Patients and study design
A prospective, double-center study, including 57 consecutive patients with hemoptysis who were referred for BAE, was conducted at Military Hospital 103 and National Lung Hospital, two national hospitals in Hanoi, Vietnam between August 2019 and July 2020. The classification of hemoptysis grades include massive hemoptysis ( 300 mL/day or respiratory failure, or hemodynamic instability), moderate hemoptysis (100-300 mL/day), and mild hemoptysis (< 100 mL/day) . All patients underwent MDCT angiography before BAE, and a paired comparison between the angiographic results of two techniques was done. Bronchoscopy was also used to diagnose the hemoptysis etiology in some patients. Enrolled patients were classified into two groups based on the history of pulmonary TB to compare the clinical and angiographic characteristics. Also, we prospectively evaluated the technical success rate and clinical outcomes of BAE with preprocedural MDCT angiography. We excluded patients with a previous history of BAE before August 2019. The protocol of this study was approved by the Institutional Review Board of Vietnam Military Medical University (number: 251/2019) and by the local ethics committees of two participating centers. All participants had provided written informed consents for this research.
MDCT angiography and image analysis
MDCT angiography was performed with a 64-MDCT scanner (SOMATOM Sensation 64; Siemens Medical Solutions, Forchheim, Germany) in 43 patients (120 kV, 320 mAs, rotation time of 0.5 second, 0.75-mm collimation, pitch of 1.5) and with a 16-MDCT scanner (Brilliance 16; Philips Medical Systems, Ohio, USA) in the remaining 14 patients (140 kV, 70-120 mAs, rotation time of 0.5 second, 0.75-mm collimation, pitch of 1.5). All patients were scanned in a supine position from lung tip to diaphragm. Patients received approximately 80 to 100 mL of contrast material (Omnipaque 300 mgI/mL; GE Healthcare, Oslo, Norway) followed by 50 mL of normal saline solution, which was injected intravenously at a rate of 4 mL/s. The automatic bolus triggering software program was performed, with a circular region of interest positioned at the level of the descending thoracic aorta. Triggered data acquisition began at the contrast enhancement level of 100 Hounsfield unit (HU). Series of images were reconstructed at 1-mm section thickness with 0.6-mm increment. All data of MDCT angiography were transferred to a workstation for post-processing. Two independent radiologists (P.A.T and T.N.T, who had more than 15 years of experience in reading MDCT angiography) analysed the CT images. The inconsistent results were discussed with a board of experts for final decisions.
We first evaluated characteristics of culprit vessels (focused on the number, origin, site of the ostium, diameter, and course), including BAs and NBSAs. BAs were divided into two groups: (a) orthotopic BAs originating from the descending aorta between the level of T5 and T6 vertebrae and (b) ectopic BAs from any level of aorta outside levels T5 and T6 vertebrae, or its branches. NBSAs were defined as arteries which enter the parenchymal through the inferior pulmonary ligament or adherent pleura and their courses were not parallel to the bronchi . BAs were considered abnormal when (a) their diameter was 2 mm, or (b) their courses were tortuous and could be identified to the hilum. NBSAs were considered abnormal if they were dilated and tortuous, within extrapleural fat related to pleural thickening. Secondly, we investigated the other radiological findings as specific lesions, which could be causes of hemoptysis, such as active/reactivated TB, bronchiectasis, aspergilloma, and malignancy. These characteristics integrated with additional tests (e.g. microscopy, bacterial/MTB/fungal culture, and Xpert MTB/RIF assay of sputum/bronchoalveolar lavage samples, serum antibody test, and histopathology test) for the definitive diagnosis. The extent of lung diseases was classified according to the number of lobes with involvement on chest CT, 1 to 3 lobes, or more than 3 lobes.
BAE procedure, including arteriography and embolization, was performed with a 5-Fr introducer sheath (Terumo, Japan) through a common femoral artery access, using the Seldinger technique. Because of the vascular map of MDCT angiography before, selective catheterization of culprit BAs and NBSAs was conducted without aortography. 5-Fr curved catheters, including Cobra, Hook, or Simmons, and right coronary artery catheters were used. Coaxial microcatheters (Carnelian 1.8/2.2-Fr; Tokai Medical Products, Japan) were guided by the 0.016/0.018-inch M guidewire (Terumo, Japan). Transcatheter embolization was completed for all abnormal arteries that met one of the following radiological characteristics: (a) tortuous enlargement of BAs and/or NBSAs which provided the region of parenchymal staining, or (b) a shunt into pulmonary vessels . Embolic agents were polyvinyl alcohol (PVA; size range, 350-710 ; Contour; Boston Scientific, USA) and a combination of PVA and gelatin sponge (Gelfoam; BioSphere Medical, USA). BAE was performed by four interventional radiologists (D.N.B and N.T.T in the first center, and T.N.T and N.D.M in the second center), all of whom have had more than ten years of experience in vascular intervention.
We recorded the number, origins, and courses of culprit BAs and NBSAs. They were considered as gold standards for diagnosis and compared to the results of MDCT angiography. Also, the rate of technical success was noted. It was defined as the embolization of all observed abnormal arteries.
Follow-up and clinical outcome
After BAE, all participating patients were followed up with a regular re-examination or by telephone to assess the clinical outcomes of BAE with preprocedural MDCT angiography. The first outcome was the rate of immediate clinical success, of which the bleeding was stopped or significantly reduced within 24 hours of BAE. The second outcome was the rates of hemoptysis recurrence, which referred to significant hemoptysis occurring after discharge. Among these, early recurrence appeared within 3 months of BAE .
Data of continuous variables were presented as mean SD [standard deviation] and that of categorical variables were presented as numbers and percentages. The baseline characteristics between history and non-history of pulmonary TB groups were compared by using t-test for continuous variables, and 2 test and Fisher’s exact test for categorical variables. The average numbers of culprit arteries per patient between two groups were analysed by the Mann-Whitney U test. To explore the risk factors associated with early recurrence, univariate and multivariate Cox proportional hazards regression models were performed. A p-value of less than 0.05 was considered as the statistical significance. All analyses were performed with SPSS version 25.0 (IBM, Armonk, NY, USA) statistical software.