Printer Friendly

Mortality by treatment in patients [greater than or equal to] 80 years of age with gastroesophageal cancer seen in a 20-year period at a single medical center.

The treatment approach to patients 80 years of age and older with gastroesophageal cancer at Baylor Scott and White in Temple, Texas, has historically favored conservative measures in the form of palliation and observation. To evaluate this trend in practice, the administered treatment(s) and subsequent patient outcomes of this group were retrospectively reviewed.


The study group included all patients 80 years of age and older with a diagnosis of gastroesophageal cancer, regardless of stage (I--IV) or performance status (Eastern Cooperative Oncology Group 0-4), seen in our facility from 1991 to 2010. The treatment course (i.e., surgery, chemotherapy, radiation therapy, or any combination of these modalities) and overall survival in months were retrospectively reviewed.

Descriptive statistics were reported using minimum, 25th percentile, median (or 50th percentile), mean, 75th percentile, maximum, and standard deviation for continuous variables. Categorical variables were described as counts and percentages. Kaplan-Meier curves were constructed for gender, age group, chemotherapy, radiation, and surgery. Cox proportional hazard models were fitted to the data. The variables included in the model were all variables with a P value < 0.25 in the univariate survival analysis. Residual diagnostics indicated a good fit of the model and no significant deviations from the model assumptions. A level of 0.05 was considered statistically significant for all other tests. SAS 9.2 was used for the statistical analysis. R software version 3.1.0 was used for the survival curves.


Between 1991 and 2010, a total of 117 patients over the age of 80 were diagnosed with gastroesophageal cancer at our facility. Table 1 depicts overall demographic and baseline characteristic information, highlighting the heterogeneity in presentation and treatment approach of gastroesophageal cancer, even in this highly specific age group. Of note, 62% of the study population was male, and 62% were between the ages of 80 and 84 at the time of inclusion. Only 29 patients (31%) presented with stage I disease, while stages II, III, and IV comprised 10%, 25%, and 34%, respectively. The most common site of primary malignancy was the gastric fundus (50%), followed by esophageal (30%) and gastric cardia (20%).

Table 2 details the frequency of treatment modalities administered. Interestingly, 57 (49%) patients received no treatment; 25% of patients underwent surgery, 11% received chemotherapy, and 27% received radiation therapy, alone or in combination with another modality. Table 3 illustrates the effects baseline characteristics may have had on treatment decisions, demonstrating that most patients who received at least one of the available treatments were between the ages of 80 and 84 and had more advanced (stage 3-4) or aggressive (grade III) disease.

The median overall survival for patients who received treatment was significantly longer relative to those without treatment, regardless of modality (Figure 1). Specifically, the median overall survival for patients who underwent surgical intervention was nearly double that of those who did not (6.8 months [95% confidence interval (CI) 3.9-19.9] vs. 3.9 months [95% CI 2.8-6.3]; P = 0.02; Figure 1a); chemotherapy offered an almost 4-fold overall survival benefit (14.8 months [95% CI 5.3-30.6] vs. 3.9 months [95% CI 3.0-5.8]; P = 0.03; Figure 1b); and radiation therapy, a greater than 3-fold overall survival benefit (11.1 months [95% CI 6.3-15.0] vs. 3.5 months [95% CI 2.4-4.6]; P = 0.04; Figure 1c).

A Cox proportional hazards model was fitted to the data. Backward, forward, and stepwise variable selection procedures chose gender, stage, and treatment with chemotherapy to be part of the model. Age group was also included in the reduced model. The model was statistically significant (P < 0.0001). Stage and chemotherapy had coefficient estimates significantly different than 0 (Table 4). Gender and age group were not significant. Patients without chemotherapy treatment were 2.4 times more likely to die than patients treated with chemotherapy. Patients with stage 4 disease were 5 times more likely to die than patients with stage 1, but otherwise no higher hazards for other stage combinations was identified. A Cox model was fitted to the data stratified by age group. Although this model seemed to indicate an interaction between age group and gender, this interaction could not be explored due to the small sample size.


The complexity of gastrointestinal oncologic resection imposes significant morbidity and mortality. A recently published study found that increasing age independently impacted outcome following esophagectomy, particularly mortality and discharge disposition, and patients who are at least 80 years of age considering esophagectomy should be recognized as a high-risk cohort (1). The same study also concluded that age should not necessarily be a contraindication for esophagectomy, but patients over the age of 80 must be carefully selected on a case-by-case basis and serious consideration of nonoperative treatment is warranted. Similarly, a systematic review published in 2007 found no evidence of inferior survival or increased treatment-related mortality in the elderly with experimental treatments compared with younger patients (2). The available data also suggest that older patients are just as willing to undergo chemotherapy as their younger counterparts, although less willing to endure severe treatment-related side effects (3). This perspective, however, is uncommonly the standpoint from which treatment strategies for older patients are formulated. In reviewing our data, almost 50% of patients 80 years of age or older with a diagnosis of gastroesophageal cancer did not receive any of the three available modalities of treatment, even though patients who received treatment fared far better.


