Definitive chemoradiation or surgery in elderly patients with potentially curable esophageal cancer in the Netherlands: a nationwide population-based study on patterns of care and survival

Abstract

Background

The aim of our study was to describe treatment patterns and the impact on overall survival among elderly patients (75 years and older) with potentially curable esophageal cancer.

Material and methods

Between 2003 and 2013, 13,244 patients from the nationwide population-based Netherlands Cancer Registry (NCR) were diagnosed with potentially curable esophageal cancer (cT2–3, X, any cN, cM0, X) of which 34% were elderly patients (n = 4501).

Results

Surgical treatment with or without neoadjuvant treatment remained stable among elderly patients (around the 16% between 2003 and 2013). However, among younger patients, surgical treatment increased from 60.2 to 67.0%. The use of definitive chemoradiation (dCRT) increased in elderly patients from 1.9 to 19.5% and in younger patients from 5.2 to 17.2%. Due to the increase in dCRT, treatment with curative intent doubled in the elderly from 17 to 37.1%. Multivariable Cox regression revealed that elderly patients with an adenocarcinoma receiving surgery alone or dCRT had a significantly worse overall survival compared to those receiving surgery with neoadjuvant chemo (radio) therapy (nCRT/CT) (HR: 1.7 95% CI 1.4–2.0 and HR: 1.9 95% CI 1.5–2.3). However, among elderly with squamous cell carcinoma overall survival was comparable between dCRT, surgery alone and surgery with nCRT/CT.

Conclusions

Survival was comparable among elderly patients with squamous cell carcinoma who underwent surgery with nCRT/CT, surgery alone or received dCRT, while elderly patients with an adenocarcinoma who underwent surgery with nCRT/CT had a better overall survival when compared with surgery alone or dCRT. Therefore, dCRT can be considered as a reasonable alternative for surgery among potentially curable elderly patients with esophageal squamous cell carcinoma. However, in elderly patients with esophageal adenocarcinoma surgery with nCRT/CT is still preferable regarding overall survival.

Introduction

The incidence of esophageal cancer, especially adenocarcinoma, has increased dramatically over the past four decades in the Western world and is still rising but at a slower rate than previously [1,2]. Esophageal cancer is mainly a disease of the elderly as a significant number of patients is aged between 60 and 85 years at the time of diagnosis [3,4]. In the Netherlands, approximately 30% of all newly diagnosed patients with esophageal cancer is 75 years or older [5].

According to the Dutch clinical practice guidelines, the preferred treatment for patients with potentially curable esophageal cancer is neoadjuvant chemoradiation followed by a subsequent esophagectomy. Early esophageal cancer (T1a) can be treated with Endoscopic Mucosal Resection (EMR) [6]. Frail patients unfit for surgery, such as some elderly patients, can be treated alternatively with a curative intention using definitive chemoradiation (dCRT) [7,8]. Furthermore, histological subtype plays a role in treatment of patients with potentially curable esophageal cancer. For example, patients with squamous cell carcinoma seem to have a better response to dCRT compared to patients with an adenocarcinoma [9–11].

Surgical treatment of esophageal cancer is complex with a high postoperative complication rate, especially in elderly patients with multiple co-morbidities, which might be an argument to withhold some patients from surgical treatment [12,13]. A previous study has shown an increase in 30 d postoperative mortality from 4.9% in patients younger than 65 years to 10.3% in patients older than 75 years [14].

However, most treatment strategies and guidelines are based on clinical trials in which elderly patients are excluded. Therefore, it is of significant importance to investigate the effect of different treatment options on survival in this specific group of patients. The aim of our study was to describe treatment patterns and the impact on overall survival in elderly patients (75 years and older) with potentially curable esophageal cancer (adenocarcinoma or squamous cell carcinoma) in the Netherlands.

Patients and methods

Data collection

Nationwide population-based data from the Netherlands Cancer Registry (NCR) were used. The NCR is based on notification of all newly diagnosed malignancies in the Netherlands by the national automated pathological archive (PALGA). Additional sources are the national registry of hospital discharge, radiotherapy institutions and diagnosis therapy combinations (specific codes for reimbursement purposes). Specially trained data managers of the NCR routinely extracted information on diagnosis, tumor stage and treatment from the medical records. Information on vital status was obtained through an annual linkage with the Municipal Administrative Database, in which all deceased and emigrated persons in the Netherlands were registered. Institutional Review Board approval was obtained from the NCR.

Patients

Between January 2003 and December 2013, 25,638 patients were diagnosed with an adenocarcinoma or squamous cell carcinoma of the esophagus or gastroesophageal junction (GEJ) in the Netherlands. The topography and morphology of the tumors were coded according to the International Classification of Diseases for Oncology (ICD-O-3) [15]. Subsite distribution was divided as: proximal (C15.0, C15.3), mid (C15.4), distal (C15.5), overlapping or not otherwise specified (C15.8, C15.9) and GEJ (C16.0). Patients diagnosed from 2003 to 2009 were staged according to TNM-6, whereas patients diagnosed from 2010 to 2013 were staged according to TNM-7 [16,17].

Patients with potentially curable esophageal tumors were eligible for this study (Figure 1). Patients were considered potentially curable in this study if they had no clinically distant metastasis (cM1b for TNM-6 and cM1 for TNM 7) (n = 8009) and no tumors infiltrating surrounding organs (cT4 according to TNM-6 and cT4A, and cT4B according to TNM-7) (n = 1368). We excluded patients with tumors infiltrating surrounding organs since it was uncertain whether or not these patients were eligible for curative treatment. For the analyses, patients with a cM1A tumor according to TNM-6 were categorized as having cN + as most patients with a cM1A tumor had a distal tumor with celiac lymph nodes which can be considered as having cN + according to TNM-7. Furthermore, patients with unknown clinical distant metastases (cMX) were included. It should be noted that as of 2010 coding regulations to register a cM0 or cM1 status into the NCR were less strict than before 2010, and, therefore, as of 2010, relatively more patients were registered with a cM0 rather than a cMX into the NCR. To account for this, we decided to include all patients with cMX. Patients with an in-situ or a cT1 tumor (n = 1002) were also excluded since these tumors are treated predominantly with an EMR rather than surgical treatment. In addition, patients with missing/unknown treatment (n = 92) and patients receiving EMR alone (n = 350) were excluded. This resulted in 13,244 patients with a potentially curable esophageal carcinoma (cT2, 3, X, any cN, cM0, X). Of these patients, 4501 (34%) were elderly patients being 75 years and older (Figure 1).

Figure 1. Flowchart of the study population. *cT4 according to TNM-6 and cT4a and cT4b according to TNM-7. **cT1 according to TNM-6 and cT1a and cT1b according to TNM-7.

Treatment

Surgery with potentially curative intent was defined as a transhiatal esophagectomy or transthoracic esophagectomy. dCRT was defined as the combination of radiotherapy and chemotherapy as primary treatment without surgery. Curative treatment was defined as dCRT, surgery alone or surgery with neoadjuvant chemoradiotherapy or chemotherapy (nCRT/CT). All other treatments were defined as ‘other’ therapy.

Statistical analysis

Differences in patient and tumor characteristics between elderly patients with an adenocarcinoma and squamous cell carcinoma were described and compared using the Pearson’s Chi-square test for nominal data. For differences in continuous variables, the independent t-test was used. Survival time was defined as time from diagnosis to death or until 1 February 2016 for patients who were still alive. Survival curves per treatment option were obtained using the Kaplan–Meier method for elderly patients according to histology. Differences in overall survival according to treatment were assessed by using log-rank tests. Multivariable Cox regression analysis was performed to evaluate independent prognostic factors for overall survival. All statistical analyses were performed using Statistical Package for Social Sciences version 22.0 (IBM Corporation, Armonk, NY, USA) and p values less than .05 were considered statistically significant.

Results

Patient characteristics

Of the potentially curable elderly patients of 75 years and older diagnosed with an esophageal carcinoma, 75.6% (n = 3402) was diagnosed with an adenocarcinoma and 24.4% (n = 1099) with a squamous cell carcinoma. There were no significant differences in age, cT-stage, cN-stage and cM-stage between both histology groups. However, patients with an adenocarcinoma had more often a distally located tumor and a poor tumor differentiation. Furthermore, elderly patients with an adenocarcinoma more often received surgical treatment (21.3%) than dCRT (7.7%), whereas patients with a squamous cell carcinoma more often received dCRT (13.1%) than surgery (10.4%) (Table 1).

Table 1. Patient characteristics of the elderly patient (≥75 years) diagnosed with potentially curable esophageal cancer in the period 2003–2013 (n = 4501).

	Adenocarcinoma	Squamous cell carcinoma
Total N = 4501	3402 (75.6%)	1099 (24.4%)	p value
Mean age	81.6 (SD 4.9)	81.3 (SD 5.0)	.051
Gender
Male	2374 (69.8%)	514 (46.8%)
Female	1028 (30.2%)	585 (53.2%)	<.001
cT-stage
2	582 (17.1%)	171 (15.6%)
3	863 (25.4%)	308 (28.0%)
Unknown/missing	1957 (57.5%)	620 (56.4%)	.163
cN-stage
N0	1167 (34.3%)	388 (35.3%)
N+	1173 (34.5%)	395 (35.9%)
Unknown/missing	1062 (31.2%)	316 (28.8%)	.302
cM-stage
M0	2790 (82.0%)	921 (83.8%)
Unknown/missing	612 (18.0%)	178 (16.2%)	.174
Tumor location
Proximal	38 (1.1%)	164 (14.9%)
Mid	208 (6.1%)	398 (36.2%)
Distal	1983 (58.3%)	454 (41.3%)
GEJ^a	1040 (30.6%)	7 (0.6%)
Overlapping/NOS^b	133 (3.9%)	76 (6.9%)	<.001
Differentiation
Well	67 (2.0%)	31 (2.8%)
Moderate	677 (19.9%)	296 (26.9%)
Poor	1147 (33.7%)	254 (23.1%)
Unknown	1511 (44.4%)	518 (47.1%)	<.001
Type of treatment
Surgery with nCRT/CT^c	250 (7.3%)	59 (5.4%)
Surgery alone	475 (14.0%)	55 (5.0%)
Definitive chemoradiation	261 (7.7%)	144 (13.1%)
Other/no treatment	2416 (71.0%)	841 (76.5%)	<.001

^a Gastro-esophageal junction.

^b Not otherwise specified.

^c Among this group of patients 77.3% received nCRT and 22.7% received CT. Two patients received CRT postoperatively.

Of all elderly patients diagnosed with potentially curable esophageal carcinoma, 6.9% received surgery with nCRT/CT, 11.8% received surgery alone, 18.6% received surgery, 9.0% received dCRT and 72.4% received other/no treatment.

Trends in treatment

From 2003 until 2013, the use of surgery with nCRT/CT among the elderly (≥75 years) and the younger patients (<75 years) increased over time from 0.5 to 13.5% and from 14.4 to 63.3%, respectively. In line with these findings, the proportion of patients with underwent surgery alone decreased among both the elderly and the younger patients from, respectively, 14.5 to 4.2% and from 45.8 to 3.7%. The use of surgical treatment (surgery with nCRT/CT or surgery alone) among all elderly patients (≥75 years) remained relatively stable over time from 15.0% in 2003 to 17.7% in 2013, whereas among the younger patients (≥75 years) the use of surgical treatment increased over time from 60.2% in 2003 to 67.0% in 2013. Furthermore, there was an increase in administration of dCRT in elderly patients from 1.9 to 19.5% as well as in the younger patients from 5.2 to 17.2% (Figure 2(a)).

Figure 2. Trends in treatment of patients with esophageal carcinoma according to age and histology. (a) Trends in treatment of all patients with esophageal carcinoma. (b) Trends in treatment of patients with a squamous cell carcinoma. (c) Trends in treatment of patients with an adenocarcinoma.

The increase in dCRT was most prominent among elderly patients with a squamous cell carcinoma in which treatment with dCRT increased from 3.5 to 30.7%, while among younger patients with squamous cell carcinoma an increase from 9.5 to 29.3% was observed (Figure 2(b)). In patients with an adenocarcinoma, the increase in use of dCRT was comparable in the elderly patients compared to the increase among younger patients (Figure 2(c)). Mainly due to the increase in dCRT, the administration of treatment with curative intent (surgery or dCRT) doubled over time in all elderly patients from 17 to 37.1%. The increase of treatment with a curative intent quadrupled over time in the elderly patient with squamous cell carcinoma from 10.5 to 41.2%. However, the increase in the use of treatment with curative intent was less prominent in the younger patients (Figure 2).

Survival

Overall, elderly patients with a potentially curable adenocarcinoma had a comparable 1- and 3-year overall survival rate compared to elderly patients with a potentially curable squamous cell carcinoma with 1-year overall survival rates of 40.8 vs. 36.5% and 3-year survival rates of 12.0 vs. 14.1%, respectively (log rank p = .621). Furthermore, the 1-year overall survival in elderly patients with an adenocarcinoma treated with surgery and nCRT/CT was 79.6% which was comparable to the overall survival of patients treated with surgery alone (64.8%) or dCRT (72.4%) whereas 3-year overall survival was significantly better for patients who underwent surgery with nCRT/CT (51.2%) compared to patients receiving surgery alone (29.5%) or dCRT (11.6%) (p < .001) (Figure 3(a)). Among elderly patients with a squamous cell carcinoma, patients receiving surgery with nCRT/CT had a better 3 year overall survival (50.2%) compared to surgery alone (40.0%) and dCRT (36.8%), however, this difference was not statistically significant (p = .267) (Figure 3(b)).

Figure 3 (a). Kaplan–Meier survival analysis among elderly patients with an adenocarcinoma (n = 3402). (b). Kaplan–Meier survival analysis among elderly patients with a squamous cell carcinoma (n = 1099).

Multivariable Cox regression analysis showed that patients with male gender, a poor tumor differentiation, an overlapping tumor/not otherwise specified tumor location, cT3 tumors, regional lymph nodes metastasis and squamous cell histology had a significantly worse overall survival. Regarding the treatment strategy, the multivariable Cox regression analysis which included both histologies showed that elderly patients who received surgery alone (hazard ratio [HR]: 1.6, 95% confidence interval [CI] 1.3–1.9), dCRT (HR: 1.7, 95% CI 1.4–2.0) or other/no treatment (HR: 4.1, 95% CI 3.5–4.8) had a significantly worse overall survival compared to patients who underwent surgery with nCRT/CT (Table 2).

Table 2. Multivariable Cox survival analysis for all elderly patients and according to histology.

Multivariable analysis*	All elderly patients (N = 4501)			Adenocarcinoma (N = 3402)			Squamous cell carcinoma (N = 1099)
	HR	95%CI	p value	HR	95%CI	p value	HR	95%CI	p value
Gender
Male	1.0			1.0			1.0
Female	0.9	0.8–1.0	.001	0.9	0.8–1.0	.015	0.8	0.7–0.9	.003
Period of diagnosis
2003–2006	0.9	0.9–1.0	.084	0.9	0.8–1.0	.058	1.0	0.8–1.1	.581
2007–2010	0.9	0.9–1.0	.085	1.0	0.9–1.0	.304	0.9	0.7–1.0	.097
2011–2013	1.0			1.0			1.0
Tumor differentiation
Well	0.8	0.6–1.0	.031	0.6	0.5–0.8	.001	1.7	1.1–2.5	.009
Moderate	0.9	0.8–0.9	<.001	0.8	0.7–0.9	<.001	1.1	0.9–1.3	.269
Poor	1.0			1.0			1.0
Unknown/missing	0.8	0.8–0.9	<.001	0.8	0.7–0.8	<.001	1.0	0.8–1.2	.922
Tumor location
Proximal	0.8	0.7–0.9	.008	1.0	0.7–1.4	.954	0.7	0.6–0.9	.002
Mid	0.9	0.8–1.0	.041	1.0	0.8–1.1	.721	0.9	0.7–1.0	.030
Distal	1.0			1.0			1.0
GEJ^a	1.0	0.9–1.1	.828	1.0	0.9–1.1	.997	1.9	0.9–4.0	.104
Overlapping/NOS^b	1.3	1.1–1.5	.001	1.2	1.0–1.4	.042	1.4	1.1–1.8	.008
Clinical T-stage
cT2	0.9	0.8–1.0	.019	0.9	0.8–1.0	.019	0.9	0.7–1.1	.411
cT3	1.0			1.0			1.0
cT unknown	1.2	1.1–1.3	<.001	1.2	1.1–1.3	.002	1.4	1.2–1.7	<.001
Clinical N-stage
cN0	1.0			1.0			1.0
cN+	1.2	1.1–1.3	<.001	1.2	1.1–1.3	<.001	1.2	1.0–1.4	.057
cN unknown	1.6	1.4–1.7	<.001	1.5	1.4–1.7	<.001	1.8	1.5–2.1	<.001
Type of treatment
Surgery with nCRT/CT	1.0			1.0			1.0
Surgery alone	1.6	1.3–1.9	<.001	1.7	1.4–2.0	<.001	1.3	0.8–2.1	.222
Definitive chemoradiation	1.7	1.4–2.0	<.001	1.9	1.5–2.3	<.001	1.4	0.9–2.0	.123
Other/no treatment	4.1	3.5–4.8	<.001	4.3	3.6–5.1	<.001	3.8	2.7–5.4	<.001
Histology
Squamous cell	1.1	1.0–1.2	.037
Adenocarcinoma	1.0

^a Gastro-esophageal junction.

^b Not otherwise specified.

* Adjusted for all variables listed in Table 2.

Comparable results were found for elderly patients with an adenocarcinoma. Among elderly patients with an adenocarcinoma, patients receiving surgery alone (HR: 1.7, 95% CI 1.4–2.0), dCRT (HR: 1.9, 95% CI 1.5–2.3) or other/no treatment (HR: 4.3, 95% CI 3.6–5.1) had a significantly worse overall survival compared to patients receiving surgery with nCRT/CT. However, among elderly patients with a squamous cell carcinoma overall survival was comparable for patients who underwent surgery alone (HR: 1.3, 95% CI 0.8–2.1), dCRT (HR: 1.4, 95% CI 0.9–2.0) or surgery with nCRT/CT (Table 2).

Discussion

This large nationwide population-based study among elderly patients with potentially curable esophageal cancer who were 75 years or older revealed an increase in treatment with a curative intent, with a consistent use of surgical treatment and a significant increase in the use of dCRT among all elderly patients in the period 2003–2013. The increase in administration of dCRT was most prominent in elderly patients with a squamous cell carcinoma. Furthermore, multivariable analysis showed no difference in overall survival for elderly patients with a squamous cell carcinoma who received surgery with nCRT/CT or surgery alone or dCRT. However, elderly patients with an adenocarcinoma who underwent surgery with nCRT/CT had a significantly better overall survival compared to patients who underwent surgery alone or dCRT.

Despite the increase in the use of treatment with curative intent among potentially curable elderly patients, explained by the increase in dCRT, there is still a large proportion of patients that were not treated with curative intent (72.4%).This study demonstrates that the elderly patients with potentially curable tumors received less often surgical treatment compared to younger patients (17.7 vs. 67.0%), whereas the use of dCRT was slightly higher in the elderly patients compared to the younger patients (19.5 vs. 17.2%). These findings may be explained by the fact that an older age is a risk factor for postoperative morbidity and mortality after esophagectomy [12,18–20]. Although other studies have shown that age alone should not be regarded as a predictor for worse overall survival after esophagectomy, in daily practice, it appears that advanced age is a significant factor in decision making whether or not patients are proposed for surgery [12,13].

Our study revealed a relatively stable use of surgical treatment and a significant increase in use of dCRT among all elderly patients during the study period especially after 2010. This striking increase in administration of dCRT is higher compared to another study in the Netherlands in an earlier period (1989–2008) in which they reported an increase from 0.19 to 2.20% [21]. The increase in use of dCRT is probably caused by the increasing awareness that dCRT has a favorable survival, especially among patients with squamous cell carcinoma and is often well tolerated, even in patients with considerable co-morbidity [7,8]. Although toxicity after chemoradiation is occurring frequently, with 75% of the patients experiencing toxicity of grade 3 or greater, especially in the elderly patients, it is often manageable [22,23].

This study showed that elderly patients with an adenocarcinoma received more often surgical treatment compared to patients with a squamous cell carcinoma which received more often dCRT especially after 2010. These results are in line with result from a large population-based study in the United States [9]. The observed difference in treatment could be explained by the fact that most studies show a better response to dCRT of squamous cell carcinomas when compared to adenocarcinoma, with a better overall survival and disease-free survival in good responders [10]. On the other hand, a study from the United Kingdom on dCRT revealed a comparable overall survival and disease-free survival between both histological subtypes. However, patients with squamous cell carcinoma had significantly more advanced stages of disease [24]. Furthermore, a significant difference in relapse pattern has been described, with adenocarcinomas being more likely to relapse in distant sites and squamous cell carcinoma more likely to recur locally [8,24].

The multivariable Cox survival analysis revealed that elderly patients with an adenocarcinoma who received surgery with nCRT/CT have a better overall survival compared to the patients receiving surgery alone or dCRT. However, among elderly patient with squamous cell carcinomas, there was no significant difference in overall survival between patients who underwent surgery with nCRT/CT, surgery alone or patients who received dCRT. Currently, there are only three randomized control trials which have directly compared dCRT with surgical treatment in patients with squamous cell carcinoma. These trials have shown comparable survival rates in patients treated with dCRT or chemoradiation followed by surgery [11,25,26]. However, in two of the three trials elderly patients were excluded and in the third trial results were not reported for elderly patients as a separate group. Furthermore, a recent Cochrane review states that there is only low-quality evidence in the literature which showed that chemoradiation appears to be equivalent to surgery in squamous cell carcinoma who are responsive to chemoradiation, however, in adenocarcinoma there is uncertainty whether or not patients receiving dCRT benefit compared to surgery [27]. Our results provide more arguments for the equivalence of dCRT to surgery in squamous cell carcinoma and confirm their statement on adenocarcinoma. The results of our study advocate for further research in which the use of dCRT and surgery are compared for disease-free survival and quality of life.

Univariable and multivariable survival analysis also revealed a similar overall survival for patients with esophageal squamous cell carcinoma who underwent surgery with or without nCRT/CT. Although there seems to be immortal time bias, the Kaplan–Meier curves for these treatment groups were parallel, assuming overall survival is comparable (Figure 3(b)). Immortal time bias exists of patients receiving nCRT/CT which takes more time to receive than surgery alone. However, no landmark analysis was performed as this would result in exclusion of many patients in the ‘other/no treatment’ group. Multivariable analysis confirmed the non-significant difference in overall survival between squamous cell carcinoma patients with and without nCRT/CT. These results are in contrast with results from the CROSS trial [28,29], which showed an improved survival for patients who received surgery with nCRT compared to patients who received surgery alone. Moreover, the difference in overall survival was higher for squamous cell carcinoma compared to adenocarcinoma. However, most elderly patients did not meet the eligibility criteria from the CROSS trial. Therefore, further research should investigate the difference in outcomes between surgery with or without nCRT/CT among elderly patients with esophageal squamous cell carcinoma.

A limitation of this study is that the NCR did not register nationwide information on co-morbidity or performance status during the study period. This might have influenced the survival analyses since co-morbidity and performance status plays an important role in the clinical decision making, especially among the elderly patients and has a significant influence on overall survival. However, the survival benefit for dCRT might even be more than observed, because especially unfit patients with multiple co-morbidities and an a priori unfavorable prognosis receive dCRT. Thus, the lack of co-morbidity data might even lead to an underestimation of the potential favorable impact of dCRT on overall survival. This study has also several strengths, such as its observational nature resulting in a representative nationwide population and, therefore, enabling the demonstration of current patterns of care and its impact on overall survival among elderly patients with esophageal cancer in daily clinical practice.

In conclusion, this large nationwide population-based study revealed that there was a consistent use of surgical treatment and a major increase in use of dCRT among all elderly patients with potentially curable esophageal cancer in the period 2003–2013. The increase in dCRT was most prominent among patients with squamous cell carcinoma. Survival was comparable among elderly patients with squamous cell carcinoma who underwent surgery with nCRT/CT, surgery alone or received dCRT, while elderly patients with an adenocarcinoma who underwent surgery with nCRT/CT had a better overall survival when compared with surgery alone or dCRT. Therefore, dCRT can be considered as a reasonable alternative for surgery among potentially curable elderly patients with esophageal squamous cell carcinoma. However, in elderly patients with esophageal adenocarcinoma surgery with nCRT/CT is still preferable regarding overall survival.

Disclosure statement

No potential conflict of interest was reported by the authors.