Usefulness of 18F-FDG PET-CT in the Management of Febrile Neutropenia: A Retrospective Cohort from a Tertiary University Hospital and a Systematic Review

Febrile neutropenia (FN) is a complication of hematologic malignancy therapy. An early diagnosis would allow optimization of antimicrobials. The 18F-FDG-PET-CT may be useful; however, its role is not well established. We analyzed retrospectively patients with hematological malignancies who underwent 18F-FDG-PET-CT as part of FN management in our university hospital and compared with conventional imaging. In addition, we performed a systematic review of the literature assessing the usefulness of 18F-FDG-PET-CT in FN. A total of 24 cases of FN underwent 18F-FDG-PET-CT. In addition, 92% had conventional CT. In 5/24 episodes (21%), the fever was of infectious etiology: two were bacterial, two were fungal, and one was parasitic. When compared with conventional imaging, 18F-FDG-PET-CT had an added value in 20 cases (83%): it diagnosed a new site of infection in 4 patients (17%), excluded infection in 16 (67%), and helped modify antimicrobials in 16 (67%). Antimicrobials could be discontinued in 10 (41.6%). We identified seven publications of low quality and one randomized trial. Our results support those of the literature. The available data suggest that 18F-FDG-PET-CT is useful in the management of FN, especially to diagnose fungal infections and rationalize antimicrobials. This review points out the low level of evidence and indicates the gaps in knowledge.


Introduction
Febrile neutropenia (FN) is a common complication of cancer, particularly in patients with hematological malignancies who receive high-intensity chemotherapy or stem cell transplantation (SCT) [1].In this setting, infections are an important complication and cause significant morbidity and mortality.The etiological workup of FN includes, in addition to microbiological tests, different imaging tests; among them, the most commonly used is computed tomography (CT).Specifically, high-resolution computed tomography is recommended for suspected invasive fungal infection (IFI) [2].
Unfortunately, in a large number of cases it is not possible to determine the cause of the fever with conventional imaging.This results in the prolonged use of empirical broad-spectrum antimicrobials, both antibiotics and antifungals [3].In addition, in almost 50% of the cases according to some studies [4], the etiology of FN is not infectious.In these situations, an early etiological diagnosis would allow the performing of targeted diagnostic tests and optimization of antimicrobial therapy, including timely discontinuation of unnecessary antimicrobials [4].Therefore, there is a need to implement better complementary tests to improve diagnosis and allow de-escalation of antimicrobials.
PET with 18 F-fluorodeoxyglucose provides functional information that correlates with the anatomical data provided by CT.Using 18 F-FDG as a radiotracer, information on the metabolism in different tissues is obtained.Unlike conventional CT, 18 F-FDG-PET-CT is capable of evaluating more than one area of the body in just one session in addition to providing metabolic information, allowing one to more easily detect clinically silent lesions [5].This imaging technique is typically used for staging cancer and has seen its use increase lately as part of the study of fever of unknown origin or FN [1,6].
In recent years, several studies have shown the potential of 18 F-FDG PET-CT to localize the source of fever and to differentiate between infection and other etiologies in patients with FN.Although 18 F-FDG-PET-CT has not shown a clear benefit in the differential diagnosis of cancer and infection, nevertheless, it could have a role in FN and in detecting the spread of infection and occult infection [7].Other authors have underlined the high negative predictive value of 18 F-FDG-PET-CT that facilitates the adjustment of antimicrobial treatment [4,8].In particular, they point out that 18 F-FDG-PET-CT could play an important role in the diagnosis of IFI and help with the withdrawal of empirically initiated antifungals [1,4,8].In a recent review on the management of patients with high-risk FN, the authors suggest that 18 F-FDG-PET-CT could be especially helpful in these patients when it comes to reducing the antibiotic spectrum without changes in ICU admissions or mortality [9].
However, the literature on this topic is scarce and mainly consists of short cases series in diverse settings.In addition, most of these studies do not compare the performance of conventional tests and 18 F-FDG-PET and the added value provided specifically by 18 F-FDG-PET-CT.Consequently, its role in routine clinical practice has not been well established so far [1,4,6,8].
Although several authors have reviewed the existing literature [1,4,[6][7][8], to our knowledge, there is not a comprehensive and systematic review on this topic.
In the present article, we analyze our center's data on the usefulness of 18 F-FDG PET-CT in hematological patients with FN and its added value compared with conventional imaging, and we perform a systematic review of the published literature on the use of 18 F-FDG PET-CT in that setting.

Single-Center Retrospective Cohort Study 2.1.1. Design, Study Period, and Subjects
Our institution is a 613-bed tertiary-care teaching hospital in Madrid, Spain.The Hematology Department has an active SCT program, including allogeneic SCT (haploidentical and cord as well).
We performed a retrospective observational study including all adult patients admitted to Puerta de Hierro-Majadahonda Hospital between 2015 and 2022 diagnosed with hematological malignancies (leukemia, aplasia, myelodysplastic syndrome, multiple myeloma, or lymphoma) undergoing chemotherapy or SCT who underwent at least one 18 F-FDG PET-CT as part of the FN management.

Data Collection
Epidemiological, clinical (including the type of hematological malignancy and SCT, the type of infection, localized or disseminated disease, and the type of pathogen), labora-tory, and imaging data were extracted from electronic medical records (SELENE System, Cerner Iberia, S.L.U., Madrid, Spain) using a standardized data collection form.The 18 F-FDG PET-CT indication and impact of the results on FN management were specifically addressed.All data were included by a primary reviewer and, subsequently, checked by two senior physicians.
2.1.3. 18F-FDG PET-CT Technique All 18 F-FDG PET-CT scans were performed according to EANM (European Association of Nuclear Medicine) guidelines, in hybrid PET/CT chamber systems [10].The CT component was non-contrast enhanced.All patients complied with a previous fasting period of at least six hours (12)(13)(14)(15)(16)(17)(18) h in cases of suspected endocarditis; in this case a dietary modification protocol was also applied).Ideally, they should maintain blood glucose levels lower than 180 mg/dL.If insulin was administered, the injection of 18 F-FDG would be spaced at least four hours apart.For infectious and inflammatory diseases, the same acquisition, reconstruction, and post-processing described in the procedures of the EANM for tumors were used [10,11].Full-body 18 F-FDG PET-CT, from the cranial vertex to the feet in a supine position, was acquired approximately 50-60 min after the intravenous injection of 370 ± 30 MBq 18 F-FDG depending on the patient's weight.When infective endocarditis was a possibility, the study was completed with dedicated cardiac 18 F-FDG PET-CT acquisition.
The 18 F-FDG PET-CT images were analyzed for increased uptake of 18 F-FDG outside the areas of physiological incorporation.A qualitative analysis was carried out, considering the uptake pattern (focal, linear, or diffuse) and the distribution of the radiotracer in the pathological area or lesion (homogeneous or heterogeneous), and a semiquantitative analysis was performed considering the intensity of the uptake.The images were interpreted as normal, equivocal, or with pathological uptake according to the standard uptake values (SUV): the visual scores were 0, no pathological uptake; 1, uptake similar to the vascular pool in the mediastinum; 2, uptake higher than the vascular pool but lower than the liver pool; 3, uptake similar to or slightly higher than the liver; 4, uptake clearly higher than the hepatic, where 0 and 1 would be negative and 2, 3, and 4 would be positive (always assessing the location and alternative causes that explain the uptake).

Other Imaging Techniques
The diagnostic workup for FN was performed at the discretion of the treating physicians.For every case, the results of conventional imaging techniques performed during the episode were compared with the 18 F-FDG PET-CT results, according to the reports by radiology specialists (or cardiologists, when applicable).This included X-ray, CT, MRI, and, in the case of bloodstream infection caused by Gram positives or yeasts, echocardiography.

Definitions
We followed the criteria for febrile neutropenia as per the NCCN guidelines [12]: -For fever, a single temperature equivalent to ≥38.3 • C orally or equivalent to ≥38.0 • C orally over a 1 h period; -For neutropenia, ≤500 neutrophils/mcL or ≤1000 neutrophils/mcL and a predicted decrease to ≤500/mcL in the next 48 h.
In addition, we also evaluated patients with persistent low-grade fever (temperature > 37.5 • C for more than 72 h).

Usual Care
For antimicrobial prophylaxis, see supplementary material S1 (for acute myeloid leukemia (AML), levofloxacin and posaconazole during the neutropenia period and acyclovir in cases who receive fludarabine; for acute lymphoblastic leukemia (ALL), cotrimoxazole and acyclovir and posaconazole when finishing vincristine; for autologous SCT, cotrimoxazole impregnation prior to transplantation, levofloxacin, fluconazole, and acy-clovir; for allogeneic SCT, cotrimoxazole impregnation prior to transplantation followed by nebulized pentamidine, levofloxacin, posaconazole, and acyclovir, plus letermovir and azithromycin in high-risk patients).
Regarding empirical antimicrobials, before 2019, empirical therapy for febrile neutropenia consisted of meropenem.From 2019 on, empirical therapy consisted of piperacillintazobactam or cefepime (the latter in cases where no anaerobic coverage was deemed necessary).No surveillance cultures were obtained routinely, but in cases known to be colonized by resistant microorganisms, antimicrobial therapy was adjusted accordingly.
In both periods, teicoplanin was added in cases of catheter infection suspicion, and amikacin was added in cases of septic shock.

