Preoperative Stereotactic Body Radiotherapy for Metastatic Bone Disease with Impending Fracture: A Nonrandomized Clinical Trial

Abstract

Background:

Surgical stabilization followed by postoperative radiotherapy is the standard of care for impending pathologic fractures due to bone metastasis. Compared with conventional radiotherapy, stereotactic body radiotherapy (SBRT) provides more durable tumor control and palliation, but treatment volume constraints often preclude postoperative use. Preoperative SBRT limited to the tumor volume offers a solution by delivering an ablative dose before intraoperative tumor dissemination, but its safety is not established.

Methods:

We conducted a phase I nonrandomized clinical trial at a quaternary care cancer center with accrual between June 2021 and April 2024. Eligible patients (N=38) had pelvic or long bone metastases with impending pathologic fracture requiring surgical stabilization. Patients received preoperative SBRT (27–30 Gy in 3 fractions or 18–24 Gy in 1 fraction) followed by surgical stabilization within 1 week. The primary endpoint was rate of wound complications within 6 weeks after preoperative SBRT and surgical stabilization. Secondary endpoints were rates of tumor recurrence and grade ≥3 treatment-related toxicities within 1 year.

Results:

Seven patients were excluded due to inability to receive treatment or death prior to primary endpoint. Thirty-one patients were evaluated at 6 weeks postoperatively, 26 at 3 months, and 16 at 1 year. Median age was 66 (interquartile range, 59–71) years and 15 patients (48%) were male. At 6 weeks, 2 of 31 patients (6.5%; 95% confidence interval [CI], 0.79%–21.4%) had experienced wound complications. At 3 months, 1 of 26 patients (3.8%; 95% CI, 0.1–19.6%) had experienced grade ≥3 treatment-related toxicity. One of 16 patients (6.3%; 95% CI, 0.16%–30.2%) had tumor recurrence at 12 months.

Conclusions:

Preoperative SBRT is associated with a low risk of wound complications and with durable local tumor control. A randomized controlled trial examining its efficacy, potential to facilitate less-invasive surgery, and impact on functional outcomes is warranted.

1. Introduction

Bone metastases that lead to pathologic fracture are a substantial source of morbidity among patients with advanced cancer.¹ Prophylactic surgical stabilization followed by postoperative radiotherapy (RT) is the standard of care for impending pathologic fracture.^2,3,4–8 While surgical fixation provides immediate skeletal stability, RT is intended to provide tumor control and pain palliation. Failure of tumor control can threaten the fixation of surgical implants and therefore the durability of palliation and functional outcomes.⁹ Although most postoperative RT failures occur within the treatment field, 29% occur outside the field but adjacent to surgical hardware, presumably because of tumor seeding within the bone during surgery.¹⁰ Expansion of the postoperative radiation field to include the entire surgical hardware is associated with a lower rate of local tumor recurrence.¹¹

Stereotactic body radiotherapy (SBRT) provides more durable tumor control and faster, longer-lasting pain palliation than conventional radiotherapy (CRT), particularly for more radioresistant histologies (e.g., renal cell carcinoma, non-small cell lung carcinoma). In a randomized trial of patients with appendicular metastases not requiring surgical stabilization, SBRT recipients had better pain response than CRT recipients, both immediately (62% vs. 36% at 2 weeks) and in the long term (77% vs. 46% at 9 months), and improved local control (100% vs. 91% at 1 year).¹² SBRT is well tolerated, with reported risks of 1%–15% for grade ≥3 toxicities¹³ and 8.5% for fracture, though it is hard to separate the effects of SBRT and the underlying tumor on fracture risk.¹⁴

While SBRT and whole-hardware radiation both have advantages, they often are mutually exclusive postoperatively due to anatomical and technological constraints on treatment volumes. However, preoperative RT has been shown to reduce field sizes and the potential intraoperative dissemination of microscopic disease for primary extremity tumors.^15,16 Furthermore, preoperative RT avoids hardware-related artifacts in CT-based treatment planning, which can introduce dosimetric inaccuracies.^17,18 Theoretically, a novel paradigm of preoperative SBRT followed by surgical stabilization would resolve these conflicting treatment goals without increasing treatment-related toxicity. However, stabilization of impending fractures is time-sensitive, and the safety of surgery promptly after preoperative SBRT, specifically regarding wound healing, is uncertain.

In this trial, we investigated the safety of preoperative SBRT followed by surgical stabilization within 1 week in the setting of impending fracture due to osseous metastasis.

