Nasopharyngeal Carcinoma (NPC), a rare disease in the world, ranks the most common cause of head and neck cancer in southern China, where the incidence is 20‒40 per 100,000 person-years.1 Owning to the ongoing improvements in radiotherapeutic techniques and chemoradiotherapy, these patients achieved excellent local control and survival, even in those with locally advanced diseases.2 Because NPC is highly sensitive to radiation, Radiotherapy (RT) alone and concurrent chemo-radiotherapy are regarded as the most efficient treatment for early and advanced stages of disease.3 However, after irradiation, the late complications, such as optic neuropathy, brachial plexus injury and endocrine dysfunction are of great concern for patients and radiation oncologists.4

In patients with NPC, RT could damage the Carotid Artery (CA) and increase the risk of carotid atherosclerosis, which may result in transient ischemic attack and stroke.5 Compared patients with non-RT, post-RT NPC patients seemed to have a thicker CA wall, higher prevalence of carotid plaque and/or greater degree of carotid stenosis.6 It is reported that the post-RT patients have twice the risk of ischemic stroke compared with the non-irradiated patients. However, the arterial wall and plaque of carotid stenosis that induced by radiation are histologically similar to spontaneous atherosclerosis. Although previous studies have investigated the differences between irradiated patients and non-irradiated controls, the independent effects of RT on carotid atherosclerosis have been inconsistent among them, especially when the conventional cardiovascular risk factors were excluded. Therefore, we conducted this meta-analysis to investigate the effects of radiation therapy on carotid stenosis in NPC patients, as well as to explore the risk factors for significant carotid stenosis.

In the present study, we performed a meta-analysis to comprehensively assess the effects of radiation therapy on carotid stenosis in NPC patients, as well as to explore the risk factors for significant carotid stenosis. Our results showed that NPC patients treated with RT had a significantly higher risk of overall/significant stenosis than the controls. Moreover, in the subgroup analysis based on the blood vessel and area affected, RT was associated with significantly higher incidences of stenosis in CCA, ICA, ECA and CCA/ICA than control. Age, smoking habit and time interval from RT were independent predictors of significant carotid stenosis.

CA stenosis is a common complication of external irradiation in head and neck cancers.20 Previous studies reported that the incidence of CA stenosis was increased more than 5-years after RT,16,30 but some other studies found that significant CA stenosis could occur as early as 1‒2-years after RT.31,32 Therefore, studies with different duration of follow-up might lead to different results. In this study, we included 16 studies of NPC patients, in which the effect of RT on carotid stenosis was assessed at different periods. Even though, our pooled data suggested that RT increased the risk of overall/significant stenosis. The exact mechanism of RT-induced CA disease is not clear, however, three possible reasons might account for this: (1) Fibrosis due to a damaged vaso vasorum, (2) Adventitial fibrosis producing obstruction, (3) Accelerated atherosclerosis.33 Cheng SW, et al.34 performed a prospective study and reported that, the carotid stenosis associated with external irradiation progressed more rapidly than nonirradiated atherosclerotic arteries.34 In that study, the annualized progression rate in irradiated arteries was higher than that in nonirradiated arteries (15.4% vs. 4.8%).16 The authors concluded that the carotid stenosis induced by external irradiation is not only attributed to pre-mature atherosclerosis alone, but also a more aggressive disease with a different biological behavior.34

In this study, we found that the incidence of overall/significant stenosis in NPC patients who underwent RT was higher than that of the controls. Our results were in accordance with the findings of previous studies.14,15,20 Lam WW, et al.20 examined the incidence of stenosis in the extracrial carotid arteries among NPC patients after RT. They found that the CCA/ICA was more common seen in RT group than in the non-RT group (77.5% vs. 21.6%, p <  0.01), followed by ECA (45.1% vs. 2.0%, p <  0.01) and Vertebral Artery (VA) (7.0% vs. 0%, p =  0.069).20 In addition, significant stenosis was only found in the RT group (CCA/ICA: 29.58%, ECA: 15.49%, VA: 5.63%).20 Therefore, the authors suggested that RT increased the risk of significant carotid stenosis.

