Insufficiency fractures of the sacrum following stereotactic body radiotherapy for sacral tumors

Abstract

Introduction

There is little data on sacral insufficiency fracture(SIF) incidence following pelvic radiotherapy, with existing studies based on conventional fractionation. Stereotactic body radiotherapy (SBRT), characterized by dose escalation with hypofractionation, may pose even greater risks to sacral integrity. This study aims to define SIF incidence and risk factors following SBRT.

Methods

Records of 43 consecutive patients who underwent sacral SBRT from September 2005-May 2009 were reviewed. Baseline patient information (age, gender, menopausal status, body mass index, use of bone-thinning agents, presence of osteoporosis), tumor characteristics (histology, lesion appearance and extent) and treatment parameters (dose/fractionation, prior radiation/surgery) were documented. Primary end-point was development of new fractures or progression of pre-existing fractures. Secondary end-points included pain scores, analgesic use, functional ability, and local tumor control.

Results

Median follow-up was 17months. Common histologies included sarcoma, renal cell, and prostate carcinoma; 47% of lesions were lytic, 37% sclerotic and the remainder mixed. Doses ranged from 18-24Gy/1fraction to 30Gy/5fractions with 45% receiving single fractions.14% had prior radiation (median dose: 30Gy/10fractions).

Five patients developed SIF. In four, fractures occurred in the context of controlled local disease. Median time to SIF was 8.2months. Symptoms varied from minimal pain requiring no intervention to severe pain impacting on function. Two patients underwent sacroplasty due to intractable pain, with both obtaining good analgesia. Low event numbers precluded meaningful univariate/multivariate analyses. One-year local tumor control rates were excellent (91.7%).

Conclusion

In this study, actuarial SIF incidence at one year was 8.2%, suggesting that SIF risk from sacral SBRT is low. However, larger prospective studies with longer follow-up are needed. In addition, novel therapies such as sacroplasty need further study to determine safety, efficacy and indications for use.

1. Introduction

Sacral insufficiency fractures are amongst the documented late toxicities of pelvic radiation that can be a source of significant morbidity for cancer patients [1,2]. Awareness and recognition of this radiation-induced complication is important for two principal reasons. Firstly, it may clinically mimic osseous metastases leading to erroneous diagnoses and unnecessary investigations and treatment with minimal benefit to the patient and potentially added toxicity. Secondly, it is often associated with significant debilitation which if correctly diagnosed can be appropriately addressed with bed rest, analgesics and in some cases, sacroplasty [3-5].

With the advent of image-guided intensity-modulated radiation therapy or stereotactic body radiation therapy (SBRT), it has become now feasible to treat sacral tumors to even higher doses with relative sparing of surrounding critical structures such as the cauda equina. While this may translate to better local tumor control without significant neural toxicity, the effect of higher radiation doses and larger fraction sizes on the skeletal integrity of the sacrum has not been quantified.

In recent years, the diagnosis of insufficiency fractures of the sacrum has been greatly aided by the increased use of imaging modalities such as bone scans, computed tomography (CT) scans and magnetic resonance imaging (MRI) as part of the routine follow-up care of cancer patients. However, there is a paucity of literature on the exact prevalence of sacral insufficiency fractures following radiation therapy and the available data is somewhat conflicting with some authors quoting rates as low as 3% while others report much higher rates of up to 50%. In addition, the existing studies are based on conventionally fractionated pelvic radiotherapy for rectal, prostate, or gynecologic cancers where the sacrum is uninvolved by malignancy [1,2,6-12]. To our knowledge, there is no published data on the risk of sacral insufficiency fractures following single-fraction or hypofractionated SBRT for primary or secondary malignancies involving the sacrum. Our study aims to address this deficiency by defining the incidence and risk factors for insufficiency fractures of the sacrum in the abovementioned cohort of patients.

2. Methods

This study was approved by the Memorial Sloan-Kettering Cancer Center (MSKCC) institutional review board with a waiver of informed consent. We reviewed the radiation oncology records of 43 patients with primary or secondary sacral malignancies who underwent single-fraction or hypofractionated SBRT to the sacrum between September 2005 and May 2009 at MSKCC. These patients were retrospectively identified from our SBRT database using appropriate search criteria. Inadequate clinical and radiographic follow-up (<3 months) was the sole factor for exclusion.

