| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Vascular and Interventional Radiology |
1 From the Department of Radiology, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-dong, Songpa-ku, Seoul 138-736, Korea. From the 2002 RSNA scientific assembly. Received May 9, 2003; revision requested July 21; final revision received November 25; accepted December 10. Supported by grant HMP-98-G-2–043 for a Highly Advanced National Project, Ministry of Health and Welfare, Republic of Korea. Address correspondence to H.Y.S. (e-mail: hysong@amc.seoul.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Sixty patients with ERFs due to esophageal or bronchogenic carcinoma and one patient with ERF due to pressure necrosis caused by initial esophageal stent placement for esophageal carcinoma were treated with covered expandable esophageal or tracheobronchial metallic stents. Information about technical success of stent placement, initial clinical success and failure, fistula reopening, and complications was obtained. Survival curves for both patient groups with initial clinical success and failure were obtained and compared with Kaplan-Meier methods and log-rank test.
RESULTS: Stent placement was technically successful in all patients, with no immediate procedural complications. The stent completely sealed off the fistula in 49 (80%) of 61 patients so that they had no further aspiration symptoms (initial clinical success). Twelve (20%) of 61 patients had persistent aspiration symptoms due to incomplete ERF closure (initial clinical failure). During follow-up, the fistula reopened in 17 (35%) of 49 patients with initial clinical success: In eight patients, the reopened ERF was sealed off successfully with stent placement or balloon dilation. In two patients with reopened ERF caused by food impaction, the reopened fistula resolved spontaneously. Seven patients did not undergo further treatment. All patients died during follow-up, and mean survival was 13.4 weeks (range, 1–56 weeks) after stent placement. Mean survival in patients with initial clinical success was significantly longer than that in patients with initial clinical failure (15.1 vs 6.2 weeks, P < .05).
CONCLUSION: Covered expandable metallic stents were placed in 61 patients with ERFs, but the initial clinical success rate was poor and the rate of reopening was high; however, interventional treatment was effective for sealing off reopened ERFs.
© RSNA, 2004
Index terms: Esophagus, grafts and prostheses, 71.1269 • Esophagus, interventional procedures, 71.1269 • Esophagus, stenosis or obstruction, 71.1432 • Fistula, gastrointestinal tract, 71.245 • Fistula, pulmonary, 60.24
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Esophageal intubation with conventional unexpandable plastic prostheses was the treatment of choice for ERF until the early 1990s; however, complications such as perforation, hemorrhage, pressure necrosis, tube obstruction, tube dislodgment, and tube migration have been reported in approximately 15%–40% of patients (3). To overcome these limitations, many investigators (3–15) have used several types of covered expandable metallic stents in the treatment of ERFs and have reported 67%–100% complete closure of ERFs following placement of various types of covered esophageal and/or tracheobronchial metallic stents. In most reports (3–15), the number of study patients was fewer than 20, and the descriptions of clinical outcomes that were based on complete follow-up data are not available in many of them. Thus, the purpose of our retrospective study was to evaluate the long-term clinical results of palliative treatment of ERF with covered expandable metallic esophageal or tracheobronchial stents.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
None of the patients were considered candidates for surgical treatment at the time of stent placement. ERFs were caused by squamous cell carcinoma of the esophagus in 56 patients, bronchogenic carcinoma in four patients, and pressure necrosis from the initial covered esophageal stent that was placed for esophageal carcinoma in one patient. Among 61 patients, nine had undergone esophageal stent placement (covered stents in eight and an uncovered stent in one) for esophageal carcinoma before detection of the ERF. The stent had covered the tumor completely at that time. Among the 61 patients, an ERF was detected on follow-up esophagograms obtained after Ivor Lewis esophagectomy (partial esophagectomy with bowel interposition) in five patients. In these five patients, the tumor recurrence was confirmed with endoscopy at the fistula site.