Data Analysis
Quantitative variables are expressed as means and standard deviations (SD) and/or medians and interquartile ranges (IQR), and qualitative variables are expressed as frequencies and percentages.Measures of central tendency (mean and SD, and median and IQR) and proportions were calculated with IBM SPSS Statistics 22.

Systematic Literature Review
The studies were identified through a systematic search in different bibliographic databases using search terms (MeSH) related to the topic, specifically, Pubmed, Embase, and Cochrane Library.These databases were searched without language or publication date restrictions (see search strategy in supplementary material S2).We did not exclude articles based on the retrospective or prospective nature of the study.The reference lists of the relevant studies were checked to identify additional relevant articles.
To be eligible, a study had to evaluate the use of 18 F-FDG PET-CT in the management of FN. References were screened by two researchers based on the title and abstract using the PICOS framework (Table 1).Irrelevant references were excluded with explicit reasons.In a second step, the remaining references were screened based on the full text.This systematic review was performed according to the recommendatio PRISMA guidelines.A checklist is included as supplementary material.

Single-Center Retrospective Cohort Study
Among 638 eligible patients with FN during the study period, 24 episodes o detected in 23 patients who underwent 18 F-FDG PET-CT as part of the FN worku 1).The characteristics of the patients who underwent 18 F-FDG PET-CT for FN displayed in Table 2. Fifty-seven percent were men.The mean age was 58.6 years Regarding the non-onco-hematological comorbidities, 17% had diabetes, 8% ha moderate chronic kidney disease, and 4% had chronic pulmonary disease (CO most frequent onco-hematological underlying disease was acute myeloid leuke patients (50%), followed by acute lymphoblastic leukemia (ALL) in 3 patients (1 tiple myeloma (MM) and myelodysplastic syndrome (MDS) in 2 patient each NK immunodeficiency in 1 patient (4%).Four patients (17%) were stem cell tra tion recipients.The reasons for the SCT were as follows: ALL in two patients, M patient, and MDS in one patient.

patients
•Admitted with haematological malignancies 184 patients  The characteristics of the patients who underwent 18 F-FDG PET-CT for FN study are displayed in Table 2. Fifty-seven percent were men.The mean age was 58.6 years (SD 19.6).Regarding the non-onco-hematological comorbidities, 17% had diabetes, 8% had at least moderate chronic kidney disease, and 4% had chronic pulmonary disease (COPD).The most frequent onco-hematological underlying disease was acute myeloid leukemia in 12 patients (50%), followed by acute lymphoblastic leukemia (ALL) in 3 patients (12%), multiple myeloma (MM) and myelodysplastic syndrome (MDS) in 2 patient each (8%), and NK immunodeficiency in 1 patient (4%).Four patients (17%) were stem cell transplantation recipients.The reasons for the SCT were as follows: ALL in two patients, MM in one patient, and MDS in one patient.
The majority of patients fulfilled criteria for persistent fever (58%), followed by patients with persistent low-grade fever (33%).Neutrophil counts on the day of the 18 F-FDG PET-CT were below 100 cells/mm 3 in 12 (50%) and between 100 and 500 cells/mm 3 in 10 (42%) patients.Two (8%) had neutrophil counts above 500 cells/mm 3 on the day the test was performed but had recently recovered from severe neutropenia during the previous week.
Patients were receiving antimicrobial prophylaxis according to risk factors, following the institution's protocols (supplementary material S1).The distribution of the empirical treatments administered at the moment of 18 F-FDG PET-CT is shown in Table 1, the most commonly used drugs being meropenem (54%) and piperacillin-tazobactam (37%).In five cases of 24 FN episodes (21%), fever was considered of infectious etiology.The etiology was bacterial infection in two cases (8%), fungal in two (8%), and parasitic in one (4%).No viral infection was diagnosed.In one case (4%) there was a clinical diagnosis of infection, but a microbiological etiology could not be determined (catheter uptake without isolation of microorganisms).Among the bacterial infections, two bloodstream infections were diagnosed: a case of catheter-related S. haemolyticus bacteremia and another of catheter-related persistent E. faecium bacteremia without catheter uptake or septic metastases.Regarding fungal infections, one patient presented an invasive fusariosis and another with possible pulmonary IFI.The remaining patient with a known infectious etiology had visceral leishmaniasis.
In all but one patient, the infection was localized.The disseminated case was a patient initially diagnosed of naso-sinusal fusariosis, where the initial 18 F-FDG PET-CT helped to diagnose the occult source of fever, and in addition, the monitoring 18 F-FDG PET-CT detected the dissemination of the infection.
The median duration of antimicrobial therapy until 18 F-FDG PET-CT was performed was 13.5 days .The fever was considered to be secondary to non-infectious causes in 20 cases (83.3%): in 11 (46%) it was secondary to the underlying hematologic malignancy, in 1 (4%) it was considered of inflammatory etiology (reduction in corticosteroids in the context of disseminated mycobacterial infection), there were 3 cases (12%) of engraftment syndrome and 2 of graft-versus-host disease (GVHD) (8%), and in 3 cases (12%) an etiology was not identified.

