Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2362031044
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Varghese, J. C.
Right arrow Articles by Mueller, P. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Varghese, J. C.
Right arrow Articles by Mueller, P. R.
(Radiology 2005;236:730-738.)
© RSNA, 2005

Vascular and Interventional Radiology

Fungus-infected Fluid Collections in Thorax or Abdomen: Effectiveness of Percutaneous Catheter Drainage1

Jose C. Varghese, MD, Peter F. Hahn, MD, PhD, Mukesh G. Harisinghani, MD, Sikandar M. Hayat, MD, Deborah A. Gervais, MD, David C. Hooper, MD and Peter R. Mueller, MD

1 From the Department of Radiology (J.C.V., P.F.H., M.G.H., S.M.H., D.A.G., P.R.M.) and Infectious Diseases Division, Infection Control Unit (D.C.H.), Massachusetts General Hospital, 55 Fruit St, White 270-E, Boston, MA 02114. From the 2001 RSNA Annual Meeting. Received July 7, 2003; revision requested September 23; revision received September 7, 2004; accepted October 15. Address correspondence to P.F.H. (e-mail: phahn{at}partners.org).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To retrospectively evaluate the effectiveness of percutaneous catheter drainage in the treatment of fungus-infected fluid collections in the thorax or abdomen and to identify any factor that may be predictive of a poor clinical outcome.

MATERIALS AND METHODS: Approval for this study was obtained from the hospital ethics subcommittee on human studies. Because the study was retrospective, patient informed consent was not required. This study was compliant with the Health Insurance Portability and Accountability Act. Retrospective analysis was performed of cases of fungus-infected fluid collections in the thorax or abdomen treated by using percutaneous catheter drainage in 60 patients (36 male and 24 female patients; mean age, 57 years; range, 2 months to 91 years) during 5 years. The patient medical records were reviewed to identify recognized factors for predisposition to fungal infection. The details of percutaneous catheter drainage and microbiologic findings were recorded. The technical success (ability of catheters placed to drain collections treated) and the clinical success (ability of patients to recover fully without surgery) of percutaneous catheter drainage were determined. A multifactor logistic regression analysis was performed to identify any clinical or microbiologic factor predictive of a poor clinical outcome.

RESULTS: Seventy-three fungus-infected fluid collections were drained in 60 patients. The collections originated from the pleura (n = 6), mediastinum (n = 2), liver (n = 3), pancreas (n = 5), obstructed biliary or urinary tract (n = 9), gallbladder (n = 1), and abdominopelvic area (n = 47). The technical success rate for catheter drainage was 79% (41 of 52 patients); the clinical success rate, 57% (34 of 60 patients). Twenty (33%) patients died from all causes during hospital admission. Multifactor logistic regression analysis was used to identify predictors of a poor clinical outcome; complexity of collection, history of malignancy, and admission to intensive care unit were significant (P < .03) and independent predictors.

CONCLUSION: Despite a moderately high technical success rate with percutaneous catheter drainage of fungus-infected fluid collections, clinical success rate was much lower. Both imaging appearance (complexity of collection) and clinical factors (history of malignancy, admission to intensive care unit) influenced prognosis.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
The incidence of fungal infection is increasing in hospitalized patients because of the ability to treat and maintain patients of increasing age and degree of debility by using the latest advances in critical care. Many patients become immunocompromised and vulnerable to opportunistic infections on account of their underlying disease and side effects from treatment (1). Results of an analysis of intensive care unit (ICU) surveillance data in the United States indicated that Candida species accounted for 10.1% of infections in this setting (2). Many fungal species coexist as part of the normal gastrointestinal flora and give rise to invasive infection when patient immunity is reduced (3). The development of fungemia is associated with a high mortality, particularly in the elderly or the critically ill patient (4,5).

Percutaneous catheter drainage is an established technique for treatment of pyogenic abscesses, yielding a success rate of 80%–90% (6–8). The treatment of deep-seated fungal infection also includes drainage of any focal fluid collection and administration of systemic antifungal therapy (7,9–12). The success rate with percutaneous drainage, however, is reported to be low (6,13–17), supposedly because of the poorly localized nature and viscous contents of fungus-infected fluid collections. This has led some investigators to advocate surgical drainage as the primary treatment for fungus-infected fluid collections (13).

In view of these considerations, the purpose of our study was to retrospectively evaluate the effectiveness of percutaneous catheter drainage in the treatment of fungus-infected fluid collections in the thorax or abdomen and to identify any medical factor that may be predictive of a poor clinical outcome.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Patient Population
We retrospectively identified all patients with fungus-infected fluid collections in the thorax or abdomen who were treated with percutaneous catheter drainage during 5 years from July 1995 to June 2000. This was achieved by cross-referencing a pharmacy-generated list of all patients who received intravenous antifungal agents during this period with a radiology database of all patients who underwent percutaneous catheter placement for drainage of abnormal fluid collections in the chest or abdomen during the same period. From the generated list of 76 patients, 16 patients were eliminated for the following reasons: (a) Systemic antifungal therapy was administered to treat a patient with a fungal infection that was elsewhere in the body and that was not related to the fluid collection drained (n = 13), and (b) systemic antifungal therapy was administered to treat a patient with a collection that became superinfected with fungi only after catheter drainage (n = 3).

The final study group of 60 patients underwent percutaneous catheter drainage and systemic antifungal therapy for eradication of deep-seated fungal infection in the thorax or abdomen. There were 36 male patients (mean age, 56 years; range, 2 months to 88 years) and 24 female patients (mean age, 57 years; range, 14–91 years), with an overall mean age of 57 years and age range of 2 months to 91 years. Approval for this study was obtained from the hospital ethics subcommittee on human studies. Because the study was retrospective, the ethics subcommittee did not require patient informed consent for inclusion in the study. This study also complied with the Health Insurance Portability and Accountability Act.