Baseline patient information such as age, gender, body mass index, menopausal status if female, concomitant use of medications that predispose to osteoporosis (eg. corticosteroids, androgen deprivation therapy and aromatase inhibitors), as well as presence of osteoporosis were recorded. In addition, tumor characteristics (tumor histology, nature of sacral lesion (ie. sclerotic, lytic, or mixed), extent of sacral involvement by tumor) and treatment parameters (site of treatment, dose/fractionation schemes, history of prior local radiation or surgery) were documented.

The primary end-point was development of a new sacral fracture or progression of a pre-existing fracture at the treatment site. To obtain this information, both pre- and post-treatment CT and MRI scans were carefully reviewed in close collaboration with an experienced neuro-radiologist. Where relevant, the event date (ie. the earliest date at which radiologic evidence of an insufficiency fracture of the sacrum was noted) was recorded. Secondary end-points included associated symptoms, visual analogue pain scores, requirement for analgesic medications, invasive procedures, impact on functional status, as well as local tumor control, all of which were obtained by further review of hospital medical records and relevant imaging studies.

3. Results

Baseline patient, tumor and treatment characteristics are shown in Tables 1, 2, and 3 respectively. Common tumor histologies included renal cell carcinoma, prostate adenocarcinoma, and sarcoma (see Table 4). The median age of patients was 64 years and the median follow-up was 17 months.

Table 4.

Histologic sub-types of primary tumor

Primary tumor histologies
Renal cell	7
Prostate	12
Colorectal	4
Breast	1
Melanoma	2
Sarcoma	5
Chordoma	2
Thyroid	1
Lung	2
Others	7

Among the 43 patients who underwent single-fraction or hypofractionated SBRT to the sacrum, 5 developed sacral insufficiency fractures. 2 of these patients had received 24Gy/1fraction, 1 had received 18Gy/1fraction and the remaining two had received 30Gy/5fractions. Figure 1 demonstrates the SBRT plan as well as the insufficiency fracture sustained by one of the patients who received 24Gy/1fraction. SBRT plans of the other 4 patients were also reviewed. In all cases, no hot spots were identified at the site of subsequent fracture development. Only one of these patients had received prior in-field radiation to a dose of 30Gy/10fractions and subsequently, 30Gy/5fractions. In one of these patients, progression of a pre-existing fracture was coincident with florid disease progression at the treated site. In the remaining four patients, fractures occurred in the setting of locally quiescent disease and appeared to be directly attributable to high-dose radiation.

Figure 1.

A. SBRT plan for patient with metastatic leiomyosarcoma involving the sacrum

B. Axial CT scan demonstrating extent of original disease (as indicated by the arrow)

C. and D. Axial and sagittal CT scans showing impacted sacral insufficiency fracture (as indicated by the arrow)

The latency period for sacral fracture development ranged from 4 months to 25 months (median: 8months). Symptoms varied widely from minimal pain requiring no analgesia to severe pain refractory to standard opiate analgesics and impacting on daily functionality. In fact, two out of the five patients who developed sacral insufficiency fractures eventually underwent sacroplasty due to intractable pain inspite of maximal pharmacologic therapy. Both patients obtained good partial relief of pain with improvement in quality of life scores.

Table 1.

Baseline patient information

Age
31-40	1
41-50	5
51-60	10
61-70	12
71-80	13
81-90	2
Gender
Male	32
Female	11
Menopausal status (if female)
Pre-menopausal	1
Post-menopausal	10
Body mass index
< 20	1
20-25	15
> 25	27
Osteoporosis documented in hospital records
Yes	1
No	42
History of prolonged (> 3 months) corticosteroid use
Yes	7
No	36
History of prolonged (> 6 months) androgen deprivation therapy
Yes	14
No	29

Table 2.

Tumor characteristics

Nature of sacral lesion
Sclerotic	16
Lytic	20
Mixed	7
Extent of bone involvement
0-20%	17
21-40%	14
41-60%	4
61-80%	7
81-100%	1

Table 3.