Stents
Since 1990, we have used two types of homemade covered expandable metallic stents: a silicone-covered modified Gianturco stent in 18 patients and a polyurethane-covered nitinol stent in 43 patients. The covered stents were constructed in our research laboratory with a 0.5-mm-diameter stainless steel wire or a 0.2-mm-diameter nitinol wire (4,16,17). The esophageal stents were classified as straight, flared, and shoulder types according to the shape of the proximal and distal ends of the stents. From these three types of stents, the stent used in each patient was selected on the basis of stent availability. Both ends of the flared and shoulder stents were 4–6 mm wider in diameter than the body of the stents to impede stent migration. Straight, flared, and then shoulder stents were used in the order of development.
The tracheal and bronchial stents were mostly the straight type. The hinged type of tracheobronchial stent was used when the fistula was located in both the carina and the left main bronchus. To make a hinged stent, the ends of a tracheal stent and a bronchial stent were connected at one point without overlap by using nylon monofilament (17). The body of each stent was 16–24 mm in diameter in the esophagus and 10–20 mm in diameter in the tracheobronchial tree. Stent length, which was determined on the basis of the length of the esophageal or tracheobronchial strictures and fistulas, ranged from 3 to 14 cm.
Stent Placement Technique
Before stent placement, the site of the ERF was evaluated by two authors (J.H.S., H.Y.S.) with either preprocedural esophagography or endoscopy. The fistula site between the esophagus and respiratory tract was at the trachea (n = 17), at the right (n = 8) or left (n = 24) main bronchus (n = 32), or at the right (n = 7) or left (n = 6) lobar bronchus (n = 13). In one patient, fistulas were at two sites involving both the trachea and the left main bronchus.
Before and after stent insertion, the same two authors evaluated patients for symptoms of aspiration, as well as for dysphagia. Dysphagia was categorized according to a grading system that was described in previous reports (11,14). Grade 0 indicated 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 experienced coughing after they swallowed food or their own saliva. Forty-three patients had grade 4 dysphagia, 10 had grade 3, seven had grade 2, and one had grade 1. The mean grade of dysphagia was 3.56.
A topical anesthetic, lidocaine hydrochloride (Dai Han, Seoul, Korea), was administered to the pharynx or larynx by means of an aerosol spray before the procedure. Neither sedatives nor general anesthetics were used. When the patient had a stricture in the esophagus with no or mild tracheobronchial stricture, the stent was placed in the esophagus. When the patient had no or mild stricture in the esophagus or had moderate to severe stricture in the tracheobronchial tree, the stent was placed in the tracheobronchial tree. When the patient had moderate to severe strictures in both the esophagus and the tracheobronchial tree, stents were placed in both sites. The stents were placed by one of three authors (J.H.S., H.Y.S., G.Y.K.), who each had more than 7 years of experience in interventional procedures.
In cases of esophageal stent placement, a 0.035-inch angled exchange guide wire (Radiofocus M; Terumo, Tokyo, Japan) was inserted through the mouth, across the stricture, and into the distal portion of the esophagus or stomach with fluoroscopic guidance. A deflated balloon catheter with a 10–12-mm-diameter balloon was then passed over the guide wire to a position across the stricture. The balloon was slowly inflated with a diluted water-soluble contrast medium until the hourglass deformity created by the stricture disappeared from the balloon contour. The 0.035-inch guide wire was replaced with a superstiff J-tip guide wire (Boston Scientific/Medi-Tech, Watertown, Mass), and then the balloon was removed. Balloon dilation was not needed before stent placement when the esophageal stricture was not severe. A stent at least 4 cm longer than the stricture was then placed so that its proximal and distal parts rested on the upper and lower margins of the stricture, as well as on the fistula opening.
In cases of tracheobronchial stent placement, all patients underwent bronchoscopic examination to localize the stricture and fistula. Then, a 0.035-inch angled exchange guide wire was inserted through the mouth, across the stricture, and into the distal portion of the stricture with bronchoscopic and fluoroscopic guidance. Bronchoscopic examinations were performed by a bronchoscopist who had 10 years of experience in bronchoscopy. Balloon dilation of the stricture was performed by using the same technique as was used for esophageal stent placement. Balloon dilation, however, was not needed before stent placement in patients with a mild tracheobronchial stricture. A stent that was at least 1 cm longer than the stricture or fistula opening was placed.
Follow-up
An esophagogram was obtained immediately after stent placement. If the stent completely sealed off the fistula, the patient was allowed to eat a soft diet 2 hours after the procedure and was encouraged to resume a tolerable diet as soon as possible. If the esophagogram showed persistent leakage through the fistula, which resulted from incomplete stent expansion, a follow-up esophagogram was obtained 2–3 days after stent placement to verify the stent expansion before food intake was started. The follow-up esophagogram was obtained by one of the authors who placed the stents.
Patients were followed up during a mean of 13.4 weeks (range, 1–56 weeks). They visited the outpatient clinic, and a follow-up esophagogram was obtained to evaluate closure of the fistula and patency or migration of the stent at 3–7 days following stent placement and then every 1 or 2 months after the procedure. If follow-up at the outpatient clinic was not practical, the patients or their families were contacted by telephone by one of two authors (J.H.S., J.O.L.) every 1 or 2 months until the patient died. Information was obtained concerning coughing while swallowing; degree of dysphagia; and complications such as pain, bleeding, food impaction, stent-related complications, granulation tissue formation, or tumor overgrowth.
We defined technical success as accurate stent placement in the target position to cover both the fistula and the stricture; initial clinical success as complete closure of the fistula, with improvement of aspiration symptoms and dysphagia within 7 days following stent placement; initial clinical failure as incomplete closure of the fistula, with persistent aspiration symptoms and dysphagia within 7 days following stent placement; and reopening of the ERF as recanalization of the completely closed fistula at the same fistula site during follow-up.
Further treatments, such as additional stent placement or balloon dilation, were performed if there were initial clinical failure, reopening of the fistula, or stent-related complications during follow-up. The patients in whom an ERF appeared after esophageal stents had been placed previously were evaluated for the cause of ERF development and treatment during follow-up. The patients who had undergone Ivor Lewis esophagectomy were evaluated with regard to their initial treatment and follow-up. Final information in regard to survival and cause of death, which was based on the review of patients’ records and information from telephone interviews, was also obtained.
Statistical Evaluation
To demonstrate the statistical validity of combining the two stent types (silicone-covered modified Gianturco stent and polyurethane-covered nitinol stent) into one group in the analysis of survival data, we performed the log-rank test to test the difference in survival between the two groups. To see if initial successful complete closure of the ERF could affect survival from the time of stent placement, cumulative survival curves for both patient groups—the one with initial clinical success and the other with initial clinical failure—were determined with the Kaplan-Meier method. The log-rank test was used to test the statistics for equality of survival distributions between the two groups. To consider the effect of the number of the stents (single vs multiple) on survival probability, a Cox proportional hazards regression model was created, because numbers of stents were thought to affect survival. Statistical software (SPSS, version 10.0; SPSS, Chicago, Ill) was used to perform the analyses. A P value of less than .05 was considered to indicate a significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Esophagograms obtained no later than 7 days after stent placement demonstrated completely occluded ERFs in 49 (80%) of 61 patients, with no further symptoms of aspiration (initial clinical success). The remaining 12 (20%) patients had persistent symptoms of aspiration because of incomplete ERF closure, although stent placement was technically successful (initial clinical failure). One week after stent placement, four patients had grade 4 dysphagia, five had grade 3, 21 had grade 2, 15 had grade 1, and 16 had grade 0. Therefore, the mean degree of dysphagia in these patients had markedly improved from 3.56 to 1.44 within 1 week following stent placement. Fifty-three patients showed improvement in grade of dysphagia: improvement of one grade, 10 patients; of two grades, 20 patients; of three grades, 13 patients; and of four grades, 10 patients. Eight patients showed no improvement in grade of dysphagia.
Follow-up Data
In the 12 patients with initial clinical failure despite well-positioned stent placement, the main cause of leakage was spillage of contrast medium through a gap between the proximal stent margin and the esophageal wall. Only five of these 12 patients received further treatment for the persistent fistula.
Among five patients who underwent further treatment, three patients who had undergone Ivor Lewis esophagectomy underwent only tracheobronchial stent placement (hinged stent in one and bronchial stent in the others) as a first treatment. In the first patient, the initially inserted hinged stent was removed 1 day after placement because of its partial migration. In the second patient, the initially inserted bronchial stent was removed and replaced with a second bronchial stent because of persistent fistula. The third patient with persistent fistula at both the trachea and the left main bronchus following initial bronchial stent placement underwent esophageal stent placement, followed by additional placement of three esophageal stents and one tracheal stent. The first and second patients showed no remarkable improvement in their symptoms related to the persistent fistulas, whereas the third patient had an eventual closure of the fistula. The fourth patient with a persistent fistula following esophageal stent (18 mm in diameter) placement underwent a second placement with a larger-diameter esophageal stent (20 mm in diameter), with closure of the fistula. The remaining patient underwent additional stent placement in the trachea 4 days following the initial esophageal stent placement, because this patient experienced severe dyspnea. Findings at fluoroscopy and endoscopy revealed tracheal compression from the esophageal stent. At insertion of an additional tracheal stent, dyspnea was successfully relieved for 2 weeks until the patient died (Fig 1).
|
|
|
|
Tumor ingrowth was detected in one patient after placement of an uncovered stent. In regard to treatment, a second covered esophageal stent was placed in eight patients who had tumor progression. In one remaining patient with pressure necrosis caused by the initial esophageal stent, a bronchial stent was placed after the initial stent was removed because the esophageal stricture was not severe. In these nine patients, the ERF remained closed following stent placement until they died.
Five of 61 patients with ERF underwent Ivor Lewis esophagectomy. Three of these patients were already described, and the other two patients are described here. In one patient of these two, placement of a hinged stent was the initial treatment, and the fistula was sealed off until the patient died. In the other patient, a large-diameter (24-mm-diameter) stent was placed in the interposed stomach as initial treatment, and the ERF remained closed until the patient died.
During the follow-up (mean survival, 13.4 weeks; range, 1—56 weeks), all study patients died of the following causes: cachexia (n = 24), diffuse metastasis (n = 12), pneumonia (n = 12), massive bleeding (n = 6), hypercalcemia (n = 1), or drug intoxication (n = 1). The cause of death was unknown in five patients. There was no significant difference in the survival of the patients who were treated with placement of the two stent types (P > .05, log-rank test). Survival following stent placement ranged from 1 to 56 weeks (mean, 15.1 ± 1.60 [SD] weeks; 95% CI: 11.67, 17.95) in patients with initial clinical success for treatment of the ERF (n = 49) and from 1 to 17 weeks (mean, 6.2 ± 1.57 weeks; 95% CI: 2.61, 8.74) in patients with initial clinical failure (n = 12) (Fig 4). There was a significant difference in survival between the patients with initial clinical success and failure (P < .01, log-rank test).
|
Complications
Twenty-eight complications occurred in 26 (43%) of 61 patients (Table 2). After stent placement, seven patients had a persistent dull chest pain that required analgesics; onset of pain occurred within 1 day, and the pain lasted for several weeks. Of the seven patients with hematemesis or hemoptysis, five died of massive bleeding. Among those seven patients, four had undergone both radiation therapy and chemotherapy and two had undergone radiation therapy prior to stent placement. Five stents (one bronchial stent, one hinged stent, and three esophageal stents) migrated in four patients. Three stents migrated completely upward or downward, and two stents migrated partially upward or downward. The migrated stent was removed and replaced with a second stent in four patients. Dyspnea from tracheal compression occurred in one patient 4 days after esophageal stent placement (Fig 1). The other complications were the same as those included in Table 2.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
In our study, although stent placement was technically successful in all patients, initial clinical success was achieved in 49 (80%) of 61 patients, and reopening of the ERF occurred in 17 (35%) of 49 patients during follow-up. Our results of initial clinical success in 80% of patients actually is at the lower end of the reported range of 67%–100% (3–15). In patients with initial clinical failure, that is incomplete closure of the ERF, persistent contrast medium leakage occurred through a gap between the dilated proximal esophageal lumen and an esophageal stent.
