vertebral body 8 ; in approximately 9% of cases were described pelvis and small bones of hand and feet localizations. 2,9e ABC treatments method has developed over time.The standard of care for ABC has been curettage with or without bone grafting of the defect but is burdened by recurrence rates of approximately 25%-31%. 10 The usage of local adjuvants has been introduced in attempt to reduce the risk of recurrence such as the usage of cryotherapy with liquid nitrogen, 11 phenol, 12 polymethylmethacrylate cementation, 13 doxycycline foam ablation, 14,15 or aggressive curettage by using high-speed burr. 16[21] Based on the assumption that ABCs usually supplied by one or more pathological feeding arteries, selective arterial embolization has been described as an adjuvant preoperative procedure to reduce intra-operative hemorrhage, and as primary treatment for lesions in difficult surgical access. 1 In the current study, we therefore asked whether (1) a single or a repeat selective arterial embolization (SAE) for treating ABCs would produce comparable healing rates compared with curettage and bone grafting; (2) evaluated the relationship of recurrence in relation to the site of the cyst, the age, and gender of the patients; (3) the two techniques differ in term of long-term complication.

| MATERIALS AND METHODS
A retrospective study was performed to identify curettage and bone grafting, or SAE performed at our institute from 1994 to 2018.
The diagnosis of ABC was always established with percutaneous CTguided biopsy or open biopsy.This series included only patients who were considered to have primary ABCs.Other conditions were considered part of the primary disease process and were not included.

Inclusion criteria were as follows:
-Primary ABC histologically confirmed.
-ABC requiring treatment treated with Curettage and bone grafting or selective arterial embolization.
-Minimum follow-up of 6 months.

Exclusion criteria were:
-Tumors that might be associated with secondary ABC.
-Latent ABC at X-Ray that did not require any treatment.

| Curettage and bone grafting
All the procedures were performed under loco-regional or general anesthesia.Skin incision was as long as the length to allow a wide exposure of the lesion.The curettage of the bone cavity stopped after obtaining healthy bleeding cancellous bone, and the cavity was filled with autologous or homologous bone grafting.No prophylactic fixation was carried out in our cohort.The total procedure lasted from 90 to 120 min.The mean hospitalization was 5 days (range 4-7).
To prevent postoperative fracture, some patients were recommended to use crutches with partial or nonweightbearing for 30 days.Patients were allowed to resume their usual activity if no pain were referred and partial ossification was observed.Intensive physical activity was allowed after 60 days on a case-by-case basis after evaluation of plain antero-posterior and lateral X-Ray or CTscan.Apart patients from being pain-free, complete intralesional ossification was required without evidence of local recurrence.

| SAE
The procedure was performed under general or local anesthesia in pediatric or adult patients respectively.All the SAE were done through the common femoral artery according to the Seldinger's technique.The procedure consists into three steps: (1) a digital subtraction angiography (DSA) of the ABC was performed obtaining a vascular mapping of the pathological feeding arteries; (2) selective and super-selective catheterizations and embolization of the lesion's feeding vessels was done; N-2-butyl cyanoacrylate (Glubran 2 surgical glue, GEM S.r.l.Italy, IT) was mixed with 33% Lipiodol ultra-fluid and the compound was injected "sandwiched" with 5% glucosate solution to prevent polymerization with blood; after every embolization a control DSA was done to evaluate occlusion of feeding vessels and residual pathological bloodstream; (3) a DSA was finally performed to evaluate the effectiveness of the occlusion.
Usually, 100-200 cc of Omnipaque ® (Iohexol 350 mg/mL; GE Healthcare) contrast medium was used in total for the diagnostic, selective, and final DSA.
The total procedure lasted from 60 to 120 min.The mean hospitalization was 2 days; patients were usually discharged the day after the SAE.Restricted activity was recommended to all patients for 30 days.After this period, they were allowed to move freely and undertake their normal everyday activities without weight bearing.
In case of local recurrence, the subsequent treatment was chosen on a case-by-case basis.Curettage and bone grafting was preferred for lesion located in the limb bones.When curettage was risky for the location of the ABC, a subsequent SAE was performed.

| Follow up
Patients were followed clinically with plain radiographs or CT scan at 3, 6, 9, and 12 months then annually in the absence of symptoms.
Treatment success was determined evaluating pre-and postprocedural imaging according to Chang classification. 22The response was sorted as follows: substantial (51%-100% of the lesion filled-in), partial (1%-50% of the lesion filled-in), and none (0% of the lesion filled-in).Local recurrence was diagnosed in case of 0% of the lesion filled-in after 6 months from the first procedure, or in case of pain.
The treatment was considered a failure when either (1) a fracture occurred; (2) there was no evidence of healing or partial healing on radiographs after 6 months; or (3) a recurrent cyst required subsequent treatment.

