The purpose of the study was to determine the incidence, risk factors, treatment, and influence on survival of patients with de novo esophageal cancer after liver transplantation (LT). From 1988 to 2006, 1,926 patients underwent LT in our institution. A total of 9 patients (0.5%) developed a de novo esophageal cancer and 1 patient a cancer of the cardia (0.05%). A retrospective analysis was performed to reveal underlying diseases, timeframes between LT and appearance of cancer, predisposing factors, cancer therapy, complications, immunosuppressive regimens, and survival. Of our 10 patients, 7 (70%) suffered from esophageal squamous cell carcinoma (SCC) and 3 patients (30%) developed an adenocarcinoma, including the patient with cancer of the cardia. A total of 9 patients were transplanted due to alcoholic cirrhosis; 1 patient suffered from hepatocellular carcinoma in nonA-nonB hepatitis-related cirrhosis. Median time to tumor diagnosis was 51 months after transplantation. A total of 5 patients were treated conservatively with combined radiochemotherapy and 5 underwent surgical resection. Patients with radiochemotherapy showed a mean survival of 14.8 months vs. 24.8 months for the patients of the surgery group. No major postoperative complication has been observed. A total of 2 patients of the surgery group are still alive after a follow-up of 15 and 89 months. In conclusion, de novo esophageal and cancer of the cardia after LT is a rare event. In spite of immunosuppression, no increased complication rate has been observed. Patients may have a survival benefit from surgical resection. Liver Transpl 13:443–450, 2007. © 2007 AASLD.
Orthotopic liver transplantation (LT) has emerged as a safe alternative to treat end-stage liver diseases, as demonstrated by the excellent long-term survival.1 While the incidence of acute rejection after organ transplantation has steadily reduced due to newer immunosuppressive regimens, de novo malignancy after transplantation has become an increasing cause of mortality; it is the second leading cause of late death after LT and closely related to long-term immunosuppression.2, 3 According to literature, cancer is expected to surpass cardiovascular complications as the primary cause of death in transplanted patients within the next 2 decades.4–7
An increased incidence was first predicted by T.E. Starzl8 in 1968 and was confirmed shortly afterward.9, 10 Since then, the frequency of de novo malignancies in immunocompromised patients ranged from 4.1 to 16%.11 Posttransplantation lymphoproliferative disorder as well as nonmelanoma skin cancer are predominant, visceral tumors are, however, rare. A nationwide cohort study from Sweden indicated a 3-fold increased risk of de novo tumors after solid organ transplantation.12
Several mechanisms are involved in the development of malignancy after organ transplantation. Immunosuppressive treatment may facilitate the replication of oncogenic viruses, downregulate natural killer cell activity, induce chronic antigenic stimulation, and decrease interferon production as well as other cytokines.13, 14 A direct carcinogenic effect has also been suggested.15 Furthermore, cyclosporine can promote cancer progression by a direct cellular effect in vitro and in vivo.16
Esophageal cancer is a disease with a poor prognosis and high complication rates after surgery. Presently only little experience exists regarding treatment and prognosis of esophageal cancer after organ transplantation. Since about 20% of the patients are transplanted for alcoholic end-stage liver cirrhosis, the incidence of esophageal cancer may be increased in this subgroup of patients. Regular alcohol consumption in combination with smoking is known to be an important etiological factor for the development of esophageal cancer. In our study we report on our experience with patients who developed a de novo esophageal or cancer of the cardia after LT.
PATIENTS AND METHODS
Patients and Study Parameters
Between January 1989 and May 2006, 1,926 patients in our institution received an orthotopic LT. In 405 patients (21%) the underlying disease was alcoholic liver cirrhosis. We retrospectively looked through the hospital records to evaluate our patients for the presence of de novo cancer of the esophagus or the cardia by focusing on endoscopy or biopsy of the esophagus and stomach. A total of 10 patients with esophageal or cardia cancer were identified (0.5%). The symptoms that led to further diagnostic test were dysphagia (n = 10), weight loss (n = 5), retrosternal or epigastric pain (n = 5), gastroesophageal reflux and pyrosis (n = 3), and dyspepsia (n = 2). Surveillance issues were complains of dysphagia and retrosternal pain as well as epigastric pain especially in time correlation with food intake. Common clinical presentation, such as weight loss, longstanding gastroesophageal reflux, dyspnea, cough, hoarseness and pain (retrosternal, back, or right upper abdominal) led to immediate diagnosis. The characteristics of these patients are summarized in Table 1. The hospital records and data of the transplant registry were reviewed retrospectively. The following parameters were studied in all patients: age at transplantation, gender, indication for LT, pretransplant alcohol and tobacco consumption, primary immunosuppression, rejection episodes, reinforced immunosuppression due to rejection treatment, clinical presentations, treatment modality of the cancer, pathological findings, recurrence-free survival after LT, overall survival after LT, cause of death, follow-up. The diagnosis of malignancy was established by histological examination of biopsies and surgical specimens. The date of the biopsy or surgical procedure was designated as the date of cancer diagnosed. All patients who underwent surgical treatment, were operated with curative intent.