This study had limitations. It failed to consider treatment bias, in that perhaps those patients selected to undergo treatment fared better because they were "healthy" enough to undergo treatment, while those who were not offered, or opted against, treatment had inferior survival because they were more "sick" as a cohort and not candidates for therapy. It also did not reflect quality of life during the gained months of survival in those who underwent treatment. Safety data and side effects were also not discussed.

With over 20 years of data, we hoped to also identify trends in mortality among the treatment and nontreatment groups over time, to perhaps demonstrate the effectiveness of modern therapies compared to those used 20 years ago. Unfortunately, our sample size was not sufficient for such an analysis, but the question would be an interesting avenue of future study. Of note, within the timeframe studied, there were no major advances largely affecting overall survival. The development of minimally invasive approaches to gastroesophageal cancer resection offered an attractive alternative to traditional transthoracic open surgery, although retrospective evaluation demonstrated identical mortality and overall surgical morbidity (4). To date, there remains no standard chemotherapy regimen for use in the neoadjuvant, combined modality setting (4). In addition, targeted therapies with known survival benefit in gastroesophageal cancer are currently limited to trastuzumab and ramucirumab, although both were approved for use after the close of our study (5).

These data represent a significant contribution to the ongoing debate over treatment strategies for those over the age of 80 diagnosed with gastroesophageal cancer. Based on these findings, while age is an important predictor of overall outcome and should play a substantial role in the decision making on approach to treatment, age alone should not be the deciding factor, but one of many patient-specific variables that warrant consideration. As our institution, the treatment of patients over the age of 80 with gastroesophageal cancer should be considered more frequently and advocated with proper selection and appropriate counseling.


The authors acknowledge the contributions of Alejandro Arroliga, MD, and Mark Holguin, MD.

(1.) Stahl CC, Hanseman DJ, Wima K, Suton JM, Wilson GC, Hohmann SF, Shah SA, Abbott DE. Increasing age is a predictor of short-term outcomes in esophagectomy: a propensity score adjusted analysis. J Gastrointest Surg 2014; 18(8):1423-1428.

(2.) Kumar A, Soares HP, Balducci B; National Cancer Institute. Treatment tolerance and efficacy in geriatric oncology: a systematic review of phase III randomized trials conducted by five National Cancer Institute-sponsored cooperative groups. J Clin Oncol 2007; 25(10):1272-1276.

(3.) Yellen SB, Cella DF, Leslie WT. Age and clinical decision making in oncology patients. J Natl Cancer Inst 1994; 86(23):1766-1770.

(4.) Lockhart CA, El-Khoueiry AB, Krasna M. Update on upper gastrointestinal cancers: what's new since the 2009 ASCO annual meeting? ASCO University 2010 Education Book. Category: Gastrointestinal (Noncolorectal) Cancer. Available at; accessed April 20, 2015.

(5.) Burtness B, Ilson D, Iqbal S. New directions in perioperative management of locally advanced esophagogastric cancer. ASCO University 2014 Education Book. Category: Gastrointestinal (Noncolorectal) Cancer. Available at ASCO%Annual%20Meeting; accessed April 29, 2015.

Tara Barnett, MD, James Mason, DO, Yolanda Munoz Maldonado, PhD, and Lucas Wong, MD

From the Division of Hematology and Oncology (Barnett, Wong), the Department of Internal Medicine (Mason), and the Office of Biostatistics (Munoz Maldonado), Baylor Scott & White Health, Scott & White Memorial Hospital, Temple, Texas.

Corresponding author: Tara Barnett, MD, Baylor Scott & White Health, Scott & White Memorial Hospital, MS-01-692, 2401 South 31st Street, Temple, TX 76508 (e-mail:
Table 1. Patient characteristics for 117 patients with
gastroesophageal cancer over the age of 80 treated at a single
institution *