Treatment characteristics

Dose/fractionation
18Gy/1#	2
21Gy/1#	1
22Gy/1#	2
24Gy/1#	14
24Gy/3#	5
30Gy/5#	19
History of prior local radiation
Yes	6
No	37

4. Discussion

Sacral insufficiency fractures are an important late side-effect of pelvic radiation that can be associated with considerable morbidity. They occur as a result of physiologic axial load stresses on weakened atrophic weight-bearing bone that has been compromised by radiation. Radiation-induced bone injury is characterized by both direct effects on the structural and functional components of bone as well as indirect ischemic effects stemming from damage and occlusion of the microvasculature. The end result is demineralization and consequently, decreased elastic resistance and an increased susceptibility to fractures.

Although the clinical presentation may be variable, patients typically present with acute, severe low back, buttock, or pelvic pain often with no antecedent history of trauma. While clinical assessment may be strongly suggestive of an underlying insufficiency fracture of the sacrum, imaging studies play a crucial role in providing a definitive diagnosis. Classic imaging findings have been well documented in the literature. Plain radiographs of the pelvis can demonstrate cortical disruption and fracture lines in the sacrum but subtle findings are often missed [3,4,13]. Bone scintigraphy can reveal the classic butterfly-shaped or “H” pattern produced by vertical fractures of both sacral alae and a horizontal component involving the body of the sacrum [14,15]. However, this distinctive distribution of radiopharmaceutical uptake is only seen in ~30% of patients, limiting its utility in distinguishing fractures from other pathologic processes such as metastases and osteomyelitis.

Characteristic CT findings include fracture lines or areas of sclerosis within the sacral ala parallel to the sacroiliac joints [16]. Frequently, there is anterior cortical disruption of the sacral ala with associated displacement of the fracture fragment. However, CTs may not detect transversely oriented fractures or fractures with a significant horizontal component. On magnetic resonance imaging, presence of fracture lines on T1-weighted images and high signal intensity parallel to the sacroiliac joints on T2-weighted images are virtually pathognomonic of insufficiency fractures [17,18]. While MRIs can detect early changes of sacral insufficiency with a reported sensitivity at or near 100%, CTs provide the most specific information for the detection of these fractures and for the exclusion of recurrent malignancy and are extremely useful in confirming equivocal findings on MR imaging. It is now recognized that CT and MRI scans are most useful in identifying insufficiency fractures of the sacrum, that they provide complementary information and that their combined use is ideal [16-18]. For the patients in our study all of whom had pre-existing neoplastic disease involving the sacrum, either primary or secondary, both CT and MRI scans were reviewed, where available.

While there is some literature on the risk of sacral insufficiency fractures following conventional pelvic radiation, we are unaware of any studies investigating the incidence of sacral fractures following single-fraction or hypofractionated SBRT. The issue of skeletal morbidity following high-dose spine stereotactic radiosurgery has only been recently addressed in a handful of studies [19-21]. One such study performed [19]at our institution demonstrated a staggering 40% risk of vertebral compression fractures. Regression analyses identified extent of vertebral involvement by tumor, presence of lytic metastases, and thoracolumbar and lumbar location as independent predictive factors for fracture development or progression. Of note, lesions located at T10 or below were approximately 4.5 times more likely to fracture than more cranially located lesions presumably due to transmission of axial load forces. Although that study did not include sacral lesions, a logical extrapolation is that the sacrum, which connects the axial skeleton to the pelvis and is subject to maximal weight-bearing forces, would be at least as susceptible to fractures. Interestingly, this was not borne out by the results of the present study which only found a 9.3% incidence of sacral insufficiency fractures attributable to radiation at a median follow-up of 17 months. There are several possible explanations for this discrepancy. Firstly, the dose/fractionation schemes utilized in the two studies were different. More than half of all patients in the current study received hypofractionated treatment (three to five fractions) whereas high-dose single-fraction SBRT was administered to all patients in the study by Rose et al. with a median dose of 24Gy/1#. In addition, a larger proportion of patients in the earlier study had lytic lesions (65%) versus 46% in the current study, a factor known to influence fracture risk. Finally, the smaller number of patients in this study may have potentially contributed to inaccuracies in estimation of fracture risk.

The temporal relationship between the incidence of sacral insufficiency fractures and time of radiation delivery is unknown. The latencies for sacral fracture development following conventionally fractionated pelvic radiation range from 1 to 87 months (median: 13.7 months) [1,2,6-12]. In our study, sacral insufficiency fractures were diagnosed after a median interval of 8 months (range: 4 months to 25 months) following SBRT. As one might expect with ultra-high doses of radiation, we observed shorter latencies to fracture development in our study compared with those involving standard fractionation but it is difficult to make any firm conclusions in this regard given the small patient numbers.