Table 1. Demographic Data Prior to Liver Transplantation
Initial immunosuppression was based on calcineurin inhibitors and prednisone, supplemented with anti-thymocyte globulin (ATG), mycophenolate mofetil, or azathioprine according to the actual study protocol. Modifications in doses or compounds were done individually, dependent on the clinical course.
Our patients were examined, which included liver function tests and routine blood tests, every 3 months for the first year, every 6 months from the second to the fourth year, and then annually after transplantation. Chest X-ray was performed individually.
Results were expressed as mean ± standard deviation. Comparison between the groups was done by the mean of the Student's t test or the Wilcoxon test for continuous variables. A P-value lower than 0.05 was considered to be statistically significant. All analyses were performed with SPSS software version 12.0 (SPSS, Inc., Chicago, IL, USA).
The demographic data of our cancer patients are shown in Table 1. All 10 patients were male, with a median age of 54.5 (range 42-63) yr at the time of transplantation. Of 10 patients, 9 underwent transplantation due to alcoholic cirrhosis, and 1 because of hepatocellular carcinoma in nonA-nonB cirrhosis. The stage of liver cirrhosis was Child-Turcotte-Pugh A (n = 2), Child-Turcotte-Pugh B (n = 4), and Child-Turcotte-Pugh C (n = 4). Consequently, 90% of the patients admitted to regularly consuming alcohol and 30% admitted being heavy smokers (up to 20 cigarettes per day). The most common clinical manifestation of the cancer was dysphagia, bleeding, and weight loss.
Of the 10 patients, 9 underwent endoscopy of the esophagus and stomach (esophagogastroduodenoscopy) pretransplantation. In a few cases, esophageal varices without bleeding as well as portal hypertensive gastropathy was seen and only in 1 case Barrett's metaplasia was obvious. As the esophagogastroduodenoscopy (2-6 months before transplantation), revealed no evidence of cancer or premalignant dysplastic lesions in any of these patients, none of them were biopsied pretransplantation.
Once a patient complained about epigastric discomfort after transplantation, immediate endoscopy was performed. All patients were diagnosed with histologically-proven esophageal or cardia cancer obtained at endoscopy, of which 3 revealed adenocarcinoma and 7 showed squamous cell carcinoma (SCC). One patient already showed lung metastases at the time of diagnosis and was therefore treated with chemotherapy.
Each patient who developed a malignancy following transplantation was evaluated for the presence of premalignant conditions before transplantation, as recorded in each patient's medical record.
Concerning the established risk factors such as alcohol and tobacco, only patients with a minimum of 6 months without alcohol intake were listed for transplantation. It was well expected that all patients quit tobacco abuse with transplant.
All patients received an initial calcineurin inhibitor–based immunosuppression in combination with prednisone (n = 10) as listed in Table 2. A total of 9 patients received cyclosporine and 1 patient tacrolimus. Cyclosporin was administered as triple therapy with prednisone and azathioprine (n = 4) or mycophenolate mofetil (n = 4). A total of 7 patients (70%) stayed on cyclosporine, and 3 (30%) changed to tacrolimus at follow-up due to acute or late rejection episodes (grade II or III). Of these 3 patients, 1 was further converted to sirolimus during follow-up because of major side-effects. Prednisone was continuously reduced and tapered in 6 patients within 6 months after LT; 4 patients required a further low maintenance dose. A total of 3 patients showed acute rejection episodes, 2 of them were treated with steroids (0.5-gm bolus of methylprednisolone for 3 days) and 1 additionally with muromonab-CD3 (OKT3). One patient with mild chronic rejection episodes, due to irregular immunosuppressive intake because of noncompliance, was not treated.
Table 2. Primary and Secondary Immunosuppressive Regimens
|6||CyA||Tac||Yes||Steroids+OKT3 = muromonab-CD3||No||Yes|
As shown in Table 3 the mean interval from LT to cancer diagnosis for all patients was 50.3 (median 51; range 1-96) months. Diagnosis of SCC (n = 7) was confirmed between 21 and 96 months after transplantation, whereas adenocarcinoma of the esophagus (n = 2) was diagnosed after 1 and 12 months. The cancer of the cardia (n = 1) developed 86 months after LT. The grading of the tumors resulted to be well differentiated (G1; n = 2), moderately differentiated (G2; n = 5), and poorly differentiated (G3; n = 3). According to the current tumor-lymph node-metastasis systems the tumors were classified as T2 (n = 2), T3 (n = 6), or T4 (n = 2). The nodal status was N0 (n = 1), N1 (n = 5), and N2 (n = 3). In 1 patient the nodal status was not classified.
Table 3. Tumor Characteristics and Follow-Up
|1||u2||u1|| ||2||25–32||SCC||Radiochemotherapy||26||11||Tumor progression||Lung metastases|
|2||u3||u1|| ||1||37–39||SCC||Radiochemotherapy||56||28||Local regression||Liver metastases|
|4||u3||u1||u1||2||20–25||SCC||Radiochemotherapy||96||13||Tumor progression||Lung metastases|
|8||p2||p0|| ||1||39||AC||Radiochemotherapy||1||4||Tumor progression||Hemorrhagic shock|
|9||p4||p2||0||2||15–20||SCC||Radiochemotherapy||84||18||Tumor progression||Lymph node metastases|
|Mean (range)|| || || || || || || ||52.6 (1–96)||14.8 (4-28)|| || |
|3||p3|| || ||2||27–35||SCC||Surgery||21||6||Tumor progression||Local tumor|
|5||p3||p2|| ||3||Cardia||AC||Surgery||86||11||Tumor progression||Peritoneal carcinomatosis|
|6||p3||p1|| ||3||34–41||SCC||Surgery||75||15||Tumor recurrence||No|
|10||p4||p2|| ||3||36–43||AC||Surgery||12||3||Tumor recurrence||Liver metastases|
|Mean (range)|| || || || || || || ||48 (12-86)||24.8 (3-89)|| || |
The decision of the treatment modality was done individually according to the general status of the patient. A total of 5 patients (50%) were treated conservatively with definite radiochemotherapy consisting of 60 Gy irradiation and cisplatin-based chemotherapy. In case of severe tumor stenosis with the inability of food intake, patients received repeated palliative dilatation or laser coagulation. The other 5 patients, who presented with better general condition, underwent surgical resection with radical lymph node dissection with curative intent. One of the surgically-treated patients was initially assigned to the group with conservative treatment. During palliative dilatation of the tumor stenosis, an iatrogenic perforation occurred that forced us to perform immediate esophageal resection. The final histological report revealed an R1 situation with residual periaortal tumor. The operating procedure was a standard transthoracic Ivor Lewis en-bloc esophagectomy (named after British surgeon Dr. Ivor Lewis who described this procedure in the 1940s) with gastric pull-up and intrathoracic stapler anastomosis in 4 patients and a gastrectomy with resection of the distal esophagus in 1 patient. Complete lymphadenectomy was part of the procedure. A transnasal feeding tube was placed routinely in all patients. Early enteral nutrition started at the day of surgery with 20 mL per hour.
Survival times after conservative and surgical treatment are shown in Table 3. A total of 2 patients of the surgery group were still alive after follow-up. The early postoperative period showed no major complications; there were no anastomotic leakages, bleeding, or severe wound infections. There was no hospital mortality. Pulmonary complications such as atelectasis and/or pneumonia could be treated conservatively with drainage and antibiotics. All patients could be dismissed within 6 weeks. Follow-up period ranged between 5 months and approximately 11 yr (135 months) after transplantation. The overall mean survival after cancer diagnosis was 19.8 (median 12; range 3-89) months. Regarding the 2 different groups with and without surgical treatment, the mean survival in the surgery group was 24.8 (median 11; range 3-89) months and in the conservative group 14.8 (median 13; range 4-28) months. A total of 8 patients died during follow-up; 7 of them died due to tumor recurrence or tumor progress. In the surgery group, 3 patients died due to local tumor recurrence (n = 1) or metastatic disease (n = 2). Two of the patients in the surgery group were still alive after 15 and 89 months of follow-up. One of these patients was tumor-free 89 months after surgical resection; the other patient experienced tumor recurrence 9 months after esophageal resection and was under radiochemotherapy at the end of follow-up. All patients of the conservative group died from tumor progression or metastatic disease (n = 4), or from hemorrhagic shock after tumor progress (n = 1). To enable enteral nutrition, the patients of the conservative group needed repeated endoscopy with dilatations. Dysphagia was a consistent symptom of the disease. On the other hand, all patients of the surgery group were able to eat regular food without the need of further endoscopy.