                 Variable          N (%)      Hazard ratio        P
                                                (95% CI)        value
Gender           Female         44 (38%)    1.6 (1.04, 2.39)    0.31
                 Male           73 (62%)
Age (years)      80-84          72 (62%)    0.7 (0.44, 0.98)    0.04
                 85+            45 (38%)
Race             White          102 (87%)   0.6 (0.33, 1.11)    0.14
                 Nonwhite       15 (13%)
Stage            0-1            29 (31%)    0.24 (0.13, 0.44)   <0.001
                 2              10 (10%)    0.41 (0.78, 0.87)   0.03
                 3              24 (25%)    0.54 (0.30, 0.94)   0.03
                 4              32 (34%)
Grade            I-II           22 (23%)    0.27 (0.12, 0.64)   0.002
                 III            63 (67%)    0.65 (0.18, 0.78)   0.005
                 IV              9 (10%)
Site             Esophagus      35 (30%)    0.99 (0.56, 1.79)   0.97
                 Gastric        58 (50%)    1.25 (0.75, 2.20)   0.41
                 Cardia         24 (20%)
Treatment        Surgery        29 (25%)      1.76 (1.1,2.9)    0.02
                 Chemotherapy   13 (11%)    0.48 (0.23, 0.89)   0.03
                 Radiotherapy   31 (27%)    1.57 (1.02, 2.48)   0.047

* Univariate hazard ratios compared with overall mortality. Baseline
references for vari-ables with more than one level in the hazard
ratio are always the last level. CI indicates confidence interval.

Table 2. Frequency of treatment modalities administered for 117
patients with gastroesophageal cancer over the age of 80 treated
at a single institution

Treatment modality                     Frequency   Percent

No treatment                              57         49%
Surgery alone                             25         21%
Chemotherapy alone                         3         3%
Radiation alone                           22         19%
Surgery and chemotherapy                   1         <1%
Chemotherapy and radiation                 6         5%
Surgery, chemotherapy, and radiation       3         3%

Table 3. Patient characteristics by treatment for patients with
gastroesophageal cancer over the age of 80 treated
at a single institution *

                   Variable    Surgery    Chemotherapy    Radiation
                               (N = 29)     (N = 13)      (N = 31)

Gender             Female      10 (35%)     2 (15%)        8 (26%)
                   Male        19 (65%)     11 (85%)      23 (74%)
Age (years)        80-84       20 (69%)     10 (77%)      17 (55%)
                   85+         9 (31%)      3 (23%)       14 (45%)
Race               White       27 (93%)     10 (77%)      24 (77%)
                   Nonwhite     2 (7%)       3 (23)        7 (23%)
Follow-up,                       6.3          14.1          10.2
months: mean, SD                (0.33,    (3.0, 48.0)    (2.2, 48.0)

Stage              0-1         9 (33%)      2 (20%)       10 (38%)
                   2           3 (11%)      1 (10%)        3 (12%)
                   3           9 (33%)      3 (30%)       10 (38%)
                   4           6 (23%)      4 (40%)        3 (12%)
Grade              I-II        3 (13%)      4 (40%)        6 (24%)
                   III         19 (79%)     6 (60%)       16 (64%)
                   IV           2 (8%)       0 (0%)        3 (12%)
Site               Esophagus   6 (21%)      3 (23%)       18 (58%)
                   Gastric     20 (69%)     5 (38%)        6 (19%)
                   Cardia      3 (10%)      5 (38%)        7 (23%)

* Tests between the treatments could not be performed since any given
patient could have one or more treatment modality.

Table 4. Coefficient estimators for the final model using all
patients in a Cox proportional hazard model *

Parameter           Parameter   Standard     P
                    estimate     error     value

Age group (80-84)     -0.34       0.24     0.17
Gender (Female)       0.45        0.26     0.08
Stage (0-1)           -1.65       0.33     <.001
Stage (2)             -0.90       0.41     0.03
Stage (3)             -0.62       0.30     0.04
Chemotherapy          -0.89       0.39     0.02

Parameter           Hazard    95% CI for
                    ratio    hazard ratio

Age group (80-84)    0.71     0.45, 1.16
Gender (Female)      1.56     0.94, 2.58
Stage (0-1)          0.19     0.09, 0.36
Stage (2)            0.41     0.17, 0.88
Stage (3)            0.54     0.30, 0.96
Chemotherapy         0.41     0.18, 0.83

* Analysis of maximum likelihood estimates. The final model
considered mortality as a function of age group, gender, stage, and
chemotherapy. The model was significant with a P < 0.001. Variables
were selected using backward, forward, and stepwise selection. All
three methods agreed on the final model. All regression diagnoses
were checked and model assumptions satisfied. CI indicates confidence
COPYRIGHT 2015 The Baylor University Medical Center
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Barnett, Tara; Mason, James; Maldonado, Yolanda Munoz; Wong, Lucas
Publication:Baylor University Medical Center Proceedings
Date:Jun 28, 2015
Previous Article:Comparison of long-term follow up of laparoscopic versus open colectomy for transverse colon cancer.
Next Article:On camel rides and Moses Maimonides.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters