Radiographic Rate and Clinical Impact of Pseudarthrosis in Spine Radiosurgery for Metastatic Spinal Disease

Purpose Pseudarthrosis within the spine tumor population is increased from perioperative radiation and complex stabilization for invasive and recurrent pathology. We report the radiographic and clinical rates of pseudarthrosis following multiple courses of instrumented fusion and perioperative stereotactic radiosurgery (SRS). Methods We performed a single institution review of 418 patients treated with non-isocentric SRS for spine between October 2002 and January 2013, identifying those with spinal instrumentation and greater than six months of follow-up. Surgical history, radiation planning, and radiographic outcomes were documented. Results Eleven patients who met criteria for inclusion underwent 21 sessions of spinal SRS and 16 instrumented operations. Radiographic follow-up was 48.9 months; 3/11 (27%) were with radiographic hardware failure, and one (9%) separate case ultimately warranted externalization due to tumor recurrence. SRS was administered to treat progression of disease in 12/21 (57%) procedures, and residual lesions in 7/11 (64%) procedures. Following first and second SRS, 8/11 (73%) and 2/7 (29%) patients were with symptomatic improvement, respectively. Conclusion Risk of pseudarthrosis following SRS for patients with oncologic spinal lesions will become increasingly apparent with the optimized management of and survival from spinal pathologies. We highlight how the need for local control outpaces the risk of instrumentation failure.


Introduction
The management of spinal column metastasis has increasingly moved towards less invasive separation surgery followed by adjuvant stereotactic radiosurgery (SRS). This decompressive approach reduces the post-operative recovery burden while preserving high levels of functional independence [1,2]. However, aggressive surgical stabilization and construct extension is still warranted among destabilizing and multiply-recurrent pathologies for preservation of function. The advantages and disadvantages of these large spinal constructs in already poor quality bone subject to further irradiation, become important in the pre-operative discussion.
Despite its radiobiological stressors, SRS is with known benefits toward pain and disease control [3][4][5]. The movement from adjuvant external radiation therapy (XRT) to SRS for spinal column metastasis provides higher doses with steeper drop-offs limiting direct radiation to the bony matrix and spinal cord [1,6,7]. Nevertheless, the long oncologic course for these patients exposes them to multiple treatment sessions and multiple modalities of radiation that additively undermine desirable fusion. Concerns for decreased periosteal osteoblastic proliferation, decreased vascularity, and increased bony pliancy subject patients to the risks of hardware failure, including implant migration, fusion failure, and biomechanical destabilization-associated pain that can warrant surgical revision [8,9].
Unfortunately, the reduced life expectancies and post-operative tolerances of the overall spinal tumor population has made an assessment of post-SRS pseudarthrosis difficult. Furthermore, published series are limited by their partial reporting of radiation treatment history or followup, leading to a wide range of pseudarthrosis rates that are difficult to compare. Thus, the collective impact of serial, surgical and adjuvant radiation treatments on surgical decisionmaking is absent. Therefore, we assessed patients with perioperative SRS for recurrent spinal column metastasis necessitating instrumentation to qualify how the rates of radiographic and clinical pseudarthrosis guide their oncologic management.

Patient population
A retrospective review of a prospectively collected database of patients treated with nonisocentric SRS between October 2002 and May 2013 was performed. We limited assessment to patients with instrumented fusion with post-surgical SRS radiation, and at least six months of radiographic follow-up. From a total of 418 patients who were treated with SRS, 42 (10%) were with prior surgery and instrumentation, 35 (8.4%) with Karnofsky performance score (KPS) > 70, and 11 (2.6%) patients who met criteria for inclusion.
Primary outcome measured for the rate hardware failure, including radiographic and clinical pseudarthrosis, as well as the rate of operative externalization. Demographic and risk factors including sex, age at treatment, weight, smoking history, osteoporosis, primary disease histology, and KPS at time of index SRS treatment were recorded. All aspects of the study were performed in accordance with relevant guidelines and regulations as were approved by the Stanford Institute Review Board. Written informed consent was not required given the retrospective study format.

