Pulmonary artery and aortic enlargement predict postoperative survival outcomes in lung cancer

Authors: Haruaki Hino, Osamu Honda, Kosuke Kashiwabara, Natsumi Maru, Takahiro Utsumi, Kento Fukumoto, Hiroshi Matsui, Yohei Taniguchi, Tomohito Saito, Tomohiro Murakawa

Abstract

OBJECTIVES

The pulmonary artery/aorta ratio can predict postoperative outcomes in patients with lung cancer; however, few studies have investigated the effects of individual changes in the pulmonary artery and aortic diameters. This study aimed to analyse the impact of pulmonary artery and aortic enlargement on outcomes following surgery for lung cancer.

METHODS

We retrospectively analysed data from 1482 patients with non-small-cell lung cancer who underwent radical surgery at our institution between 2006 and 2022. The maximum diameters of the pulmonary artery and ascending aorta were measured using preoperative computed tomography, and patients were divided according to cutoff values set for the receiver operating characteristic curve. All patients were then stratified into three non-enlarged pulmonary artery and aorta (‘normal’ group, n = 244), enlargement of either the pulmonary artery or aorta (‘either enlargement’ group, n = 689) and enlargement of both the pulmonary artery and aorta (‘both enlargement’ group, n = 549). The preoperative clinical characteristics and postoperative outcomes were analysed in these three groups.

RESULTS

Five-year overall survival rates in the ‘normal’, ‘either enlargement’ and ‘both enlargement’ groups were 80.2%, 77.3% and 71.0% (*P *= 0.002), respectively. Enlargement of both vessels was an independent negative prognostic factor for both overall survival (hazard ratio, 1.56; *P *= 0.04) and cancer-specific death (hazard ratio, 1.83; *P *= 0.03). However, this association was not significant for non-cancer-specific deaths (hazard ratio, 1.39; *P *= 0.35).

CONCLUSIONS

Measurement of the pulmonary artery and aortic diameters might be a useful predictor of postoperative survival in patients with lung cancer.

INTRODUCTION

The pulmonary artery (PA)/aorta ratio and degree of PA dilation reflect the severity of pulmonary hypertension (PH), heart failure and postoperative prognosis after cardiac surgery [1–3]. PH, one of the many causes of PA dilation, can be assessed using invasive and non-invasive examinations such as cardiac catheterization and echocardiography [4, 5]. These measurements have previously been applied to predict outcomes in patients with chronic obstructive pulmonary disease (COPD) or those undergoing cardiac surgery [6, 7]. Slightly differing from PA dilation, ascending aortic dilation is caused by several factors, such as ageing, hypertension and smoking,

COPD, and connective tissue diseases (e.g. Marfan syndrome) [8]. The PA/aorta ratio is used to predict the outcomes of patients with lung cancer [9, 10]; however, to our knowledge, the effects of both the individual and combined changes in the PA and aorta have not yet been investigated. Therefore, this study aimed to analyse how PA and aortic enlargement reflect postoperative outcomes in patients with lung cancer.

PATIENTS AND METHODS

Ethical statement

This study was conducted in accordance with the Declaration of Helsinki and was approved by the research review board of Kansai Medical University, Osaka, Japan (20 May 2023; approval no. 2023006). The requirement for informed consent was waived, given the retrospective nature of the study.

Patients

This retrospective study included patients who underwent surgery for primary lung cancer at our institution between 2006 and 2020. The collected data included the age at surgery, sex, percentage of vital capacity (%VC), forced expiratory volume in one second/forced vital capacity (FEV1/FVC), body mass index (BMI), smoking pack-years, preoperative carcinoembryonic antigen (CEA) level, maximum standardized uptake value (SUVmax) of the tumour, Charlson comorbidity index (CCI) [11], operative procedure, histology, clinical stage, postoperative complications defined by the Clavien–Dindo classification [12], and PA and aortic diameters.

