Palliative Treatment of Malignant Esophagopulmonary Fistulas With Covered Expandable Metallic Stents

Transcription

1 Vascular and Interventional Radiology Original Research Vascular and Interventional Radiology Original Research FOCUS ON: Kyung Rae Kim 1,2 Ji Hoon Shin 2 Ho-Young Song 2 Gi-Young Ko 2 Jin Hyoung Kim 2 Hyun-Ki Yoon 2 Kyu-Bo Sung 2 Kim KR, Shin JH, Song HY, et al. Keywords: bronchogenic carcinomas, esophageal carcinomas, esophageal fistula, oncologic imaging, respiratory tract fistula, stents DOI: /AJR Received November 27, 2008; accepted after revision March 3, Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 388-1, Pungnap-2 dong, Songpa-gu, Seoul , Korea. Address correspondence to J. H. Shin (jhshin@amc.seoul.kr). 2 Present address: Department of Radiology, University of Texas Health Science Center at San Antonio, San Antonio, TX. WEB This is a Web exclusive article. AJR 2009; 193:W278 W X/09/1934 W278 American Roentgen Ray Society Palliative Treatment of Malignant Esophagopulmonary Fistulas With Covered Expandable Metallic Stents OBJECTIVE. The objective of our study was to evaluate the safety and clinical effectiveness of covered expandable metallic stents for palliative treatment of malignant esophagopulmonary fistulas. MATERIALS AND METHODS. Between November 1990 and January 2008, 14 patients with fistulas between the esophagus and pulmonary parenchyma were treated with covered expandable metallic esophageal stents. The fistulas were caused by esophageal (n = 9) or bronchogenic (n = 5) carcinomas. At the time of stent placement, all patients had aspiration pneumonia and 11 had lung abscesses (79%). Technical and clinical success, fistula reopening, complications, and survival rates were evaluated. RESULTS. Stent placement was technically successful in all cases, and there were no immediate procedural complications. Complete fistula sealing resulting in resolution of aspiration symptoms (i.e., clinical success) occurred in 12 patients (86%). During follow-up (mean survival, ± 79.9 days; median survival, 65.5 days; range, days), the fistula reopened in two of 12 clinical success patients. One patient (7%) experienced complications resulting from dyspnea due to tracheal compression by the esophageal stent. Although 13 patients died of aspiration pneumonia and the remaining patient died of cancer bleeding, none of the mortalities was related to the stent placement procedure. CONCLUSION. The use of covered expandable metallic stents appears to be safe and feasible for the palliative treatment of malignant esophagopulmonary fistulas. D evelopment of malignant esophagorespiratory fistulas in patients with esophageal or bronchogenic carcinoma is a devastating and life-threatening complication. Untreated fistulas cause continued respiratory tract contamination that can result in pulmonary sepsis and death, making early detection and prompt treatment critical [1, 2]. Although the term esophagorespiratory fistula is used to describe all fistulas between the bronchial tree and esophagus, the fistula site is esophagotracheal in 52 57% of patients and esophagobronchial in 37 40% [2 4]. In the remaining patients (3 11%), communication is established peripherally through the lung parenchyma, forming an esophagopulmonary fistula [2 4]. Reported treatments for esophagorespiratory fistulas include placement of various types of covered esophageal or tracheobronchial metallic stents [5 14]. However, there have been no reports to our knowledge regarding ther apeutic stenting for esophagopulmonary fistulas. For the current study, we investigated the safety and clinical effectiveness of covered expandable metallic esophageal stents for the palliative treatment of malignant esophagopulmonary fistulas. Materials and Methods This study was performed with the approval of our institutional review board. All patients provided written informed consent for the procedure. Patient Population The baseline characteristics of the patients are detailed in Table 1. From November 1990 through January 2008, 85 patients with symptoms of aspiration and dysphagia secondary to malignant esophagorespiratory fistulas were treated with covered expandable metallic stents. Of those 85 patients, 14 (16%) had fistulas between the esophagus and lung parenchyma including one patient having both esophagopulmonary and bronchoesophageal fistulas. All 14 patients were male. The W278 AJR:193, October 2009