Wang et al (6) used the term "funnel phenomenon" to describe a space between the esophageal wall and the proximal part of the stent that resulted from the loss of apposition of the stent against the esophageal wall. To treat this phenomenon, Saxon et al (9) performed additional stent placement in the esophagus; Wang et al (6) and Deviere et al (22) performed injection of tissue glue into the space around the stent with endoscopic guidance, but results were unsatisfactory. On the other hand, the presence of marked stricture of the esophagus at the site of the ERF can help seal the ERF following stent placement (9). Although the degree of esophageal stricture is important, we predicted that this phenomenon might cease when the placed esophageal stent was fully expanded; therefore, we paid attention to follow-up esophagograms obtained within 1 week following stent placement in these patients. Although the esophageal stents were nearly fully expanded within a week following stent placement, we observed that the dilated proximal esophageal lumen did not return to normal in some of these patients and thereby resulted in persistent contrast medium leakage. This may have been caused by chronic irreversible dilation of the proximal esophageal wall due to neural plexus injury of the esophageal wall associated with an underlying carcinoma (6). Thus, placement of a covered expandable metallic stent may be insufficient to completely occlude an ERF when patients have a less severe esophageal stricture or a markedly dilated proximal esophageal lumen.
The reopening rate of the ERF of as much as 35% (17 of 49 patients) is high compared with the reported incidence of reopening of the fistula, which ranged from 0% to 20% (3–15). This discrepancy might be caused by the stent type or stent placement technique; however, a different length of follow-up and the thoroughness of follow-up to detect the fistula might actually be the most important determining factors. In this study, thorough follow-up with esophagography with barium on an outpatient basis or with a telephone interview until all study patients died would have revealed a more accurate evaluation regarding the incidence of reopened ERFs. The reported causes of reopening following stent placement were stent occlusion (caused by tumor overgrowth or ingrowth, food impaction, or granulation tissue formation), stent migration, and stent covering disruption (6,8–10,13–15).
Among 17 patients with reopened fistulas in this study, the reopening was mostly caused by complications associated with the placed stent. Eight of these patients were successfully treated with placement of another stent or balloon irrigation. In two patients, the reopened fistulas spontaneously resolved. Seven patients did not receive further treatment. We can say that interventional treatment for the reopened ERF was effective, because palliation of symptoms was accomplished in all eight patients until they died. Treatment of reopened ERFs, however, can cause problems if the reopening is caused by stent migration, is associated with a gap between the proximal margin of the stent and esophageal wall, or is spontaneous following treatment such as radiation therapy. In these cases, there is a high possibility of reopening of the ERF following placement of an additional esophageal stent because of the disappearance of or the improvement in a previous esophageal stricture. Placement of a tracheobronchial stent can be a good alternative, especially when the location of a fistula is above the aortic arch or when there is associated dyspnea or tracheobronchial stricture (10,13,23).
Tracheobronchial stent placement was useful in selected patients with ERF as a first treatment or during follow-up in this study. Morgan et al (13) placed two covered stents in the trachea when ERF of the upper esophagus was difficult to treat with esophageal stents; they obtained encouraging results. Several investigators (23,24) reported placement of double stents in both the esophagus and trachea for the treatment of esophageal or bronchogenic carcinoma with both esophageal and tracheobronchial stricture. However, Nomori et al (23) documented that there is a high risk of fistula development because of necrosis of both the esophageal and tracheobronchial walls from the pressure of the stents.
