Imaging characteristics and clinical management of intraosseous hemangioma in tubular bones: a retrospective case series

Abstract

Objectives

Intraosseous hemangioma (IH) is a rare benign vascular tumor of bone, infrequently observed in tubular bones. The imaging characteristics of IH in these locations remain poorly defined. This study aimed to characterize the imaging features and clinical management of IH in tubular bones.

Methods

This retrospective analysis reviewed the imaging and clinical data of 10 patients with histopathologically confirmed IH of tubular bones.

Results

Of the 10 patients, seven were female (70%) and three were male (30%), with ages ranging from 10 to 75 years (mean ± SD, 40.1 ± 22.8 years). Nine patients (90%) presented with localized pain and swelling. Radiographic findings included six purely osteolytic lesions (60%), three with a soap-bubble appearance (30%), and one lesion (10%) was heavily calcified with a lobulated morphology. Among intramedullary cases, three showed cortical destruction, two had pathological fractures. Two lesions (20%) were confined to the cortex, demonstrated cortical thickening, sclerosis, and internal lytic changes. MRI revealed iso- or hypointense on T1WI and hyperintense on T2WI/STIR in three patients (30%). One patient (10%) displayed patchy hyperintensity on both T1WI and T2WI, with partial hypointensity on T2-STIR. Flow voids corresponding to low-signal blood vessels were seen in four patients (40%), while one (10%) exhibited a hypervascular lesion. Among the four patients who underwent contrast-enhanced CT or MRI, all marked heterogeneous enhancement. Nine patients (90%) were treated with curettage of the lesion, while one patient (10%) underwent complete lesion excision. In one hypervascular case, preoperative DSA embolization was performed.

Conclusion

Intraosseous hemangioma of tubular bones predominantly present as osteolytic lesions with characteristic soap-bubble morphology on radiographs. Recognition of intralesional vascular structures is critical for diagnosis. Imaging modalities, particularly CT and MRI, enable precise lesion characterization and assist in surgical planning. In hypervascular cases, preoperative embolization guided by vascular imaging can effectively minimize intraoperative bleeding risk.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-025-09443-9.

Introduction

Intraosseous hemangioma (IH) is a benign vascular neoplasm accounting for less than 1% of all primary bone tumors. It most frequently involves the vertebrae and craniofacial skeleton, whereas its occurrence in long tubular bones is exceedingly rare [1–4]. IH typically present as solitary lesions, although approximately one-third of cases demonstrate multifocal involvement [5, 6]. These tumors can occur at any age but most often affect women in the fourth to fifth decades of life [1]. Clinical manifestations range from incidental, asymptomatic findings to localized pain or pathologic fracture. Notably, IH of the appendicular skeleton are predominantly symptomatic, with more than 90% of extremity lesions producing clinical symptoms [3, 7]. Because of their rarity and highly variable imaging characteristics, intraosseous hemangiomas of tubular bones pose a persistent diagnostic challenge [8].

Histopathologically, IH are typically classified into four subtypes: cavernous, capillary, venous, and mixed. Cavernous hemangioma represents the most common subtype in the tubular bones of the extremities [6, 9–11]. These lesions consist of thin-walled, blood-filled vascular channels lined by a single layer of flattened endothelial cells without cytologic atypia [11–13]. Most tubular bone IH arise in the diaphysis or metadiaphysis and can be medullary, periosteal, or intracortical depending on their site of origin [1, 9]. Medullary IHs most commonly involve the diaphysis of tubular bones [14].