Data Collection
Data collection was performed by means of a review of the patient medical records by one of the authors (S.M.H.) to obtain clinical details, which included medical history, drug therapy history, and symptoms at presentation. Also recorded was a history of any well-recognized factors that may have led to a predisposition to fungal infection: malignancy; diabetes mellitus; acquired immune deficiency syndrome; and prior treatment with broad-spectrum antibiotics, steroids, or chemotherapy and/or radiation therapy (Table 1).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Factors of Predisposition to Fungal Infection in 60 Patients

 
From a review of data about the patient's clinical history and of imaging findings, the cause of the fungus-infected fluid collection was classified as that which resulted from a gastrointestinal leak or that which arose from diseases of the pulmonary, pancreatic, hepatobiliary, or renal systems. Causes of a gastrointestinal leak were further classified as those that resulted from spontaneous perforation, surgical anastomotic breakdown, or trauma.

Radiologic Drainage Technique
Percutaneous catheter drainage was performed by any of 12 attending staff interventional radiologists, a group that included three of the authors of this study (P.R.M., P.F.H., D.A.G.), who were on duty at the time of the procedure or by one of the many closely supervised fellows who were undergoing interventional radiology training during this period. The experience of the staff radiologists who performed these procedures ranged from 3 to 25 years at the time of their initial involvement with patient treatment.

A tandem trocar technique (18) was used for catheter insertion in a majority of patients. Catheters used for drainage included 8–14-F locking polyurethane drainage catheters (Dawson-Mueller Drainage Catheter; Cook, Bloomington, Ind) for abdominal collections and 16-F multipurpose polyvinyl catheters (Mueller Multipurpose Drainage Catheter Set; Cook) for pleural collections. Computed tomographic (CT) guidance was used for all instances of percutaneous catheter drainage except biliary and renal interventions, for which ultrasonography (US), with or without fluoroscopy, was used.

After an initial diagnostic aspirate was obtained, all collections were aspirated as much as possible through the drainage catheters, and the cavity was irrigated with normal saline. Abdominal catheters were used for gravity drainage, and pleural catheters were attached to a water seal vacuum pump (Pleur-Evac; Teleflex Medical, Research Triangle Park, NC) for drainage. All catheters were routinely flushed with 10 mL of normal saline every 8 hours to maintain patency. Intracavitary antifungal or fibrinolytic agents were never used.

Catheters were removed when drainage decreased to less than 20 mL/24 h, together with an improvement in patient sepsis. If there was no improvement in patient sepsis within 3–4 days of percutaneous catheter drainage, repeat CT was performed, and further drainage of any residual collections was performed. The catheters of patients with continued high outputs (>50 mL/24 h) were injected with water-soluble contrast material under fluoroscopy to diagnose any underlying fistula.

Technical Success
The technical success of percutaneous catheter drainage was defined as the ability of the placed catheters to evacuate the treated fluid collections. This was determined by measuring the size of the collections with imaging studies performed before and after drainage. For this purpose, the CT (n = 53) and US (n = 7) hard-copy films or images on the radiology picture archiving and communications system that were obtained in all patients before drainage were retrieved and reviewed by one of the authors (J.C.V.), who had more than 6 years of experience in abdominal imaging. With CT studies, the maximum axial dimension of the fluid collection was measured by using electronic calipers, and with US studies, the maximum fluid collection diameter as given on the images was recorded.

Imaging studies performed after treatment were available for analysis in only 52 patients (CT, n = 46; US, n = 6) because routine postprocedural studies were not performed in all patients. All imaging studies were performed within 2 weeks of completion of the drainage. These were again retrieved and analyzed, in a fashion similar to that used previously, by a second radiologist (M.G.H.) who had more than 5 years of experience in abdominal imaging and was unaware of the findings from the initial reading.

On the basis of the change in size before and after drainage, the collections were classified as adequately drained if there was greater than 90% reduction in size of the collection; partially drained if there was a reduction in size of the collection but the reduction was less than 90%; and poorly drained if there was no change in the size of the collection. Percutaneous catheter drainage was considered a technical success if collections were adequately drained and a technical failure if they were partially or poorly drained.

The collections for which drainage was performed were specifically evaluated for complexity of collection: A collection was classified as simple if it was well-defined and unilocular, with no evidence of an enteric communication, and as complex if it was multilocular or had an enteric fistula. For the purposes of this study, multiple small collections clustered together in the same anatomic area were considered a single multilocular collection.

Clinical Success
Patients were considered to have responded clinically to percutaneous catheter drainage if there was defervescence, stabilization of vital signs, or reduction in the need for cardiovascular and/or respiratory support within 3–4 days of drainage. The leukocyte response to drainage was determined by noting the white blood cell count immediately before and 1 week after drainage. Patient admission to the ICU for ventilatory support at any period during the hospitalization was recorded. Patient clinical outcomes, which included surgery for failed catheter drainage, deaths that occurred during hospitalization, and details of discharge from the hospital were also recorded.

On the basis of these findings, percutaneous catheter drainage was considered a clinical success if the patient recovered from sepsis and was discharged from the hospital without surgical drainage and a clinical failure if the patient required surgical drainage because of continued deterioration in clinical condition or died from sepsis or other causes. Those patients who recovered from sepsis as a result of catheter drainage but required elective surgery to correct the underlying cause of the collection were included in the clinical success group.

Microbiologic Techniques
With use of standard methods, initial diagnostic aspirates from collections in all patients and samples from other sites of fungal infection were analyzed microscopically and cultured semiquantitatively for aerobic and anaerobic bacteria and for fungi. Growth of fungus was reported as rare, moderate, or abundant. Candida species were identified by growth characteristics and color and fluorescence of colonies on blood and molybdate agar.

