Aneurysmal bone cysts are benign, locally aggressive bone tumors with high rates of recurrence after treatment that usually affect children and adolescents [1, 4, 5]. They most commonly occur as primary lesions in the metaphyses of long bones, though they often also present in the spine . Approximately 70% of ABCs are reported to be primary lesions, while 30% are thought to be secondary to associated pathologies . In approximately 75% of cases there is a balanced chromosomal translocation with involvement of the ubiquitin carboxyl-terminal hydrolase 6 (USP6) gene on chromosome 17p13 [4, 22]. This translocation event increases TRE17 production, which increases matrix metallopeptidase (MMP)-9 and MMP-10 activity . These MMPs are responsible for preventing osteoblast maturation and increasing vascular endothelial growth factor (VEGF) release, which leads to increased vascularization that is characteristic of ABCs .
Preliminary diagnosis can be achieved via imaging [22, 23]. Biopsies allow for more definitive diagnosis, with specimens marked by fibroproliferative stroma, multinucleated giant cells (osteoclasts), hemosiderin laden macrophages and vascular spaces as hallmarks of ABCs [13, 22]. Further diagnostic certainty can be obtained by assaying for a USP6 translocation. Needle biopsy of primarily cystic bone lesions is difficult because specimens have only scant cellular elements from the cyst margins.
Cervical spine ABCs are some of the most challenging of ABCs to treat. Complete surgical resection of ABCs with or without adjuvants is associated with the lowest recurrence rates [1, 22]. Practically speaking, this strategy is limited to smaller, more isolated ABCs such as those in the spinous process or a lamina at one level. This is the minority of cases in this report’s cohort where, on average, the ABCs extended 50% of the way around the spinal canal (Table 1). Most successful resections of anything but the simplest of ABCs still result in spinal instability requiring multilevel spinal fusions and instrumentation with resultant lifelong limited neck mobility . Even with complete resections, cervical spine ABCs have an 8% recurrence rate . Rather than complete resection, curettage with or without bone void fillers or adjuvants, coupled with spinal fixation, remains the standard of care despite its 10–60% recurrence rate [1, 4, 9, 23].
Alternative treatments such as radiotherapy or systemic medical treatment with Denosumab have been reported [1, 23, 24]. Despite their favorable initial outcomes, these treatment modalities have limitations. Radiotherapy is typically used for recurrent ABCs or those in hard-to-reach locations because its longer-term effects remain incompletely known but include myelopathy and secondary malignancies [1, 23]. Denosumab, which blocks osteoclast-stimulated bone resorption, risks development of hypocalcemia and rebound hypercalcemia at the start and end of treatment, respectively [13, 25]. Also, insufficient safety data on systemic use of Denosumab in children warrants further caution when considering it for pediatric ABC treatment because it has potential long-term side effects on the immature skeleton away from the ABC [11, 13]. Denosumab is known to affect the physes of a child’s growing skeleton in addition to the ABC, even after relatively short exposures, resulting in dense metaphyseal bands on radiographs during treatment, and there are insufficient data to know what, if any, more permanent physeal injury could occur with longer medication exposure . Last, recurrence risk of the ABC following cessation of Denosumab is also unknown.
The literature cites sclerotherapy with various agents, including Ethibloc, polidocanol, or calcitonin and methylprednisolone, as a potential treatment option. Ethibloc is no longer available, polidocanol has not been approved for use in the United States, and methylprednisolone can worsen lesions [1, 4, 15]. A recent report cited success in adding transarterial embolotherapy to sclerotherapy for treatment of unresectable or recurrent ABCs . Our only institutional experience with embolotherapy in ABC management in the spine is as a preoperative measure prior to excision of extraosseous disease related to a massive lumbar spine ABC. Also, while embolotherapy has been shown to be a useful adjunct to diminish intraoperative bleeding, it has not significantly reduced recurrence risk following the surgery . Embolotherapy of cervical spine ABCs would add significant risks and challenges to sclerotherapy alone, with questionable gains over what we have demonstrated in this cohort.
Doxycycline has been presented as a sclerotherapy agent for spinal ABC treatment [22, 23, 28]. Doxycycline enhances bone formation by osteoblasts and inhibits MMPs, angiogenesis and osteoclast bone resorption . Its effects should therefore counter some of those resulting from increased TRE17 production inherent in ABC pathophysiology [4, 28]. Prior studies have demonstrated its success as a sclerotherapy agent treating primarily appendicular ABCs, with complication rates <5% and recurrence rates between 6% and 11% [15, 16]. Most patients in those studies were clinically asymptomatic after doxycycline treatment. More recently, studies have assessed the success of doxycycline sclerotherapy as a treatment for spine ABCs [22, 28]. Although these studies comprised relatively few patients, the authors concluded that doxycycline sclerotherapy was a viable treatment for ABCs in difficult locations [22, 28]. Our study supports this conclusion with an 86% cure rate in treatment of 14 cervical spine ABCs using image-guided doxycycline sclerotherapy.