On radiographs, medullary IH of tubular bones typically appear as well-circumscribed osteolytic lesions with internal trabeculation, producing a characteristic “honeycomb” or “soap-bubble” appearance [3, 14]. Computed tomography (CT) demonstrates a lytic lesion with internal trabeculae forming the classic “polka-dot” or “honeycombed” pattern and may also reveal a “sunburst” configuration of radiating trabeculations [9, 10, 15, 16]. On magnetic resonance imaging (MRI), lesions show hypointense to intermediate signal on T1-weighted images and markedly hyperintense signal on T2-weighted and short-tau inversion recovery (STIR) sequences, reflecting low fat content and sluggish blood flow; hypointense septa are often visible on these sequences [17, 18]. Post-contrast MRI typically shows intense or heterogeneous enhancement [19, 20]. CT remains the optimal imaging modality for evaluating IH, providing superior delineation of cortical and trabecular patterns [21].

Because of the scarcity of reported cases, particularly in tubular bones, most available literature consists of isolated case reports. In this study, we analyzed the clinical and imaging characteristics of 10 patients with IH of tubular bones to enhance diagnostic recognition of this uncommon entity.

Materials and methods

This retrospective study analyzed the imaging characteristics and clinical data of 10 patients with histopathologically confirmed IH of tubular bones. Digital radiography (DR) was performed in seven patients, CT in nine patients, and MRI in eight patients. Four patients underwent contrast-enhanced CT or MRI. Surgical approaches and clinical outcomes were also recorded.

Radiographs were assessed for lesion location, morphology, margins, internal matrix, and cortical involvement. CT provided detailed evaluation of cortical destruction, internal trabecular patterns, and calcification. MRI included multiplanar T1-weighted, T2-weighted, and short tau inversion recovery (STIR) sequences, assessing signal intensity, internal vascularity, and soft-tissue extension.

Protocols

CT imaging was performed using two 64-slice multidetector spiral scanners. Cases 1, 3, 5, and 8 were scanned using a Siemens Somatom Definition Flash Dual-Source CT system, while Cases 2, 4, 6, 7, and 10 were scanned using a Philips Ingenuity 64-slice CT system. Scanning parameters were as follows: matrix = 512 × 512, FOV = 300–500 mm, tube voltage = 120 kV, automatic tube current modulation, slice thickness = 1 mm, and pitch = 1 mm. The scanning range was defined by the CT topogram to encompass the entire lesion and at least one adjacent joint. For contrast-enhanced CT, Ultravist 370 was administered intravenously at 1.5 ml/kg through the antecubital vein at a rate of 5.0 ml/s.

MRI examinations were performed using two 3.0-T scanners. Cases 2, 3, 4, 6, 7, and 10 were scanned on a Siemens MAGNETOM Verio 3.0T (Siemens AG, Germany), and Cases 5 and 8 on a GE Discovery MR750 3.0T system (GE Healthcare, Milwaukee, WI, USA). An appropriate surface coil was used. Routine sequences included T1-weighted (TR/TE = 400–830/6–17), T2-weighted (TR/TE = 3000–4800/83–136), and STIR (TR/TE = 2877–7290/48–78) imaging, acquired in coronal, sagittal, and axial planes. The matrix ranged from 256 × 256 to 512 × 512. For enhanced imaging, Gd-DTPA was administered intravenously at 0.1 mmol/kg via the antecubital vein. Field of view, slice thickness, and interslice gap were adjusted according to lesion size and anatomical region.

To facilitate early histopathological diagnosis, all patients underwent percutaneous needle biopsy before surgery. Surgical specimens were subsequently examined pathologically. Postoperative imaging follow-up was conducted, and three patients were monitored for 1–3 years after surgery.

All imaging data were independently reviewed by two senior radiologists with over 10 years of experience. In cases of disagreement, a chief radiologist adjudicated the findings. The evaluated parameters included lesion location, signal or density intensity (relative to skeletal muscle), periosteal thickening, cortical destruction, enhancement pattern, and presence of soft-tissue masses.

Results

Clinical features

Clinical characteristics of all participants are summarized in Table 1. The cohort comprised seven females (70%) and three males (30%), aged 10–75 years (mean ± SD: 40.1 ± 22.8 years). Nine patients presented with varying degrees of pain, local swelling, or pathological fractures, while one was asymptomatic. The disease duration ranged from 1 month to 20 years.

