Background
Esophageal cancer is the seventh most common cancer and the sixth leading cause of cancer-related death worldwide. In 2020, an estimated 604,100 new cases of esophageal cancer and 544,076 deaths were reported [
1]. In 2019 in Japan, esophageal cancer was the eighth most common cancer among men and the seventh leading cause of cancer-related death, with a mortality rate of 9.4 deaths per 100,000 people [
2]. In Europe and the United States, the incidence of esophageal adenocarcinoma is increasing and that of esophageal squamous cell carcinoma (ESCC) is decreasing. In Japan, however, the prevalence of ESCC remains high, and approximately 90% of esophageal cancers are ESCC [
3].
In Japan, ESCC is usually detected by upper gastrointestinal (GI) endoscopy. ESCC is an extremely malignant tumor that predominantly affects the mid to lower esophagus, and as it progresses, it causes esophageal obstruction. In the early stages of the disease, esophageal symptoms such as dysphagia may be mild. As the cancer progresses and the esophageal lumen narrows, severe dysphagia may occur. However, a detailed evaluation of the relationship between subjective symptoms and tumor progression as well as prognosis has not been performed in ESCC until now.
Although many studies have shown that age, male sex, body mass index, smoking, alcohol intake, foods containing N-nitroso compounds, hot drinks, red meat intake, and human papillomavirus are the main risk factors for ESCC [
4,
5], most patients are diagnosed at an advanced stage. Limited therapies are available for these patients, and their prognosis is very poor. The 5-year overall survival rate is less than 40%, and many patients die within 1 year of diagnosis despite advances in therapy [
6]. However, because the prognosis of early-stage ESCC has been shown to be very favorable [
7], its detection at an early stage is considered to be of great clinical importance.
In Japan, there is currently no mass screening program for ESCC. This may be due to the fact that the number of patients with ESCC is relatively small compared with the number of patients with gastric, colorectal, lung, breast, and cervical cancer, for which there are organized screening programs. There is also the possibility that endoscopic screening does not improve the prognosis of ESCC because of its extremely rapid progression. The doubling time is often used as a guide to determine the rate of tumor growth. Our previous study showed that the doubling time of gastric cancer was 10.1 months, whereas that of ESCC was 3.3 months, indicating an extremely high tumor growth rate [
8].
As described above, ESCC is potentially curable if detected early. With the exception of a study from China in 2020 [
9], no report to date has addressed the importance of endoscopic screening for ESCC. The first aim of this study was to investigate the clinical factors that affect the prognosis of patients with ESCC, and the second aim was to elucidate in detail how the subjective symptoms are related to their prognosis.
Methods
Patients
In this retrospective study, we evaluated the clinical records of 503 consecutive patients with esophageal cancer who underwent a thorough examination, including definitive diagnosis by pathologists and staging based on the current American Joint Committee on Cancer TNM system [
10], at Kawasaki Medical School General Medical Center and Kawasaki Medical School Hospital from April 2011 to December 2019. The first lesion detected during the study period was included in the analysis. If multiple lesions were simultaneously detected, the largest lesion was included. Patients who had previously been treated for ESCC at other hospitals and/or for whom the details of the diagnostic procedure were unknown were excluded. The following data were retrospectively reviewed: clinical manifestations, age, sex, body mass index, history of alcohol consumption, smoking history, reason for endoscopic examination, underlying disease, tumor site, tumor size, histopathological grading, macroscopic tumor type, primary treatment, current American Joint Committee on Cancer TNM classification [
10], vital status of patients (survival, death, or loss to follow-up), and duration of follow-up. Six established prognostic factors for ESCC (body mass index, alcohol drinking, cigarette smoking, sex, clinical stage, and age) [
11] and subjective symptoms were used to subgroup patients and analyze survival differences. As shown in Table
1, the 5-year survival rate of stage ≥ III ESCC and stage < III ESCC patients differs greatly. Therefore, we divided the clinical stage into two groups: stage ≥ III ESCC and stage < III ESCC.