| Statistical analysis
Descriptive statistics were performed for all variables.We determined differences between the two groups of patients.Chi-square test was used to test the association between categorical variables.Kaplan-Meier survival analysis with 95% confidence intervals (CI) was undertaken with the Mantel-Cox log-rank test to evaluate the influence of the gender, age, and stage of the ABC.For abnormally distributed data, the Mann-Whitney test was used to analyze two independent populations.Multivariate Cox regression with Wald backward method was performed to identify the more predictive model for healing.We performed the survival to failure (defined above) using the Wilcoxon (Gehan) estimator of survival function.

| RESULTS
Out of 354 patients affected by primary ABC registered in our database, 89 cases didn't fit the inclusion criteria and were excluded (43 were latent ABC, 24 had a follow-up less than 6 months, and 22 were secondary ABC area on another bone tumor).Of the remaining 265 patients, 219 were treated with curettage and bone grafting (curettage group), and 46 with selective arterial embolization (SAE Group).There were 134 (50.6%) males and 151 (49.4%) females.The mean age was 15 years (range: 2-77 years).The cysts were located in the tibia in 68 cases (25.7%), in the femur in 61 cases (23%), in the humerus in 29 (10.9%), in the fibula in 25 cases (9.4%), in the pelvis in 23 cases (8.7%), in the clavicular in 11 cases (4.2%), in the radius in eight cases (3%), in the metatarsus in 8 cases (3%), in the metacarpal in two cases (0.8%), in the ulna in seven cases (2.6%), in the calcaneus and sacrum five cases for each (1.9%), in the vertebra in 10 cases (3.8%), and in the scapula in three cases (1.1%).The stage of the cist was considered active in 79.6% of the cases (211 patients) and aggressive in 20.4% (54 patients).There were no statistical differences among the two groups with respect to age, site of the lesion, stage, and clinical presentation (Table 1).
The mean follow-up was 40 months (range: 18-230 months) for the curettage group and 121 months (range: 6-280 months) for the SAE group.In those cases that required more than one treatment, only the response to the first procedure was reported.

| Comparison between curettage and bone grafting and SAE
Of the 219 patients treated with curettage and bone grafting (curettage group), 165 out of 219 (75.3%) experienced bone healing, while local recurrence was observed in 54 cases (24.7%) after 12 months on average (range: 3-120 months) from surgery.Fourteen of them were treated with additional curettage and bone grafting, six with a single injection of bone marrow concentrate and demineralized bone matrix, and five with selective arterial embolization.In two cases, a further treatment was not advisable given the absence of pain, the minimal risk of pathological fracture, and the absence of increase size at imaging.Ten patients presented local recurrence after the second treatment.Two of them needed a third treatment to be healed.All these 10 patients underwent curettage and bone grafting from the second treatment onwards.
After the first SAE, bone ossification was seen in 27 out of 46 patients (58.7%), without needing any further treatment (Figure 1).

| Factors associated with local recurrence
Univariate and multivariate analysis showed a significant difference in local recurrence rates in patients younger than 10 years; patients older than 10 years had a risk of recurrence 47% lower than patients younger than 10 years (hazard ratio [HR] 0.47, 95% CI, 0.29-0.76,p = 0.002).
Gender, the site of the ABC had no influence on the outcome.
The mean time of recurrence was 12 ± 16.6 months for the curettage group and 5 ± 4.4 months for the SAE group.Gender, the site, and the size of the ABC had no influence on the outcome.

| Long-term complication rate
Twenty-six patients (12%) in the curettage Group had long-term complications.Twelve presented chronic pain, 11 limb-length discrepancies (5% of all cohort), three had an infection that was treated surgically and with antibiotic suppression, and one case had a heterotopic ossification.SAE group presented a lower complication rate (6%) where two patients experienced skin necrosis, and one limb-length discrepancies (2% of all cohort).Treatment of ABCs is a challenge.Even if curettage with or without bone grafting persist to be the most widely used treatment for ABCs, [23][24][25] many surgical and nonsurgical options have been proposed over the years to reduce the risk of recurrence as high-speed bur, 16 large cortical window, 26 adjuvant therapy, 22,27,28 doxycycline foam ablation, 14,15 and bone marrow injection technique. 29,30These new therapies are based on an improved understanding of the nonneoplastic nature of the lesion.However, until now, there was no consensus regarding the best procedure.Moreover, approaching patients with ABCs is still case-by-case, based on the location, the risk of fracture or bleed.Assumed a benign biology of the lesions, a nonsurgical treatment is reasonable; SAE has been reported as a safe and effective procedure for ABC. 1,31We therefore asked whether (1) a single or a repeat SAE for treating ABCs would produce comparable healing rates compared with curettage and bone grafting; (2) evaluated the relationship of recurrence in relation to the site of the cyst, the age, and gender of the patients; (3) the two techniques differ in term of long-term complication.

| Limitation
This study had several limitations.First, we present a retrospective study that includes patients from a very wide period.Second, the two groups included a large variety of site of ABC.Curettage was performed for aggressive lesion in the long bone, while SAE in the pelvis or in the vertebral ABCs.However, to the best of our knowledge, this is first study that compare the efficacy of two different technique for treating ABC.

| Comparison between curettage and bone grafting and SAE
Avoidance of local recurrence persist to be a challenge in the treatment of ABC.The only treatment free of recurrence is "en bloc" resection.However, due to the benign histology of the tumor and the good results achieved with other less aggressive treatments, "en bloc" resection should be considered an overtreatment and performed only in very selected cases.
We observed that, after one procedure, the percentage of healing was higher in the curettage group than in the SAE group.
Otherwise, the final healing rate independent from the number of SAE was higher in the SAE group than in the curettage group.These data are consistent with data reported in the literature (Table 2).
Rossi et al. 1  In fact, we reported that the rate of recurrence after a single treatment was 19% in a series of 21 patients. 30rettage is still considered the standard of cure for ABC, but his recurrence rate ranging from 3% to 37% in the literature. 25,26,36,37We found a recurrence rate of 24.7% in the curettage group.Ramírez et al. 38 got a global rate of recurrence of 27.5% in 29 patients treated with curettage plus bone grafting or bone resection, while Mankin et al. revised a series of 150 primary ABCs treated mainly with curettage and implantation of chips of allograft or autograft or polymethylmethacrylate reporting a recurrence rate of 20%. 34 found that SAE had a twice as high risk of local recurrence than curettage, but the recurrence occurs sooner in the curettage group than in the SAE group.Best of our knowledge, these data have never been previously described in the literature.These data are difficult to compare.However, our cohort has a selection bias that may influence the outcome of the treatment.Further study will be required to verify the findings.
The strength of the SAE is given by the minimally invasively and lower complications rate.Mean hospitalization was 2 days, the patients were usually discharged the day after SAE, and a second treatment such as curettage, concentrated bone marrow injection, or a subsequent SAE may be performed in case of recurrence.The contribution of SAE to the management of ABC has the unquestionable advantage of being easily repeatable in case of recurrence, consisting in a less invasive, lower cost and simpler procedure than surgery.Moreover, doesn't require any further medication, representing an optimal choice in case of difficult anatomical sites or potentially harmful surgical procedures, and patient may resume the physical activity as before the treatment from the day after.

| Factors associated with recurrence
Wald backward method was performed to identify the more predictive model for failure.The odds ratio (OR) for each parameter with 95% CI was used to express the Cox regression results.Statistical significance was set at p-value of 0.05 or less.All statistical analyses were performed with IBM SPSS Statistic 25.0 (IBM™ Corp).
Local recurrence was observed in 19 patients after 18 months on average (range: 3-145 months) from the first procedure.Eleven recurred patients were treated with SAE (four patients need two while seven need three SAE to heal), and eight patients with curettage and bone grafting.Thirty-eight out of 46 (82%) patients experienced bone ossification regardless the number of SAE, and 8 of 46 local recurrence (14.2%).Patients needed a mean of 1.2 ± 0.45 SAE for healing.The overall rate of local recurrence for all patients was 26.7% (71 out of 265 patientets).Kaplan-Meier survival estimate, with recurrence as endpoint, showed 75.8% and 60.9% 5-years recurrence free survival (p = 0.02, Wilcoxon-Gehan test) for the curettage group and SAE Group respectively (Figure2).

T A B L E 1
Descriptive baseline data.
scan of the pelvis of a 9 years-old child show an osteolytic area in the left pubis and extending into the soft tissue.A CT-guided needle biopsy was carried out and the diagnosis of aneurysmal bone cyst (ABC) was verified.Patient referred persistent pain in the left pubic area and underwent selective arterial embolization (SAE) 4 months after diagnosis.Axial CT scan of the pelvis 3 months (B), 1 (C), and 3 (D) years after SAE shows progressive ossification and a progressive tumor size reduction.Patient experienced pain relieve 2 months after SAE.F I G U R E 2 Lifetime table analysis of the treatment survival in the two groups.The end point is the failure of the first procedure.The treatment was considered failed when a local recurrence occurred within 5 years from surgery.The curettage group showed fewer failures compared to the SAE group (Wilcoxon-Gehan test; p = 0.02).
Recurrence of ABC has been previously associated with younger age of the patients, juxta-epiphyseal location of the lesion, and female gender.39Based on the results of the current study, the predictor factors for local recurrence were patients' age younger than 10 years.No difference in recurrences was observed with respect to the gender of the patients and the site of the ABC.SAE group showed aF I G U R E 3The propose algorithm for treatment of histology proven primary ABCs is shown.*Safely curettage is advisable in lesion located in the upper or lower limb with a low risk of intraoperative bleed or nerve damage.ABC, aneurysmal bone cysts; SAE, selective arterial embolization.
Comparative overview of the literature.
35We previously consider the injection with bone marrow concentrated, demineralized bone matrix and PRF for treating ABCs, unfortunately they have not shown lower local recurrence rates.T A B L E 2a Final healing rate independent from the number of SAE.CEVOLANI ET AL.|1433