Treatment
Surgical and SRS treatment history were characterized for analysis. All instrumented and noninstrumented fusions to the spine were documented. Individual surgical interventions were further assessed for indication, need for instrumentation, time from any preceding SRS or instrumented surgery, location relative to prior surgery, symptoms prior to treatment, response to treatment, surgical approach as well as incorporation of corpectomy, junctional levels, and graft.
Regarding radiation treatment, all sessions with exposure to the spine by either conventional radiation therapy or SRS were documented. Individual SRS interventions, which were all performed at the authors' institution, were further assessed for indication, levels treated, time from index instrumented-surgery, time from prior SRS, location relative to prior SRS, symptoms prior to SRS, and response to treatment. Biological equivalent doses were estimated with an alpha/beta of 10.

Statistical analyses
Differences in means and proportions were determined by Mann-Whitney/un-paired t-test. Probabilistic univariate and multivariate analyses were performed by logistic regressions. Statistical analyses were performed using Stata/SE 13.1 Software (StataCorp, College Station, TX, USA). Statistical significance was targeted for p < 0.05.

Discussion
Current improvements in oncologic care will yield prolonged survival that demands recurrent considerations for instrumented stabilization and risks of revision [10]. Here we characterized the heterogeneous and complex clinical histories of patients with extended follow-up and determine that pseudarthrosis remains a prominent risk, as 27% were with radiographic evidence of lucency. However, a complete survey of their clinical courses suggests that patients with spinal metastasis tolerate aggressive surgical management with perioperative SRS, and may present with progression and urgent decompression needs that outpace competing hardware risks. hardware failure rate at a median 19 months, all with compression fractures perhaps related to long and/or posterior-only stabilization [11,12]. Aside from our broader criteria, a higher pseudarthrosis rate in our series may be attributed to longer follow-up and inclusion of patients' multiple courses of radiation, thereby capturing the lifelong pseudarthrosis rate.
Higher instrumentation failure rates were reported in earlier XRT studies where perhaps dosing regimens had not yet been optimized [8,11]. Emery et al. noted that patients with pseudarthrosis were all with total XRT > 40 Gy, but the statistical significance of this threshold was eliminated once converted to BED [8]. The more contemporary option for SRS with higher prescription doses is largely preferable to XRT to achieve local control [1,7,11]. This could preserve bony integrity and preclude implant externalization, as was necessary in one of our patients and which our BED analysis supports, perhaps obviating stress-imposing extension of constructs as in one of our cases [11].
Most importantly, our review further qualifies the role of perioperative SRS despite hardware risks [13,14]. We see that radiographic pseudarthrosis is more prominent than symptomatic failure. Time to pseudarthrosis occurred at 36, 54, and 117 months, which was longer than the median time to retreatment. The benefits to quality of life are evident in the rate of symptomatic improvements and the rate at which they discharge to home. In the same vein, Laufer et al. reported an optimistic functional status following repeat posterolateral decompression (without radiation), with 65% still ambulatory at the time of last follow-up and with unchanged rates of estimated survival [6].
Our results and prior work will help counsel patients with recurring pathology and spinal surgeries. Local recurrence is estimated to affect a quarter of patients [6,14,15]. Likewise, reoperation is anticipated for 25% of spinal metastasis patients. Although pseudarthrosis currently does not appear to fully correlate with symptomatic failure, there is the future risk that with prolonged survival, patients will increasingly experience pain associated with instrumentation and perioperative SRS. Post-operative radiation is known to have detrimental effects on fusion, particularly in the early post-operative period [16,17]. If patients are anticipated to experience multiple procedures, such as those with favorable pathologies, perhaps more frequent spinal imaging should be obtained in select populations to provide neoadjuvant SRS.

Conclusions
Here we provide a complete longitudinal assessment of patients with instrumented metastatic spinal disease who required re-operation and re-irradiation, not captured in prior works. We suspect symptomatic pseudarthrosis rate is an overestimate of the problem given the number of patients lost to follow-up due to survival. We anticipate with the improvement of individualized therapy and increased availability for SRS treatment, future studies will offer greater follow-up and sample size for statistical assessment. While we reaffirm a genuine risk for pseudarthrosis following perioperative SRS, the associated symptoms are minimal and should not deter aggressive radiation and surgical management if with good prognosis.

Additional Information Disclosures
Human subjects: All authors have confirmed that this study did not involve human participants or tissue. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work.