Table 1
Patients’ clinicopathological features
Age | 68.8 ± 9.4 |
Sex (men/women) | 385/84 (4.6/1) |
Clinical stage of disease (0/I/II/III/IV) | 81/76/68/118/126 |
Treatment* (ESD or EMR/OPE/CT and or RT/BSC) | 90/257/150/42 |
Outcome (alive/dead/unknown) | 160/212/97 |
Observation period (months) | 29.8 |
Drinking (%) | 392 (83.6%) |
Smoking (%) | 376 (80.2%) |
DM/HT/DL | 80/211/86 |
Diagnosis opportunity (medical checkup/outpatient consultation due to symptoms) | 129/270 |
History of cancer of other organs (%) | 154 (32.8%) |
Lesion location (Ce/Ut/Mt/Lt/Ae) | 25/91/271/69/10 |
Macroscopic tumor type (0/1/2/3/4/5) | 207/18/124/89/8/23 |
Histopathological grading (well/moderate/poor/unknown) | 91/243/71/64 |
Cases of stenosis (%) | 75 (16.0%) |
5-year survival rate |
Stage 0 | 0.98 (95% CI 0.86–0.99) |
Stage I | 0.92 (95% CI 0.77–0.98) |
Stage II | 0.69 (95% CI 0.49–0.83) |
Stage III | 0.35 (95% CI 0.24–0.47) |
Stage IV | 0.16 (95% CI 0.07–0.29) |
Next, the patients were divided into two groups: a symptomatic group and an asymptomatic group. Each clinical stage group was further divided. The symptomatic group comprised patients who visited one of our hospitals for any upper GI symptom, such as dysphagia, chest pain, or chest discomfort, and were found to have ESCC during upper GI endoscopic examinations, whereas the asymptomatic group comprised patients who underwent upper GI endoscopy for screening purposes and were incidentally found to have ESCC. First, the proportions of the symptomatic and asymptomatic groups among all eligible patients were investigated, and differences in clinical characteristics, pathological findings, and prognosis between the two groups were compared. Next, in the symptomatic group, differences in the incidence of subjective symptoms according to tumor size, tumor location, macroscopic tumor type, and clinical stage were examined. Finally, the patients in the symptomatic group were further divided into two groups: patients with swallowing-related symptoms and patients with other symptoms. Swallowing-related symptoms were defined as symptoms of either dysphagia or discomfort during swallowing. The prognosis was compared between the two groups.
This study was approved by the ethics committee of our hospital (Institutional Review Board no. 3806-1) and was conducted in accordance with the ethical principles related to the Declaration of Helsinki. Neither the patients nor the general public were involved in the design, conduct, reporting, or dissemination of the study plan.
Statistical analysis
Continuous and categorical variables are presented as mean ± standard deviation and number (%), respectively. Continuous data were compared using the Mann–Whitney U test. Pearson’s χ2 test or Fisher’s exact test was used to analyze categorical data and compare proportions. The survival rates of patients were plotted using Kaplan–Meier curves, and differences were evaluated using the log rank test. Cox regression analysis was used to estimate the hazard ratio (HR) and calculate the 95% confidence interval (CI). All statistical analyses were performed using SPSS Version 17.0 (SPSS Inc., Chicago, IL, USA). A p-value of < 0.05 (two-sided) denoted a statistically significant difference.
Discussion
This study showed three clinically important results. First, male sex, TNM classification of ≥ 3, and subjective symptoms were independent risk factors for overall survival of ESCC in Japan. Second, the mean time from symptom appearance to diagnosis was short (2.4 ± 4.3 months), and most patients were difficult to treat endoscopically at the time of symptom appearance. Third, the prognosis of patients with swallowing-related symptoms was significantly worse than that of patients with other symptoms.
In our study, TNM classification of ≥ 3 was identified as a prognostic factor. Mariette et al. identified three favorable prognostic factors: low pT stage, pN0 stage, and surgery during the study period [
12]. In a multivariate analysis by Akgul et al. progression of pathological stage had a negative effect on survival [
13]. These reports support our results. However, in the present study, there was a sex difference in prognosis. A recent study by Luo et al. indicated that sex was an independent prognostic factor in Chinese patients with ESCC who underwent definitive radiotherapy, with better survival outcomes in women than men [
14]. In our study, 108 (23.0%) patients received radiotherapy. Previous studies have shown that androgen exposure can facilitate the proliferation of human ESCC cells and that activation of androgen receptors may promote progression of ESCC [
15,
16]. The relationship between androgen levels and prognosis in ESCC patients will be determined in future prospective studies. However, Zhang et al. assessed the role of sex in the prognosis of ESCC in Chinese and in Caucasian patients in the United States and found that it was not an independent prognostic factor in these patients [
17], although they did not provide specific information about variables and treatment modalities, which might have affected the analysis. Therefore, the prognostic significance of sex in ESCC should be interpreted cautiously at present, and a further large prospective study will be necessary to clarify this point.
In this study, subjective symptoms, especially dysphagia, were identified as a prognostic factor. A recent study by Zang et al. in residents aged 40–69 years of a region with a high incidence of upper GI cancer showed that subjects with high livestock/poultry meat intake, low fruit and vegetable consumption, and typical symptoms were at higher risk of ESCC, and they speculated on an association between tooth loss and oral hygiene and esophageal disease [
18]. We did not examine tooth loss or oral hygiene in our present study, and these points will be investigated in the future. Although it has been empirically shown that many patients with ESCC have already progressed by the time subjective symptoms appear, there have been no reports detailing the relationship between subjective symptoms and their prognosis. In this study, we reconfirmed for the first time that endoscopic screening according to patients’ subjective symptoms is not effective. The survival rate of patients diagnosed after the onset of symptoms has reportedly remained unchanged during the past few decades despite advances in treatment methods [
6], which is consistent with our study results. The risk of ESCC is high in men of advanced age, and even higher in heavy drinkers, people who develop flushing of the face after drinking, and smokers. Patients with head and neck cancer, achalasia, and corrosive esophagitis are also considered to be at high risk [
4,
19]. Patients with ESCC in this study had also these risk factors. Therefore, endoscopic screening of high-risk groups, without excessive reliance on subjective symptoms, may be important for the early detection of ESCC, which can be treated effectively and with minimal invasion [
20].
Unfortunately, most cases of ESCC are detected at an advanced stage of the disease; the 5-year survival rate is less than 20% in developed countries and less than 5% in many developing countries, where most cases of ESCC occur [
7]. In this study, 52.0% of consecutive patients with ESCC had advanced cancer, and their 5-year survival rate was 0.211 (95% CI 0.155–0.274), which is considered very poor. However, research has shown that patients with early-stage cancer, which can be treated endoscopically or surgically, have a much better 5-year survival rate of 80%–90% than those with advanced stage [
21]. When the patients with ESCC in this study were limited to those with stages 0 or I cancer, the 5-year survival rate was 0.92–0.98, which is consistent with previous reports. Therefore, the main goal of ESCC screening is to detect ESCC at an early stage, allowing for curative treatment. Several studies have demonstrated that endoscopic screening programs in high-risk populations are associated with early detection, reduced mortality, and improved cost-effectiveness. A community assignment study with 10 years of follow-up in an endemic area of China showed a 33% reduction in ESCC-related cumulative mortality in intervention communities that offered one Lugol’s screening session to adults aged 40–69 years [
22]. Similar results were observed in another large retrospective study of a high-risk Chinese population [
23]. In response to these findings, screening programs have been proposed for high-risk populations worldwide [
24]. Screening programs are being further promoted for the treatment of esophageal squamous dysplasia and early ESCC. Unfortunately, because of the low incidence of ESCC, the effectiveness of population-based cancer screening for ESCC has not been demonstrated in Japan; therefore, no organized screening for ESCC is being conducted at present. We believe that our finding that the prognosis of patients varies greatly depending on the reason for upper GI endoscopy in 469 patients with ESCC in Japan is clinically significant. The results of this study may provide evidence to promote efficient endoscopic screening of Japanese patients with ESCC in the future.
This study had several limitations. First, it was a retrospective cohort study conducted at two institutions. Therefore, it was not possible to investigate individual physicians’ experiences or access the upper GI endoscopy findings in individual residential areas. Prospective studies must be designed to address this issue. Second, the possibility of referral filter bias cannot be ruled out. Such bias would increase the number of patients in the asymptomatic group. However, the characteristics of the patients in the present study were not significantly different from those of the patients in the national survey of the Japan Esophageal Society [
25], and we believe that our data represent the current status of ESCC without bias. Third, among all patients, 38 (8.1%) had risk-free ESCC, i.e., women with no history of alcohol consumption or smoking. The number of patients with subjective symptoms at the time of consultation was 29 (76.3%), which was significantly higher than that of patients with ESCC with any risk. We believe that screening of these patients will also be a clinical issue to be considered in the future. Fourth, in the present study, in patients with multiple lesions, only the largest tumor was included in the analysis. This could be considered a factor that might affect the results of the study. The number of patients with simultaneous multiple lesions in this study group was 72 (15.4%). This figure was consistent with a previous large study in which a proportion of patients had simultaneous multiple lesions in ESCC [
26], confirming that there was no bias in our study population. A previous study by Pesko et al. showed that the tumor stage in patients with multiple tumors was determined by the main tumor stage and was not affected by the associated lesions [
27]. Thus, our inclusion of only the largest tumor of patients with multiple lesions in the analysis is reasonable.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.