Pelvic Insufficiency Fracture After Pelvic Radiotherapy for Cervical Cancer: Analysis of Risk Factors

Oh D. IJROBP 2008;70:1183

To investigate the incidence, clinical characteristics, and risk factors of pelvic insufficiency fracture (PIF) after pelvic radiotherapy (RT) in cervical cancer. Medical records and imaging studies, including bone scintigraphy, CT, and MRI of 557 patients with cervical cancer who received whole-pelvic RT between January 1998 and August 2005 were reviewed.

Results

Eighty-three patients were diagnosed as having PIF after pelvic RT. The 5-year cumulative incidence of PIF was 19.7%. The most commonly involved site was the sacroiliac joint. Pelvic pain developed in 48 patients (57.8%) at diagnosis. Eleven patients (13.3%) needed admission or narcotics because of severe pain, and others had good relief of symptoms with conservative management. In univariate analysis, age ≥55 years (p < 0.001), anteroposterior/posteroanterior parallel opposing technique (p = 0.001), curative treatment (p < 0.001), and radiation dose ≥50.4 Gy (p = 0.005) were the predisposing factors for development of PIF. Concurrent chemotherapy (p = 0.78) was not significant. Multivariate analysis showed that age ≥55 years (p < 0.001), body weight <55 kg (p = 0.02), curative treatment (p = 0.03), and radiation dose ≥50.4 Gy (p = 0.04) were significant predisposing factors for development of PIF.

Conclusion

The development of PIF is not rare after pelvic RT. The use of multibeam arrangements to reduce the volume and dose of irradiated pelvic bone can be helpful to minimize the risk of fracture, especially in elderly women with low body weight.

Discussion 

Radiation-induced bone injury occurs in clinical conditions such as radio-osteonecrosis, insufficiency fracture, abnormality of bone growth, and development of secondary malignancy. Emami  reported a tolerance dose (TD5/5–TD50/5) of mature bone as 60–77 Gy for radio-osteonecrosis of mandible. However, there have been no data for insufficiency fracture, and some have reported that the incidence of insufficiency fracture increased above the threshold dose of 40 Gy. In conventional pelvic RT, irradiated dose of the pelvic bone is usually 45–50 Gy, and the development of PIF after pelvic RT at this level has been considered a rare complication, especially in the era of mega-voltage equipment

However, several recent studies showed that the incidence of PIF after pelvic RT might have been underestimated in gynecologic patients. They reported the cumulative incidence of symptomatic PIF at 5 years as 8.2–17.9%. In our series, the cumulative incidence of PIF was 19.7% at 5 years in all patients and 11.1% in symptomatic patients, which was in accordance with other recent studies This higher-than-previously-believed incidence may be due to wider use of imaging modalities (e.g., CT, MRI, and bone scintigraphy) that are more sensitive to detect PIF than conventional radiography. Blomlie evaluated the incidence of PIF using MRI prospectively and showed that 89% of patients had findings compatible with PIF after pelvic RT. Abe  showed a 34% prevalence of PIF after pelvic RT using bone scintigraphy. We tried to perform CT or MRI during the follow-up and bone scintigraphy at least one time per year, so as to detect asymptomatic patients (35 of 83, 42.2%) with PIF. The characteristics of irradiated patients can affect the incidence of PIF. As revealed in our study, older patients receiving pelvic RT are more susceptible to the development of PIF. In our study, the cumulative incidence of PIF at 5 years in patients aged ≥55 years was 31.8%. In the study of Ogino  PIF patients were all postmenopausal, and the Baxter  included patients aged >65 years.

The sacral ala adjacent to SI joints is the most commonly involved site of PIF. Solitary pubic bone or acetabulum fracture is rare (none in our study), and most lesions are accompanied with SI joints. This indicates that initial mechanical failure of the sacrum causes other subsequent pelvic bone fracture. In addition, a single site of fracture can progress to multiple sites. In our study, initial single-site fracture subsequently progressed to multiple fractures in 28.1% of patients (9 of 32).

As has been reported by many investigators, our study showed that symptoms of most patients were mild and resolved after conservative management based on analgesics and rest, but some patients (13.3%) had severe pain and were managed with narcotics or admission. Extent of lesions may correlate with severity of symptoms. Blomlie showed that all patients without pain had smaller lesions (<1 cm2) on MRI and indicated the small fractures might be not painful. In our study, symptomatic patients were more likely to involve multiple sites of pelvic bone than asymptomatic patients.

