Introduction

Cardiovocal syndrome is characterized by recurrent laryngeal nerve (RLN) paralysis (RLNP) due to cardiovascular disease such as aortic aneurysm (AA) [1]. It is reported that AAs causing hoarseness were most frequently located in the distal part of the aortic arch [2]. Although the possible mechanisms of RLNP induced by AA include RLN stretching and compression between the aortic arch and pulmonary artery [3, 4], detailed reports on this topic are lacking. The left RLN courses proximal to the subclavian artery origin, descends anterior to the aortic arch, passes lateral to the ligamentum arteriosum, interposes between the aortic arch and pulmonary artery, and ascends within the tracheoesophageal groove [3]. Viewtify® software (SCIEMENT, Tokyo) [5] enables three-dimensional (3D) cardiovascular evaluation and may be useful for the evaluation of aortic aneurysms. We tried to elucidate the anatomical characteristics of AA-associated RLNP using 3D computed tomography (3DCT).

Methods

We conducted a retrospective single-center study using medical data from patients diagnosed with vocal-fold paralysis (VFP) between 2013 and 2022 at a voice outpatient clinic at the University of Tokyo Hospital. Patients presenting with hoarseness as the primary complaint, subsequently diagnosed with VFP after a thorough investigation, were categorized into the preoperative group. Subsequently, we identified a postoperative group comprising patients who did not experience hoarseness prior to AA surgery but developed postoperative hoarseness, leading to a diagnosis of VFP upon referral to our department. Additionally, a control group was established, consisting of patients with VFP but lacking evidence of AA. Following the extraction of the preoperative group, the postoperative and control groups were organized by matching them with the preoperative group based on age, sex, and patient count. Patients with thoracic deformities affecting mediastinal structures due to scoliosis or other factors were excluded from the study, as variations in the course of the RLN may occur compared to those without such deformities.

To evaluate the relationship of the aortic aneurysm to the surrounding tissue, 3DCT images were reconstructed using Viewtify® to confirm pulmonary artery (PA) and ligamentum arteriosum compression by the AA, where the left RLN runs. The importation of Digital Imaging and Communications in Medicine (DICOM) data from the original CT is completed within seconds. Viewtify® instantaneously generates 3DCT images of areas within a predefined CT value range using this data. Viewpoint, cross-section, and rendering thresholds can be adjusted in real time, facilitating the immediate extraction of the required images without the need for pre- or post-processing of the original DICOM data. The ligamentum arteriosum was identified by remaining duct-like structure [6] or calcification [7]. Preoperative or pre-vocal treatment CT images were used for 3D image reconstruction (Fig. 1). Next, to measure the AA circumference as an indicator of RLN stretching, a cross-sectional view was constructed using the ImageJ software (version 1.5108; National Institutes of Health, Maryland, USA) [8], and the AA circumference was compared between the groups.

Fig. 1
figure 1

Representative three-dimensional digital graphic models for each group. (A) The region around the aortic arch and pulmonary artery in the control group. (B) The region around the aortic aneurysm in the preoperative group. The pulmonary artery is compressed by the aortic aneurysm. (C) The region around the aortic aneurysm in the postoperative group. The yellow arrow indicates the ligamentum arteriosum. Abbreviations: Ao, aortic arch; AA, aortic aneurysm; PA, pulmonary artery; SCA, subclavian artery

Full size image

Cross-sections along the path of the left RLN were obtained to evaluate the anatomical relationship with the AA or aortic arch in each group. In the control group, a cross-section along the path that passed near the origin of the subclavian artery, outside the ligamentum arteriosum and tracheoesophageal groove, was reconstructed (Fig. 2A, Supplementary Movie S1). In the preoperative and postoperative groups, the ligamentum arteriosum was not identified on CT because of AAs. Therefore, we obtained the cross-section with the smallest AA circumference along the path that passed near the origin of the subclavian artery and tracheoesophageal groove in all cases (Fig. 2B,C). Three doctors (JS, TG, and RU) independently evaluated the circumference, and the inter-rater reliability between them was determined using Cronbach’s alpha, with results showing a significant correlation coefficient of r > 0.8.

Statistical comparisons were performed using BellCurve for Excel (version 4.04; Social Survey Research Information Co. Ltd. Tokyo, Japan). Comparisons between groups were made by one-way ANOVA and Bonferoni’s post hoc test. Statistical significance was set at P < 0.05. The Human Ethics Committee of the University of Tokyo approved the study protocol (No. 2487, 2022179NI).

Results

Six patients were included in each of the following groups: patients with VFP without AA (control group), patients with VFP before AA surgery (preoperative group), and patients with no preoperative hoarseness and VFP after AA surgery (postoperative group).

In the control group, no PA compression by the aorta was observed and the ligamentum arteriosum was detected in all cases (Fig. 1A). In the preoperative group, all patients showed PA compression findings; however, the ligamentum arteriosum was not detected in all cases (Fig. 2B, Supplementary Movie S2). In the postoperative group, no PA compression was observed despite the presence of AAs and the ligamentum arteriosum was not detected in all cases (Fig. 2C).

Fig. 2
figure 2

Representative cross-sections along the path of the left recurrent laryngeal nerve. (A) In the control group, no compression was observed between the aortic arch and the pulmonary artery. (B) In the preoperative group, significant compression is observed between the aortic arch and pulmonary artery (black arrows). The compression between the ligamentum arteriosum and the aortic aneurysm is also suspected (purple dotted line). (C) In the postoperative group, no compression findings are observed between the aortic arch and pulmonary artery despite the presence of the aortic aneurysm. The blue dotted lines indicate the tracheoesophageal groove. The red dotted lines indicate the AA circumference. Abbreviations: Ao, aortic arch; AA, aortic aneurysm; PA, pulmonary artery

Full size image

AA circumference was compared between the groups. The mean circumference of the aortic arch was 86.0 ± 12.7 mm, 178.3 ± 30.2 mm, and 117.5 ± 30.0 mm in the control, preoperative, and postoperative groups, respectively (Fig. 2), and it was significantly greater in the preoperative group than in the control and postoperative groups (P < 0.001, P = 0.0015, respectively) (Fig. 3).

Fig. 3
figure 3

Mean circumference of the cross-section in each group. The aortic aneurysm circumference in the preoperative group is significantly larger than those in the control and postoperative groups. **P < .01

Full size image

Discussion

This study demonstrated that 3DCT image analysis along the RLN is useful for assessing the RLNP risk in patients with AA. The very clear 3D images support the possibility of nerve stretching or compression by the AA as a cause of RLNP. We estimated the RLN course from the anatomy around the aortic arch using 3DCT software because the RLN was not visualized on CT. Compared with standard CT, 3DCT is much more accurate for assuming the RLN course because the display angle of an object can be freely changed to create an evaluation cross-section at the most suitable site in three dimensions according to the assumed course of the RLN.

The 3DCT images of the preoperative group showed PA compression, but not the ligamentum arteriosum. Considering that the PA wall is supple and the effect of RLN compression between the PA and AA is thought to be low, compression between the ligamentum arteriosum and AA can cause nerve traction and chronic constriction, which may result in nerve palsy. In previous studies, AA has been considered to stretch [9] or compress [4] the RLN, leading to nerve-conduction defects and dysfunction. Although our findings support both hypotheses, whether RLNP is caused by stretching or compression of the RLN by AA remains unclear. Autopsy studies have shown that the distance between the aorta and pulmonary artery within the aortic window is only 4 mm and have suggested that compression of the RLN between the two structures is responsible for palsy [10]. The average length of the left RLN is approximately 43 cm [11]. A nerve loses its function when it is suddenly stretched by 8–20% of its total length [12]. However, if a nerve is stretched slowly over a few years, the length could increase without any functional disturbance [12]. Aortic aneurysms usually enlarge slowly, in years, but can be paralyzed when considering the distance of extension.

Recent studies on RLNP associated with AAs include two systematic reviews [2, 4] and one non-systematic review [3]. However, only the diameter of the AA was considered in multivariate analysis [2], and no study has assessed the extension of the RLN as undertaken in the present study. Therefore, a substantial strength of this study lies in its utilization of 3DCT to generate a cross-section along the RLN, enabling evaluation of its extension.

This study had some limitations. Firstly, patients may not have complained of hoarseness because the vocal folds were fixed in the median position before surgery. Therefore, it was impossible to determine whether VFP occurred after AA surgery. Secondly, The RLN course was inferred from the surrounding anatomy and may not reflect the actual course since the RLN was not confirmed by surgery or autopsy. Thirdly, the data utilized in this study suggested compression between the AA and the PA or ligamentum arteriosum, but this remains speculative and not a definitive indicator. Fourthly, a definitive cutoff value for AA circumference indicating the potential onset of RLNP was not established during this study. Determining specific cutoff values for the circumference of the AA that can lead to RNP is crucial. ROC curves and other analytical methods can serve as valuable tools in achieving this objective. Future prospective studies conducted in collaboration with cardiovascular surgeons are warranted to conclusively address this issue.

Conclusion

In conclusion, in patients with aortic aneurysms, 3DCT along the RLN is useful for assessing the risk of RLNP. Measurement of the circumference of the aneurysm and verification of compression at the ligamentum arteriosum using 3DCT are important in assessing the risk of RLNP.

Movie S1. The video demonstrates how to create the cross-section along the running of the recurrent laryngeal nerve in the control group. The cross-section is reconstructed to pass through the subclavian artery, outside of the ligamentum arteriosum, and tracheoesophageal groove. The red line represents the expected running line of the recurrent laryngeal nerve. Movie S2. The video illustrates how to create the cross-section along the running of the recurrent laryngeal nerve and the whole view of the aneurysm in the preoperative group.

The cross-section is developed with the smallest circumference along the path that passes near the origin of the subclavian artery and tracheoesophageal groove. The red line represents the expected running line of the recurrent laryngeal nerve. Compression of the recurrent laryngeal nerve between the aortic arch and pulmonary artery can be observed. The blue circle indicates the supposed compression of the recurrent laryngeal nerve between the ligamentum arteriosum and aortic aneurysm.