Characteristics of the 18 F-FDG PET-CT as Compared with Conventional Imaging
The indication for 18 F-FDG PET-CT was the study of FN in all patients.The median time of neutropenia before 18 F-FDG PET-CT was 13.5 days (3-74), and the median time to 18 F-FDG PET-CT from the beginning of fever was 13 days (3-28).
The 18 F-FDG PET-CT showed pathological uptake in 20 (83%) cases.The most frequent location of this uptake was intra-abdominal visceral (37%, mainly hepato-splenic uptake), followed by bone marrow and lung uptake (21%).The most common distribution was multifocal in 71% of cases followed by focal uptake in 12%.The remaining 17% did not show any 18 F-FDG uptake.
Table 3 shows the results of conventional imaging compared with those of 18 F-FDG PET-CT and the added value of 18 F-FDG PET-CT in the management of febrile neutropenia.The 18 F-FDG PET-CT provided added value to the previous conventional imaging study in 20 patients (83%) considering as such the diagnosis of new infection or the exclusion of infection that led to the modification of antimicrobials or the initiation of treatment of the underlying hematological disease.It contributed to the diagnosis of new sites of infection in 4 patients (17%), ruled out infection in 16 patients (67%), and helped modify antimicrobial treatment in 16 patients (67%).It also allowed starting chemotherapy or immunomodulators in four patients: two patients started chemotherapy, one patient received specific treatment for graft-versus-host disease, and corticosteroid doses were increased in another patient.
Pathological uptake in 18 F-FDG PET-CT helped perform targeted diagnostic tests in 58%: four patients (17%) underwent fine-needle aspiration/biopsy of the pathological uptake area, and one patient (4%) underwent bronchoscopy, among other complementary tests.
The results of 18 F-FDG PET-CT implied the removal of the venous catheter in one case and surgical debridement in another one for source control.
In the cases that were eventually diagnosed as infection, the most common site of infection identified by 18 F-FDG PET-CT was hepatosplenic and biliary (8%), followed by catheter and pulmonary (4%).
Among the non-infectious causes, the most common reason for pathological uptake in 18 F-FDG PET-CT was the underlying disease in 11 patients (46%).Concerning antimicrobial therapy, in 16 patients (67%) the antimicrobial spectrum was modified based on 18 F-FDG PET-CT results.In 2 (8%) patients, antimicrobials were de-escalated; in 1 (4%) case the spectrum was expanded; and in 10 (42%) patients it allowed the discontinuation of antimicrobials.There was a need to start new antimicrobial treatment in three (12%) patients, in one of the cases also accompanied by surgical debridement.
Only one patient had more than one 18 F-FDG PET-CT during the study of the episode of FN.The patient diagnosed with sinus IFI due to Fusarium had a control 18 F-FDG PET-CT scan performed one month after the first one, which, as aforementioned, detected dissemination of the infection to the lungs as well as persistence of the sinusal infection.

Systematic Literature Review Search Strategy and Inclusion
The literature search retrieved 341 references that were de-duplicated, and non-English, Spanish, and French references were excluded (Figure 2).The remaining references were screened for eligibility based on the titles and abstracts (of which 160 were excluded).Based on full-text evaluation of the remaining publications, 16 articles that evaluated the use of 18 F-FDG PET-CT in the management of FN were included.This resulted in a sample of one RCT and 15 publications.Among them, five narrative reviews and a survey were excluded from the present analysis.Two-case reports were excluded to avoid publication bias.Eight articles were selected (Table 4).

Quality appraisal
Microorganisms 2024, 12, x FOR PEER REVIEW 22 of 31

Search Strategy and Inclusion
The literature search retrieved 341 references that were de-duplicated, and non-English, Spanish, and French references were excluded (Figure 2).The remaining references were screened for eligibility based on the titles and abstracts (of which 160 were excluded).Based on full-text evaluation of the remaining publications, 16 articles that evaluated the use of 18 F-FDG PET-CT in the management of FN were included.This resulted in a sample of one RCT and 15 publications.Among them, five narrative reviews and a survey were excluded from the present analysis.Two-case reports were excluded to avoid publication bias.Eight articles were selected (Table 4).

Quality appraisal
The quality of the eight included publications was evaluated as moderate to poor.The methodology used was heterogeneous.Because of the nature of the interventions, blinding of the patients and staff was not possible.Three of the studies were retrospective and, thus, non-random by design.Among the prospective studies, in one case, there was no comparison of 18 F-FDG PET-CT and conventional techniques, whereas the remaining studies performed both techniques on the same patients sequentially; therefore, there was no randomization to one or the other study.Likewise, there was no randomization of the sequence in which the techniques were performed.The quality of the eight included publications was evaluated as moderate to poor.The methodology used was heterogeneous.Because of the nature of the interventions, blinding of the patients and staff was not possible.Three of the studies were retrospective and, thus, non-random by design.Among the prospective studies, in one case, there was no comparison of 18 F-FDG PET-CT and conventional techniques, whereas the remaining studies performed both techniques on the same patients sequentially; therefore, there was no randomization to one or the other study.Likewise, there was no randomization of the sequence in which the techniques were performed.
Only one open-label randomized controlled trial was identified.Although masking of the randomization was not possible, the clinical impact of the randomized scans and the cause of neutropenic fever were assessed by an independent adjudicating committee to reduce the risk of bias.

•
Results of the studies according to the methodology (Figure 2) 1.

Clinical trial
Recently, a multicenter phase 3 controlled clinical trial [3] was published that randomized patients with high-risk NF 1:1 to perform CT vs. PET-CT.The primary endpoint was a combination of starting, stopping, or changing the spectrum of antimicrobial therapy as a result of the information provided by the imaging technique.They included a total of 134 patients (PET-CT 65; CT 69).Antimicrobial rationalization occurred in 82% of patients in the 18 F-FDG PET-CT group and 65% in the CT group.The most frequent action was the reduction in the spectrum of antimicrobial therapy, 43% for 18 F-FDG PET-CT compared with 25% for CT (p = 0.024).The authors concluded that 18 F-FDG PET-CT was associated with better optimization of antimicrobial therapy and could help decision making in this type of patient.The drawback of this clinical trial was the lack of direct comparison of 18 F-FDG-PET-CT and conventional imaging in the same patient.The authors did not provide information about whether the differences in baseline characteristics or the final fever etiology between patients who underwent 18 F-FDG PET-CT or conventional CT were statistically significant.

Original articles
Seven original articles that studied the usefulness of 18 F-FDG PET-CT in FN were found through the systematic search, five of them prospective and two retrospective [14][15][16][17][18][19][20].The characteristics of the original articles that evaluated 18 F-FDG-PET-CT's usefulness for FN management are summarized in Table 3.

a.
Studies comparing the results of conventional tests and 18 F-FDG PET-CT in the same patient Four articles [14,15,19,20] compared conventional imaging and 18 F-FDG-PET-CT performed in the same patient as part of the study of FN, in order to assess which one provides more information to improve management.Only one of these provided individual data with a head to head comparison of conventional imaging with 18 F-FDG-PET-CT in the same patients [19].A total of 161 patients with FN were evaluated in these studies (147 adults and 14 children).The median time to 18 F-FDG PET-CT from the beginning of fever to the performance of 18 F-FDG PET-CT varied between 6 and 14 days.The median time from conventional imaging to 18 F-FDG PET-CT was not provided in any of these studies.The final diagnosis was infectious in a high proportion of cases, varying from 55% to 79%.
Camus et al. [14] carried out a prospective single-center study to investigate the ability of 18 F-FDG-PET-CT to find the source of infection in patients with FN.In this study, among the 38 patients with a final clinical diagnosis of infection (79%), 23 had a pathological FDG uptake, resulting in a 18 F-FDG-PET-CT sensitivity of 61%.Among the 17 patients diagnosed with pneumonia by conventional evaluation, 18 F-FDG-PET-CT detected pulmonary uptake in 11 (64.7%) and uptake at multiple levels in 6 (35.3%).Gafter-Gvili et al. also evaluated in a prospective study the performance of 18 F-FDG-PET-CT for the diagnosis and treatment of infections in high-risk patients with FN [15].In this case, the sensitivity of 18 F-FDG-PET-CT was 79.8% compared with 51.7% for chest/sinus CT alone.The specificities were 32.14% versus 42.85%.Furthermore, in more than 50% of patients, 18 F-FDG-PET-CT changed the pre-test diagnosis and helped modify patient management.Both studies concluded that 18 F-FDG PET-CT had the ability to assist in the evaluation and management of these patients.To assess the impact of 18 F-FDG-PET-CT on persistent or recurrent fever.n = 14.In 11 of them (79%), 18 F-FDG-PET-CT had a clinical impact: in three, treatment was de-escalated, and in five, antibiotics were discontinued. 18F-FDG-PET-CT scans identified new foci in seven patients. 18 A third prospective study, carried out by Guy et al. [19], which included 20 patients with NF who underwent 18 F-FDG-PET-CT in addition to conventional techniques, revealed that 18 F-FDG-PET-CT was able to identify nine infections that CT was not able to identify and had a clinical impact in 75% of patients since it inducedtreatment changes.Like previous articles, it concluded that 18 F-FDG-PET-CT was useful in the assessment of NF.
A last, retrospective study [20], carried out in a pediatric population that included 14 patients, observed that 18 F-FDG PET-CT had a positive impact in 11 patients (79%), favoring the rationalization of antimicrobials in three (21%) and their discontinuation in five (36%).Furthermore, compared with conventional tests, it helped identify new sites of infection in seven (50%) patients and contributed to the final diagnosis in six (43%) patients.As in previous articles, the authors consider the potential usefulness of 18 F-FDG-PET-CT as part of the study of FN.
Another aspect to highlight is the usefulness of 18 F-FDG-PET-CT to assess the dissemination of infections.In the articles by Camus and Gvili discussed previously, 18 F-FDG-PET-CT was used to detect occult lesions that led to the diagnosis of disseminated infection in 27% and 1.3% of cases, respectively [14,15].
All of these studies [14,15,19,20], emphasize the usefulness of 18 F-FDG-PET-CT in the diagnosis of fungal infection and the rationalization of antifungal treatment.a.
Studies that do not compare conventional tests and 18 F-FDG PET-CT in the same patient Three articles evaluated the contribution of 18 F-FDG PET-CT in the diagnosis of infection without doing a head to head comparison with conventional tests in the same patient.These studies either performed conventional tests and 18 F-FDG PET-CT in different patients [17] or performed only 18 F-FDG PET-CT [16,18].A total of 92 patients with FN were evaluated in these studies.
The retrospective study by Koh KC et al. [17] evaluated the impact of 18 F-FDG-PET-CT on the use of antimicrobials in FN.They identified two groups: one that had undergone 18 F-FDG-PET-CT (n 37) and another that had had conventional imaging (n 76).There were no significant differences between cases and controls with respect to age, sex, underlying malignancy, and chemotherapy.The 18 F-FDG-PET-CT determined the cause of FN in 94.6% of patients compared with 69.7% in the conventional imaging group.Furthermore, 18 F-FDG PET-CT had a significant impact on antimicrobial use compared with conventional imaging (35.1% vs. 11.8%;p 0.003) and was associated with a shorter duration of antifungal therapy.The authors stated that 18 F-FDG PET-CT improved diagnostic performance and allowed the rationalization of antimicrobials in these patients.In this study as well, the usefulness of 18 F-FDG-PET-CT for the diagnosis of IFI and the rationalization of antifungal treatment was evidenced.
The two remaining studies did not compare 18 F-FDG-PET-CT with conventional imaging.The retrospective study by Mahfouz T et al. [18] reviewed the contribution of 18 F-FDG-PET-CT performed to 248 patients with multiple myeloma (MM) for the staging or diagnosis of infection where there was an uptake atypical for myeloma that could be suggestive of infection.A total of 165 infections were identified in 143 adults with MM, 27 of these episodes being in the context of neutropenia.The 18 F-FDG PET-CT detected 46 infections not detectable by other methods, helped determine the extent of infection in 32 episodes, and led to modification of the diagnosis and therapy in 55.In patients with staging 18 F-FDG PET-CT, twenty silent infections were detected.They concluded that 18 F-FDG PET-CT in MM was a useful technique for diagnosing infection; unfortunately, the authors did not provide specific results for the subset of patients with FN.
The prospective study by Vos FJ et al. [16] included 28 hematological patients with neutropenia who underwent 18 F-FDG-PET-CT in cases of CRP levels greater than 50 mg/L.In 26 out of 28 (92.9%)patients, that increase in CRP levels was accompanied by fever.The median time from starting chemotherapy to 18 F-FDG PET-CT was 14 days.They found pathological FDG uptake in 26 of 28 cases (92.9%).The authors did not specify in how many cases 18 F-FDG-PET-CT guided the performance of diagnostic tests.In this study, pulmonary uptake was significantly associated with the presence of IFI (p = 0.04).They determined that 18 F-FDG-PET-CT in the context of increased CRP was capable of detecting infection in situations of severe neutropenia.An evaluation of the impact of 18 F-FDG-PET-CT on antimicrobial prescriptions was not provided.

Discussion
Our data indicate that 18 F-FDG-PET-CT is useful in the management of FN.In 87% of the cases it helped to confirm or rule out infection, allowing optimization of empirical antimicrobial treatment including de-escalation or discontinuation of unnecessary antimicrobials in 16 cases (67%).These results support data from the 318 total cases with 18 F-FDG-PET-CT for FN analyzed in the studies included in the present review.
To the best of our knowledge, the present study is one of the few that compares the performance of conventional tests and 18 F-FDG PET-CT in the same patient during the FN episode.When comparing conventional imaging with 18 F-FDG-PET-CT performed on different patients, the differences in underlying diseases or in fever etiology could account for the differences observed in the yields of these tests.Comparing both techniques in the same patient overcomes this limitation.In addition, only patients who were still neutropenic at the moment of 18 F-FDG-PET-CT were selected (with the exception of two patients who had recovered neutrophils very recently, fewer than 3 days before PET), so that we cannot attribute the better performance of 18 F-FDG-PET-CT to neutrophil recovery.
The 18 F-FDG PET-CT was especially relevant in the diagnosis of uncommon fungal and parasitic infections, such as fusariosis or leishmaniasis.Interestingly, in the present series the proportion of infectious etiology was low, only 16.7%.Being a retrospective study, we cannot exclude that only patients with a lower probability of infectious cause (non-responders to antimicrobials, with already negative prior tests) underwent 18 F-FDG-PET-CT.In any case, the ability to rule out infection in these cases is the reason why antimicrobial therapy could be adjusted, similar to what other authors report (in spite of having a much higher proportion of infectious etiologies, ranging from 55% to 79%).Another important aspect is that thanks to the 18 F-FDG PET-CT results in cases in which infection was ruled out, patients were able to resume the chemotherapy treatment necessary for their underlying hematological disease, similar to other works [21].
Another benefit that 18 F-FDG-PET-CT provides is its potential to detect dissemination of infection, especially in cases of IFI.Several articles state that 18 F-FDG-PET-CT may have greater sensitivity than conventional tests to detect dissemination and occult lesions in the context of IFI [1,4,22].In the analysis of our data, 18 F-FDG-PET-CT was key in the diagnosis of disseminated fungal disease in one of the patients, which led to changes in therapeutic management.
Limitations to stating the role of 18 F-FDG-PET-CT in febrile neutropenia workup are access and cost.
The systematic search retrieved several very heterogeneous articles that intended to evaluate the benefits of 18 F-FDG-PET-CT in FN.The different methodologies, the lack of direct comparison between the techniques, and the different populations of patients studied precluded the performance of a metanalysis.With the exception of the only randomized controlled trial, the quality of the retrieved studies was poor.The lack of randomization together with the impossibility of masking increases the risk of bias.In this systematic review we analyze the results of these studies and point out knowledge gaps and unanswered questions.
First, although some of them are prospective studies, all are single-center studies that include only a small number of patients.The study by Mahfouz T et al. [18], even if it analyzes a large sample, is a retrospective study that only includes patients with MM, with a small proportion of FN.The studies by Koh et al. [17] and Guy et al. [19] were performed at the same hospital during the same period and might thus include in part the same patients.
We consider that the most relevant limitation of the majority of the articles is that they do not compare conventional tests and 18 F-FDG-PET-CT to better discern what value the 18 F-FDG-PET-CT adds in patients with FN, and among those that do (only five small studies) [14,15,17,19,20], not all perform both techniques on the same patient [17].In this sense, the clinical trial by Douglas A. et al. [3] is a significant contribution in this area, but similar to others, its main limitation is not performing both tests in the same patient.As aforementioned, differences in underlying baseline characteristics or even in the cause of fever are difficult to address with this design and could explain, at least to some degree, the differences observed in the yields of the techniques that are being evaluated.When evaluating a diagnostic test, we believe it should be compared with other tests performed on the same patient.
Moreover, among those that compared 18 F-FDG-PET-CT with conventional imaging, only one provided individual patient data [19].This fact, in addition to the aforementioned methodologic heterogeneity, made it impracticable to perform a metanalysis.
In spite of these limitations, according to the results of the aforementioned articles that altogether include a total of 344 cases of FN, 18 F-FDG-PET-CT seems to provide relevant information for the management of FN in a high proportion of cases.
In several of the studies the information provided by 18 F-FDG PET-CT is especially relevant in the case of difficult to diagnose infections such as IFI [1,3,15,17,20] or parasitic infections.In addition to helping diagnose IFI and unveil dissemination, it also seems to be more useful to monitor the response to treatment than CT alone [23] since CT in some cases continues to show radiological lesions corresponding to scar tissue that do not show pathological uptake in 18 F-FDG PET-CT, allowing the ending of antifungal treatment [1].
Of even greater importance is 18 F-FDG PET-CT's contribution to optimizing antimicrobial use in FN.Due to the high negative predictive value of 18 F-FDG PET-CT, it allows reducing the use of broad-spectrum antimicrobials, favoring in many cases de-escalation and even discontinuation of empirical treatment, mainly of antifungals [1,3,17].
The clinical trial by Douglas et al. provides relevant information about the safety of basing clinical decisions on 18 F-FDG PET-CT results.However, a formal cost-effectiveness analysis is pending to justify better access to 18 F-FDG PET-CT in FN high-risk patients [3].
Clinicians experienced with the use of 18 F-FDG PET-CT for the study of infection favor its use for prolonged FN and for an IFI diagnosis, according to a survey carried out by the Australian group.In particular, physicians who treat onco-hematological patients are likely to use 18 F-FDG PET-CT in patients with FN to optimize the diagnosis and therapeutic management [24].
Many unanswered questions remain.In many cases, 18 F-FDG PET-CT is considered when fever persists despite empirical antimicrobial treatments.But how long should we wait before performing 18 F-FDG PET-CT?When will it perform better during the course of FN?Is there a basic workup that should be performed before considering 18 F-FDG PET-CT?Should this basic set of studies include conventional imaging, or should 18 F-FDG PET-CT replace them?Is there a particular subset of patients who would benefit more from 18 F-FDG PET-CT?More studies with an adequate design are needed to clarify these points.We believe that a large multicentric prospective study that selects a well-categorized and homogeneous population of high-risk FN patients, using a protocolized workup for FN that includes both conventional imaging tests and 18 F-FDG PET-CT performed on the same patient during a short pre-established time window, would be an appropriate model to clarify the role of 18 F-FDG PET-CT and thus a diagnostic protocol that could include 18 F-FDG-PET-CT.