2. Methods

2.1. Study Design

This was a single-institution phase I trial at a quaternary cancer center. Patients with metastatic disease in the pelvis or long bones and impending fractures requiring surgical stabilization were eligible. Recruitment began in June 2021 and ended in April 2024. Enrolled patients underwent expedited radiation simulation and planning, and then received SBRT to the target lesion. Within 1 week of completing SBRT, patients underwent prophylactic surgical stabilization of the impending fracture. They were then closely monitored for wound complications, tumor recurrence, and any Common Terminology Criteria for Adverse Events (CTCAE) grade ≥3 treatment-related adverse events. Radiographs were obtained at each postoperative study timepoint, and any available oncologic systemic surveillance imaging that included the target lesion was reviewed. Any radiographic suspicion for local recurrence was further investigated. This study was approved by our institutional review board (protocol 21–216), and patients provided written informed consent.

2.2. Participants

Patients were eligible if they were ≥18 years old, had a tissue diagnosis of cancer, and had a metastasis in the pelvis or long bones at risk for pathologic fracture that required prophylactic stabilization. Candidate patients were evaluated by an orthopaedic surgeon and deemed at risk for fracture if they exhibited high-risk radiographic features (e.g., cortical lysis, pertrochanteric location, large tumor size) and/or concerning signs on examination (e.g., pain with weight-bearing). As per the authors’ typical practice, a Mirel’s score¹⁹ threshold was not used, but rather fracture risk and urgency were clinically estimated using clinical, radiographic, and systemic factors. Other inclusion criteria included life expectancy >3 months, ability to tolerate radiation simulation and immobilization, and target lesion in the humerus, radius, pelvis, sacrum, femur, or tibial diaphysis.

Patients were excluded if they had diffuse bone metastasis, tumor size incompatible with SBRT, prior RT or surgery at the target lesion, anticipated skin dose to the planned surgical incision ≥9 Gy, or an imminently impending fracture requiring immediate surgery. Within 30 days before radiation simulation, patients had cross-sectional imaging delineating the target lesion and radiographs of the entire bone.

2.3. Procedures

2.3.1. Radiation

Following enrollment, patients underwent computed tomography (CT) simulation. Immobilization was performed using a device from CDR Systems (Alberta, Canada). The gross tumor volume (GTV) was contoured by a radiation oncologist. The clinical target volume (CTV) was defined as a 0.5-cm expansion of the GTV. For tumors without extraosseous extension, the CTV was confined to the bone. The CTV was expanded by 0.2–0.3 cm for the planning target volume (PTV).

Patients received 1 of 2 RT dose-tiers delivered in either 1 or 3 fractions. We limited the regimens to 3 fractions to avoid delays in surgical stabilization that would increase the risk of fracture completion. The first dose-tier (BED₁₀ of 50.4–51.3, α/β =10) was delivered as 27 Gy in 3 fractions or 18 Gy in 1 fraction and used for tumor histologies considered to have intermediate or high radiosensitivity (e.g., non-small cell lung cancer). The second dose-tier (BED₁₀ of 60–81.6 Gy, α/β =10) was delivered as 30 Gy in 3 fractions or 20–24 Gy in 1 fraction and used for histologies considered radioresistant (e.g., renal cell carcinoma, cholangiocarcinoma). Regimens were determined by the treating radiation oncologist based on histology, tumor volume, and anatomic structures in the radiation field. Single-fraction regimens were given preference when clinically appropriate to expedite surgical stabilization of lesions requiring more urgent intervention. We used routine cone beam CT and intrafractional tumor tracking to ensure accurate treatment delivery.

2.3.2. Surgery

To minimize vulnerability to pathologic fracture, patients were asked to avoid weight-bearing on the affected limb prior to surgery, and surgical stabilization was performed within 7 days of RT completion. The type of surgery was determined by the treating orthopaedic oncologist based on tumor and patient characteristics. Examples of stabilization include cement augmentation, intramedullary nail fixation, plate and screw fixation, and joint arthroplasty. At the surgeon’s discretion, stabilization was sometimes combined with partial or total lesion excision. Placement of a surgical drain, wound closure, and dressing were performed according to the surgeon’s standard practice. Per our institutional practice, perioperative antibiotics were administered until the drain had been discontinued, or for 24 hours postoperatively if no drain was required. Surgery that fully exposed the tumor and bone was classified as open, while percutaneous surgery that involved fluoroscopically-guided instrumentation without open exposure was classified as minimally invasive. We did not alter preoperative or postoperative management based on the surgical approach.