To assess the PA and aortic diameters, the maximum diameter was manually evaluated using preoperative computed tomography (CT) (Fig. 1). Owing to limited health insurance coverage, a 5-mm slice thickness was used. According to detailed methods described in a prior publication [1], the PA and ascending aortic diameters were measured at the level of the PA bifurcation, ideally where both the right and left PA appear to be similar in size. The diameter of each mean value was measured independently by two authors (H.H. and O.S.), who were blinded to all clinical information. To clarify the impact of the PA and aortic enlargement and assess its correct correlation with survival, the use of continuous variables, such as the diameters of the PA and aorta, was more desirable than using categorical variables. However, owing to the complicated etiologies of PA and aorta enlargement, categorical variables were used to understand the broad trends and significance of PA and aorta enlargement. We divided the patients into high- and low-diameter groups according to the cutoff values set for receiver operating characteristic curves for the PA and aorta diameters, of which the end-point was overall death; 27.0 mm for PA (95% confidence interval [CI] 0.487–0.552, area under curve [AUC] 0.519) and 33.5 mm for aorta (95% CI 0.506–0.572, AUC 0.539).

Subsequently, all patients were stratified into three normal PA and aortic diameter (‘normal’ group, n = 244), either PA or aortic enlargement (‘either enlargement’ group, n = 689), and enlargement of both the PA and aorta (‘both enlargement’ group, n = 549). The preoperative clinical characteristics and postoperative outcomes were compared among these three groups.

Initially, we calculated the PA/aorta ratio, which was the predominant method used in previous studies [9, 10]; however, both diameters were relatively small, and the calculated errors increased. Furthermore, since differing etiologies could have impacted outcomes, we used the diameters of both the PA and the aorta for the analyses. Regarding survival data, the overall survival (OS) time was calculated as the time from surgery to the date of death or last follow-up, while non-cancer-specific survival time was calculated as the time from surgery to the date of non-cancer death or last follow-up within the entire follow-up period. The tumour stage was determined according to the 8th edition of the Union for International Cancer Control TNM staging system [13], and the histological tumour type was determined according to the 3rd edition of the World Health Organization Classification of Tumors [14].

Statistical analysis

Continuous and categorical variables were compared between the three groups using Kruskal–Wallis tests, respectively. Spearman’s correlation analysis was performed to assess the association between the variables. Survival time and differences in survival time were assessed using the Kaplan–Meier method and log-rank test over the entire available follow-up period, respectively. To evaluate OS, univariate and multivariate analyses were performed using the Cox proportional hazards model. Multivariate analysis was performed for each variable in the univariate analysis with a *P < *0.2. Moreover, the Fine–Gray model was applied to elucidate true independent risk factors for cancer or non-cancer death in a competing risk analysis using the total available follow-up period. Those variables were selected following prior publications [9, 10].

JMP software ver. 12 (SAS Institute, Inc., Cary, NC, USA) was used for statistical analyses such as the Kaplan–Meier method, log-rank test and Cox regression analysis, whereas competing risk and Spearman’s correlation analyses as well as the Kaplan–Meier method were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [15]. Statistical significance was set at a *P *< 0.05.

RESULTS

A total of 1803 patients who underwent surgery for lung cancer were initially enrolled. However, 321 patients were excluded due to incomplete resection (123 patients), multiple lesions (68 patients), pathological adenocarcinoma in situ (35 patients), preoperative treatment (54 patients), small-cell lung cancer or combined small-cell lung cancer (35 patients) and unknown PA and aorta diameter (20 patients). Finally, 1482 patients were included in the analysis (Supplementary Fig. S1).

Figure 2 shows the distribution of the PA and aortic diameters. The mean (range) diameters of the PA and aorta were 2.65 (1.81–4.54) cm (Fig. 2A) and 3.30 (1.97–4.82) cm (Fig. 2B), respectively, showing normal distribution. Supplementary Figure S2 demonstrates a relative correlation between the aortic diameter and age (ρ = 0.263, *P *= 7.17e−25) and the PA and aortic diameters (ρ = 0.210, *P *= 3.55e−16).

Table 1 shows the pre- and perioperative clinical characteristics of the patients. Compared with the ‘normal’ and ‘either enlargement’ groups, the ‘both enlargement’ group was significantly associated with a higher age, CCI, BMI, pack-years, male sex, lower respiratory function (*P *< 0.05) and a marginal difference in preoperative complications of cerebral-cardiovascular disease (*P *= 0.06). However, the ‘both enlargement’ group did not have any significant correlation with total postoperative complications, cardiopulmonary complications and pulmonary fistulae, as well as histology, procedure, standard uptake value and clinical staging (P > 0.05).