Confirmation of species was performed with a fully automated microbial identification system with advanced software (Vitek; BioMerieux, Hazelwood, Mo) and by morphology on rice infusion agar (Oxgall Tween 80; Med-Ox Diagnostics, Ogdensburg, WY). Additional confirmatory tests for identification of C glabrata included morphology of growth on Sabouraud agar and the results of the rapid assimilation broth test. Candidemia was diagnosed when at least one blood culture yielded Candida species. The lysis centrifugation system (Isolator; DuPont, Wilmington, Del) was used for some blood cultures.

Statistical Analysis
The means of continuous variables were compared by using t tests, and the relationship between two categorical variables was investigated by using the {chi}2 test of the null hypothesis of statistical independence. A multifactor logistic regression analysis was performed to identify which, if any, of 15 clinical and microbiologic factors were related to the probability of a failed clinical outcome. These 15 factors were age, sex, history of diabetes, history of acquired immunodeficiency syndrome, history of malignancy, history of solid-organ transplantation, steroid therapy, antibiotic therapy, chemotherapy and/or radiation therapy, use of total parenteral nutrition, admission to the ICU, complexity of collection, semiquantitative measurement of fungal growth in the culture, the finding of pure fungal growth in the culture, and fungemia.

The classic measures of sensitivity and specificity were computed, and the null hypothesis of equally directional discordance was investigated by using a {chi}2 test. The conventional 5% level of statistical significance was used as a nominal reporting level throughout this article. For the statistical analysis, a software package (SAS, version 9; SAS, Cary, NC) was used.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Collections
Seventy-three fungus-infected fluid collections were drained in 60 patients: one in each of 50 patients, two in each of eight patients, three in one patient, and four in one patient. The fluid collections were from the pleura (n = 6), mediastinum (n = 2), liver (n = 3) (Fig 1), pancreas (n = 5) (Fig 2), obstructed biliary or urinary tract (n = 9), gallbladder (n = 1), and abdominopelvic area (n = 47) (Fig 3). A further 17 fluid collections that were drained in 14 of these patients were not included for analysis, because fungal organisms were not isolated from seven collections, and needle aspiration was the only procedure performed in 10 collections.



View larger version (107K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. Hepatic candidiasis in 48-year-old woman with duodenal carcinoma and common bile duct obstruction. (a) Transverse CT image obtained through the dome of the liver with intravenous injection of contrast material shows a dominant multilocular fluid collection (arrow). Multiple microabscesses scattered throughout the remainder of the liver are not shown. (b) Transverse CT image obtained through the liver with intravenous injection of contrast material after percutaneous catheter drainage of the dominant hepatic fluid collection shows almost complete resolution of the collection. Aspirated fluid grew mixed organisms including Candida and Enterococcus species. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through a draining percutaneous transhepatic biliary catheter (long white arrow) shows patency of a biliary metallic stent (arrowheads) with free flow of contrast material into the duodenum. With the combined use of common bile duct stent placement to relieve biliary tract obstruction, percutaneous catheter drainage (black arrow) of the dominant liver collection, and systemic antifungal therapy, patient had complete recovery. A second catheter (short white arrow) drains a paraduodenal perforation arising from the duodenal tumor.

 


View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. Hepatic candidiasis in 48-year-old woman with duodenal carcinoma and common bile duct obstruction. (a) Transverse CT image obtained through the dome of the liver with intravenous injection of contrast material shows a dominant multilocular fluid collection (arrow). Multiple microabscesses scattered throughout the remainder of the liver are not shown. (b) Transverse CT image obtained through the liver with intravenous injection of contrast material after percutaneous catheter drainage of the dominant hepatic fluid collection shows almost complete resolution of the collection. Aspirated fluid grew mixed organisms including Candida and Enterococcus species. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through a draining percutaneous transhepatic biliary catheter (long white arrow) shows patency of a biliary metallic stent (arrowheads) with free flow of contrast material into the duodenum. With the combined use of common bile duct stent placement to relieve biliary tract obstruction, percutaneous catheter drainage (black arrow) of the dominant liver collection, and systemic antifungal therapy, patient had complete recovery. A second catheter (short white arrow) drains a paraduodenal perforation arising from the duodenal tumor.

 


View larger version (118K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1c. Hepatic candidiasis in 48-year-old woman with duodenal carcinoma and common bile duct obstruction. (a) Transverse CT image obtained through the dome of the liver with intravenous injection of contrast material shows a dominant multilocular fluid collection (arrow). Multiple microabscesses scattered throughout the remainder of the liver are not shown. (b) Transverse CT image obtained through the liver with intravenous injection of contrast material after percutaneous catheter drainage of the dominant hepatic fluid collection shows almost complete resolution of the collection. Aspirated fluid grew mixed organisms including Candida and Enterococcus species. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through a draining percutaneous transhepatic biliary catheter (long white arrow) shows patency of a biliary metallic stent (arrowheads) with free flow of contrast material into the duodenum. With the combined use of common bile duct stent placement to relieve biliary tract obstruction, percutaneous catheter drainage (black arrow) of the dominant liver collection, and systemic antifungal therapy, patient had complete recovery. A second catheter (short white arrow) drains a paraduodenal perforation arising from the duodenal tumor.

 


View larger version (169K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a. Fungus-infected pseudocysts caused by recurrent pancreatitis in 27-year-old woman with history of acquired immunodeficiency syndrome. (a) Transverse CT image obtained through the upper abdomen with orally and intravenously administered contrast material shows two well-defined pseudocysts (arrows) arising from the tail of the pancreas. (b) Follow-up transverse CT image obtained through the upper abdomen with orally administered contrast material shows both pseudocysts to be completely treated after percutaneous catheter (arrow) drainage of the larger cyst and needle aspiration of the smaller cyst. Cultures of the aspirates grew C albicans. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through the draining percutaneous catheter (long arrow) shows communication of a residual pseudocyst (curved arrow) with pancreatic tail duct (short arrow). Multiple small calculi within a dilated pancreatic head duct (arrowheads) caused obstruction to the flow of contrast material into the duodenum. Endoscopic retrograde cholangiopancreatography and papillotomy were subsequently performed, with stone clearance. This result allowed complete resolution of the pseudocyst, with removal of the percutaneous catheter.