Long-term complications after LT have become more important in the last years since modern immunosuppression has led to an increased survival. One of the major complications is the development of de novo malignancies after organ transplantation. Lymphoproliferative disorders as well as nonmelanoma skin cancer are predominant malignant diseases. Analysis from our department confirm these findings, with a total of 7.2% de novo malignancies after LT. Furthermore, intraepithelial neoplasias of the cervix uteri, breast carcinoma, and lung carcinoma were observed.17 Other gastrointestinal cancers in our department were gastric (n = 3) and colorectal cancer (n = 2). A nationwide cohort study from Sweden indicated a 3-fold increased risk of de novo tumors after solid organ transplantation.12 Tumors of the gastrointestinal tract such as esophageal and gastric cancer are observed only occasionally.18, 19 Four single cases after LT showed rapid progression of Barrett's esophagus to high-grade dysplasia and adenocarcinoma.20–23 One case showed SCC 6 yr after LT due to alcoholic cirrhosis.24 Two cases have been reported in patients after renal transplantation.25, 26 These patients had none of the more commonly recognized risk factors such as extensive tobacco or alcohol abuse, which seems to implicate the immunosuppressive therapy as a possible cause of esophageal cancer. A 3-fold increased risk of esophageal cancer after organ transplantation, which reached statistical significance, has been described.12 Many etiological factors of de novo tumors are known, the most important being the immunosuppressive treatment, which leads to increased replication of oncogenic viruses, downregulation of natural killer cell activity, induction of chronic antigenic stimulation, and decrease of interferon production as well as other cytokines.13, 14 A direct carcinogenic effect has also been suggested.15 Cyclosporine has been shown to promote cancer progression by a direct cellular effect in vitro and in vivo.16 Another series of patients revealed a low CD4+/CD8+ ratio during cyclosporine-based therapy, which increased their long-term chance of de novo cancer, whereas conversion to tacrolimus for treatment of graft rejection emerged as a safe option.17 The tumor-promoting effect of cyclosporine may be present in our patients. Of 10 patients, 9 received cyclosporine-based immunosuppression, while a tacrolimus-based immunosuppression is predominant in patients after LT in our department. New immunosuppressive regimens tend to include rapamycin, which has been shown to inhibit tumor growth in vitro27 and in the clinic.28 Although azathioprine was related to the development of neoplasia in other series,29 it appeared to be independent of an increased cancer risk.30 This confirms the results of our study, in which treatment with azathioprine (n = 4) showed no significant effect on the development of neoplasia (median 85 months; range 26-96 months) nor on the surveillance (median 12 months; range 11-18 months).
Patients with esophageal cancer face an additional risk due to the habit of regular alcohol and tobacco consumption. The occurrence may be related to a chronic exposure of alcohol and tobacco prior to transplantation, and was shown to play a more important role in patients with adenocarcinoma than in those who underwent transplantation for nonalcoholic cirrhosis.31 These findings conform to our results, in which 9 out of 10 patients suffered from alcoholic liver cirrhosis. In another series, the relative risk of esophageal cancer was significantly higher in men than in women.12 As reported in earlier studies,31, 32 oropharyngeal SCC, and, as seen in our study, also adenocarcinoma of the esophagus and cardia, were diagnosed exclusively in men. However, this finding must be interpreted cautiously, since there was a high ratio (4.0) of male/female patients who underwent transplantation due to alcohol-induced cirrhosis in our center, as well as in other reports (ratio 8.0).10, 11 Alcohol intake, especially in combination with smoking, greatly increases the risk of SCC.33 The lifetime risk of esophageal cancer is 0.8% for men and 0.3% for women, as previously reviewed.34 The incidence of de novo esophageal cancer after LT seems to be increased in patients with alcoholic cirrhosis, but not in patients with different underlying disease. However, our series is too small to draw a definite conclusion. Any factor that causes chronic irritation and inflammation of the esophageal mucosa appears to increase the incidence of SCC. Immunosuppression may enhance the oncogenic effects of pretransplantation alcohol and tobacco consumption, as well established risk factors.35 Although the pathogenesis of esophageal cancer remains unclear, data from animal studies suggest that oxidative damage from factors such as smoking, alcohol consumption, or gastroesophageal reflux, which cause inflammation, esophagitis, and increased cell turnover, may initiate the carcinogenic process.36 The question arises whether tumors of the gastrointestinal tract, in particular esophageal cancer, developed after organ transplantation or whether small tumor nodules were already present at the time of transplantation. There is no doubt that the patient with an apparent esophageal cancer only 1 month after LT must have had this tumor already at the time of transplantation, although the last endoscopy performed 2 months prior to LT due to bleeding esophageal varices did not reveal anything. All patients received repeated endoscopy; however, malignant or premalignant lesion could not be detected. The cell-doubling time of the cancer cells is not known, least of all during continuous immunosuppression. We cannot exclude, even in the patient with the longest tumor-free period of 96 months after LT, that perhaps a small tumor nodule was present at the time of LT. Similar to our observation, a rapid development of tumors between 2 and 48 months after LT due to alcohol-related cirrhosis has been observed.31, 32 Not only immunosuppressive drugs play a major role in the development of posttransplant tumors, but there are also well-accepted additional risk factors such as sun exposure, Epstein-Barr virus infection,19 or human herpes virus-8 infection.37 In patients who underwent transplantation for adenocarcinoma, immunosuppressive drugs could similarly enhance the effect of alcohol and tobacco, which is a well-established risk factor for esophageal SCC.38 If a tumor was present at the time of transplantation, the growth could be stimulated by the various aforementioned factors. A total of 3 of our patients suffered from adenocarcinoma. A correlation between these tumors and alcohol abuse is not known, but often they arise due to premalignant lesions. Pretransplantation screening has recently been recommended for patients with Barrett's esophagus, which is detected in up to 2% of LT candidates.39 After the diagnosis had been confirmed, the individual treatment modality had to be selected. In case of good general condition, we performed an en-bloc esophagectomy with the intent to cure. Patients with severe general disease were treated conservatively. If one looks at the literature, the overall 5-yr survival rates for esophageal carcinomas lie between 10% and 15%. According to the tumor-lymph node-metastasis classification the values fluctuated after complete surgical removal between 50 to 80% for stage I, 30 to 40% for stage IIA, 10 to 30% for stage IIB, and 10 to 15% for stage III.34 Regarding SCC, the survival ranges from 50% for T0/T1, under 20%40 to 40% for T2, and 20% for T3/T4.41 In our institution we had similar 5-yr survival rates of 45% for T2 and 22% for T3/T4 tumors after R0 resection (n = 183).
Insufficiency of the esophageal anastomosis varies from 5 to 20% in large cohort studies, with a range of 0 to 50% in special selected groups.42–49 We experienced no anastomotic leakage, but 3 patients developed pneumonia and pleural effusion, which were treated conservatively. Despite the elevated risk of our patients after LT, our hospital mortality was zero. Compared to the literature with 3 to 10% hospital mortality,42, 50, 51 it becomes evident that en bloc esophagectomy in patients after LT is feasible with a low risk. A summary of studies regarding different treatment modalities in patients with esophageal cancer without organ transplantation showed a benefit for patients who received neoadjuvant radiochemotherapy and surgical resection vs. radiochemotherapy alone.52 The 5-yr survival rate could reach 6 to 27% with radiochemotherapy or radiotherapy alone. However, radiochemotherapy with subsequent surgical resection improved the outcome with 5-yr survival rates of 17 to 49%.52 Our small number of 10 patients did not allow a statistically significant conclusion in the comparison of survival between the conservative and the surgery group. However, we could recognize a longer time range of survival in the surgery group (3-89 months) with one long-term survivor. The quality of life has improved, as the patients after surgery were able to eat alone with no need of a feeding tube. Furthermore, the psychological benefit of being apparently tumor-free was part of the improved quality of life.
No experience exists about a neoadjuvant radiochemotherapy and esophageal resection after LT. Since the complication rate after esophageal resection appears to be independent of a LT, patients might have a beneficial effect with this approach.
We thank Mrs. Sylvia Albrecht for excellent technical assistance in preparing the manuscript.