A 76-year-old man was admitted to our emergency department by ambulance at 19:37 because of his sudden speech disturbance for 40 min. Medical records revealed a previous diagnosis of hypertension for eight years and chronic atrial fibrillation for six years. He was on oral antihypertension drugs but did not receive anticoagulant treatment. His family history was negative for any neurological disorder, and he denies histories of tinnitus, head trauma, or cranial surgery. On neurological examination, the patient was alert but notable for fluent aphasia with paraphrastic errors and inability to name, read, or write. The cranial nerves were intact, and all modalities of sensation or muscle strength were normal. Physical examination revealed body temperature of 36.8 °C, heart rate of 84 beats per minute, respiration rate of 20 breaths per minute, and blood pressure of 158/76 mm Hg (1 mm Hg = 1.33 kPa). Emergency cranial non-contrast CT was ordered, and the result showed neither intracranial hemorrhagic nor early ischemic changes. Extensive laboratory examinations were unremarkable, including complete blood count, glucose, electrolytes, coagulation studies, and hepatic and renal function tests. Given the acute onset of the clinical manifestation and a history of atrial fibrillation without anticoagulation, cerebrovascular embolism involving the sensory speech center in the dominant hemisphere’s temporal lobe was considered the most probable diagnosis. At 36 min from admission, the patient received intravenous thrombolysis on a standard dose of alteplase (0.9 mg/kg body weight), with the onset-to-needle time being 73 min. Before thrombolysis, the patient had a score of 6 on the National Institutes of Health Stroke Scale (NIHSS), and the modified Rankin scale (mRS) was 2. In order to determine whether there is intracranial large vessel occlusion and the necessity of bridge thrombectomy, multimodal CT imaging, including a head-and-neck CT angiography (CTA) and cerebral CT perfusion, was immediately scheduled after initiation of thrombolysis. The results indicated that the intracranial vessels were clearly visualized, with no large vessels or branches occluded. Regional hyperperfusion status in the left temporal lobe was recognized on the CT perfusion, with characteristics of an increase both in cerebral blood flow and cerebral blood volume. As the hemodynamic change of hyperperfusion was sometimes compatible with seizure activity, a bedside electroencephalogram (EEG) was performed to eliminate the possibility of focal seizures. On examination, the patient was still symptomatic, but synchronized EEG revealed only intermittent polymorphic delta wave changes in the left temporal region without typical epileptiform discharge. Considering the clinical symptoms and signs could be explained by the involvement of the left temporal lobe, the patient was regarded as having cerebral infarction but experienced early vessel recanalization accompanied by post-recanalization hyperperfusion. Intravenous thrombolysis was halted 39 min into thrombolysis in fear of hemorrhagic transformation, with the actual dosage of alteplase being 65% of the standard. The patient’s vital signs were stable during the thrombolysis, with no new neurological signs or headaches occurring. The patient was transferred to the stroke unit ward for further monitoring. Twelve hours after thrombolysis the following day, the patient’s speech disturbance was partially recovered, with the NIHSS score decreased to 3. And a cranial MRI was scheduled to localize the lesions. Unexpectedly, heterogeneous signal abnormalities in the left temporal lobe were presented, subsequently confirmed by the cranial CT being the intraparenchymal hemorrhage. Surgical intervention was not warranted as the hematoma was localized with no significant occupying effect. Thus the patient was on continuous monitoring and symptomatic treatment. On the seventh day, cranial CT showed obvious absorption of the hemorrhage. The symptoms of aphasia were further improved, leaving only occasional spelling mistakes. Cranial susceptibility-weighted imaging (SWI) was also performed during the hospitalization to evaluate the possible cerebral microbleeds burden as post-thrombolysis hemorrhage occurred. The results indicated no evidence of cerebral microbleed but conspicuous venous vasculature in the left temporo-occipital lobe and cerebellum. When reevaluating the initial CT angiography imaging, early visualization of the left transverse sinus segment, previously described as venous contamination, had regained our attention. Cerebral digital subtraction angiography (DSA) was ordered, which revealed a left transverse sinus and sigmoid sinus DAVF (TS-SS DAVF) supplied mainly by the ipsilateral occipital artery and branches of the middle meningeal artery. The increased shunt flow from the fistula and combined downstream stenosed sigmoid sinus caused venous hypertension and retrograde venous drainage directly into cortical veins. The patient refused endovascular embolization and was free of new symptoms on a 3-month clinical revisit.