 


View larger version (158K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b. Fungus-infected pseudocysts caused by recurrent pancreatitis in 27-year-old woman with history of acquired immunodeficiency syndrome. (a) Transverse CT image obtained through the upper abdomen with orally and intravenously administered contrast material shows two well-defined pseudocysts (arrows) arising from the tail of the pancreas. (b) Follow-up transverse CT image obtained through the upper abdomen with orally administered contrast material shows both pseudocysts to be completely treated after percutaneous catheter (arrow) drainage of the larger cyst and needle aspiration of the smaller cyst. Cultures of the aspirates grew C albicans. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through the draining percutaneous catheter (long arrow) shows communication of a residual pseudocyst (curved arrow) with pancreatic tail duct (short arrow). Multiple small calculi within a dilated pancreatic head duct (arrowheads) caused obstruction to the flow of contrast material into the duodenum. Endoscopic retrograde cholangiopancreatography and papillotomy were subsequently performed, with stone clearance. This result allowed complete resolution of the pseudocyst, with removal of the percutaneous catheter.

 


View larger version (141K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c. Fungus-infected pseudocysts caused by recurrent pancreatitis in 27-year-old woman with history of acquired immunodeficiency syndrome. (a) Transverse CT image obtained through the upper abdomen with orally and intravenously administered contrast material shows two well-defined pseudocysts (arrows) arising from the tail of the pancreas. (b) Follow-up transverse CT image obtained through the upper abdomen with orally administered contrast material shows both pseudocysts to be completely treated after percutaneous catheter (arrow) drainage of the larger cyst and needle aspiration of the smaller cyst. Cultures of the aspirates grew C albicans. (c) Supine abdominal radiograph obtained during water-soluble contrast material injection through the draining percutaneous catheter (long arrow) shows communication of a residual pseudocyst (curved arrow) with pancreatic tail duct (short arrow). Multiple small calculi within a dilated pancreatic head duct (arrowheads) caused obstruction to the flow of contrast material into the duodenum. Endoscopic retrograde cholangiopancreatography and papillotomy were subsequently performed, with stone clearance. This result allowed complete resolution of the pseudocyst, with removal of the percutaneous catheter.

 


View larger version (169K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. Unilocular diverticular fluid collection treated with percutaneous catheter drainage in 83-year-old woman with history of giant cell arteritis. (a) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material shows a moderately large unilocular fluid collection (long arrow) at the right side of the pelvis, associated with moderately severe sigmoid diverticulosis (short arrows). Culture of aspirated fluid grew C glabrata and Enterococcus species. (b) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material after transgluteal catheter (arrow) drainage shows complete resolution of the fluid collection. Eleven days after drainage, the patient underwent successful sigmoid resection with primary anastomosis.

 


View larger version (157K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. Unilocular diverticular fluid collection treated with percutaneous catheter drainage in 83-year-old woman with history of giant cell arteritis. (a) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material shows a moderately large unilocular fluid collection (long arrow) at the right side of the pelvis, associated with moderately severe sigmoid diverticulosis (short arrows). Culture of aspirated fluid grew C glabrata and Enterococcus species. (b) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material after transgluteal catheter (arrow) drainage shows complete resolution of the fluid collection. Eleven days after drainage, the patient underwent successful sigmoid resection with primary anastomosis.

 
Predisposing Factors and Features
Fifty-eight (97%) of 60 patients in this study had clinical factors that predisposed them to fungal infection (Table 1). Most patients had indwelling tubes, which included urinary catheters, arterial catheters, and central venous catheters, inserted at some time during their admission. Forty-one (68%) patients had undergone surgery at a mean of 15.8 days (range, 2–208 days) prior to percutaneous catheter insertion. In the majority, this consisted of gastrointestinal surgery performed to treat a variety of underlying bowel conditions, which included a neoplasm, inflammatory bowel disease, and perforation. In three patients, the surgery was performed primarily to drain abdominal fungus-infected fluid collections, but they subsequently required catheter drainage because of recurrence. Thirty-seven (62%) patients were admitted to the ICU at some time during their hospital stay.

The commonest cause of fungus-infected fluid collection in this study was a leak from the gastrointestinal tract, which was found in 41 (68%) of 60 patients. The full list of causes of collection is given in Table 2. The collections ranged in maximum diameter from 2.2 to 19.0 cm, with a mean of 7.1 cm. Eighteen (30%) patients had a complex collection because of the presence of loculi (n = 9), fistula (n = 7), or both (n = 2). The anatomic location and imaging features of the collections are given in Table 3.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Causes of Fungus-infected Fluid Collections in 60 Patients

 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Imaging Features of Collections according to Anatomic Location

 
Catheters
The catheters used for drainage ranged in size from 7 to 22 F, with a 10- or 12-F catheter used in 34 (57%) of 60 patients. A total of 89 catheters were placed; a single catheter was placed in 42 (70%) patients, and a maximum of five catheters were placed in one patient. Secondary catheter interventions were required in 13 patients because catheters fell out in two patients, upsizing and/or repositioning for better drainage was required in six patients, and additional catheters were placed for complete drainage in five patients. The catheters were removed electively because of adequate treatment in 35 patients, removed at surgery following drainage in 10 patients, left in till death in 12 patients, and left in position for long-term drainage in three patients. Excluding the latter three patients, mean duration of catheter drainage was 17.4 days (range, 1–128 days).