In our study, tracheobronchial stent placement with or without esophageal stent placement was performed in seven patients because of previous Ivor Lewis esophagectomy (n = 4), combined tracheal and esophageal stricture (n = 1), tracheal compression following esophageal stent placement (n = 1), and ERF secondary to pressure necrosis caused by previous placement of a covered esophageal stent (n = 1). Five of these seven patients had occlusion of ERF following tracheal or bronchial stent placement. Thus, we think that the placement of a tracheobronchial stent is useful when there is associated substantial tracheobronchial stricture or after previous Ivor Lewis esophagectomy. After Ivor Lewis esophagectomy, the replaced stomach or colon shows a large lumen compared with the lumen of the original esophagus. Therefore, a stent with a large diameter is necessary to cover the large lumen when we insert a stent into the replaced stomach or colon. Otherwise, tracheal or bronchial stent placement can be a good alternative, especially in cases of associated tracheobronchial stricture.
When we compared the survival of the 49 patients in whom the procedure was an initial clinical success with that of the 12 patients in whom it was an initial clinical failure, we found a significant increase in mean survival in those with initial clinical success (15.1 vs 6.2 weeks). Although the survival is not long, it is important that control of pulmonary contamination can provide the opportunity for secondary treatment, as well as the opportunity for improved survival and quality of life. This result has a connection with the observation of Burt et al (18) that delaying therapy longer than a week after diagnosis of ERF dramatically decreased the ability to achieve palliation.
Major and minor complications may occur following esophageal or tracheobronchial stent placement (5–15,25–27). These complications include persistent chest pain (3.8%–7.7%), food impaction (0%–6.5%), stent migration (4.0%–15.0%), covered membrane disruption (0%–8.3%), tumor ingrowth or overgrowth (3.0%–36.0%), granulation tissue formation (2.0%–13.4%), tracheal compression (0%–6.1%), hemorrhage (0%–19.0%), and perforation or fistula formation (1.9%–7.3%). Some of these complications can be prevented or overcome by means of stent design modification or placement of additional stents. Potentially fatal complications such as perforation or hemorrhage, however, cannot be completely avoided despite careful stent placement technique, because the placed stent itself may act as a chronic irritant to the esophageal wall (28).
This study had limitations. First, it was a retrospective case analysis. Second, we used two types of homemade esophageal stents: the modified Gianturco stent and the polyurethane-covered nitinol stent with three shapes. The use of several kinds of stents, however, was inevitable because the duration for collection of the patients was approximately 12 years.
In conclusion, placement of a covered expandable metallic stent into the esophagus or tracheobronchial tree in patients with ERF was technically feasible, but the rate of initial clinical success was poor and the rate of reopening of the ERF during follow-up was high. The rate of reopening of the ERF was high and occurred in as many as 17 (35%) of 49 patients with initial clinical success. Most reopening of the ERF was caused by complications related to placement of stents. The interventional treatment was useful to successfully seal off the ERF.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, H.Y.S.; study concepts and design, J.H.S., H.Y.S.; literature research, J.H.S., G.Y.K.; clinical studies, all authors; data acquisition, J.H.S., G.Y.K., J.O.L.; data analysis/interpretation, J.H.S., G.Y.K., H.Y.S.; statistical analysis, J.H.S.; manuscript preparation, J.H.S., G.Y.K.; manuscript definition of intellectual content and editing, J.H.S., H.Y.S.; manuscript revision/review, K.B.S., H.K.Y.; manuscript final version approval, K.B.S., H.K.Y., H.Y.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S-M. Lin, T-Y. Lin, C-L. Chou, H-C. Chen, C-Y. Liu, C-H. Wang, H-C. Lin, C-T. Yu, K-Y. Lee, and H-P. Kuo Metallic stent and flexible bronchoscopy without fluoroscopy for acute respiratory failure Eur. Respir. J., May 1, 2008; 31(5): 1019 - 1023. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Kvale, P. A. Selecky, and U. B. S. Prakash Palliative Care in Lung Cancer: ACCP Evidence-Based Clinical Practice Guidelines (2nd Edition) Chest, September 1, 2007; 132(3_suppl): 368S - 403S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