Regarding survival analysis, the 5-year OS rates of the enlarged and non-enlarged PA groups were 74.2% (95% CI 71.3–76.8%) and 80.4% (95% CI 75.3–84.5%), respectively (*P *= 0.0051) (Fig. 3A), while those of the enlarged and non-enlarged aorta groups were 72.2% (95% CI 68.1–75.8%) and 78.0% (95% CI 74.9–80.8%), respectively (*P *= 0.007) (Fig. 3B). Based on preoperative comorbidity, the 5-year OS rates of patients with and without cerebral-cardiovascular disease were 68.0% (95% CI 59.6–75.0%) and 76.5% (95% CI 73.9–78.9%), respectively (*P *= 0.006) (Supplementary Fig. S3). Furthermore, the 5-year OS rates of the ‘both enlargement’, ‘either enlargement’ and ‘normal’ groups were 71.0% (95% CI 66.6–75.0%), 77.3% (95% CI 73.7–80.5%) and 80.2% (95% CI 74.4–84.8%), respectively (*P *= 0.002) (Fig. 4 and Graphical Abstract).

Univariate and multivariate analyses using the Cox proportional hazards model demonstrated that enlargement of both vessels was a significant negative prognostic factor for OS (hazard ratio [HR] 1.56, 95% CI 1.03–2.43, *P *= 0.03). Other clinical characteristics were also highlighted as significant negative prognostic factors, including a higher age (HR 1.04, *P *= 0.0004), male sex (HR 1.67, *P *= 0.005), lower %VC (HR 0.98, *P *< 0.0001), lower BMI (HR 0.91, *P *< 0.0001), higher CEA value (HR 1.01, *P *= 0.01), higher CCI (HR 1.22, *P *= 0.002), higher SUVmax (HR 1.06, *P *= 0.0001), clinical Stage II (HR 1.79, *P *= 0.003) and clinical Stage III (HR 2.04, *P *= 0.02) (Table 2).

Supplementary Figure S4 summarizes the number and causes of death. Notably, the ‘both enlargement’ group was significantly associated with higher rates of total death (175/549 [31.9%], *P *= 0.0019) and cancer-specific death (110/549 [20.0%], *P *= 0.03) than the other groups; however, no significant difference in the number of cerebral-cardiopulmonary death (28/549 [4.3%], *P *= 0.20) was observed. Supplementary Figure S5A and B shows the death-related results stratified by the individual enlargement status of the aorta and PA. Notably, enlargement of the aorta and PA was individually either significantly or relatively associated with higher rates of total and cancer-related deaths; however, they were not significantly correlated with cerebral-cardiopulmonary death.

Regarding the competing risk analysis for cancer- or non-cancer-specific death, the following variables were independent negative risk factors for cancer-specific death (*P *< 0.05): male sex; lower %VC and BMI; higher CEA value, clinical stage and SUVmax; and enlargement of both the PA and aorta. In contrast, older age, higher CCI and lower %VC and BMI were independent poor risk factors for non-cancer-specific death (*P *< 0.05), whereas enlargement of both vessels was not (*P *= 0.35) (Table 3).

DISCUSSION

PA dilation is a risk factor for poor survival in various populations. According to the Rotterdam Study, the PA/aorta ratio is a risk factor for all-cause mortality in the general population with moderate-to-severe COPD [16]. Moreover, in patients with coronary artery disease, left heart failure and congenital heart failure, a greater PA/aorta ratio is a risk factor for mortality [17–19]. Concerning the outcomes of patients with COPD, a PA/aorta ratio of >1 or a large PA diameter has been shown to be a reliable predictor of PH [7, 20], acute exacerbation and mortality [21]. A higher mean PA/aorta ratio is also associated with unfavourable prognoses in patients with idiopathic pulmonary [22] and cystic fibrosis [23]. These outcomes suggest that a larger PA diameter or PA/aorta ratio could significantly predict the risk of morbidity and mortality among patients with cardiovascular or secondary pulmonary disease. In a previous study, complex structural cardiac changes were found to promote PA enlargement, consequently worsening the prognosis after cardiac surgery [18]. As for COPD and interstitial pneumonia, secondary PH is triggered by elevated pulmonary vascular resistance, and PA dilation itself is representative of the underlying cardiopulmonary burden [7, 20]. Therefore, complications such as respiratory exacerbation or acute respiratory failure may lead to poorer survival.