Organisms
The details of the contents of the fluid collections that were drained and their microbiologic findings are given in Table 4. Candida albicans was the most frequent fungal organism isolated, being cultured from 58 (79%) of 73 collections. C glabrata was the second most frequent fungal organism and was isolated from 17 (23%) of 73 collections. This organism occurred significantly (P < .01) more commonly in pelvic collections; it was found in seven (50%) of 14 collections within the pelvis, compared with 10 (17%) of 59 collections outside of the pelvis. Other forms of Candida and fungi were isolated much less frequently.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Nature of Aspirates and Microbiologic Findings according to Anatomic Location in 73 Collections

 
Of the 60 patients, fungal growth was reported as abundant in 18 (30%) patients, moderate in 18 (30%) patients, and rare in 24 (40%) patients. Pure fungal growth was found in 15 (25%) patients, and bacterial coinfection, in the remaining 45 (75%) patients. Aerobic Gram-negative Enterococcus species originating from the gastrointestinal tract were the commonest bacterial species isolated. All patients with bacterial coinfection were also treated with intravenous antibiotics as determined with drug sensitivity studies. Fungemia was present in seven (12%) of 60 patients in this study, and catheter drainage provided the first isolate of deep fungal infection in 29 (48%) patients.

Fungal organisms were isolated from one or more distant sites in 45 (75%) of 60 patients. The commonest sites of isolation were sputum (n = 23), urine (n = 19), wounds (n = 11), and surgical drains and/or central venous catheters (n = 9). C albicans and C glabrata were again the commonest organisms isolated. In any patient, the organism isolated from the distant site differed from that isolated from the drained collection in 26 (58%) of 45 patients. All patients were treated with systemic antifungal agents: Fluconazole (Diflucan; Pfizer Roerig, New York, NY) was used in 49 (82%) patients; amphotericin B (Abelcet; Enzon Pharmaceuticals, Piscataway, NJ), in seven patients; and a combination of the two, in the remaining four patients. Amphotericin B bladder irrigation was also used in two patients with fungal urinary tract infection.

Technical Success
Collections were adequately drained in 41 (79%) of the 52 patients who underwent follow-up imaging, partially drained in 10 (19%) patients, and not drained at all in one (2%) patient. These results yielded a technical success rate of 79% for percutaneous catheter drainage of fungus-infected fluid collections. Because of the increased viscosity of contents or the presence of a fistula, catheters were upsized or repositioned in six (12%) patients for improved drainage.

At specific analysis to determine the effect of loculation on technical success of drainage, percutaneous drainage was found to be successful in four (44%) of nine patients with imaging evidence of a multilocular collection (Fig 4), compared with 37 (86%) of 43 patients with a unilocular collection (Fig 3). Thus, the presence of loculi significantly (P < .01) reduced the success of percutaneous catheter drainage.



View larger version (173K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a. Multilocular fungus-infected fluid collection at right side of pelvis in 62-year-old woman with acute leukemia and perforated acute diverticulitis. (a) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material shows multiple small separate air and fluid collections (arrows) clustered together at the right side of pelvis. (b) Transverse CT image obtained through the pelvis with orally administered contrast material shows residual inflammatory soft-tissue thickening and fluid collection (arrow) at right side of pelvis, despite prolonged drainage and insertion of a second percutaneous catheter. Because of continued patient sepsis, a laparotomy was performed, with surgical drainage of the residual fluid collection and fashioning of a defunctioning colostomy.

 


View larger version (151K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b. Multilocular fungus-infected fluid collection at right side of pelvis in 62-year-old woman with acute leukemia and perforated acute diverticulitis. (a) Transverse CT image obtained through the pelvis with orally and intravenously administered contrast material shows multiple small separate air and fluid collections (arrows) clustered together at the right side of pelvis. (b) Transverse CT image obtained through the pelvis with orally administered contrast material shows residual inflammatory soft-tissue thickening and fluid collection (arrow) at right side of pelvis, despite prolonged drainage and insertion of a second percutaneous catheter. Because of continued patient sepsis, a laparotomy was performed, with surgical drainage of the residual fluid collection and fashioning of a defunctioning colostomy.

 
When we compared patients with pure fungal growth (n = 15) with those with mixed fungal and bacterial growth (n = 37), loculi were present in two (13%) versus seven (19%) patients, and complexity of collection, in four (27%) versus 11 (30%) patients; catheter drainage was technically successful in 12 (80%) versus 29 (78%) patients. At statistical analysis, there was no significant difference (P > .05) in the presence of loculi or the complexity of collection between these two groups of patients; in both, the collections were equally well drained by using percutaneous catheters.

Clinical Success
In 60 patients, indication for drainage was as follows: fever, in 50 (83%) patients; multiorgan failure that required respiratory and/or cardiovascular support, in 14 (23%) patients; and a miscellaneous group of other symptoms, in 11 (18%) patients. The patient mean white blood cell count before drainage was 17.3 x 109/L (range, [3.3–48.3] x 109/L). Within 3–4 days of percutaneous catheter drainage, the clinical condition of the patient improved, with defervescence and reduced respiratory and/or cardiovascular support in 44 (73%) patients. The patient mean white blood cell count at 1 week after drainage was 15.0 x 109/L (range, [1.3–30.7] x 109/L), a significant (P < .01) reduction when compared with that before drainage.

Thirty-four (57%) of 60 patients who underwent percutaneous catheter drainage had a complete clinical recovery, without the need for surgical intervention; however, the remaining 26 (43%) patients required open surgical drainage of the infected collection and/or died. Thus, the clinical success rate with catheter drainage of fungus-infected fluid collections was determined to be 57%, and the clinical failure rate was 43%. The mean patient hospital stay was 41 days, with a range of 2–120 days.