Sclerotherapy of spinal ABCs, particularly cervical spine ABCs, requires a higher level of attention to technical detail than most extremity ABCs. Targeting the many cystic areas in the ABCs requires accurate needle placement and sclerosant injection to avoid non-target tissue injury given a higher density of sensitive tissues near cervical spine ABCs than in most extremity ABCs (Fig. 1). We find that double-needle technique is necessary in all ABC treatments to prevent pressure build-up within the ABC and leakage into adjacent tissues. Also, test injections with dilute contrast agent prior to sclerosant injections can give a preview of where the sclerosant will go once injected. There is a case report journaling the death of a child after Ethibloc injection into a C2 ABC using single-needle technique ; the authors postulated there was retrograde flow of the sclerosant through a feeding vessel to the vertebral arteries with resultant vertebrobasilar embolization and brainstem infarction.
Certainly, the risk of sclerotherapy in cervical spine ABCs is higher than in extremity ABCs, but the reward for the patient is also higher because successful sclerotherapy obviates morbidity and mortality risks associated with open surgical procedures on the cervical spine. Cervical spine ABCs require more intensive clinical decision-making than extremity ABCs before, during and after treatment. The treating physician needs to be willing to move on to surgical referral for tumor removal and operative fixation in cervical spine ABCs if it appears the treatment is not working or working too slowly and leaving the patient at risk of collapse, paralysis or death. These nuances of cervical spine ABCs with their higher risk-to-reward ratios than many other ABCs is why we present only cervical spine ABCs in this report. We hope that in later reports we can expound on other neuroaxis ABCs we have treated in the skull, thoracic spine, lumbar spine and sacrum.
In this study, 12 of 14 patients were successfully treated with doxycycline sclerotherapy as a standalone therapy. All except one successfully treated patients were asymptomatic at last clinical follow-up. Using doxycycline sclerotherapy, we successfully treated three patients who received prior treatment at outside institutions. Treating spine ABCs can be a much greater challenge when working on a recurrent tumor in a postoperative bed, especially if one must work around fixation devices such as plates, screws, cages, cement and rods, which make targeting difficult on any image-guided procedure.
Of the 12 successful cases we present, 2 patients suffered cyst recurrence identified on follow-up imaging and these were treated successfully in each patient with one additional doxycycline treatment. This highlights the need to follow these patients with imaging for years after successful treatment. Our protocol is to follow all our patients for 5 years after treatment to search for any recurrence of cystic bone destruction. While it is not our routine clinical practice to score our follow-up imaging using numerical modified Neer scoring, we did use it in this paper to give a more precise description of the outcomes. The two reviewers scored all successful cases as Neer 1 or 2 and both failures as Neer 3 or 4, which fits with our more qualitative assessment of success in these patients, which is always a combination of the imaging and any clinical symptoms. Predicting which patients need more sclerotherapy and which can be observed is difficult, and other authors have also presented a grading system that correlates with their outcomes ; the radiographic and CT components of their grading system for ABC healing after sclerotherapy were nearly identical to those in the Neer system we used, though theirs had five grades and ours had four. More important, the other authors found, as we do, that a combined radiographic analysis of healing together with clinical parameters such as persistent pain are needed to predict who would benefit from more sclerotherapy.
Doxycycline sclerotherapy failed in treating 2 of our 14 patients. While we have no definitive explanation for these failures it has been our experience with many other ABCs in all areas of the axial and appendicular skeleton that lesions that appear particularly aggressive by imaging are often just that. Lesions that are very expansile, have innumerable sub-centimeter cysts (Fig. 3), have many fluid-fluid levels indicating repeated episodes of bleeding, and cause greater than average pain from rapid growth and fracture tend to be resistant to therapy. These lesions tend to have brisk bleeding for both the interventional radiologist and the surgeon. For this reason, our practice when working with these lesions is to decrease the interval of serial therapy from 3 months to 2 months in hopes of more effective treatment. Despite this approach, with these two patients one of the lesions did not respond and one nearly doubled in size (Fig. 3) after the first treatment and they were referred for surgery.
The primary limitation of this study derives from its being retrospective, such that all patients were not treated precisely the same, though most procedures performed in the 14 patients, 59/65 (91%), were completed with the same technique and injected agents. The exceptions were detailed earlier in the paper. Three of the patients had failed treatments of different kinds at other hospitals prior to coming to us. The contribution of those prior treatments to the final outcomes cannot be fully known, but in all three cases the patients were ultimately successfully managed with doxycycline sclerotherapy. This treatment variability inherent in a retrospective series such as this one could pollute the data, but we do not believe it substantively alters the conclusion of this paper. Of the 14 total patients in this reported cohort, 9 came for treatment from out-of-town, with most coming from out-of-state or -country. This made obtaining follow-up sometimes difficult, leading to 2 early patients being lost to follow-up immediately following successful completion of doxycycline therapy. Over time we have learned to be more vigilant in getting imaging and clinical follow-up and we hope to soon start a prospective registry to track all ABC patients’ experiences during treatment and follow-up.