Ascending aortic dilation is attributable to ageing, long-term hypertension, smoking, systemic arteriosclerosis change, hyperlipidemia, aortic valve disorders and connective tissue diseases [8]. Moreover, upregulation of proinflammatory cytokines in the vascular endothelial layer may promote cancer malignancy and worsen survival [8, 24]. However, few reports have focused on the association between ascending aortic dilation and cancer survival. We considered the aetiology of aortic dilation to be slightly different from that of PA; PA dilation was considered to be attributable to COPD and PH. Table 1 shows that patients with ‘both enlargement’ were significantly correlated with lower FEV1/FVC and higher pack-year, which suggested that COPD occurred more in patients in the ‘both enlargement’ group. Furthermore, COPD also had a considerable association with higher cancer aggressiveness [25]. In addition, we showed the association between the enlargement of both vessels and higher preoperative cerebral-cardiovascular disease complications. This association may indirectly affect poor postoperative survival. Based on our results, the use of both assessments may predict lung cancer survival, especially cancer-specific survival, more precisely. However, further large-scale studies involving multiple institutions are required to elucidate the significance of PA and aortic dilation on postoperative outcomes in patients with lung cancer.

Recent studies have demonstrated a unique association between PA dilation and lung cancer. For example, Eul et al. [26] confirmed that a high PA/aorta ratio was a strong surrogate parameter for PH and a predictor of survival in patients with Stage I–IV lung cancer. Similarly, Gong et al. reported that an increased PA/aorta ratio was associated with a shorter OS and recurrence-free survival in patients with Stage III–IV advanced lung cancer who received immune checkpoint inhibitors [27]. Furthermore, in patients undergoing surgery for lung cancer, the PA size on the surgical side is a significant predictor of increased major postoperative complications [28], and a PA/aorta ratio of >1 significantly affects postoperative cardiopulmonary complications [9]. Regarding long-term outcomes, Kanzaki et al. demonstrated that a larger main PA diameter relative to the body surface area was a significant predictor of poor postoperative survival in patients with lung cancer having COPD [10]. This may be due to increased vascular remodelling and perivascular inflammatory cell accumulation derived from the nuclear factor kappa B chemokine cascade, which might cause more aggressive lung cancer and significantly increased cancer mortality [29]. Furthermore, since tumour-associated macrophages play a central role in lung cancer growth and metastasis, cancer malignancy might be accelerated via microenvironmental inflammation based on tumour cell–immune cell crosstalk in patients with lung cancer having PH [30]. However, a large-cohort validation study is required to clarify the precise aetiology.

We demonstrated that the clinical use of PA and aorta enlargement predicted postoperative survival in patients undergoing curative surgery. In the background, patients with ‘both enlargements’ had more cerebral-cardiovascular disease complications (Table 1) and significantly poorer postoperative survival than those without cerebral-cardiovascular diseases (Supplementary Fig. S2). Therefore, cautious follow-ups, especially for cancer recurrence or cerebral-cardiovascular events, should be performed even in patients with Stage I lung cancer having enlarged PA and aorta. Nonetheless, the significance of both PA and aorta enlargement for predicting postoperative cancer survival has rarely been reported. Large-scale multicentre studies are desirable to ascertain the utility of PA and aorta enlargement in lung cancer surgery.

This study had some limitations. First, we retrospectively analysed a dataset of patients with all types of lung cancer from a single institution. Second, selection bias may have been introduced when selecting patients, operative procedures, and measuring PA and aorta sizes using 5-mm CT sections, not for our study but for preoperative cancer staging. Moreover, a considerable number of patients dropped out, the observation time was long (over 10 years), and post-recurrence treatment and postoperative adjuvant chemotherapy were not uniform. Third, we could not establish a precise causal relationship between PA and aortic dilation and survival. Fourth, a validation cohort was not used to justify the impact of the PA and aortic diameter. Nonetheless, this study analysed robust data for as many as 1482 lung cancers with various PA and aortic diameters, with or without COPD; it demonstrated the utility of both PA and aortic diameters in predicting postoperative lung cancer survival. In the future, a large global cohort study is required to prove the prognostic value of the PA and aortic diameters in patients with lung cancer. This may have significant implications regarding the decision-making process for surgical procedures and cautious follow-ups, especially for cancer recurrence or cerebral-cardiovascular events, in patients with Stage I lung cancer having both PA and aorta enlargement.

Supplementary Material