2 mean age was 59.8 ± 7.8 (SD) years (range, years). None of the patients was considered a surgical candidate at the time of stent placement. Ten patients had received combined chemotherapy and radiation therapy, two patients received only chemotherapy, and the remaining two did not receive any form of treatment before the esophagopulmonary fistula developed. One patient developed an esophagopulmonary fistula 1 month after undergoing esophageal stent placement and chemotherapy and radiation therapy for malignant esophageal stricture, whereas the others did not undergo esophageal stent placement before the development of an esophagopulmonary fistula. Diagnoses of esophagopulmonary fistulas were established using esophagography in all patients. The confirmatory esophagogram finding was a direct communication between the esophagus and lung parenchyma without visualization of the trachea or bronchial trees. The underlying cause of the esophagopulmonary fistulas was esophageal squamous cell carcinoma in nine patients or bronchogenic carcinoma in five patients, three with adenocarcinoma and two with squamous cell carcinoma. Diagnoses of malignancy were established using endoscopic biopsy in all patients. All patients reported dysphagia to solids, liquids, or both secondary to severe esophageal stricture and aspiration. All patients underwent contrast-enhanced CT before stent placement. Aspiration pneumonia was detected in all 14 patients and lung abscesses in 11 patients. In all 11 lung abscess cases, CT showed a fistulous communication between the abscess and the esophagus. Techniques of Stent Placement Before stent placement, the site of the esophagopulmonary fistula was evaluated on esophagography. Fistulas were located in the upper thoracic (n = 3) or middle thoracic (n = 11) esophagus. Patients were evaluated for symptoms of both aspiration and dysphagia before and after stent placement. Dysphagia was classified according to a previously published grading system [10, 11, 14]: grade 0, normal swallowing; grade 1, ability to swallow semisolids; grade 2, ability to swallow soft foods; grade 3, ability to swallow liquids only; and grade 4, complete dysphagia. All patients reported coughing while swallowing food or saliva. Dysphagia was grade 4 in 13 patients and grade 2 in one patient. The mean dysphagia score was Polyurethane or polytetrafluoroethylene-covered esophageal stents (Niti-S esophageal covered stent, Taewoong) were used. The body of each stent was 18 mm in diameter and mm in length when fully expanded. Both ends of the stents were 6 8 mm wider in diameter than the body of the stents to impede stent migration. A topical anesthetic, lidocaine hydrochloride, was administered to the pharynx via aerosol spray before the procedure. A inch angled exchange guidewire (Radifocus M, Terumo) was inserted through the patient s mouth, the esophageal stricture, and then into the stomach under fluoroscopic guidance. After the length of the stricture was measured, a stent that was at least 4 cm longer than the stricture was placed so that its proximal and distal parts rested on the upper and lower margins of the stricture. Because all fistulous communication occurred at the esophageal stricture site, the stent could cover the fistula opening and the esophageal stricture. An esophagogram was obtained immediately after stent placement. If the stent completely sealed the fistula, the patient was allowed to eat a liquid diet 1 hour after the procedure and was encouraged to resume a tolerable diet gradually. If the esophagogram showed persistent leakage through the fistula that resulted from incomplete stent expansion, a follow-up esophagogram was obtained 1 3 days after stent placement to verify stent expansion before food intake was permitted. Follow-Up A follow-up esophagogram was obtained to evaluate fistula closure, luminal patency, and stentrelated complications at 7 days and 1 month after stent placement. Information regarding coughing while swallowing, the degree of dys p hagia, or complications such as bleeding or dyspnea due to tracheal compression by the stent was obtained from each patient in the outpatient clinic or by telephone interviews every 1 or 2 months. Technical success was defined as accurate stent placement in the target position with coverage of the fistula and the stricture. Clinical success was defined as complete closure of the fistula with improvement of aspiration symptoms and dysphagia within 7 days after stent placement. Clinical failure was defined as incomplete closure of the fistula with persistent aspiration symptoms and dysphagia within 7 days after stent placement. Recurrence of the completely closed esophagopulmonary fistula within 7 days after stent placement was also considered clinical failure. Placement of an additional stent, percutaneous feeding gastrostomy, or both were performed in cases of clinical failure, fistula recurrence, or stent-related complications during follow-up. Final information with regard to survival and to the cause of death was obtained through review of patient medical records and telephone interviews. In each patient with a lung abscess, the diameter of the abscess cavity was measured on CT scans obtained before and after stent placement and the difference was evaluated. Results Stent placement outcomes are detailed in Table 1. Stent placement in the esophagus was technically successful in all patients, with no procedure-related complications. All patients required placement of a single stent in the esophagus. Esophagograms obtained no later than 7 days after stent placement showed complete occlusion of the esophagopulmonary fistula in 12 (86%) of the 14 patients with no further symptoms of aspiration (i.e., clinical success) (Fig. 1). The remaining two patients (14%) had persistent symptoms of aspiration due to incomplete closure of the esophagopulmonary fistula despite successful stent placement (i.e., clinical failure). One week after stent placement, three patients had grade 4 dysphagia, one had grade 3, seven had grade 2, and the remaining three had grade 0 dysphagia. The mean dysphagia grade in those patients decreased from 3.86 to 2.07 within 1 week of stent placement. Improvement of dysphagia scores was noted in 10 patients, whereas four patients showed no improvement. Two patients with dysphagia that did not improve could not eat at all because of their poor general medical condition secondary to a lung abscess and pneumonia despite the absence of fistula recurrence. In the remaining two patients who showed clinical failure despite successful stent placement, the main cause of leakage was a gap between the proximal stent margin and the esophageal wall. These two patients underwent percutaneous feeding gastrostomy. All study patients died during the followup period (mean survival, ± 79.9 days; median survival, 65.5 days; range, days). Thirteen patients died of aspiration pneumonia and the remaining patient died of massive bleeding from the esophageal cancer. A complication occurred in one (7%) of the 14 patients: dyspnea 48 days after stent placement due to tracheal compression by the esophageal stent. The dyspnea symptoms were relieved after placement of a covered stent (20 mm in diameter and 80-mm long) in the trachea at the same level as the esophageal stent. In 10 of the 11 patients with a lung abscess, palliative stent placement was successful in closing the esophagopulmonary fistula. The mean longest diameter of the abscess cavity changed from 8.0 cm (range before stent placement, cm) to 5.4 cm (range after stent placement, cm) according to follow-up CT scans obtained a mean of 35 AJR:193, October 2009 W279

3 TABLE 1: Baseline Characteristics of Patients and Treatment Outcomes Patient a No. Age (y) Diagnosis Site of Fistula(s) 1 69 Esophageal SCC, lung abscess that coexisted with EBF Mid esophagus, LLL; mid eso phagus, right main bronchus Treatment Before Stent Placement Clinical Result (Complication) Treatment of Complications Survival (d) Chemotherapy Success None Esophageal SCC, lung abscess Upper esophagus, RUL Chemotherapy and RT Success None Esophageal SCC, lung abscess Mid esophagus, RLL Chemotherapy and RT Success None Lung SCC Mid esophagus, RLL None Success None Esophageal SCC, lung abscess Upper esophagus, RUL Chemotherapy and RT Failure (tracheal compression) Tracheal stent, gastrostomy 6 61 Esophageal SCC, lung abscess Mid esophagus, RLL Chemotherapy and RT Success None Lung adenocarcinoma, lung abscess Mid esophagus, RLL Chemotherapy and RT Success None Lung SCC Mid esophagus, RLL Chemotherapy and RT Success None Esophageal SCC, lung abscess Mid esophagus, RLL Chemotherapy and RT Success None Lung adenocarcinoma Upper esophagus, RUL Chemotherapy and RT Failure Gastrostomy Esophageal SCC, lung abscess Mid esophagus, RUL None Success None Esophageal SCC, lung abscess Mid esophagus, RUL Chemotherapy and RT Success None Lung adenocarcinoma, lung abscess Mid esophagus, RUL Chemotherapy and RT Success None Esophageal SCC, lung abscess Mid esophagus, RUL Chemotherapy Success None 95 Note SCC = squamous cell carcinoma, EBF = bronchoesophageal fistula, LLL = left lower lobe of the lung, RUL = right upper lobe of the lung, RLL = right lower lobe of the lung, RT = radiation therapy. a All patients were male. A Fig year-old man (patient 1 in Table 1) with esophagopulmonary and bronchoesophageal fistulas due to esophageal cancer. A, Chest CT scan shows midesophageal mass (arrow) and abscess in lower lobe of left lung (arrowhead). B, Esophagogram shows esophagopulmonary fistula (arrow) between mid esophagus and left lower lobe and esophagobronchial fistula (arrowhead) between mid esophagus and right main bronchus. C, Esophagogram obtained 2 days after stent placement shows occluded fistulas. B 61 C W280 AJR:193, October 2009

4 days (range, days) after stent placement (Fig. 2). However, despite the decrease in size, the lung abscesses persisted and resulted in clinical condition deterioration. All patients with lung abscesses died days after stent placement. Discussion One of the largest clinical investigations into the incidence of esophagorespiratory fistulas found that they occurred in 4.94% of 1,943 esophageal cancer patients, 0.16% of 5,714 lung cancer patients, and 14.75% of 41 tracheal cancer patients [2]. Esophagorespiratory fistulas can arise spontaneously due to tumor invasion, radiation therapy, laser treatment, or pressure necrosis caused by a previously placed stent [8]. In the current study, 10 patients received combined chemotherapy and radiation therapy and two patients received chemotherapy before the esophagopulmonary fistula developed; therefore, chemotherapy and radiation therapy appear to be highly associated with fistula development. A C Fig year-old man (patient 11 in Table 1) with esophagopulmonary fistula and lung abscess due to esophageal cancer. A and B, Axial (A) and coronal (B) chest CT scans show midesophageal mass and lung abscess (arrows) in upper lobe of right lung. C and D, Follow-up axial (C) and coronal (D) CT scans obtained 45 days after placement of covered esophageal stent show abscess cavity (arrow, C) has decreased in size. One patient who developed an esophagopulmonary fistula 1 month after esophageal stent placement had also undergone combined chemotherapy and radiation therapy before fistula development. Although the primary cause was considered to be pressure necrosis due to a previously placed esophageal stent, chemotherapy and radiation therapy may also have contributed to fistula formation in this patient. Francis and Goldstein [15] reported one case of asymptomatic esophageal carcinoma with an esophagopulmonary fistula masquerading as a primary lung abscess. In the current study, all patients had aspiration pneumonia and 11 patients (79%) had a lung abscess. We surmise that the high incidence of lung abscess was caused by direct flow of esophageal content into the lung parenchyma not passing through the trachea or bronchial trees. On the basis of the high incidence of lung abscess in our patients with esophagopulmonary fistula, we believe that esophagopulmonary fistula should be considered separately from other forms of esophagorespiratory fistula in terms of diagnostic and therapeutic approach. Shin et al. [11] reported the long-term outcomes of 61 patients with esophagorespiratory fistulas after palliative treatment with covered expandable metallic stents. The clinical success (80%) and failure (20%) rates reported in that study are comparable to those in the current study (86% and 14%, respectively). Those authors reported a mean survival of 93.8 days (13.4 weeks), which is shorter than the mean survival of days in the current study. Given that pneumonia and lung abscesses are more common in patients with esophagopulmonary fistula than those with esophagotracheal or bronchoesophageal fistula, the longer survival may reflect the differences in the number of patients in each study (61 vs 14 patients, respectively). In most esophagorespiratory fistula patients, oral intake is limited due to paroxysmal coughing, often leading to profound malnutrition and death from recurrent pulmonary B D AJR:193, October 2009 W281

5 infections and sepsis. Therefore, fistula closure requires greater attention than treatment of the underlying malignancy. Although we were able to close esophagopulmonary fistulas using palliative stent placement, persistence of preexisting lung abscesses often led to patient death. Therefore, placement of a covered stent in the esophagus for palliative treatment of esophagopulmonary fistula in patients with lung abscesses may be considered ineffective because fistula closure means closed natural drainage of lung abscesses into the esophagus. Furthermore, the covered stent can exacerbate the severity of the lung abscess by closing the natural drainage tract. We did not attempt percutaneous abscess drainage in any of the patients with a lung abscess because the abscesses were located deep in the thoracic cavity and we believe that these patients would not be able to cope with such an invasive procedure. Although follow-up CT showed all abscess cavities had decreased in size, stent placement did not completely resolve cavities. To increase survival, the diagnosis of esophagopulmonary fistulas needs to be established early ideally before the onset of pneumonia or development of lung abscesses. High-risk patients with esophageal or bronchogenic carcinomas who undergo chemotherapy or radiation therapy should be closely followed up for symptoms associated with fistula formation. In conclusion, placement of covered expandable metallic esophageal stents appears to be safe and feasible for the palliative treatment of malignant esophagopulmonary fistulas. However, persistence of lung abscesses and deterioration of clinical condition are frequently observed even after successful stent placement. References 1. Buess G, Schellong H, Kometz B, Grussner R, Junginger T. A modified prosthesis for the treatment of malignant esophagotracheal fistula. Cancer 1988; 61: Martini N, Goodner JT, D Angio GJ, Beattie EJ Jr. Tracheoesophageal fistula due to cancer. J Thorac Cardiovasc Surg 1970; 59: Angorn IB. Intubation in the treatment of carcinoma of the esophagus. World J Surg 1981; 5: Duranceau A, Jamieson GG. Malignant tracheoesophageal fistula. Ann Thorac Surg 1984; 37: Abadal JM, Echenagusia A, Simo G, Camunez F. Treatment of malignant esophagorespiratory fistulas with covered stents. Abdom Imaging 2001; 26: Do YS, Song HY, Lee BH, et al. Esophagorespiratory fistula associated with esophageal cancer: treatment with a Gianturco stent tube. Radiology 1993; 187: Han YM, Song HY, Lee JM, et al. Esophagorespiratory fistulae due to esophageal carcinoma: palliation with a covered Gianturco stent. Radiology 1996; 199: Morgan RA, Ellul JP, Denton ER, Glynos M, Mason RC, Adam A. Malignant esophageal fistulas and perforations: management with plastic-covered metallic endoprostheses. Radiology 1997; 204: Saxon RR, Barton RE, Katon RM, et al. Treatment of malignant esophagorespiratory fistulas with silicone-covered metallic Z stents. J Vasc Interv Radiol 1995; 6: Saxon RR, Morrison KE, Lakin PC, et al. Malignant esophageal obstruction and esophagorespiratory fistula: palliation with a polyethylene-covered Z-stent. Radiology 1997; 202: Shin JH, Song HY, Ko GY, Lim JO, Yoon HK, Sung KB. Esophagorespiratory fistula: long-term results of palliative treatment with covered expandable metallic stents in 61 patients. Radiology 2004; 232: Wang MQ, Sze DY, Wang ZP, Wang ZQ, Gao YA, Dake MD. Delayed complications after esophageal stent placement for treatment of malignant esophageal obstructions and esophagorespiratory fistulas. J Vasc Interv Radiol 2001; 12: Weigert N, Neuhaus H, Rosch T, Hoffmann W, Dittler HJ, Classen M. Treatment of esophagorespiratory fistulas with silicone-coated self-expanding metal stents. Gastrointest Endosc 1995; 41: Wu WC, Katon RM, Saxon RR, et al. Siliconecovered self-expanding metallic stents for the palliation of malignant esophageal obstruction and esophagorespiratory fistulas: experience in 32 patients and a review of the literature. Gastrointest Endosc 1994; 40: Francis PB, Goldstein J. Asymptomatic esophageal carcinoma with esophagopulmonary fistula masquerading as a primary lung abscess. South Med J 1979; 72:75 77 W282 AJR:193, October 2009