2.4. Outcomes

The primary endpoint was the rate of major wound complications within 6 weeks of surgery. This included surgical revision of a wound, surgical site infection (documented by physical examination, laboratory evaluation, and/or imaging), readmission for wound care, or prolonged (>3 weeks) active wound care. Secondary outcomes included rates of local recurrence at 3, 6, and 12 months; rates of CTCAE grade ≥3 adverse events at those timepoints; incidence of pathologic fracture completion between protocol enrollment and surgery; and incidence of delay of surgery due to RT or its complications. As an exploratory analysis, we calculated the anticipated radiation volume that would have been required for postoperative radiation encompassing the entire implant and compared this to the actual volume delivered preoperatively to the tumor. We defined the postoperative GTV by contouring the extent of bone containing surgical hardware on postoperative imaging, expanded this by 2 cm, and removed in-air regions of the expansion to define the PTV.

2.5. Statistical Analysis

For the primary endpoint, a decision rule with 30 evaluable patients was implemented, with each patient contributing a binary endpoint (event vs. no event) within 6 weeks. Patients who died or were lost to follow-up without the event within 6 weeks of surgery were not counted. A 2-stage decision rule similar to the minimax Simon 2-stage design²⁰ was used to investigate the event rate. If there were >4 events among the first 17 evaluable patients or >5 events among the 30 evaluable patients, the proposed treatment would be declared unsafe and the trial would stop. These thresholds were determined to represent acceptable levels of safety, based on historical rates of surgical site infection following radiation.^21–24

For the secondary endpoints, local recurrence rates at 3, 6, and 12 months were evaluated using the Kaplan-Meier method. Patients who died without local progression, were lost to follow-up, or withdrew from the study were censored. For the 12-month timepoint, the sole local recurrence on study was not captured in the Kaplan-Meier analysis because diagnosis occurred later than 12 calendar months but within the 12-month clinical follow-up timeframe prescribed by the protocol. We therefore reported an additional analysis for local recurrence using binary outcomes and exact Clopper-Pearson confidence interval (CI) at the 12-month timepoint. We also analyzed local recurrence using a competing-risks framework. Death without local recurrence was a competing risk. The rate of grade ≥3 toxicity was estimated by sample proportions and 95% CIs. Median on-study follow-up was estimated using the reverse Kaplan-Meier method. We used paired Student’s t-test to compare preoperative and theoretical postoperative radiation volumes.

3. Results

3.1. Baseline Characteristics

Thirty-eight eligible patients enrolled in the trial. Seven were excluded from analysis because they could not be evaluated for the primary endpoint: two before SBRT, 2 after SBRT but before surgery, and 3 after surgery but before 6-week follow-up (Figure 1). Table 1 lists baseline patient and tumor characteristics. The median age of evaluable patients was 66 (interquartile range [IQR], 59–72) years; 16 (52%) were female. The most prevalent histologies were lung cancer (n=7), renal cell carcinoma (n=6), and breast cancer (n=5). The most common locations of bony metastasis were the proximal femur (n=15) and the femoral and humeral shafts (n=5 each). Figure 2 shows a Kaplan-Meier plot of overall survival.

Figure 1.

TREND participant flow diagram

Table 1.

Baseline Characteristics of Enrolled Patients

Characteristic	No. (%) or Mean (IQR)
	Included in Analysis (N=31)	Excluded from Analysis (N=7)
Age, years	66 (59–72)	53 (47–67)
Sex
Female	16 (52)	4 (57)
Male	15 (48)	3 (43)
Tumor histology
Lung	7 (23)	3 (43)
Renal cell	6 (19)	1 (14)
Breast	5 (16)	0 (0)
Sarcoma	3 (10)	1 (14)
Thyroid	2 (6)	0 (0)
Prostate	2 (6)	0 (0)
Other	6 (19)	2 (29)
Location of metastasis
Proximal femur	15 (48)	3 (43)
Femoral shaft	5 (16)	2 (29)
Humeral shaft	5 (16)	1 (14)
Proximal humerus	3 (10)	1 (14)
Distal humerus	2 (6)	0 (0)
Acetabulum	1 (3)	0 (0)

IQR = interquartile range.

Figure 2.

Overall survival among evaluable patients

Among patients receiving first dose-tier SBRT, 6 underwent treatment with 27 Gy in 3 fractions and 7 with 18 Gy in 1 fraction (Table 2). Among those receiving second dose-tier SBRT, 8 received 30 Gy in 3 fractions, 5 received 20 Gy in 1 fraction, and 5 received 24 Gy in 1 fraction. Supplementary Table 1 reports frequency of regimens by histology. The median time between RT simulation and treatment initiation was 7 days. Median time between completion of radiation and surgical stabilization was 2.6 days. Eleven patients had open surgery and 20 minimally invasive surgery. Eight patients underwent gross total tumor resection, and 23 had partial or no tumor resection. Median on-study follow-up was 373 days (95% CI, 353–391).

Table 2.

Treatment Characteristics of Patients Included in Analysis

Characteristic	No. (%)
Radiotherapy regimen (Gy × fractions)
Standard dose-tier
9×3	6 (19)
18×1	7 (23)
High dose-tier
10×3	8 (26)
20×1	5 (16)
24×1	5 (16)
Surgical approach
Open	11 (35)
Minimally invasive	20 (65)
Tumor resection
Gross total	8 (26)
Partial/none	23 (74)

3.2. Primary Outcome

At 6 weeks, 2/31 patients (6.5%; 95% CI, 0.79%–21.4%) had experienced major wound complications. One had delayed healing of a percutaneous incision over the ischium that had been used to instrument the target acetabular lesion. The wound was treated with dressing changes until healed (at the 6-month exam). The other patient developed an infection at the entry site for a humeral nail; it was initially treated with oral antibiotics and wound care but required surgical debridement at 3 months postoperatively. Both wounds were outside of the radiation field and the skin dose was 0 Gy.

3.3. Secondary Outcomes

The local recurrence rate was 0% at 3 and 6 months. At 12 months, 1 of the 16 patients who remained alive experienced local recurrence (6.3%; 95% CI, 0.16%–30.2%). This patient had been diagnosed with an undifferentiated pleomorphic sarcoma metastasis in the proximal femur and treated with 24-Gy SBRT (in 1 fraction) followed by prophylactic plate stabilization. The recurrence, at the distal margin of the radiation field, was asymptomatic, noted on routine surveillance x-ray (Figure 3A), and confirmed with positron emission tomography (PET)/CT (Figure 3B). It was treated with image-guided percutaneous thermal ablation and remained controlled at the latest (33-month) clinical follow-up. As the recurrence was within the span of the stabilization hardware, there was no skeletal instability and no revision surgery was required. Figure 4 shows the cumulative incidence of local recurrence. The estimated cumulative incidence at 12.3 months—the time of the sole local recurrence—was 3.8% (95% CI, 0.25%–17%)

Figure 3.

Cumulative incidence of local recurrence

Figure 4.

(A) Poststabilization surveillance x-ray and (B) PET Maximum Intensity Projection image demonstrating local recurrence

The incidence of treatment-related grade ≥3 adverse events was 3.8% (95% CI, 0.1%–19.6%) at 3 months, 0% (95% CI, 0%–16.1%) at 6 months, and 0% (95% CI, 0%–20.6%) at 12 months, rates that fell within our prespecified threshold of acceptability. The sole event, at 3 months, was the aforementioned wound complication requiring debridement. Grade ≥2 complications are listed in Supplementary Table 2.

One patient sustained a pathologic humeral fracture after SBRT during sterile preparation of the arm in the operating room (a known complication of surgery). No surgeries were delayed because of RT or its complications.

The mean anticipated PTV associated with postoperative radiation (1,573.6 cc, standard deviation [SD]=586.3) was significantly greater than the preoperatively delivered volume (122.6 cc, SD=70) (p<0.001).

4. Discussion

This is the first study, to our knowledge, to demonstrate the safety and feasibility of a novel paradigm for treating metastatic bone disease with preoperative SBRT followed by short-interval surgical stabilization. Despite the use of higher biologically effective doses than in conventional paradigms and a short interval between radiation and surgery, our rate of postoperative wound healing complications (6.5%) was within the threshold specified by the predetermined treatment rule, demonstrating sufficient safety for our approach to advance to a higher-phase clinical trial. Notably, both wound complications were outside the radiation field and therefore likely unrelated to the preoperative SBRT but were included to ensure we were capturing the worst-case scenario. While data for appendicular metastases are lacking, one study found infection rates of 6% for spine surgery after image-guided RT and 13% after CRT (difference nonsignificant).²¹ A separate study reported an 8% rate of delayed healing after SBRT.²² For patients undergoing en bloc resection for extremity sarcoma, the rate of wound complications was higher among those receiving preoperative radiation (35%) than those receiving postoperative radiation (17%).¹⁵

For patients surviving 12 months, our local recurrence rate was 6.3%, which compares favorably with rates reported with conventional postoperative RT (17%–25%),^10,11 and is similar to reported control rates with postoperative SBRT of 95% at 6 months and 87% at 2 years.¹⁴ Although this finding should be interpreted cautiously, it is promising considering that most tumors in our study (71%) were of radioresistant histologies and 74% were not excised. The recurrence rate is not entirely surprising, as similar results have been reported in the literature on spinal SBRT, including 1-year cumulative incidence of local failure of 4.6% for renal cell carcinoma²⁵ and 9.5% for a heterogenous group of tumors, 77% of which were of radioresistant histology.²⁶

The durable control in our study may be due to the combination of higher BED (compared with conventional postoperative RT) and treatment of the entire active tumor volume. SBRT improves local control and palliation more effectively than CRT,¹² and escalating the BED in an SBRT treatment plan can provide faster, more durable pain relief²⁷ and further increase local control.^26,28 For instance, in a randomized controlled trial of SBRT for bone metastases, treatment with 24 Gy in 1 fraction (BED = 81.6 Gy) resulted in a 5.8% rate of local recurrence at 3 years, compared with 22% after 9 Gy in 3 fractions (BED = 51.3 Gy) (p=0.005).²⁸ Although grade 3 toxicity was higher with increased BED (7.8% vs. 3.9%), the difference was nonsignificant (p=0.49). Nevertheless, further studies are required to validate the efficacy of this approach.

The rate of high-grade adverse events in our trial was low. While we believe a 3.8% rate of grade ≥3 radiation-related complications was acceptable in this setting, the average rates are poorly defined in the literature.²⁹ Our protocol allowed two dose-tiers of SBRT, reflecting the risk-benefit decision that radiation oncologists make in clinical practice, since dose escalation improves tumor control²⁸ but may increase the risk of radiation-related adverse events. Radiation volume is a contributing factor to toxicity. We found that the treatment volume for preoperative SBRT was less than the anticipated volume for whole-hardware postoperative CRT.

The protocol did not prescribe a standard surgical treatment, as no consensus exists even amongst experienced orthopaedic oncologists regarding specific anatomic locations.³⁰ This resulted in various approaches to surgical tumor excision and mode of stabilization, and consequently to variation in surgical invasiveness or morbidity. We could not assess the impact of these factors on treatment outcomes due to the absence of surgical hardware failure and low local recurrence rate. However, it is reassuring that despite the high percentages of radioresistant tumors and patients with high-risk anatomic sites, such as the proximal femur (48%), most patients were able to avoid the additional morbidity of complete tumor resection or open surgical stabilization (which were performed for only 26% and 35% of patients, respectively). While a comparison of functional and oncologic outcomes associated with various surgical techniques is beyond the scope of this study, our findings suggest preoperative SBRT coupled with less-invasive surgery could be a safe paradigm that warrants further study. While surgery may achieve the dual goals of tumor ablation and skeletal stabilization, the durability of local control provided by histology-specific preoperative SBRT may allow surgeons to decouple these goals and pursue less-invasive interventions focused on skeletal stabilization alone.

4.1. Limitations

This study had several limitations. It was a nonrandomized, uncontrolled trial focused on safety and not designed to assess efficacy. Patient attrition was high early in the study due to early cancer progression–related deaths. This reveals an important limitation in patient prognostication and selection, as those with very short life expectancies would not be expected to benefit from this intervention.

The low complication rate should be interpreted cautiously. The inclusion criteria and clinical decision-making within the study may have introduced selection bias for patients likely to respond well to this treatment paradigm. Most patients had minimally invasive surgery, and the complication rate might have been higher with more invasive surgery. Likewise, it is possible that higher single-fraction SBRT doses would have increased the surgical complication rate, and this endpoint will need to be monitored in future trials. In addition, the high rate of early attrition may have selected out patients who were at elevated risk for complications.

Given the heterogeneity of anatomic locations and types of surgical interventions, we did not gather data on or compare functional outcomes. Subsequent studies with narrower inclusion criteria would benefit from evaluation of functional outcomes. Due to the low incidence of tumor recurrence in this study, we could not evaluate the potential additive impact of surgical tumor excision on local control and seeding. While this trial demonstrates that clinician judgment associated with a preoperative SBRT paradigm is safe and potentially effective, a phase III trial would provide the opportunity to randomize patients to these treatment elements.

4.2. Conclusion

In this nonrandomized clinical trial, patients with bone metastases treated with preoperative SBRT followed by short-interval surgical stabilization experienced low rates of wound complication, local recurrence, and adverse events. This paradigm has the potential to reduce treatment volumes and provide more rapid, complete, and durable palliation to this critical patient population. It warrants further investigation in a randomized controlled trial.