The risk factors of osteoporosis are closely correlated with the development of PIF. Blomlie  showed that 95% of patients with PIF reported in the literature were postmenopausal women. Ikushima reported that the mean age of patients who developed PIF was significantly higher than that of other patients (69 years vs. 59 years). Ogino reported 57 cases of PIF after pelvic RT in postmenopausal women. They showed that low body weight (≤49 kg) and more than three deliveries were significant factors for the development of symptomatic PIF. In our study both low body weight (<55 kg) and older age (≥55 years) were significant predisposing factors for PIF in multivariate analysis. Many medical illnesses or medications, such as rheumatoid arthritis, hyperthyroidism, and corticosteroids, are also reported as risk factors for osteoporosis. Ogino investigated Type II diabetes mellitus and showed no significant correlation with PIF. In our study 1 patient had a history of rheumatoid arthritis, and 5 had a history of hyperthyroidism; none of them developed PIF, so we did not perform statistical analysis.

It is well known that radiation toxicity is strongly correlated with irradiated volume and dose. In our study, both the four-field box technique and the AP/PA parallel opposing technique were used. In the four-field box technique, lateral portals could spare the irradiated volume of small bowel and rectum and also spare the irradiated volume of the posterior portion of the sacrum and SI joints. The incidence of PIF was higher in patients receiving the AP/PA technique (35.8% vs. 17.1% at 5 years) in univariate analysis, but there was no significant difference between the two techniques in multivariate analysis. This volume effect of the sacrum could be confirmed more accurately by dose–volume histograms of the irradiated sacral bone, but we could not get this information because we did not use CT planning.

Patients who received more irradiated dose to the sacrum had more risk of PIF (21.7% vs. 2.1% at 5 years, p = 0.005) in both univariate and multivariate analysis. It was uncertain whether the small difference (median 45 Gy vs. 50.4 Gy) could affect the development of PIF, but there might be a threshold dose for PIF at approximately 45 Gy, as reported by Fu . Our results showed that patients receiving curative RT had more risk of PIF than patients receiving postoperative adjuvant RT (30.1% vs. 11.1% at 5 years, p < 0.001). The patients who received curative RT received an additional dose of high-dose-rate intracavitary brachytherapy. Fu  calculated the brachytherapy dose contribution to the pelvic bone and estimated it to be approximately 10% of the central brachytherapy dose. It was uncertain whether this small additional dose of sacral bone by high-dose-rate intracavitary brachytherapy could be one of the causes of this difference.

Concurrent chemotherapy is used frequently in gynecologic cancer for increasing tumor control, but it is well known that it also increase the toxicity of radiation. Thus many investigators have thought that combined therapy with chemotherapy might increase the risk of PIF, but there have been few studies to evaluate this. One report showed that combined treatment with radiation and chemotherapy might predispose to pelvic fracture in cervical cancer. In our study, although we had the limitation of being a retrospective study, we showed that the use of cisplatin-based concurrent chemotherapy did not increase the risk of pelvic bone fracture.

To minimize the risk of PIF, we can make an effort by two approaches. The first approach is to improve the osseous environment by combined treatment of osteoporosis. Bisphosphonate has been used as an effective agent for treatment of osteoporosis, and it has also been shown to be effective to reduce cancer-induced bone loss in several studies. Further study is required to determine whether it can reduce the risk of PIF in patients with high-risk factors such as older age and lower body weight.

The second approach is to reduce radiation toxicity. As mentioned above, our results indicated that irradiated volume and dose to sacrum and SI joints might correlate with the risk of PIF. Ogino suggested that a multibeam arrangement by CT planning could shield the posterior portion of the sacrum and SI joints without inadequate coverage of the target volume. Intensity-modulated radiotherapy (IMRT) can reduce the irradiated dose and volume of normal tissue. Many studies showed that IMRT was associated with decreased radiation toxicity. Sacral bone-sparing IMRT may reduce the radiation dose to the sacrum and SI joints. However, it may be difficult to achieve significant sparing to reduce the risk of PIF because of its proximity to the target volume. Radiation-protecting agents such as amifostine (WR2721) also may reduce the risk of PIF. One animal study showed that amifostine improved bone density after RT compared with a control group.

In conclusion, the development of PIF is not a rare complication after pelvic RT when CT-based planning is not used. The use of a multibeam arrangement to reduce the volume and dose of irradiated pelvic bone can be helpful to minimize the risk of fracture, especially in elderly women with low body weight. We will perform a further prospective study in such patients to evaluate dose–volume effects on pelvic bone using three-dimensional conformal RT or IMRT planning and to evaluate whether combined treatment of osteoporosis using bisphosphonate can reduce the risk of development of PIF.