Fluordeoxyglucose positron emission tomography contributes to management of pediatric liver transplantation candidates with fever of unknown origin

Systemic infections are considered to be contraindications for liver transplantation.1 Therefore, pediatric candidates for liver transplantation presenting with febrile episodes need to be evaluated thoroughly. Fever of unknown origin (FUO) is defined as a temperature higher than 38.3°C on several occasions and lasting longer than 3 weeks, with a diagnosis that remains uncertain after 1 week of investigation.2 In patients awaiting liver transplantation, FUO often causes considerable delay in adequate treatment of serious infections, leading to deteriorated general condition of the transplant candidate and prolonged time and increased mortality on the waiting list. Sensitive, reliable diagnostic tools are needed to support a timely and accurate diagnosis of infection. In particular, the diagnosis of an intrahepatic focus of infection is important for the management of the candidate on the waiting list, as the presence of a localized intrahepatic infection without systemic involvement does not render the patient unsuitable for transplantation. In the absence of overt pathology, e.g., abscess formation reliably identified by ultrasound, computed tomography (CT) or magnetic resonance imaging (MRI), alternative modes of imaging, e.g., scintigraphy, have been tested to adequately identify a focus of intrahepatic infection. In general, 3 scintigraphic methods have been applied. The use of ⁶⁷Ga-citrate has been largely abandoned because of unfavorable dosimetry and image quality. The alternative has become radiolabeled autograft leukocytes. ¹¹¹In-labelled (isotope labeling with ¹¹¹Indium) or ^99mTc-labeled (isotope labeling with ^99mTechnetium) leukocytes (white blood cells) are routinely used in clinical practice. However, they have limitations, such as instability of the labeling and high radiation burden,3 the cumbersome collection of blood in pediatric patients, and the relatively long time between injection and diagnosis. In the recent years, [18 F]fluordeoxyglucose (FDG)–positron emission tomography (PET) has become a standard procedure in the workup of patients with oncological diseases.4 The high sensitivity of FDG-PET coincides with a relatively low specificity, due to accumulation of FDG in inflammatory processes.3 This drawback has, however, led to a new development, i.e., the use of FDG-PET in inflammatory conditions including FUO. PET has been known for some time to be useful in localization and imaging of infectious foci.5-8 Infection imaging with FDG-PET relies on the fact that granulocytes and mononuclear cells use glucose as an energy source specifically during their metabolic burst upon activation by triggers.3, 9 Therefore, understandably, FDG accumulates in many different types of inflammatory lesions as well as in malignant tumors. The advantages of FDG-PET are early imaging after injection,8 higher resolution and higher target-to-background ratio,3 sensitivity to chronic low-grade infections,3, 6, 10, 11 and high interobserver agreement.4, 10, 12 Here we present our experience using FDG-PET to positively identify intrahepatic infections in children on the waiting list for liver transplantation while other imaging modalities failed. FUO is considered a contraindication for transplantation; however, hepatic infections are frequently only cured by removal of the infected organ during transplantation. Therefore the potential of FDG-PET to identify intrahepatic infections may be crucial for management of patients with FUO awaiting liver transplantation.

Abbreviations

FDG, [18 F]fluordeoxyglucose; PET, positron emission tomography; FUO, fever of unknown origin; CT, computed tomography; MRI, magnetic resonance imaging; BA, biliary atresia.

MATERIALS AND METHODS

RESULTS

Five of 11 patients had abnormal FDG-PET results with increased focal uptake in liver parenchyma as compared to surrounding cirrhotic tissue. In all 5 patients undergoing liver transplantation with previously detected intrahepatic positive FDG-PET signals, findings in the explanted liver were abnormal (Table 1). Structures were identified that were macroscopically or histologically compatible with local infections (patients 1, 2, and 5) or cultures taken intraoperatively from the explanted organ grew bacterial strains, which may infect chronically diseased biliary trees in cirrhosis (patients 1-5). Thus, all of these FDG-PET findings were considered true-positives. In 6 patients with FUO, FDG-PET scans did not reveal any increased focal uptake. The uptake was isointense on background level throughout the whole organ. In these patients' cultures, macroscopic or microscopic findings at the time of transplantation did not reveal signs of focal infection. These FDG-PET results were considered true-negatives.

All 11 patients in this study suffered from biliary cirrhosis.

Clinical Characteristics of Patients With FUO and True-Positive Intrahepatic FDG-PET Findings

All patients with positive intrahepatic FDG-PET results were diagnosed with primary biliary-tract diseases: 4 of 5 patients had been diagnosed with biliary atresia (BA) and had undergone a Kasai operation (portoenterostomy); 1 patient had primary sclerosing cholangitis. All patients with BA had experienced 1 or more bouts of cholangitis earlier in the disease course. One patient (diagnosed with BA) had received steroids in the postoperative course after a Kasai operation. Patient 3 with primary sclerosing cholangitis had been treated for cholangitis, hydropic gallbladder, and renal failure 6 weeks before transplantation. Following bile duct decompression through papillotomy, he recovered clinically but developed persistent fever in the later course. Ultrasound and/or CT or MRI were performed in all patients and revealed irregular, partially dilated bile ducts in 1 (patient 3) and 3 inhomogeneous cystic lesions in the right and the left liver lobe in another patient (patient 1), which were found to accumulate FDG-PET signal. None of these lesions showed signs of an abscess in standard imaging. In the period of 3 months to 2 weeks before FDG-PET was performed, E. coli bacteremia was detected in patient 2, for which he received antibiotics according to susceptibility. Bile was cultured in 1 case and revealed a streptococcus species coinciding with biliary obstruction, which was treated with decompression (patient 3) and antibiotics according to susceptibility. Patients 2 and 3 recovered initially following antibiotic treatment. Febrile episodes recurred, however, necessitating further diagnostic procedures. Repeatedly, an enterococcus species was cultured in feces from 1 patient (patient 1). Single or multiple intrahepatic FDG-PET signals were detected in segments 2, 4, and 8 and diffusely in the left liver lobe (Table 1; Figs. 1 and 2). Superposition of FDG-PET and previously taken MRI images allowed a co-localization of presumed sites of infection with irregularly distended bile ducts (patient 3; Fig. 2). Sites of physiologic tracer accumulation included brain, kidneys, bladder, and frequently sites of increased muscle activity, e.g., the diaphragm and perilingual regions (patient 4). Positive, isolated, intrahepatic FDG-PET signals were interpreted as potential cholangitis and treated with broad-spectrum antibiotics until liver transplantation.

Four patients underwent liver transplantation during periods of recurrent fever due to presumed cholangitis without signs of systemic sepsis. In 1 patient fever did not recur after long-term treatment with antibiotics. Intra- and postoperatively, all patients received antibiotic treatment according to susceptibility testing of the detected strain. One patient died following primary nonfunction of the graft. Three patients are now in good clinical condition, 2 in their second yr after liver transplantation. In 1 patient, biliary strictures occurred postoperatively that required transient stenting 2 months after liver transplantation.

DISCUSSION

Clearly, morbidity and mortality before and after liver transplantation correlate with frequency and severity of infections of hepatic and extrahepatic origin.13 Cholangitis is a frequent complication in patients with biliary cirrhosis awaiting liver transplantation. For detection of hepatic infections diagnostic tools are limited. Positive identification of infections such as cholangitis in the liver parenchyma may be difficult and diagnoses are frequently made on the basis of exclusion.

Focal infectious processes can be detected by several radiological techniques, including CT, MRI, and ultrasound. However, positive identification of infectious and inflammatory foci may be difficult in their early phases because of the lack of substantial anatomical changes at that time. Also, discrimination of active infectious or inflammatory lesions from residual changes due to chronic disease (e.g., liver cirrhosis) or surgery remains difficult. All of our patients presented with elevated levels of serum C-reactive protein as an indicator of ongoing infection. Other parameters such as procalcitonin have been shown to be advantageous in discriminating infectious from inflammatory foci14 and may also be of value in patients with liver cirrhosis and FUO. Risk of infection of the biliary tract is high in patients after Kasai procedure due to BA15-17 and may be observed in patients with strictures of the biliary tract such as in primary sclerosing cholangitis.13, 18 Recurrent biliary infections in these diseases are considered to be risk factors for deterioration of liver function prior to liver transplantation.13 FUO is considered a relative contraindication, and systemic infections should be treated before liver transplantation. However, infection originating from within the liver may not be curable without removal of the diseased organ.1 For positive identification of patients who are potential candidates for liver transplantation in spite of recurrent fever episodes during ongoing localized hepatic infections, sensitive tools for diagnosis are mandatory. In the patients presented here, imaging by ultrasound, CT, or MRI had failed to demonstrate infectious lesions in the cirrhotic liver. Although the number of patients in our retrospective study is limited, our results confirm those from other studies that FDG-PET is a sensitive test for detecting infectious foci in FUO.2, 19 When correlated with anatomic or histologic signs of hepatic infections and/or intraoperative cultures of the excised liver, all 5 abnormal intrahepatic FDG-PET signals can be considered true-positives. We cannot exclude that Corynebacterium and Staphylococcus epidermidis, identified in intraoperative cultures from explanted livers (patients 2, 3, and 5), are contaminants. However, both strains have been described as opportunistic pathogens leading to hepatic abscesses in immunocompromised patients.20, 21 The abscesses found in the resected liver of patients 2 and 5 underscore the potential role of staphylococcus as a liver pathogen in patients with biliary cirrhosis. An obvious problem of FDG-PET imaging is low specificity.22 This is demonstrated in patients with positive FDG-PET signals originating from extrahepatic structures. In this case, differentiation between tissue with physiologic uptake (e.g., muscle tissue) and enhanced signal due to infection may be difficult. In a single case, culture results correlated with signal accumulation in the urinary tract. In patients with positive FDG intrahepatic signals and thus suspected hepatic infections, antibiotic treatment was continued until transplantation while the patient had no signs of systemic involvement of infection, except for recurrent febrile episodes. Based on the FDG-PET result the patients remained active candidates on the waiting list and underwent liver transplantation. During the immediate postoperative period and initiation of immunosuppressive therapy no complications arose from systemic infection after removal of the affected organ. The number of patients examined in our study is too small to conclude whether this approach can be considered as safe. Further studies are needed to address the risks of performing liver transplantation in this setting. However, as we think that positive detection of intrahepatic infectious lesions has strong implications for management of these patients with regard to their suitability for transplantation, we propose an algorithm for diagnostic procedures for pretransplantation patients with FUO as depicted in Figure 3. This algorithm should be tested in larger studies to confirm the role of FDG-PET in febrile patients with cirrhosis. In conclusion, FDG-PET may contribute to optimized management of pediatric patients with FUO on the waiting list for liver transplantation. However, extrahepatic signal accumulation must be interpreted with caution. Further prospective controlled studies are warranted to evaluate FDG-PET as a diagnostic tool to identify hepatic infection in liver transplant candidates with FUO.