Table 1.

Clinical data and imaging features of 10 patients

ID	Gender/Age(years)	Symptom/sign	Anatomical location	Histopathological Classification	Radiograph/CT	MRI
1	F/41	Pain	Tibia, diaphyseal; Intramedullary	cavernous	Purely lytic lesion, significant heterogeneous enhancement, cortical erosion	N/A
2	F/68	-	Ulna, metaphyseal; Intramedullary	mixed	soap-bubble appearance with thin sclerotic margin	T1WI: iso; T2WI and STIR: hyper, internal linear low signal
3	F/10	Pain, soft tissue swelling	Humerus, diaphyseal; Intramedullary	cavernous	Purely lytic lesion, local cortical absence and vascular passage	Dilated and tortuous blood vessels in the lesion with low signal in T1WI, T2WI and STIR; only the periphery of the lesion significantly enhanced
4	F/28	Pain, soft tissue swelling	Femur, diaphyseal; Intramedullary	Arteriovenous hemangioma	Purely lytic lesion, cortical erosion, periosteal reaction; significant heterogeneous enhancement, dilated and tortuous blood vessels in the lesion and peripheral soft tissue	Dilated and tortuous blood vessels in the lesion and peripheral soft tissue with low signal in T1WI, T2WI and STIR, peripheral soft tissue swelling
5	M/48	Pain	Femur, metaphyseal; Intramedullary	cavernous	osteolytic lesion with severe calcificationa, high-low hybrid density	high-low hybrid signal in T1 and T2/STIR
6	F/32	Pain, pathological fracture	Foot, 5st metatarsal; Intramedullary	venous	soap-bubble appearance with thin sclerotic margin; cortical erosion, with pathologic fracture	T1WI: iso, T2WI and STIR: hyper, blood vessels in the lesion and peripheral soft tissue with low signal in T1WI, T2WI and STIR, peripheral soft tissue swelling
7	M/13	Pain	Ulna, metaphyseal; Cortex	capillary	Cortical thickening, osteolytic destruction within the cortex, cortical erosion, the lesion extend to peripheral soft tissue	T1WI:iso, T2WI and STIR: hyper, peripheral soft tissue swelling
8	F/75	Cutting pains	Tibia, metaphyseal; Cortex	mixed	Cortical sclerosis and thickening, osteolytic destruction within the cortex, cortical erosion	T1WI: hyper, T2WI andSTIR: hyper, with thin hypointense margin, peripheral soft tissue swelling; significant heterogeneous enhancement
9	M/36	Soft tissue swelling	Hand, 3st middlephalanx; Intramedullary	mixed	Expansion of the third phalanx, with soap-bubble appearance, peripheral soft tissue swelling	N/A
10	F/56	Pathological fracture	Humerus, diaphyseal and metaphyseal; Intramedullary	cavernous	Expansion of the humerus, purely lytic lesion; with pathologic fracture	T1WI iso; T2WI and STIR: with mixed signals of hyper and low

F female, M male, -: asymptomatic

All cases were confirmed as IH by histopathological examination. The lesions contains proliferative vascular and fibrous tissue, with dilated vessels exhibiting sinusoidal changes. Some lesions contain hemorrhage, calcification, fat, or thrombus (Figs. 1 and 3). Five lesions involved the lower extremities, and five involved the upper extremities. Eight cases were located in long tubular bones and two in short tubular bones. Among the long tubular bones lesions, four were metaphyseal, three were diaphyseal, and one extended across both regions. All patients underwent surgical resection with biopsy, and histopathological examination confirmed intraosseous hemangioma in every case (Figs. 1 and 3). Of these, four (40%) were cavernous, three (30%) mixed, one (10%) capillary, one (10%) venous, and one (10%) arteriovenous—a rare subtype (Table 1).

Fig. 1.

Case 6 A 32-year-old woman with intramedullary intraosseous hemangioma in the fifth metatarsal; a-c Internal oblique radiograph, coronal, and axial CT images of the left foot demonstrate a lytic lesion with a soap-bubble appearance, thin sclerotic margin, and cortical erosion, accompanied by a pathologic fracture; d T1-weighted MR image shows the lesion as isointensity; e-f Axial T2WI and coronal T2-short tau inversion recovery MR images demonstrate the lesion as hyperintensity, with low-signal blood vessels inside the lesion and in the peripheral soft tissue, and peripheral soft tissue swelling appears hyperintensity; g Histological examination shows hyperplastic and dilated blood vessels between bone trabeculae (hematoxylin and eosin (H&E) stain, × 100)

Fig. 3.

Case 5 A 48-year-old man with intramedullary intraosseous hemangioma in the femur. a-b AP radiograph and coronal CT demonstrate a mixed high-density appearance; c-e Coronal T2-short tau inversion recovery MR image, axial T1WI, and T2WI show a mixed high- and low-signal lesion. f Histological examination shows tumor-like hyperplasia of the vascular tissue with focal fibro-fatty tissue involving hemorrhage, along with vascular hyperplasia and dilatation

Imaging features

The imaging features of the 10 IH cases in tubular bones are summarized in Table 1.

DR Findings

Seven patients underwent digital radiography (DR). Six lesions were medullary and one intracortical. Among the medullary lesions, three lesions exhibited a soap-bubble appearance (Figs. 1 and 2), two lesions showed purely lytic, and one lesion showed heavily calcified lobular lesion (Fig. 3). The intracortical lesion exhibited cortical sclerosis and thickening.

Fig. 2.

Case 2 A 68-year-old woman with intramedullary intraosseous hemangioma in the ulna, a-d lateral radiograph, sagittal CT, and axial CT image demonstrate a lytic lesion with a soap-bubble appearance, displaying a small amount of fatty density within the lesion (CT value=-32 hounsfield unit), thin sclerotic margin, and cortical erosion; e-g Axial T1WI, axial T2WI, and sagittal T2-short tau inversion recovery MR images show the lesion as isointense on T1WI and hyperintense on T2WI/T2-STIR, exhibited patchy hyperintense lesions on both T1WI and T2WI sequences and hypointensity on short tau inversion recovery (STIR) sequence (arrow). Low-signal blood vessels within the lesion

CT findings

Nine patients underwent CT scans, revealing seven medullary and two intracortical lesions (Figs. 1, 2, 3, 4 and 5 ). Among medullary cases, six appeared purely lytic (Figs. 1, 2, 4 and 5), among of which one case showed punctate fatty density within the lesion (Fig. 2). One case presented as a heavily calcified lobular mass (Fig. 3). The intracortical lesions demonstrated osteolytic destruction with cortical thickening, sclerosis, and internal osteolysis (Fig. 5). Contrast-enhanced CT, performed in two patients, revealed marked heterogeneous enhancement (Fig. 4). In one case, both CT and DSA showed clusters of dilated, tortuous vessels within the lesion and adjacent soft tissues; following preoperative embolization, abnormal vascularity was significantly reduced (Fig. 4).

Fig. 4.

Case 4 A 28-year-old woman with intramedullary intraosseous hemangioma in the femur, (a-b) AP radiograph and coronal CT demonstrate a lytic lesion with dilated and tortuous blood vessels in the lesion, with periosteal reaction. c Contrast-enhanced CT image shows significant heterogeneous enhancement, with dilated and tortuous blood vessels in the lesion and the surrounding soft tissue; (d-e) Sagittal T2-short time inversion recovery MR image DSA show a mass of dilated and tortuous blood vessels within the lesion and peripheral soft tissue. f After preoperative DSA embolization, the abnormal blood vessels decreased significantly.

Fig. 5.

Case 7 A 13-year-old man with cortical intraosseous hemangioma in the ulna. a Coronal CT demonstrates as cortical thickening of the ulna with internal osteolytic destruction. b-e Coronal T1WI, axial T1WI, and T2WI, and coronal T2-STIR MR images demonstrate the lesion as isointense on T1WI and hyperintensity on T2WI/T2-STIR, with peripheral soft tissue swelling with hyperintensity

MRI findings

Eight patients underwent MRI. Three lesions exhibited iso- to hypointensity on T1WI and hyperintensity on T2WI/STIR (Figs. 1 and 5). Among of them, one lesion displayed patchy hyperintensity on both T1WI and T2WI, and hypointensity on short tau inversion recovery (STIR) sequence (Fig. 2). One lesion appeared hyperintense on all sequences, while another demonstrated mixed high and low signals (Fig. 3). Low-signal vascular channels were observed within four lesions (Figs. 2 and 4). Peripheral soft-tissue swelling was present in four cases (Figs. 1 and 5), and one lesion displayed a thin hypointense rim. The two intracortical lesions were associated with cortical thickening and sclerosis, mirroring CT findings (Fig. 5). Contrast-enhanced MRI, performed in two patients, One case showed significantly heterogeneous enhancement, while the other exhibited only peripheral rim enhancemen-a pattern previously described in the literature [22].

Treatment and prognosis

Nine patients underwent curettage of the lesion followed by bone grafting and/or internal fixation, while one patient received complete lesion excision. In one case, preoperative CTA revealed multiple dilated and tortuous vessels within the lesion and adjacent soft tissues. To mitigate the risk of intraoperative hemorrhage, preoperative DSA embolization was performed (Fig. 5), effectively minimizing intraoperative blood loss. All patients received postoperative radiographic evaluations, and none experienced postoperative complications (Figs. S1 and S2). Three patients were followed for 1–3 years, and no evidence of recurrence was observed (Figs. S1 and S2).

Discussion

We comprehensively characterized the imaging features and clinical outcomes of IH in tubular bones. Among the 10 cases, two involved short tubular bones and two were intracortical. Radiographically, these lesions typically appeared osteolytic with a soap-bubble-like configuration, although extensive calcification mimicking osteoblastic lesions was occasionally observed. MRI findings were variable but generally demonstrated intermediate T1 and high T2 signal intensities. Abnormal intralesional vascularity was detectable on both CT and MRI. All patients underwent surgical curettage and achieved favorable postoperative recovery.

IH is an uncommon bone tumor, and its occurrence in tubular bones is exceptionally rare [8]. Although most IH are asymptomatic, previous studies have reported that more than 90% of lesions in the extremities cause symptoms [7, 23]. In our series, nine patients (90%) presented with pain or soft-tissue swelling, consistent with prior observations. While IH most frequently affect women aged 40–50 years, our cohort showed a broader age range, with only two patients within this typical age group-likely reflecting the limited sample size. Histopathologic evaluation revealed that cavernous and mixed types comprised 70% of cases, whereas only one case (10%) was a pure capillary hemangioma, aligning with previous reports. IH in tubular bones often manifests as osteolytic lesions or demonstrates a characteristic “soap-bubble” appearance, consistent with previous reports [9].

In this study, we found that IH in tubular bones often presents as osteolytic lesions or exhibits a soap bubble-like appearance. This appearance results from expansive proliferation of dilated vascular channels and thickened, remodeled bone trabeculae [1, 9]. Notably, one of our cases involved a heavily calcified femoral lesion, consistent with prior documentation of a similar humeral epiphyseal hemangioma [1]. IH arising in short tubular bones is exceedingly rare; in our cohort, two such lesions were confined to the medullary cavity and exhibited a soap-bubble appearance with surrounding soft-tissue swelling—radiologic features similar to those seen in long tubular bones. Furthermore, intracortical IH demonstrated cortical thickening, sclerosis, and internal lytic changes, in agreement with earlier studies [1, 9]. Additional findings such as cortical destruction, periosteal reaction, and soft-tissue swelling were also observed. Importantly, these imaging features did not adversely affect patient outcomes; all patients recovered well postoperatively, with no recurrence during follow-up.

The MRI characteristics of IH in tubular bones are variable, reflecting differences in lesion composition—such as venous flow velocity, fat content, hemorrhage, and the presence of coarse trabeculae [4, 24]. In our series, three lesions displayed iso- or hypointensity on T1WI and hyperintensity on T2WI/STIR, consistent with the low fat content characteristic of these lesions [1]. One lesion exhibited hyperintensity on all sequences (T1WI, T2WI, and short tau inversion recovery), the pathology indicates local bleeding within the lesion. Of particular note, low-signal abnormal vessels were detected in four cases—a relatively uncommon but diagnostically valuable feature. Previous studies have classified vascular malformations into low-flow and high-flow types based on hemodynamics, and any malformation with an arterial component is considered high flow [25–27]. By combining DSA and CTA images, we found that these lesions contained a large number of arterial vessels, all of which were high flow vascular malformation. Therefore, we believe that the low signal displayed by the blood vessels within the lesion on MRI was caused by signal voids of fast-flow vessels. MRI thus plays a crucial role in assessing lesion extent, internal composition, and extraosseous involvement.

Contrast-enhanced CT and MRI features of tubular bone IH are seldom reported. In our study, four cases exhibited significant heterogeneous enhancement, the pathology indicates that the lesions contain proliferated vascular and fibrous tissues, and in two cases there are a large number of inflammatory cells in the lesions. Additionally, previous study found that the enhancement pattern of the IH was associated with endothelial cell proliferation [9, 19, 28]. Another case demonstrated peripheral rim enhancement, a pattern also described in the literature [20, 22]. Such enhancement characteristics aid in the diagnosis of IH and highlight abnormal blood vascularity, which is essential for preoperative evaluation and surgical planning.

Overall, the imaging manifestations of IH are heterogeneous, often complicating diagnosis. However, the imaging features of different subtypes also share some commonalities. Except for the the calcification subtype, the other subtypes all present with osteolytic bone destruction, and the boundaries of the tumors are generally clear. Moreover, due to the lack of fat components within the lesion, which often appears as an iso- or hypo-signal on T1WI and high signal on T2WI.The imaging profile of appendicular bone IH differs from that of the axial skeleton. On CT, the appendicular bone IH usually showed lytic lesion, which was caused by the tumor directly eroding, absorbing and replacing the normal bone marrow tissue and trabeculae of bone; whereas the IH of axial skeleton usually exhibit the palisade or polka-dot sign, which was the projection of compensatory thickened longitudinal trabeculae on the image. On MRI, the appendicular bone IH appendicular lesions typically show iso- or hypointensity on T1-weighted images and hyperintensity on T2-weighted or short tau inversion recovery sequences, whereas the IH of axial skeleton usually exhibit high signal intensity on both T1WI and T2WI, and showed high signal intensity on STIR. Due to fat components are rarely present in tubular bone IH, while the IH of axial skeleton with higher fat and fluid contents [1, 29, 30]. Some lesions may mimic aggressive bone pathology—exhibiting periosteal reaction or soft-tissue extension—and must therefore be differentiated from other neoplastic and infectious entities [31–33]. (1) Osteolytic IH should be distinguished from simple bone cysts, eosinophilic granuloma, and bone metastases. Simple bone cysts are typically unilocular intramedullary lesions with smooth margins and cortical thinning without disruption; contrast enhancement is limited to the cyst wall, with the cavity remaining non-enhancing [34]. Eosinophilic granuloma usually affects children or adolescents and presents as focal medullary destruction with periosteal reaction, extensive marrow edema, and soft-tissue swelling; chronic lesions may show fibrous repair and sclerosis [35]. Bone metastases, in contrast, are often multifocal, occur in older patients, and are associated with a known primary malignancy. (2) Intracortical IH should be differentiated from non-ossifying fibroma (NOF) and chronic localized osteomyelitis. NOF commonly affects individuals under 20 years of age and typically involves the metaphyseal region of long bones in the lower limbs. Radiographically, NOF presents as an eccentric, well-defined, scalloped lesion with a sclerotic rim; cortical outlines may appear attenuated or obscured. MRI signal variability reflects differing amounts of fibrous tissue, collagen, histiocytes, hemorrhage, hemosiderin, and bone trabeculae [36, 37]. Chronic localized osteomyelitis (Brodie’s abscess) often follows acute infection in children or adolescents and manifests as a small, round or oval, sclerotic lesion with central bone destruction, typically in the metaphysis. (3) Periosteal IH should be distinguished from osteoid osteoma, myositis ossificans, and periosteal osteosarcoma. Osteoid osteoma is characterized by marked sclerosis surrounding a central nidus and presents clinically with severe nocturnal pain [38]. Myositis ossificans develops rapidly after trauma and demonstrates peripheral ossification separated from the bone cortex by a radiolucent zone [39]. Periosteal osteosarcoma, which predominantly affects young adults, manifests as a diffuse soft-tissue mass with aggressive periosteal reaction extending into surrounding soft tissues, cortical thickening, and vertical spiculated bone formation, though medullary invasion is uncommon [40].

The management of IH remains controversial, although it is generally accepted that asymptomatic and small lesions do not require intervention [41]. Curettage and bone grafting are typically recommended for symptomatic lesions or those at risk of pathologic fracture [22, 42]. In our series, nine patients presented with pain or swelling, and two sustained pathological fractures. Consequently, all patients underwent surgical curettage and achieved excellent postoperative outcomes, with no recurrence among those followed up—findings consistent with previous reports [20]. In one case, preoperative contrast-enhanced CT revealed numerous dilated and tortuous vessels within the lesion and adjacent soft tissues. Given the high risk of intraoperative hemorrhage associated with IH [24], preoperative digital subtraction angiography (DSA) embolization was performed, effectively minimizing intraoperative bleeding. The patient recovered well postoperatively, in line with prior findings [33]. Preoperative CT and MRI evaluations are crucial for identifying highly vascular lesions and guiding embolization planning, thereby reducing surgical risk and improving outcomes.

Limitation

This study has several limitations. First, due to the rarity of IH in tubular bones, the sample size was relatively small. Future work will aim to collect additional cases to enhance the comprehensiveness of the analysis. Second, some patients had incomplete imaging data because certain studies were performed at other institutions prior to referral. Finally, postoperative follow-up was limited to a small subset of patients, restricting the assessment of long-term prognosis. Future studies will strengthen follow-up protocols to better characterize postoperative outcomes.

Conclusion

Intraosseous hemangioma of tubular bones demonstrates diverse imaging characteristics, including lytic morphology, soap-bubble appearance, and variable calcification. MRI plays a pivotal role in revealing internal vascularity and differentiating IH from other lytic lesions. Comprehensive preoperative imaging assessment is essential for accurate diagnosis, optimal surgical planning, and improved patient outcomes.

Authors’ contributions

YJ and YD had a major contribution in planning and formulating the study design, and data interpretation. YD had a major contribution in writing the manuscript. YZ performed the statistical analysis. YD and NZ contributed to data collection. YD, ZL and YJ had an important contribution in formulating the research plan. All authors approved the final version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding

The authors state that this work has not received any funding.

Data availability

The data of the 10 patients are in-house dataset and are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Research Ethics Committee of the Honghui Hospital Affiliated Xi’an Jiaotong University (number: 2025-KY-096-01). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The informed consent to participate was obtained from all of the participants in this study. To protect patients’ privacy, the real names and medical record numbers of the patients were all concealed in the research, it did not cause any harm to the participants, and did not adversely affect the rights of participants.

Consent for publication

The participants have provided written informed consent for their personal or clinical details to be published in this study.

Competing interests

The authors declare no competing interests.