During their hospital admission, 20 of 60 patients (including three patients who had undergone surgical drainage after failed initial catheter drainage) died, resulting in a hospital mortality of 33%. The majority of patients died from protracted fungal and/or bacterial sepsis or from other systemic causes that resulted in multisystem organ failure. All deaths occurred in patients who were admitted to the ICU; this finding resulted in a mortality of 54% (20 of 37) in patients with fungal infection and ICU admission. Mortality was higher in patients with pure fungal growth, with deaths occurring in seven (47%) of 15 patients, compared with deaths in 13 (29%) of 45 patients with mixed fungal and bacterial growth, although this difference did not reach statistical significance (P = .27).

Eighteen (30%) patients required surgery after percutaneous catheter drainage: In nine patients, the surgery was performed emergently to treat inadequately drained fungus-infected fluid collections (Fig 4); in the other nine patients, the surgery was performed electively to correct underlying causes of collection. The causes included urinary tract obstruction in three patients (cystoscopic resection of a bladder tumor with stent placement, n = 1; megaureter repair, n = 1; and reconstruction of ileal-urinary conduit for stenosis, n = 1) and sigmoid diverticular disease in two patients (Fig 3). The elective surgery was performed with a mean delay of 50.2 days (range, 2–175 days) after catheter insertion.

The results of the multifactor logistic regression analysis of 15 clinical and microbiologic factors related to a poor clinical outcome are given in Table 5. Of the 15 factors evaluated, admission to the ICU (P < .001), complexity of collection (P < .029), and history of malignancy (P < .011) were found to be significantly and independently related to an outcome of clinical failure. Of these three factors, admission to the ICU was most strongly related to an outcome of clinical failure.


View this table:
[in this window]
[in a new window]

  		TABLE 5. {chi}2 Test Results for 15 Clinical and Microbiologic Factors Related to Poor Clinical Outcome

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
In the literature, there are multiple reports of experience with percutaneous catheter drainage of fungus-infected fluid collections in the thorax (12) and abdomen (6,7,11,13–15,19–24). Some authors have commented on the reduced success with percutaneous drainage (6,13–15) caused by the ill-defined nature and increased viscosity of fungal collections. With improvements in imaging technology and interventional techniques, however, radiologists are now able to identify and treat most fungus-infected fluid collections in the body.

In general, there are two types of fungal infections: the diffuse miliary type acquired through hematogenous spread, which predominantly involves the solid viscera (16,25,26), and the localized fluid type that occurs in the peritoneal and/or pleural spaces, which is caused by a leak from the gastrointestinal tract. The first type usually manifests as hepatosplenic candidiasis in the immunocompromised patient after a period of neutropenia. At imaging, the organs involved typically show minimal textural changes or microabscesses that require percutaneous aspiration and/or biopsy for diagnosis (16,25). These are almost always adequately treated with systemic antifungal therapy (16), with percutaneous treatment (aspiration or catheter drainage) reserved for the larger collections (24). In our study, there were two such patients with hepatic candidiasis who had dominant collections treated successfully with this regimen.

The second type of infection is characterized by localized fluid collections that occur in the peritoneal and/or pleural spaces because of a leak from the gastrointestinal tract and often occurs in the postoperative period. This cause accounted for infection in the majority (68%) of patients in this study. In addition to fungi isolated from these collections, coinfection with bacteria was also found in 75% of the patients in the overall study group. Aerobic Gram-negative Enterococcus was the commonest bacteria identified; this organism is commonly associated with the immunocompromised patient (27). This finding of a mixed fungal and bacterial infection in the postsurgical patient is a common occurrence, with a reported incidence of 24%–87% (12,28).

The main cause of technical failure in our study was related to the finding of loculi within the collection at imaging. Even with early imaging and reintervention in patients who did not immediately respond to percutaneous catheter drainage, the presence of loculi and enteric fistula resulted in an increased failure rate. Loculi and enteric fistula are well recognized causes of failure with percutaneous catheter drainage of pyogenic collections (29–31). There is evidence that loculi occur with an increased incidence in fungal infections (12,21). In a study of splenic fungal abscesses, Nelken et al (21) found a significantly (P < .001) higher number of fungal collections to be multilocular when they were compared with bacterial collections (90% vs 25%).

Not all researchers in some studies, however, found that complexity of the collection was associated with failure of percutaneous catheter drainage (21,32). Results of a recent multivariate analysis by Benoist et al (32) indicated no association with loculi or enteric fistula, but the maximum diameter of the abscess (<5 cm) and the absence of antibiotic therapy were the determinants of failure in this group of patients with postoperative abdominal fluid collections. Although Nelken et al (21) found an increased incidence of loculi in fungus-infected fluid collections compared with pyogenic collections in their study of splenic collections, they did not find a significant (P > .8) difference between the clinical successes with drainage of unilocular collections compared with those with drainage of multilocular collections (70% vs 50%).

We, on the other hand, found that complexity of the collection significantly (P < .01) impaired the technical success of drainage, with success decreasing from 86% in patients without loculi to 44% in patients with loculi. We did not use intracavitary fibrinolytic or antifungal therapy to try to improve the success of drainage. In a study of Ko et al (12), however, a high incidence of loculate fungal empyema was reported and occurred in up to 41% of their patients, with a resultant high mortality of 73%. However, all four patients who were treated with intrapleural instillation of antifungal agents survived.

Upsizing or repositioning of catheters was required to improve drainage in a small number of patients (n = 6) in our study. Increased viscosity of collections and the presence of enteric fistula were the main reasons for this requirement, which was found particularly with paraduodenal and pancreatic collections. Intermittent use of normal saline flushes was found to be useful in making the contents easier to aspirate. The one complete failure of drainage that occurred was in a patient with a posttraumatic pleural collection that turned out to be a hematoma infected with Candida organisms. This patient died of multisystem organ failure despite subsequent thoracotomy and surgical evacuation of the infected hematoma. Although surgical drainage has been advocated as the primary treatment for complex fluid collections (31), we do not feel that an initial attempt with radiologic management compromised the use of surgery in our group of patients. In patients with incomplete catheter drainage and continued sepsis, surgical drainage was performed to aid patient recovery. We used percutaneous catheter drainage and surgical drainage as complementary rather than as competitive techniques, an approach that also has been used by other investigators (7).

The relatively high technical success rate of 79% with percutaneous catheter drainage was associated with the expected reduction in fever and reduced respiratory and/or cardiovascular support in 73% of our patients, together with a significant decrease in their white blood cell counts. Despite this finding, the clinical success, which was defined as the patient's ability to recover without surgery, was much lower at 57%. The all-cause hospital mortality was 33%. This finding of a high clinical failure rate associated with high mortality in fungus-infected patients is not unique; in most studies dealing with treatment of patients who had undergone surgery and had fungal infections, researchers (5,9,10,12,19,33,34) also reported a high mortality ranging from 31% to 75%. This is consistently higher than that which is typically found in patients with catheter drainage of pyogenic collections (29,32). We believe that this discrepancy is due to the severe underlying debility of fungus-infected patients who eventually succumb from overwhelming sepsis or other systemic complications.

At statistical analysis, the findings of loculi at imaging, history of malignancy, and admission to the ICU were independent and significant predictors of a poor clinical outcome. Of these findings, admission to the ICU had the strongest association. This is not surprising, as this finding is simply a reflection of the severity of the underlying debility of the patient. Cornwell et al (33) also showed that in patients in the ICU, mortality was significantly higher in patients with systemic fungal infection (36.6%) compared with those without fungal infection (10.5%). In our study, all patients who died had been admitted to the ICU, and this finding resulted in a mortality of 54% in this group. Therefore, the finding of invasive fungal infection in a surgical patient being cared for in the ICU is often a premorbid condition and a predictor of poor clinical outcome.

Other medical factors, which included age of patient (34), severity of underlying medical disease (5), and semiquantitative estimation of fungal load (9), have all been previously identified as factors predictive of clinical outcome in patients with invasive fungal infections. We did not find that the age of the patient influenced clinical outcome. We found an association between patients who had pure fungal growth and a higher mortality when these patients were compared with those who had mixed flora, although this association did not reach statistical significance. In a study of candidal isolates from the peritoneum of surgical patients, Calandra et al (9) found a high mortality in patients with intraabdominal candidal infection who had a high or increasing amount of Candida in the semiquantitative culture. The confirmation of these findings may well be worth pursuing in a larger study.

There are at least two limitations to our study, and these limitations are primarily related to its retrospective nature. First, there is a selection bias that was introduced because we included nonconsecutive patients who were selected to undergo percutaneous catheter drainage on the basis of the referral pattern, patient medical condition, and imaging characteristics. Therefore, we were unable to evaluate the effectiveness of percutaneous catheter drainage in all patients who had fungus-infected fluid collections at presentation during the study period. Second, we were unable to stratify patients according to the severity of their underlying disease by assigning a physiologic score, and, therefore, we were unable to directly evaluate the influence of this factor on patient clinical outcome.

In conclusion, percutaneous catheter drainage is moderately effective in the treatment of patients with fungus-infected fluid collections, with a technical success rate of 79% and a clinical success rate of 57%. Technical failure is significantly related to the finding of loculi within the collection at imaging. Despite a relatively high technical success rate (79%) with percutaneous catheter drainage of fungus-infected fluid collections, the clinical success rate (57%) remains much lower because of the severity of the underlying debility of the patient. Complexity of collection, history of malignancy, and admission to ICU are significant and independent predictors of a poor clinical outcome.


   ACKNOWLEDGMENTS
 
We acknowledge the help given by Elkan F. Halpern, PhD, in the statistical analysis for this study.


