In the present study, we found that postradiotherapy SIF was a relatively common occurrence for patients with cervical cancer, and patients have some certain clinical characteristics, such as older ages with postmenopausal status, developed with SIFs within 2 years after RT, involved bilateral sacrum, concomitant fractures. MRI, especially coronal FS-T2W imaging was useful to detect and character SIFs.
The actual incidence of RT-induced is unclear, the incidence in our study was 10.8%, which was relatively higher than previous studies reported between 2.9% and 9.6% [1, 8, 10], this partly ascribed to the fact that SIFs were all detected by MRI. It has been implied that the sacrum is the initial site of fracture and then lead to increased stress on other sites of the pelvis [3, 14]. In our study, almost 53.6% of patients had concomitant fractures, suggested that identification of SIFs should arouse clinical suspicion of fractures in other sites. However, SIFs are most frequently associated with pubis fractures, with a reported coincidence of 78% [25]. We found that only 14.3% patients had coexistent pubis fractures, it was lower than that of the presences of fracture in L5 (14.3%), acetabulum (17.8%), ilium (21.4%).
The clinical presentation of SIF is vague and non-specific. We observed almost two-thirds of SIFs were clinically symptomatic, the incidence was higher than a study reported about one third [22]. First, this may be most likely explainable by the fact that 53.6% of patients with concomitant fractures were included, and 72.2% of the symptomatic patients developed with concomitant fractures. Multiple site fractures may be contributed to some patients with pain [9]. Second, 82.9% of SIFs presented severe bone marrow edema patterns in present study, which was the same as Blomlie et al. [26] revealed, larger lesions (> 1 cm2) on MRI tended to be more likely painful.
Older patients with postmenopausal status after RT are more susceptible to the development of IF [1, 2, 9, 17]. We found that SIFs frequently occurred in postmenopausal patients, and the median age was 60 years. Based on our results, most of affected patients developed with SIFs within 2 years, and the median interval time from RT to SIFs was 10 months, which is almost similar with previous studies with a median interval time between 8 and 14 months [1, 6, 22]. Supporting a previous study demonstrated that 88.9% of patients with RT-induced SIFs were bilateral [21], we observed over two-third of SIFs have arosen bilaterally.
Although SIFs are rarely life-threatening, they should be deserved special attention because they can influence the quality of life [27]. Only 20–38% of SIF could be identified on plain films [28, 29]. Recently, modern imaging modalities during follow-up have been applied in the detection and characterization of SIFs. Both BS and CT have some limitations, BS is one of the most sensitive imaging modalities for detecting SIF and the “H-sign” regarded as a gold standard for diagnosing SIFs[16]. However, this sign is frequently not presented, reported in just 40% of SIFs[29]. In our study, 53.3% (8/15) of patients who had additional BS showed the typical “H-sign”. Although, CT may be useful to ascertain equivocal findings on BS or MR imaging, which has been shown to be less sensitive to detect SIFs than BS or MRI, with a recorded sensitivity between 60 and 75% [11, 28, 30]. In this study, 83.3% of SIFs were detected by CT for patients with had additional CT examinations.
MRI is an alternative technique due to its high soft-tissue contrast, multiplanar imaging, and avoidance of ionizing radiation. MRI is one of the most sensitive imaging techniques to detect RT-induced IF by visualizing the bone marrow edema. Cabarrus et al[11], showed that the overall sensitivity of MRI was significantly higher than CT (100% vs 74.6%), although the sensitivity for detecting fracture lines was similar (93.3% for MRI vs 89.7% for CT). We found all SIFs showed bone marrow edema, and fracture lines were visualized in 64.6% of lesions. Furthermore, MRI is also helpful to differentiate SIF from the bone metastasis, because MRI is very useful to identify the soft-tissue component, the absence of focal or discrete soft-tissue mass around fracture sites is an important sign for distinguishing SIF from malignancy [3, 7, 20, 21]. In present study, no soft-tissue tumor was detected in fracture sites.
The FS-T2W imaging is especially sensitive for visualizing early bone marrow edema, and coronal imaging of sacrum is recommended to be included in suspected cases[3, 30]. Gupta et al. demonstrated that coronal STIR sequence had additional value to the L-spine MRI by increasing significant findings detection in 6.8% of patients, including SIF or sacroiliitis [31]. Compared with previous studies [7, 11, 21, 32], the advantage of our study was that both coronal FS T2-weighted and gadolinium-enhanced T1-weighted imaging were performed for all patients. In line with the previous studies, we found coronal FS-T2WI was more sensitive to detect both bone marrow edema and fracture lines than that of T1WI or enhanced T1WI. In addition, this study revealed that T1WI detected the least fracture lines because of bone marrow edema pattern was also hypointensity, and enhanced T1WI detected least bone marrow edema due to some SIFs with no or mild enhancement, which may be difficult to identify.
There were several limitations to our study. First, our study was retrospectively performed in a single-institution. Second, the follow-up period was inconsistent (range, 25–72 months), which might have resulted in underestimation of the true prevalence of SIF. Third, only a small number of patients undergone simultaneous BS or CT examinations, we were unable to compare the diagnostic ability of MRI with BS or CT. Fourth, none of the IF lesion was diagnosis based on histopathology, as a pathologic diagnosis was generally impractical and actually unnecessary.