Clinical and 18F-FDG PET/CT Imaging Characteristics of Post-radiotherapy Sacral Insufficiency Fractures in Cervical Cancer Patients

Introduction: Sacral Insufficiency Fractures (SIFs) are a common yet frequently misdiagnosed late complication following pelvic radiotherapy for cervical cancer. Accurate differentiation from bone metastases is crucial to avoid unnecessary interventions. While MRI is sensitive for early marrow edema, integrated 18F-FDG PET/CT offers a unique simultaneous assessment of bone metabolism and systemic tumor status. However, comprehensive studies detailing the qualitative and quantitative PET/CT characteristics of post-radiotherapy SIFs are lacking. This study aims to systematically define these features and establish their discriminative value. Materials and Methods: In this retrospective study, we analyzed 32 cervical cancer patients who developed SIFs following pelvic radiotherapy and underwent 18F-FDG PET/CT imaging between January 2018 and January 2024. Diagnosis was based on characteristic radiologic findings, clinical correlation, and a minimum 12-month follow-up. Qualitative (fracture patterns, FDG uptake morphology) and quantitative (SUVmax, SUR-BP ratio, CT densitometry) parameters on 18F-FDG PET/CT were evaluated results: 3.Results 3.1 Clinical features In this study, 32 patients with ages ranging from 47 to 83 years (mean age 63.88±9.38 y, median age 66 y). SIFs were frequently occurred in patients with a postmenopausal status, accounting for 93.7% (30/32) of patients (Figure 1). The median time from the start of radiotherapy to the detection of SIFs on imaging was 14 months (range, 2-32 months), and the great majority of affected patients developed with SIFs within 2 years after RT (90.6%, 29 of 32 patients). Among the patients, 18 (56.3%) presented with low back or hip pain, 5 (15.6%) were asymptomatic and detected during routine oncological surveillance, while the remaining 9 (28.1%) were identified on imaging due to tumor recurrence (including irregular vaginal bleeding in 3 cases, abdominal pain in 3, low back pain in 2, and fecal discharge from the vagina in 1). All 32 cervical cancer patients had no history of trauma or malignant tumors in other systemic sites. Of these, 22 patients underwent hysterectomy with bilateral salpingo-oophorectomy followed by radiotherapy, while 10 patients received radical radiotherapy alone. 3.2 PET/CT findings Among the 32 patients, sacral involvement patterns were as follows: unilateral left sacral ala in 9 cases (28.1%), unilateral right sacral ala in 8 cases (25%), bilateral sacral alae in 12 cases (37.5%), left sacral ala with sacral body involvement in 1 case (3.1%), and bilateral sacral alae with sacral body involvement in 2 cases (6.2%). The concomitant fractures were observed in 9 patients (28.1%) with pubic fractures, 8 cases (25%) with iliac fractures, and 2 patients (6.2%) with bilateral L5 transverse process fractures. The affected sacral vertebrae ranged from S1 to S3, and no associated soft-tissue masses were identified. Regarding imaging characteristics, 28 patients (87.5%) exhibited heterogeneous bone density elevation adjacent to the sacroiliac joint, characterized by ill-defined margins around the hyperdense regions; the remaining 4 patients showed normal bone density. Irregular, serrated, longitudinal fracture lucencies with low density were observed in 22 patients (68.7%). Hypermetabolism was observed in all patients, with the maximum standard uptake value (SUVmax) ranging from 1.5 to 5.1 and an average of 2.45±0.74. In one case, a follow-up PET/CT scan performed two years later demonstrated progressive osteosclerosis with increased bone density and an expanded lesion area. Notably, the SUVmax declined markedly from 3.4 to 1.4 (Figure 2), suggesting reduced metabolic activity. A diffuse or patchy uptake pattern parallel to the sacroiliac joint was seen in 31 patients (96.8%), which did not fully coincide with the areas of increased bone density; the remaining patient presented with focal metabolic hyperactivity. The standardized uptake ratio relative to liver blood pool (SUR-BP) ranged from 1.3 to 2.0, with a mean value of 1.46 ± 0.38. CT analysis revealed significant differences in vertebral body attenuation values between the regions within and outside the SIF radiotherapy field, with mean CT values of 36.8±28.6 HU and 78.0±37.3 HU, respectively (paired t-test: t=-4.950, P<0 xss=removed>max</sub> 0.56 vs. 1.65; arrow, B). Similarly, the 65-year-old female (13 months post-RT) shows more pronounced osteopenia in irradiated vertebrae (5 HU vs. 59 HU; arrow, C) and suppressed metabolic activity (SUV_max 0.70 vs. 2.01; arrow, D), both delineated by the L3-L4 interface. Figure 2 A 72-year-old female with sacral insufficiency fracture (SIF) at 9 months after radiotherapy for cervical cancer, presenting with low back pain for more than 2 months. Axial CT images (A, E, I) showed low-density fracture lines or cortical buckling in the bilateral sacral alae, bilateral transverse processes of L5, left pubic ramus and pubic symphysis; among these, the fracture lines in the sacral alae were serrated and parallel to the sacroiliac joints. Multiple ill-defined patchy slightly hyperdense shadows were observed around the fracture lines. The corresponding axial PET/CT fusion images (B, F, J) demonstrated diffusely mild increased FDG uptake with a SUVmax of 3.4 (arrow, ↑). At the follow-up examination at 32 months after radiotherapy, axial CT images (C, G, K) revealed that the range of abnormal bone changes had significantly enlarged compared with the previous findings, showing multiple punctate and patchy hyperdense shadows with obvious proliferative sclerosis and clear boundaries in the bilateral sacral alae, bilateral ilia, bilateral transverse processes of L5, bilateral pubic rami and pubic symphysis; only the fracture ends of the left pubic ramus and pubic symphysis were obviously separated, while the fracture lines in other locations had disappeared, showing post-repair changes. The corresponding axial PET/CT fusion images (D, H, L) showed a decreased SUVmax of 1.4 (arrow, ▲). Figure 3: A 68-year-old female with SIF at 30 months after radiotherapy for cervical cancer, presenting with hip pain for more than 1 month. At 10 months post-radiotherapy, axial and coronal CT images (A, C) showed decreased bone mineral density of the pelvis, presenting as osteoporotic changes, and the corresponding axial and coronal PET/CT fusion images (B, D) revealed no abnormal increased FDG uptake. Axial and coronal CT images (E, G) at 30 months post-radiotherapy demonstrated irregular low-density fracture lines in the left sacral ala and sacral body, accompanied by multiple ill-defined patchy slightly hyperdense/high-density shadows in the surrounding area, while the corresponding axial and coronal PET/CT fusion images (F, H) showed diffusely mild increased FDG uptake with a SUVmax of 3.0. Figure 4: A 67-year-old female with SIFs at 16 months after radiotherapy for cervical cancer, presenting with vaginal fecal fluid leakage for more than 1 month. Axial CT images (A-C) showed fracture lines or cortical buckling in the bilateral sacral alae and bilateral ilia, with multiple nodular and patchy hyperdense shadows in the surrounding area, presenting as obvious proliferative sclerosis with relatively clear boundaries. The corresponding axial PET/CT fusion images (D-F) revealed multiple scattered patchy areas of mildly increased FDG uptake, with a SUVmax of 2.0. Results: SIFs predominantly occurred in postmenopausal women (93.7%, 30/32) at a median of 14 months post-radiotherapy. Sacral involvement was observed as follows: unilateral ala (53.1%, 17/32), bilateral alae (37.5%, 12/32), and extension to the sacral body (9.4%, 3/32). The affected segments were primarily located at S1-S3. Concomitant pelvic fractures were also frequently identified, including pubic (28.1%, 9/32), iliac (25%, 8/32), and bilateral L5 transverse process fractures (6.2%, 2/32). Common CT findings included ill-defined osteosclerosis near the sacroiliac joint (87.5%, 28/32) and linear or curvilinear hypoattenuating fracture lines (68.7%, 22/32). PET revealed characteristic mild, diffuse/patchy FDG uptake parallel to the sacroiliac joint (96.8%, 31/32) with low metabolic activity (mean SUVmax 2.45±0.74, mean SUR-BP 1.46±0.38). Quantitative CT confirmed significant osteopenia within the radiation field (mean HU 36.8±28.6 vs. 78.0±37.3 outside, p<0.001). Discussion: Post-radiation SIFs predominantly affect postmenopausal cervical cancer patients due to radiotherapy-induced osteoporosis and bone vulnerability. These fractures often present with nonspecific pain and require differentiation from bone metastases, for which 18F-FDG PET/CT is essential due to its ability to detect characteristic metabolic patterns and associated osteoporotic changes. Key diagnostic features include linear or curvilinear hypoattenuating fracture lines, mild diffuse/patchy FDG uptake parallel to the sacroiliac joint, and background osteoporotic changes within radiation fields. Integrated PET/CT outperformed single-modality imaging by enabling simultaneous assessment of bone metabolism and systemic tumor status, a critical advantage over MRI (superior for marrow edema but unable to evaluate systemic disease) and standalone CT (lacking metabolic discrimination of benign vs malignant lesions). Early recognition of SIFs through integrated imaging is critical to avoid misdiagnosis and unnecessary invasive procedures, thereby guiding appropriate conservative management. Conclusion: SIFs represent a prevalent post-radiotherapy complication in cervical cancer patients, with a particular predilection for postmenopausal women. 18F-FDG PET/CT demonstrates high diagnostic reliability for diagnosing SIFs, which typically present as linear fractures parallel to the sacroiliac joints on a background of osteoporotic changes, accompanied by mild diffuse or patchy FDG uptake and frequently co-occurring with pelvic fractures at other sites. Integrated PET/CT is crucial for early recognition, preventing misdiagnosis as metastasis, and guiding appropriate conservative management.