Although a variety of studies have confirmed the relationship between RT and carotid stenosis, the incidence of significant stenosis induced by RT varies greatly among them. Carmody BJ, et al.31 and Lam WW, et al.20 reported that the rate of significant stenosis (70%‒90% and ≥ 50% stenosis, respectively) in RT group was 21.7% and 77.5%, respectively, which was higher than that in control groups (4% and 14.3%). The mean radiation dose in these studies was approximately 6000 cGy. Whereas, Chang YJ, et al.15 reported the rate of significant stenosis induced by RT (6225 cGy) was 19.8%, which was lower than the findings of the prior two studies. The different results in the three studies might be caused by the time interval between RT and examination, which was 6.5 years, 9.2 years, and 4.9 years in the studies of Carmody BJ, Lam WW and Chang YJ, respectively.

Another reason for the difference in incidence of carotid stenosis might be the different measurement methods. Zhou L et al.,14 reported that the incidence of CA and VA stenosis in RT group was 73.6% and 81.9%, which was higher than that of 45%/38% and 51.25%/7.04% in the two studies conducted by Law WW, et al.20,29 In the two latter studies, ultrasonography was used to detect the degree of vessel stenosis by measuring the percentage reduction in the diameter of true lumen.20,29 Whereas, in the prior study, the authors used Contrast-Enhanced MR Angiography (CE-MRA) to assess the degree of vessel stenosis by measuring the percentage reduction in the area of true lumen.14 It is reported that area measurement is closely related to the results of Digital Subtraction Angiography (DSA), which is more accurate than the diameter assessment, especially for arteries with irregular lumens.35 Moreover, ultrasonography cannot provide 3D image or detect intracranial CA stenosis. However, MIP images of CE-MRA can provide multiple projections of the CAs and can display a panoramic review of the Cas.36

In this study, we also investigated the risk factors for significant stenosis in patients with NPC after RT. Our results demonstrated that age, smoking habit and time interval from RT were independent predictors of significant carotid stenosis. These findings were in consistent with the results of previous studies.24 Liao W, et al.24 found that age was correlated with significant carotid stenosis in patients undergoing RT, and the age in RT group (54.1 ± 9.5 years) was significantly older than those in non-RT group (50.7 ± 8.0 years) (p =  0.015). Similarly, Cheng SW, et al.17 revealed that patients who had undergone RT more than 5-years earlier had a 15 times higher risk of developing significant carotid stenosis than those with less than 5-years (26% vs. 6%, p <  0.01). Chu PY, et al.26 found that smoking increased the incidence of extracranial artery stenosis in patients with NPC who underwent RT (adjusted Odd Ratio [OR = 4.472], 95% CI: 2.057‒9.725; p <  0.001). For the time interval from RT, both the studies conducted by Liao W, et al.24 and Zhou L, et al.14 showed that, it was associated with higher carotid stenosis risk (OR = 1.068, 95% CI: 1.033‒1.105, p =  0.001; β = 1.076, 95% CI: 0.998‒8.621, p =  0.05). Other risk factors, such as RT dose, hypertension, diabetes mellitus, heart disease, hyperlipidemia, and surgery, were not found to be related with the incidence of significant stenosis in this study. These findings were also in accordance with the results of other studies.23,26 Even so, management of these risk factors is still beneficial for the patients with NPC who underwent RT since it is applicable to the general population.37

There were several potential limitations to note when interpreting our findings. First, the studies included in this meta-analysis were performed with observational design (cohort study, case-control study, or cross-sectional study). Although observational study is more likely to reflect the real-world, it is limited by the selection bias and potential for confounding factors. Second, there were moderate or substantial heterogeneity observed among the included studies. However, the source of heterogeneity for overall carotid stenosis was not identified after sensitivity analysis. This might be explained by the differences across the included studies, such as study design, selection of control, imaging techniques used for carotid stenosis, RT methods, and radiation dose. These factors may account for the heterogeneity and have an impact on the effects. Third, the sample size varied greatly among the included studies, some of which were relatively small. It is reported that studies with small sample size were more likely to overestimate the treatment effect as compared with larger trials. Therefore, large-scale trials are needed to draw definitive conclusions.

The present meta-analysis suggested that patients who underwent RT of NPC had a higher risk of developing carotid stenosis. Thus, patients and physicians should be aware of the complications after RT so that every effort can be made to avoid the development of these complications. Considering the potential limitations in this study, more large-scale, well-designed trials are needed to verify our findings as well as to explore the mechanism for radiation-induced carotid stenosis.

The authors declare no conflicts of interest.