As previously described, there was considerable variation in symptomatology among the study patients who sustained sacral insufficiency fractures from SBRT, ranging from minimal pain requiring no pharmacologic intervention to severe pain refractory to standard analgesics. Two patients in our study underwent sacroplasty for intractable pain, with both obtaining excellent results, with a significant reduction in analgesic use and a return to ambulation. This is supported by the bulk of the available literature on sacroplasty which indicates that it is a highly efficacious procedure associated with rapid and sustained pain relief in well-selected patients [22-25]. In addition, the risk profile appears to be relatively low in experienced hands when performed with rigorous image guidance and an appreciation of the complex anatomy of the sacrum. However, sacroplasty remains a relatively novel technique and is an area of active investigation and development. More robust studies are needed to definitively establish safety, efficacy, as well as indications for its use.

As with all retrospectively performed studies, our study does have some limitations. Firstly, the sample size was small and there were too few events in this study to perform meaningful statistical analyses to assess dose response as well as to identify other risk factors for SIF. Secondly, the follow-up period was highly variable (range: 4.6 to 39.3 months) and the number of follow-up imaging studies for each patient also varied. Further, although the majority of patients had a combination of CT and MRI scans as part of their follow-up care, a significant minority underwent CT scans alone. This may have led to an underestimation of the true prevalence of sacral insufficiency fractures due to the inferior sensitivity of CT scans in detecting subtle or evolving sacral fractures. Thirdly, as with any retrospective study, there was heavy reliance on the availability and accuracy of medical records to evaluate the impact of certain potential confounding factors, such as osteoporosis. Although the presence of documented osteoporosis was recorded, we did not have the means to ascertain the true prevalence or severity of osteoporosis in this group of patients. Finally, our study lacked a control group and hence, comparisons with a similar subgroup of patients who underwent either no treatment or conventionally fractionated radiotherapy were not possible. In addition, both the clinician and radiologist who evaluated all follow-up imaging were aware of the patients’ prior history of SBRT. However, selection bias which is an inherent component of most retrospective reports was to a large extent eliminated in this study as all patients who received single-fraction or hypofractionated SBRT to the sacrum at our institution between Sep 2005 and May 2009 were identified from our meticulously maintained prospective SBRT database and included. In addition, imaging studies were generally obtained at three-monthly intervals prior to each clinic visit as per our pre-defined protocol for SBRT patients, minimizing bias that arises from ad-hoc imaging performed primarily on the basis of symptoms.

Irrespective of the aforementioned limitations, this study provides much needed information in an area which is essentially devoid of data. With long-term survivorship in oncologic patients as a result of advancements in therapeutic strategies, issues affecting quality of life such as skeletal morbidity are gaining increasing importance. Recognition of its occurrence, quantification of risk, and identification of contributory factors are crucial. Studies such as these allow radiation oncologists to carefully weigh the use of relatively novel techniques such as SBRT against the potential burden of treatment-related toxicity and to have frank discussions with patients regarding anticipated clinical outcomes in order to avoid future conflicts stemming from unrealistic expectations and underappreciated adverse effects.

5. Conclusion

In this study, the actuarial incidence of sacral insufficiency fractures at one year was 8.2%, suggesting that SIF risk from sacral SBRT is acceptably low. Nevertheless, it is important to include information about this potential toxicity in the SBRT consent-taking process. Clinical awareness of this entity is also vital as it may mimic bone metastases leading to erroneous diagnoses and unnecessary investigations and treatment. It is also often associated with significant debility which, if correctly diagnosed, can be appropriately addressed with bed rest, analgesics and in some cases, sacroplasty.

In addition, prospective studies with larger patient numbers and longer follow-up are warranted to better characterize the incidence, clinical course and risk factors of sacral insufficiency fractures. Moreover , relatively new therapeutic interventions such as sacroplasty need further study in order to determine their safety and long-term efficacy as well as to establish indications for their use. Other areas that are open to investigation include the utility of medications such as bisphosphonates in the prophylactic setting in reducing the risk of insufficiency fractures of the sacrum among high-risk groups.