Inferior phrenic artery-to-pulmonary artery fistula after surgery for catamenial pneumothorax: a case report

A 56-year-old woman was referred to our respiratory department after a vascular abnormality was detected on chest CT performed for malignancy screening. She had undergone a routine health checkup, during which serum tumor markers were found to be elevated. As part of a comprehensive evaluation to exclude malignancy, contrast-enhanced chest CT was performed and revealed a suspected abnormal vascular connection in the right lower lung field, raising suspicion of a pulmonary arteriovenous fistula. During further clinical evaluation, it was confirmed that the patient had undergone thoracoscopic wedge resection of the right middle lobe at the site of air leakage and partial diaphragmatic resection for catamenial pneumothorax at the age of 46 years, approximately 10 years earlier. The complete medical chart from the outside hospital where the surgery had been performed was not available; however, the retrieved operative record indicated that no pleural adhesion procedure or reinforcement was performed. Detailed information regarding the exact extent and location of the diaphragmatic resection and the specific repair technique could not be confirmed from the available records. Her postoperative course had been uneventful, and there had been no recurrence of pneumothorax. At presentation, she was asymptomatic, without dyspnea, hemoptysis, or chest pain. Physical examination and routine laboratory findings were unremarkable. Chest CT revealed aneurysmal dilatation of the right A9 segmental pulmonary artery without identifiable drainage into the pulmonary vein, along with tortuous dilatation of the right inferior phrenic artery (Fig.  1 a, b). Three-dimensional CT reconstruction demonstrated a direct connection between these vessels (Fig.  1 c). Furthermore, four-dimensional CT angiography revealed absence of contrast opacification of the dilated A9 pulmonary artery during the pulmonary arterial phase, followed by retrograde opacification from the right inferior phrenic artery during the systemic arterial phase, favoring a systemic artery–to–pulmonary artery fistula rather than a pulmonary arteriovenous fistula (Fig.  1 d, e). Four-dimensional CT angiography was performed as a dynamic study in which a fixed anatomical region centered on the lesion was repeatedly scanned during rapid intravenous contrast injection, allowing time-resolved assessment of contrast flow and shunt hemodynamics. The scan coverage was 160 mm, the acquisition interval was 0.275 s, contrast material was administered at 4.0 mL/s, and the estimated radiation dose was 14.11 mSv. Pulmonary perfusion scintigraphy with technetium-99 m macroaggregated albumin (MAA) revealed a small intrapulmonary shunt fraction of 2.2%, while ventilation scintigraphy revealed no defects. Although the shunt was hemodynamically small and conservative management was considered reasonable, selective angiography later demonstrated early pulmonary venous drainage, raising concern about a potential risk of systemic embolization. In addition, future lesion progression or hemoptysis could not be excluded. Therefore, after multidisciplinary discussion and in accordance with the patient’s preference for definitive risk reduction, endovascular treatment was selected. Fig. 1 Multimodal computed tomography (CT) evaluation of an inferior phrenic artery–to–pulmonary artery fistula. a Chest CT with lung window setting shows aneurysmal dilatation of the right A9 segmental pulmonary artery without identifiable drainage into the pulmonary vein. b Contrast-enhanced chest CT with mediastinal window setting demonstrates tortuous dilatation of the right inferior phrenic artery adjacent to the right hemidiaphragm. c Three-dimensional CT reconstruction demonstrates a direct connection between the right A9 segmental pulmonary artery and the right inferior phrenic artery (arrow and arrowhead). d Four-dimensional CT angiography (pulmonary arterial phase) shows absence of contrast opacification of the dilated peripheral right A9 pulmonary artery. e Four-dimensional CT angiography (systemic arterial phase) demonstrates retrograde opacification of the dilated peripheral right A9 pulmonary artery supplied by the right inferior phrenic artery, consistent with an inferior phrenic artery–to–pulmonary artery fistula Multimodal computed tomography (CT) evaluation of an inferior phrenic artery–to–pulmonary artery fistula. a Chest CT with lung window setting shows aneurysmal dilatation of the right A9 segmental pulmonary artery without identifiable drainage into the pulmonary vein. b Contrast-enhanced chest CT with mediastinal window setting demonstrates tortuous dilatation of the right inferior phrenic artery adjacent to the right hemidiaphragm. c Three-dimensional CT reconstruction demonstrates a direct connection between the right A9 segmental pulmonary artery and the right inferior phrenic artery (arrow and arrowhead). d Four-dimensional CT angiography (pulmonary arterial phase) shows absence of contrast opacification of the dilated peripheral right A9 pulmonary artery. e Four-dimensional CT angiography (systemic arterial phase) demonstrates retrograde opacification of the dilated peripheral right A9 pulmonary artery supplied by the right inferior phrenic artery, consistent with an inferior phrenic artery–to–pulmonary artery fistula Subsequent celiac arteriography confirmed opacification of the pulmonary artery via the right inferior phrenic artery. Selective angiography demonstrated early filling of the right pulmonary vein, indicating a potential risk of systemic embolization with liquid embolic agents; therefore, coil embolization was selected (Fig.  2 a). Post-embolization angiography demonstrated complete occlusion of the right inferior phrenic artery with disappearance of the shunt (Fig.  2 b). The postprocedural course was uneventful, and follow-up CT performed one month later demonstrated regression of the dilated vascular structures without recurrence. Additional follow-up at 6 months also demonstrated no evidence of recurrence or newly developed abnormal vascular findings. The patient remained asymptomatic during follow-up, with no evidence of recurrent hemoptysis or other related symptoms. The patient understood that the lesion had been incidentally detected during evaluation for a separate clinical concern and that it was hemodynamically small at presentation. At the same time, she was concerned about the possibility of future complications and expressed a preference for treatment to reduce potential risk. After discussion of the available management options, she expressed satisfaction that treatment could be performed and felt reassured after the procedure. Fig. 2 Angiographic confirmation and endovascular treatment of the inferior phrenic artery–to–pulmonary artery fistula. a Celiac arteriography demonstrates opacification of the pulmonary artery via the right inferior phrenic artery. Selective angiography further reveals early filling of the right pulmonary vein (arrows), indicating a potential risk of systemic embolization with liquid embolic agents; therefore, coil embolization was selected. b Post-embolization angiography confirms complete occlusion of the right inferior phrenic artery with disappearance of the arteriovenous shunt Angiographic confirmation and endovascular treatment of the inferior phrenic artery–to–pulmonary artery fistula. a Celiac arteriography demonstrates opacification of the pulmonary artery via the right inferior phrenic artery. Selective angiography further reveals early filling of the right pulmonary vein (arrows), indicating a potential risk of systemic embolization with liquid embolic agents; therefore, coil embolization was selected. b Post-embolization angiography confirms complete occlusion of the right inferior phrenic artery with disappearance of the arteriovenous shunt

Systemic artery–to–pulmonary artery fistula (SAPAF) is a rare vascular abnormality characterized by an abnormal shunt between a systemic artery and the pulmonary arterial circulation [ 1 – 6 ]. SAPAF can be congenital or acquired, and in the latter case it may develop after thoracic surgery, pneumothorax [ 7 ], infection [ 8 ], or chronic inflammation. Involvement of the inferior phrenic artery is particularly rare [ 6 ], and in most of the reported cases, the SAPAF originated from the bronchial or intercostal arteries. As SAPAF may mimic a pulmonary arteriovenous fistula or other pulmonary vascular disorders on conventional imaging [ 4 ], the accurate identification of the feeding artery and flow direction is essential for enabling appropriate management. Four-dimensional CT angiography (4D-CTA) is a dynamic imaging technique in which a fixed anatomical region is repeatedly scanned during rapid intravenous contrast injection, allowing time-resolved visualization of contrast flow. Unlike conventional static CT angiography, this approach enables assessment of flow direction and shunt hemodynamics in a clinically intuitive manner. In the present case, this temporal information was critical for distinguishing a systemic artery–to–pulmonary artery fistula from other vascular abnormalities and for selecting an appropriate embolization strategy [ 4 , 8 , 9 ]. We describe an inferior phrenic artery–to–pulmonary artery fistula that was incidentally identified nearly a decade after thoracoscopic surgical treatment of catamenial pneumothorax. This case suggests a possible contribution of endometriosis-related angiogenesis to postoperative vascular remodeling and highlights the complementary roles of 4D-CTA and selective angiography in diagnosis and treatment planning.

To the best of our knowledge, this appears to be the first reported case of an acquired SAPAF originating from the inferior phrenic artery as a late postoperative complication of surgically treated catamenial pneumothorax. Although the lesion was considered likely to be acquired in relation to prior surgery, a pre-existing vascular abnormality cannot be completely excluded. SAPAF is an uncommon but clinically important cause of abnormal pulmonary vascular findings and, in some patients, hemoptysis or hemothorax [ 2 , 4 , 5 ]. Inferior phrenic artery–to–pulmonary artery fistulas are particularly rare and have been described after pulmonary resection, pneumothorax surgery [ 7 ], and chronic inflammatory lung disease [ 2 ]. In this context, the pathogenesis of SAPAF is likely multifactorial, involving postoperative diaphragmatic remodeling, chronic inflammation, and angiogenic stimulation that is associated with thoracic endometriosis [ 10 – 12 ]. In our patient, the fistula appeared 10 years after thoracoscopic surgery for catamenial pneumothorax, suggesting that postoperative scarring at the diaphragm and lung base may have promoted collateral formation from the inferior phrenic artery to the pulmonary circulation [ 2 , 7 ]. Catamenial pneumothorax is strongly associated with thoracic endometriosis [ 10 – 13 ], and endometriosis-related tissue has been linked to angiogenic activity [ 10 – 13 ]. Thus, both postoperative mechanical collateralization and endometriosis-related angiogenesis may have contributed, although this remains speculative because histologic confirmation was not available. An alternative explanation is purely mechanical collateralization secondary to postoperative adhesion and scarring [ 2 , 7 ]. Previously reported inferior phrenic artery–to–pulmonary artery fistulas have been associated with pulmonary resection, prior pneumothorax surgery, and chronic inflammatory lung disease [ 2 , 6 , 7 ]. Most were diagnosed by conventional angiography or static contrast-enhanced CT, and reported treatments included transcatheter embolization and surgery, generally with favorable short-term outcomes [ 1 – 3 , 6 , 16 , 17 ]. In this context, the present case is notable for its occurrence after surgery for catamenial pneumothorax and for the use of 4D-CTA to noninvasively characterize shunt hemodynamics before angiographic confirmation. A key strength of this case is the use of 4D-CTA, which enabled noninvasive visualization of flow direction and accurate identification of the feeder artery. In contrast to conventional CT, 4D-CTA clarified the hemodynamic distinction between SAPAF and pulmonary arteriovenous fistula [ 1 , 4 , 8 , 14 , 15 ]. Selective angiography further demonstrated early pulmonary venous drainage, which influenced treatment selection because liquid embolic materials could have posed a risk of systemic embolization. Although the shunt fraction was small and the patient was asymptomatic, treatment was selected on the basis of angiographic hemodynamic findings, potential for future progression, and concern for systemic embolic or hemorrhagic complications. Coil embolization provided a controlled and safe therapeutic approach [ 1 , 3 , 16 , 17 ] and resulted in complete angiographic and clinical success. Several limitations of this report should be acknowledged. First, this is a single-case report, which limits the generalizability of the findings. Second, although four-dimensional CT angiography was informative in this case, its routine clinical use may be constrained by increased radiation exposure and limited availability. Finally, although no recurrence was detected at 6 months, longer-term radiological follow-up remains limited, which should be acknowledged as a limitation when interpreting the durability of coil embolization. In conclusion, inferior phrenic artery–to–pulmonary artery fistula should be considered in patients with prior thoracic surgery who present with unexplained vascular abnormalities near the diaphragm. In this context, 4D-CTA is a powerful noninvasive diagnostic tool for evaluating patients with suspected SAPAF and can play a crucial role in guiding safe and effective endovascular management.