   FOOTNOTES
 

Abbreviations: ICU = intensive care unit

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, J.C.V., P.R.M.; study concepts, P.R.M., P.F.H., D.C.H.; study design, J.C.V., D.A.G., P.F.H., M.G.H.; literature research, J.C.V., S.M.H.; clinical studies, J.C.V., P.F.H., P.R.M.; data acquisition and analysis/interpretation, S.M.H., D.C.H., P.F.H., J.C.V.; statistical analysis, P.F.H.; manuscript preparation, P.F.H., D.A.G., J.C.V., P.R.M., D.C.H.; manuscript definition of intellectual content, editing, revision/review, and final version approval, P.F.H., J.C.V., P.R.M.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Hart PD, Russel E Jr, Remington JS. The compromised host and infection. II. Deep fungal infection. J Infect Dis 1969; 120:169–191.
  2. Jarvis WR, Martone WJ. Predominant pathogens in hospital infections. J Antimicrob Chemother 1992; 29(suppl A):19–24.[Abstract/Free Full Text]
  3. Marshall JC, Christou NV, Horn R, Meakins JL. The microbiology of multiple organ failure: the proximal gastrointestinal tract as an occult reservoir of pathogens. Arch Surg 1988; 123:309–315.[Abstract/Free Full Text]
  4. Eubanks PJ, de Virgilio, Klein S, Bongard F. Candida sepsis in surgical patients. Am J Surg 1993; 166:617–620.[CrossRef][Medline]
  5. Komshian SV, Uwaydah AK, Sobel JD, Crane LR. Fungemia caused by Candida species and Torulopsis glabrata in the hospitalized patient: frequency, characteristics, and evaluation of factors influencing outcome. Rev Infect Dis 1989; 11:379–390.[Medline]
  6. Cinat ME, Wilson SE, Din AM. Determinants for successful percutaneous image-guided drainage of intra-abdominal abscess. Arch Surg 2002; 137:845–849.[Abstract/Free Full Text]
  7. Huang CJ, Pitt HA, Lipsett PA, et al. Pyogenic hepatic abscess: changing trends over 42 years. Ann Surg 1996; 223:600–607.[CrossRef][Medline]
  8. vanSonnenberg E, Ferrucci JT, Mueller PR, Wittenberg J, Simeone JF. Percutaneous drainage of abscesses and fluid collections: technique, results, and applications. Radiology 1982; 142:1–10.[Abstract/Free Full Text]
  9. Calandra T, Bille J, Schneider R, Mosimann F, Francioli P. Clinical significance of Candida isolated from peritoneum in surgical patients. Lancet 1989; 2:1437–1440.[CrossRef][Medline]
  10. Solomkin JS, Flohr AM, Simmons RL. Indications for therapy for fungemia in postoperative patients. Arch Surg 1982; 117:1272–1275.[Abstract/Free Full Text]
  11. Wise GJ, Talluri GS, Marella VK. Fungal infections of the genitourinary system: manifestations, diagnosis, and treatment. Urol Clin North Am 1999; 26:701–718.[CrossRef][Medline]
  12. Ko SC, Chen KY, Hsueh PR, Luh KT, Yang PC. Fungal empyema thoracis: an emerging clinical entity. Chest 2000; 117:1672–1678.[Abstract/Free Full Text]
  13. Pruett TL, Rotstein OD, Crass J, Frick MP, Flohr A, Simmons RL. Percutaneous aspiration and drainage for suspected abdominal infection. Surgery 1984; 96:731–737.[Medline]
  14. Aeder MI, Wellman JL, Haaja JR, Hau T. Role of surgical and percutaneous drainage in the treatment of abdominal abscesses. Arch Surg 1983; 118:273–280.[Abstract/Free Full Text]
  15. Hesse UJ, Sutherland DR, Najarian JS, Simmons RL. Intraabdominal infections in pancreas transplant recipients. Ann Surg 1986; 203:153–162.[Medline]
  16. Marcus SG, Walsh TJ, Pizzo PA, Danforth DN Jr. Hepatic abscess in cancer patients: characterization and management. Arch Surg 1993; 128:1358–1364.[Abstract/Free Full Text]
  17. Lent WM, Goldman MJ, Bizer LS. An objective appraisal of the role of computed tomographic (CT) guided drainage of intra-abdominal abscesses. Am Surg 1990; 56:688–690.[Medline]
  18. Varghese JC. Percutaneous drainage of abdominal abscesses and fluid collections. In: Kandarpa K, Aruny JE, eds. Handbook of interventional radiologic procedures. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins, 2001; 332–341.
  19. Marsh PK, Tally FP, Kellum J, Callow A, Gorbach SL. Candida infections in surgical patients. Ann Surg 1983; 198:42–47.[Medline]
  20. Escalante-Glorsky S, Youssef AI, Chen YK. Torulopsis glabrata–infected pancreatic pseudocysts: diagnosis and treatment. J Clin Gastroenterol 1995; 21:230–232.[CrossRef][Medline]
  21. Nelken N, Ignatius J, Skinner M, Christensen N. Changing clinical spectrum of splenic abscess: a multicenter study and review of the literature. Am J Surg 1987; 154:27–34.[CrossRef][Medline]
  22. Fitzgerald EJ, Lyons K. Candida abscess of the pancreas: diagnosis and treatment by computed tomography–guided percutaneous drainage. Br J Radiol 1986; 59:1121–1123.[Abstract/Free Full Text]
  23. Khardori N, Wong E, Carrasco CH, Wallace S, Patt Y, Bodey GP. Infections associated with biliary drainage procedures in patients with cancer. Rev Infect Dis 1991; 13:587–591.[Medline]
  24. Friedman E, Blahut RJ, Bender MD. Hepatic abscesses and fungemia from Torulopsis glabrata: successful treatment with percutaneous drainage and amphotericin B. J Clin Gastroenterol 1987; 9:711–715.[Medline]
  25. Shirkhoda A. CT findings in hepatosplenic and renal candidiasis. J Comput Assist Tomogr 1987; 11:795–798.[Medline]
  26. Pastakia B, Shawker TH, Thaler M, O'Leary T, Pizzo P. Hepatosplenic candidiasis: wheels within wheels. Radiology 1988; 166:417–421.[Abstract/Free Full Text]
  27. Catalano OA, Hahn PF, Hooper DC, Mueller PR. Efficacy of percutaneous abscess drainage in patients with vancomycin-resistant enterococci. AJR Am J Roentgenol 2000; 175:533–536.[Abstract/Free Full Text]
  28. Rutledge R, Mandel SR, Wild SR. Candida species: insignificant contaminant or pathogenic species. Am Surg 1986; 52:299–302.[Medline]
  29. Lang EK, Springer RM, Glorioso LW, Cammarata CA. Abdominal abscess drainage under radiologic guidance: causes of failure. Radiology 1986; 159:329–336.[Abstract/Free Full Text]
  30. Gerzof SG, Johnson WC, Robbins AH, Nabseth DC. Expanded criteria for percutaneous abscess drainage. Arch Surg 1985; 120:227–232.[Abstract/Free Full Text]
  31. Malangoni MA, Shumate CR, Thomas HA, Richardson JD. Factors influencing the treatment of intraabdominal abscesses. Am J Surg 1990; 159:167–171.[CrossRef][Medline]
  32. Benoist S, Panis Y, Pannegeon V, et al. Can failure of percutaneous drainage of postoperative abdominal abscesses be predicted? Am J Surg 2002; 184:148–153.
  33. Cornwell EE, Belzberg H, Berne TV, et al. The pattern of fungal infections in critically ill surgical patients. Am Surg 1995; 61:847–850.[Medline]
  34. Tang E, Tang G, Berne TV. Prognostic indicators in fungemia of the surgical patient. Arch Surg 1993; 128:759–763.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Varghese, J. C.
Right arrow Articles by Mueller, P. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Varghese, J. C.
Right arrow Articles by Mueller, P. R.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE