2026 年 13 巻 p. 245-251
Eight years after resection of hypopharyngeal carcinoma with neck dissection and ligation of the left internal jugular vein, a man in his 50s presented with an incidentally discovered isolated left transverse sinus dural arteriovenous fistula. Transvenous embolization was performed to achieve complete obliteration. The patient likely developed venous hypertension due to stenosis of the left transverse sinus, limiting contralateral drainage and the absence of collateral communication between the internal jugular vein and the suboccipital cavernous sinus. A retrospective review of serial imaging revealed progression of venous drainage from the initial development of transverse sinus dural arteriovenous fistula to an eventual isolated sinus configuration. Although dural arteriovenous fistula formation is associated with venous hypertension, this case suggests that anatomic venous variants may predispose individuals to delayed dural arteriovenous fistula formation, with abrupt alterations in venous flow serving as triggers. The serial imaging demonstrates the progressive nature of this pathology. Assessment of venous anatomy in patients undergoing planned occlusion of major venous pathways may help predict long-term complications.
Dural arteriovenous fistulas (dAVFs) are acquired vascular malformations characterized by abnormal arteriovenous shunts within the dura mater. Their development is often associated with predisposing factors such as venous sinus thrombosis, head trauma, surgery, infection, radiation therapy, and hemodynamic changes secondary to tumors.1-6) Among these factors, venous hypertension and impaired venous drainage represent key mechanisms that promote angiogenesis.
Although internal jugular vein (IJV) resection is occasionally performed during head and neck cancer surgery, subsequent intracranial dAVF formation remains rare. Notably, the rich collateral venous network in the head and neck region typically provides adequate compensatory drainage following unilateral IJV resection, which may explain the rarity of clinically significant venous complications.7-10) Nonetheless, in rare instances, alterations in venous drainage pattern following IJV resection may predispose patients to dAVF formation.
We present herein a rare case of an isolated dAVF in the left transverse sigmoid sinus detected on imaging 8 years after left IJV resection during radical neck dissection for head and neck malignancy. The patient's anatomic venous variant may have contributed to dAVF formation in this case. Published serial imaging findings demonstrating the progressive course from dAVF formation to complete sinus isolation remain limited. The present case, through contrast-enhanced computed tomography (CT) findings spanning 8 years, provides illustrative insight into this progressive course and may help improve the understanding of the pathophysiological mechanisms underlying isolated sinus formation.
A man in his 50s underwent total pharyngolaryngectomy with bilateral neck dissection for left hypopharyngeal carcinoma (T4aN2b). During the procedure, the left IJV was firmly adherent to metastatic lymph nodes at level III (Figure S1) and was resected en bloc. The patient subsequently underwent cisplatin-based chemotherapy combined with postoperative radiotherapy (66 Gy). Two years after the initial surgery, the patient developed pulmonary metastasis in the left lung, which was treated via thoracoscopic surgery. Two years later, right renal carcinoma was detected, and a right nephrectomy was performed. Following these treatments, the patient remained recurrence-free, and regular surveillance was continued.
Eight years following the initial neck surgery, findings suggestive of dAVF were incidentally detected on follow-up contrast-enhanced CT, and the patient was referred to our department. At referral, the patient reported no headaches or neurologic deficits. Brain magnetic resonance angiography revealed abnormal time-of-flight (TOF) signals within the left transverse sinus and vein of Labbé. Susceptibility-weighted magnetic resonance imaging demonstrated multiple prominent leptomeningeal veins and cerebral microbleeds.
Cerebral angiography showed that the fistula was fed by the petrosquamous branches of the left middle meningeal artery, the jugular and mastoid branches of the left occipital artery, and the tentorial artery from the left internal carotid artery (Figure 1A and B, and Figure S2). These feeders drained into the isolated left transverse sinus. The isolated sinus exhibited retrograde cortical venous drainage through the vein of Labbé into the superficial middle cerebral vein and vein of Trolard. Intracranial venous return occurred predominantly through the right transverse sinus into the right IJV. Cone-beam CT revealed feeding arteries clustered within the parasinus of the left transverse sinus, forming a shunt pouch. Based on these findings, the patient was diagnosed with a Borden type III dAVF of the isolated left transverse sinus, and transvenous embolization was planned.
Perioperative angiography.
Anteroposterior (A) view of the left external carotid artery and lateral (B) view of the left common carotid artery before endovascular treatment demonstrate an arteriovenous shunt with a venous pouch (arrow). The arteriovenous shunt is supplied by the left middle meningeal artery, the left occipital artery, and the tentorial artery and drains into the vein of Labbé via retrograde flow. The left internal jugular vein and the medial segment of the left transverse sinus were occluded, and intracerebral venous flow drained into the superior sagittal sinus and right transverse sinus. The lateral view (C) during endovascular treatment shows that a microcatheter (arrowheads) had advanced through the right transverse sinus and the occluded left transverse sinus, and coils were deployed into the venous pouch. The lateral view of the left common carotid angiogram (D) after endovascular treatment shows the disappearance of the arteriovenous shunt.
Under general anesthesia, the right radial artery was punctured, and a 4-Fr Axcelguide sheath (Medikit, Tokyo, Japan) was advanced and positioned within the left external carotid artery for diagnostic angiography. Subsequently, a 6-Fr FUBUKI XF guiding sheath (Asahi Intecc, Aichi, Japan) was inserted into the right cubital vein. A 6-Fr Cerulean DD6 MK 2 catheter (Medikit, Tokyo, Japan), TACTICS catheter (Technocrat Corporation, Tokyo, Japan), and 0.035-inch SURF guidewire (PIOLAX, Kanagawa, Japan) were coaxially advanced through the right subclavian vein into the right IJV. The FUBUKI catheter was positioned within the high cervical IJV, and the DD6 catheter was positioned within the right proximal transverse sinus. The Tactics catheter was navigated close to the occluded segment of the left transverse sinus. Subsequently, the SURF guidewire was advanced via the torcular herophili into the occluded transverse sinus, and the Tactics catheter was advanced to the medial vicinity of the isolated sinus. The SURF guidewire was then exchanged for an SL-10 microcatheter (Stryker, Kalamazoo, MI, USA) and a CHIKAI Nexus 014 (Asahi Intecc, Aichi, Japan) to enable navigation into the isolated sinus. The catheter length limited the initial attempt to reach the shunt pouch, so the SL-10 was exchanged for a Headway Duo microcatheter (Terumo, Tokyo, Japan), which was successfully advanced to the shunt pouch.
Eight coils were deployed into the shunt pouch, resulting in the disappearance of shunt flow (Figure 1C and D). Postoperative magnetic resonance imaging performed the following day showed no evidence of cerebral infarction, and the abnormal TOF signal in the left transverse sinus disappeared.
Dural arteriovenous fistulas are predominantly acquired lesions, and their formation is associated with various factors such as venous sinus thrombosis, trauma, surgery, infection, and radiation therapy.1-5) Among these, venous hypertension and subsequent angiogenesis are recognized as key pathogenic mechanisms.11-13) Although venous hypertension due to IJV ligation is a recognized cause of dAVF, with several cases reported in the literature,4,6) this complication remains rare despite the relative frequency of IJV ligation in head and neck surgery. In the present case, the patient received cisplatin-based chemotherapy in conjunction with 66 Gy of radiation therapy to the head and neck region following IJV resection, with the radiation field including the sigmoid sinus and cervical area. Cisplatin chemotherapy has been reported to increase the risk of deep vein thrombosis.14) Postoperative chemoradiotherapy may have contributed to thrombotic progression and venous hypertension. However, because postoperative chemoradiotherapy is routinely administered to most patients undergoing IJV resection for head and neck malignancy, and dAVF formation remains exceedingly rare in this patient population, radiotherapy and chemotherapy alone are unlikely to account for the development of dAVF in this case. This suggests that additional patient-specific factors, beyond simple venous outflow obstruction, may determine individual susceptibility to dAVF formation. We speculate that, in the present case, anatomical variation in collateral venous drainage may have contributed to sustained venous hypertension in the left transverse sinus following IJV ligation, thereby facilitating dAVF development.
Rarity of dAVF following IJV resection and the role of anatomic venous variantsUnder normal circumstances, the primary compensatory mechanism relies on venous rerouting through the torcular to the contralateral transverse sinus and IJV. However, pre-ligation imaging in this patient revealed anatomical features that likely limited this primary compensatory mechanism. The left transverse sinus demonstrated focal stenosis medial to the vein of Labbé (Figure 2A and Figure S3A-D). Imaging findings revealed nodular filling defects within the sinus lumen, suggesting stenosis caused by arachnoid granulations. The pre-existing characteristics would have restricted collateral flow from the left transverse sinus toward the torcular, potentially impairing efficient drainage to the contralateral system following left IJV occlusion. In addition to the transcranial route, collateral drainage via the suboccipital cavernous sinus (SCS) serves as an important auxiliary pathway following IJV occlusion.9,15,16) The SCS pathways typically drain from the sigmoid sinus to the SCS via the posterior condylar vein (PCV) and mastoid emissary vein or via the anterior condylar confluence (ACC) through the anterior condylar vein or lateral condylar vein (Figure 2B). When the primary transtentorial pathway is compromised-as may have occurred in this patient due to left transverse sinus stenosis-the SCS pathway assumes greater functional importance.
Craniocervical collateral drainage patterns and venous variants.
Pre-ligation contrast-enhanced CT (A) demonstrates a stenotic left transverse sinus medial to the vein of Labbé (arrowhead). Schematic illustration (B) of typical left venous structures at the craniocervical junction viewed from the medial side demonstrates the communication between the IJV and the SCS. The schema of collateral flow during IJV occlusion demonstrates that sigmoid sinus flow drains into the SCS through the PCV, the mastoid EV, and the ACC. The maximum intensity projection image of this patient, obtained using contrast-enhanced 3-dimensional time-of-flight magnetic resonance angiography (C), and a schema (D) of the left side show no communication between the IJV and SCS due to the absence of the PCV, EV, and ACC.
ACC: anterior condylar confluence; ACV: anterior condylar vein; EV: emissary vein; IJV: internal jugular vein; IPS: inferior petrosal sinus; JB: jugular bulb; JF: jugular foramen; LCV: lateral condylar vein; PCV: posterior condylar vein; SCS: suboccipital cavernous sinus; SS: sigmoid sinus; TS: transverse sinus
The patient demonstrated a unique anatomic venous variant characterized by the absence of communication between the IJV and the SCS. Contrast-enhanced 3-dimensional TOF magnetic resonance angiography revealed an underdeveloped PCV and emissary vein, as well as the absence of the ACC connecting the inferior petrosal sinus (IPS) to the jugular bulb (Figures 2C and 3), resulting in poor communication between the IJV and the SCS (Figure 2D). Furthermore, CT imaging revealed the absence of the foramina of the emissary vein and PCV (not shown). Pre-treatment vertebral artery angiography demonstrated that the IPS received venous drainage from the posterior fossa and drained into the SCS via the lateral condylar vein (Figure S4). Although no reports have systematically evaluated the presence or degree of communication between the IJV and the SCS, previous studies have examined anatomic variations in the junction between the IPS and the jugular bulb or IJV, with 1 report identifying the absence of the ACC in 3.6% of cases.17) Embryologically, the caudal portion of the IPS is derived from the ventral myeloencephalic vein. This vein is thought to form through the anastomosis of the vagal vein, hypoglossal vein, and upper cervical intersegmental veins during dorsolateral migration of the anterior cardinal vein, thereby determining the venous architecture of the IPS-IJV junction and the SCS.17,18)
Imaging findings of absent anterior condylar confluence.
Axial images are shown. Contrast-enhanced 3D time-of-flight magnetic resonance angiography demonstrates a space (dotted arrowhead) between the inferior petrosal sinus (white arrowhead) and the jugular bulb or internal jugular vein (white arrow), and the left anterior condylar confluence is absent. The development of connections between the internal jugular vein and suboccipital cavernous sinus (asterisk) was considered poor.
3D: 3-dimensional
Thus, the combination of left transverse sinus stenosis and absence of an IJV-SCS collateral pathway likely resulted in dual restriction of compensatory venous drainage. This multifactorial anatomical limitation provides a plausible explanation for localized venous hypertension in the left transverse sinus. Prolonged exposure to venous hypertension may have induced endothelial injury and pathologic angiogenesis within the sinus wall, ultimately leading to arteriovenous shunt formation.11) Of note, contrast-enhanced high-resolution black-blood magnetic resonance images demonstrated circumferential wall thickening with enhancement surrounding the left transverse sinus (Figure S3E and F).
Progression to an isolated sinus in the transverse sinus dAVFReports documenting the progression of transverse sinus dAVF to isolated sinus configurations are scarce. This case provides insight into this progression through annual serial CT performed during oncologic surveillance (Figure 4).
Progressive changes in intracranial venous drainage after internal jugular vein ligation.
Follow-up contrast-enhanced computed tomography findings are displayed before IJV ligation (A) and at 2 years (B), 6 years (C), and 8 years (D) after left IJV ligation. (B) Two years later, enhanced contrast uptake is observed at the resected left IJV stump (arrowhead). (C) Six years later, the medial site of the left transverse sinus (arrows) shows decreased enhancement, and mild intracranial venous dilatation (dotted area) is observed. (D) Eight years later, contrast enhancement at the IJV stump has disappeared, and intracranial venous congestion is significant. A schematic illustration (E) of each phase shows the progression of dAVF formation. Gradual changes in venous drainage may result in an isolated sinus with intracerebral venous drainage via the cortical vein.
dAVF: dural arteriovenous fistula; IJV: internal jugular vein; SSS: superior sagittal sinus
Two years following IJV resection, contrast enhancement was observed at the resected IJV stump (Figure 4B), suggesting early arteriovenous shunting and probable dAVF formation. Six years after resection, the left transverse sinus demonstrated decreased enhancement (Figure 4C), indicating progressive thrombosis accompanied by mild intracranial venous dilatation, while contrast enhancement of the IJV remained visible. At 8 years, contrast enhancement at the IJV stump disappeared (Figure 4D), and marked intracranial venous congestion had become evident.
A schema illustrating the progressive pathophysiologic sequence based on serial imaging is presented in Figure 4E. As mentioned, the patient had underdeveloped alternative venous drainage pathways via the IPS and SCS; however, we believe that marginal antegrade drainage through small cervical venous collaterals was likely. This observation can be explained by the plexiform nature of the IJV, which may form through anastomoses of various embryologic venous systems,17) potentially allowing marginal antegrade drainage through these delicate venous branches. Although the imaging resolution at that time made precise tracing of detailed venous pathways difficult, the patient was likely predisposed to venous hypertension following IJV resection.
Regarding the rationale for changes in venous drainage patterns, several studies have reported on mechanisms observed in cavernous sinus dAVFs.19-21) The authors stated that venous sinuses encased in dura, such as the IPS, are prone to occlusion because of limited space for expansion, whereas cortical veins, which are capable of dilation, remained patent, leading to pathology involving intracranial reflux. Assuming a similar mechanism, we propose that the transverse sinus occluded over time due to dual restriction of compensatory drainage-focal left transverse sinus stenosis limiting transtorcular compensation and poor IJV-SCS connections further restricting alternative pathways-while the vein of Labbé, being capable of dilation, remained patent and continuous with the affected sinus. Of note, the medial boundary of the isolated sinus corresponded to the site of sinus stenosis identified prior to dAVF development, further supporting this hypothesis.
A limitation of this case is reliance on serial contrast-enhanced CT findings rather than angiography to infer the progressive evolution of venous drainage patterns, precluding precise characterization of shunt hemodynamics at each time point. However, isolated sinus dAVFs are frequently diagnosed only after becoming symptomatic with intracranial complications such as hemorrhage, and serial imaging documentation of progression to isolated sinus configuration remains limited in the literature. The present case demonstrates serial changes in intracranial venous drainage following IJV resection, providing observational support for the progressive nature of dAVF evolution and enabling early diagnosis and treatment before symptom onset.
Clinical implicationsThis case highlights the importance of recognizing anatomic venous variants as potential risk factors for delayed vascular complications following cervical venous occlusion. Preoperative assessment of venous anatomy, particularly the transtorcular route to contralateral drainage, the IPS-IJV connection, and PCVs providing primary SCS-IJV communication, may help identify patients at higher risk of intracranial venous complications. Long-term surveillance may be warranted in selected high-risk cases, especially for patients with multiple risk factors, including anatomic variants, radiation therapy, and chemotherapy-induced vascular injury. Understanding disease progression and underlying anatomic predispositions is crucial for preventing and optimally managing these rare but potentially serious complications.
ConclusionsWe report herein a case of ipsilateral isolated transverse sinus dAVF following IJV ligation. Serial imaging findings suggest that alternative venous pathways associated with anatomic venous variants may be involved in the onset and progression of the dAVF. The details of this case underscore the importance of dAVF as a progressive lesion. Therefore, preoperative anatomical assessment and long-term postoperative follow-up are crucial.
All authors have no conflict of interest.
This manuscript does not require ethical review by an Institutional Review Board as it is a single case report. Written informed consent was obtained from the patient for publication of this case report and accompanying images.
The original contributions presented in the study are included in the article/figure; further inquiries can be directed to the corresponding author.