Institutional Review Board Statement:
The authors confirm that the study was approved by the Institutional Review Board (CEIm) at Hospital Universitario Puerta de Hierro (Majadahonda) (PI 109/23), and a waiver for informed consent was granted.The study complied with the provisions in European Union (EU) and Spanish legislation on data protection and the Declaration of Helsinki.
Data Availability Statement: After publication, the data will be made available to others upon reasonable requests to the corresponding author.A proposal with a detailed description of study objectives and statistical analysis plan will be needed for evaluation of the reasonability of requests.It might also be required during the process of evaluation.Deidentified participant data will be provided after approval from the principal researchers of Hospital Universitario Puerta de Hierro (Majadahonda).

P
(population): Patients of all ages with febrile neutropenia in the setting of oncohematologic malignancy I (intervention): 18 F-FDG PET-CT in the setting of febrile neutropenia management C (comparison): Conventional diagnostic tests used for febrile neutropenia workup Discrepancies in the scores were resolved through discussion.

Table 1 .
Research question in PICOS framework.

Table 2 .
Characteristics of FN episodes from our hospital.

Table 3 .
Evaluation of conventional imaging and added value of18F-FDG PET-CT in 24 patients with febrile neutropenia.

Table 4 .
Summary of the eight articles selected by means of the systematic search.