Nuclear Medicine Imaging Tools in Fever of Unknown Origin: Time for a Revisit and Appropriate Use Criteria

Abstract

Fever of unknown origin (FUO) is a clinical conundrum for patients and clinicians alike, and imaging studies are often performed as part of the diagnostic workup of these patients. Recently, the Society of Nuclear Medicine and Molecular Imaging convened and approved a guideline on the use of nuclear medicine tools for FUO. The guidelines support the use of 2-¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET)/computed tomography (CT) in adults and children with FUO. ¹⁸F-FDG PET/CT allows detection and localization of foci of hypermetabolic lesions with high sensitivity because of the ¹⁸F-FDG uptake in glycolytically active cells that may represent inflammation, infection, or neoplasia. Clinicians should consider and insurers should cover ¹⁸F-FDG PET/CT when evaluating patients with FUO, particularly when other clinical clues and preliminary studies are unrevealing.

The Society of Nuclear Medicine and Molecular Imaging recently convened and approved a guideline that support the use of 2-¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography in adults and children with fever of unknown origin.

BACKGROUND

The majority of human fever episodes are transient and do not need specialized treatment or diagnostic testing. Some are manifestations of more serious illnesses, most of which can be readily diagnosed and effectively treated. Despite major advances in medical tools, a small subgroup of fevers are both persistent and challenging to diagnose. Such puzzling fevers have fascinated and frustrated clinicians since the earliest days of clinical thermometry, resulting in a welter of clinical publications [1]. The most important of these, from a historical perspective, is the article “Fever of unexplained origin: report on 100 cases” by Petersdorf and Beeson (1961) to prospectively develop the first set of formal diagnostic criteria [1, 2]. Subsequent investigators modified these criteria to reflect practice changes in medicine, such as the 4 categories of fever of unknown origin (FUO; ie, classic, nosocomial, neutropenic, and human immunodeficiency virus [HIV]-associated), move toward outpatient evaluations, and present qualitative defining criteria [3–5].

The diagnostic defining criteria for FUO are primarily based on 3 main criteria that vary based on the source being referenced: (1) temperatures ≥100.9°F (≥38.3°C) on several occasions with; (2) an illness of more than 3 weeks’ duration in adults or at least more than 1 week duration in children; and (3) an uncertain diagnosis after either a predetermined time frame (ie, 1 week of hospital-based study, 3 days of hospital study, or 3 outpatient visits), or on completion of a recommended set of minimal laboratory and imaging studies (Table 1) [2, 4–6]. There is a wide plethora of FUO causes, with diagnoses falling into 1 of 5 categories: infectious diseases, noninfectious inflammatory disorders, oncologic conditions, noninflammatory miscellaneous conditions, and undiagnosed FUO illnesses (7). A heterogeneous entity such as FUO often requires an extensive diagnostic workup and in contemporary medicine, the use of 2-deoxy-2-¹⁸F-fluoro-D-glucose (¹⁸F-FDG) positron emission tomography (PET) has been recommended for patients with FUO [7]. In modern practice, ¹⁸F-FDG PET is usually combined with computed tomography (ie, ¹⁸F-FDG PET/CT), which not only provides an anatomic reference for the PET images but also improves the accuracy of the study, especially when performed with intravenous contrast enhancement [8].

Table 1.

Definitions for Fever of Unknown Origin in Adults and Children

Adults	Petersdorf and Beeson [2] Illness of more than 3 weeks’ duration, Fever higher than 100.9°F (38.3°C) on several occasions, and Diagnosis uncertain after 1 week of study in hospital
	Durack and Street [4] Illness of more than 3 weeks’ duration, Fever higher than 100.9°F (38.3°C) on several occasions, and Diagnosis uncertain after 3 days of study in hospital or 3 outpatient visits
	de Kleijn et al [3]a Illness of more than 3 weeks’ duration, Fever higher than 100.9°F (38.3°C) on several occasions, and Diagnosis uncertain after obligatory investigationsb
Children	Chow and Robinson [7] Illness of more than 1 week duration, Fever higher than 100.9°F (38.3°C) on several occasions, and Diagnosis uncertain after 1 week of initial investigations

^aPatients are excluded from this fever of unknown origin criterion if they have neutropenia (neutrophil count <0.5 × 10⁹/L for at least 1 week within 3 months before the start of the fever, receive immunosuppressive drugs because of solid organ or hematologic stem cell transplant, known hypogammaglobulinemia or use of 10 mg prednisone or equivalent for ≥2 wk in the 3 mo before the start of the fever, uncontrolled human immunodeficiency virus (HIV) infection or CD4 < 200 cells/mL, and/or patients who receive biologic therapies (eg, anti-tumor necrosis factor or monoclonal antibody). Adapted with permission from Wright WF et al [9]. Am J Med. 2022.

^bThe standard minimum diagnostic protocol includes the following components:

(1) Laboratory tests—complete blood count including differential, comprehensive metabolic panel including calcium and liver function tests, erythrocyte sedimentation rate, C-reactive protein, and ferritin.

(2) Microbiology tests—blood cultures (minimum 3 sets with spacing, incubation 5 d), urinalysis and reflex to urine culture (minimum 1 set), and tuberculin skin test or interferon-gamma release assay.

(3) Imaging tests—must include both abdomen ultrasonography and posteroanterior-lateral view chest plain-film or chest/abdominal/pelvic computed tomography.

The first consensus guideline for the appropriate use criteria of nuclear medicine for adults and children with FUO was recently published by the Society of Nuclear Medicine and Molecular Imaging (SNMMI) and included members from the Infectious Diseases Society of America [10]. In this viewpoint article, we explore the challenges with FUO evaluation, highlight the recent SNMMI guidelines supporting the use of ¹⁸F-FDG PET/CT, and the implications for clinical practice.

Evaluating the FUO Patient

FUO remains one of the most commonly accepted but challenging medical diagnoses for several reasons. For example, there is no single set of defining criteria that all patients with this condition must meet, no universally accepted postdiagnostic evaluation method, and no agreement regarding a uniform set of FUO-associated disease category classifications [3, 6, 11, 12]. Additionally, the prevalence of FUO diagnostic disease categories differs among age groups [13], geographical location [12, 14, 15], country-level gross national income per capita [13, 16], fever pattern and duration [15], prospective or retrospective study design [12, 13, 16], differences in applied FUO defining criteria [13], and variations associated with the type of medical care available [12].

Most patients described in early FUO studies exhibited atypical manifestations of common illnesses in an era when advanced diagnostic laboratory or imaging studies were not available [1, 2]. Over the past 60 years, data from FUO studies have not only included common disorders but also rare conditions [1]. Therefore, the concept of potential diagnostic clues (PDCs), harnessing possibly useful aspects of the individual history, physical examination, and laboratory or imaging test results are used as a guide [5, 17]. However, PDCs have limitations as illustrated by De Kleijn et al [1, 5], where as many as 48% were misleading. Similarly, Bleeker-Rovers et al [18] reported an average of 15 PDCs per patient, of which the diagnostic yield (defined as the proportion of patients in whom the PDCs contributed to the diagnosis) was only 19%. Moreover, traditional laboratory tests, including cultures, serologic examination, and conventional imaging studies, yield a diagnosis in approximately one fourth of the cases, with low to moderate sensitivity for FUO-associated diseases [1, 5, 18]. Of the commonly employed diagnostic imaging tests, 1 small prospective series reported diagnostic yields of 60% for plain-film chest radiography, 82% for chest CT, 86% for an abdominal ultrasound (US), and 92% for abdominal CT [18], although other studies do not support such high diagnostic yield for these imaging tests. Yields for bone marrow aspirate and biopsy, which contribute to the diagnosis of FUO in approximately one fourth of cases, are usually worthwhile only when there are abnormal complete blood cell counts [19, 20]. Furthermore, histopathologic examination of tissues obtained by excisional biopsy, needle biopsy, or laparotomy establish a diagnosis in fewer than half of the cases [1, 21]. Finally, data on the use of molecular assays (ie, nucleic acid amplification tests such as 16s ribosomal sequencing, multiplex polymerase chain reaction assays, and next-generation sequencing) in FUO are limited and are often reserved for selective cases that remain undiagnosed, and may require a particular type of specimen for diagnosis [9]. Despite the vast challenges in evaluating these patients, there is growing evidence suggesting that the nuclear medicine test ¹⁸FDG PET/CT is useful early in the evaluation of FUO [1, 7, 18, 22]. Although no universally accepted definition of early use exists, we propose early use of ¹⁸FDG PET/CT once patients have completed the minimal standardized set of investigative tests that serves as the foundation for the qualitative criteria [3] or once patients complete the length of evaluation (7 or 3 days) as required by the quantitative criteria [2, 4] (Table 1).

¹⁸F-FDG PET/CT in FUO

In contrast to other nuclear medicine imaging methods such as gallium-67 (⁶⁷Ga) citrate (binds to leukocytes and also extravasates at sites with inflammation) and radiolabeled leukocyte scans (based on migration of labeled leukocytes to sites of inflammation), data indicate that ¹⁸F-FDG PET (detects foci of hypermetabolism based on ¹⁸FDG absorption in glycolytically active cells that may indicate inflammation, infection, or neoplasia [7]), contributes better diagnostically useful information than anatomic imaging such as CT alone [1, 7]. Although the use of contrast significantly enhances the diagnostic yield of CT, ¹⁸F-FDG PET/CT generally provides higher sensitivity than other approaches [7].

In an early comparative study with ⁶⁷Ga-citrate among 58 consecutive cases of FUO, Blockman et al [23] reported that ¹⁸F-FDG PET and ⁶⁷Ga-citrate scintigraphy were normal in 23% and 33% of these cases, helpful in the final diagnosis in 35% and 25%, respectively, and noncontributory in 42% each.

A March 2008 memorandum from the Centers for Medicare & Medicaid Services reaffirmed the policy of denying coverage for ¹⁸F-FDG PET for the diagnosis of FUO [7]. This decision was based on a small number of studies that suggested the test contributed to a final diagnosis in 33%–35% of cases [18, 23, 24]. Pooled data from several meta-analyses published subsequently [25–29] demonstrate a much higher diagnostic yield, with sensitivities ranging from 86% to 98%. In a subgroup analysis of 5 studies that addressed the contribution of ¹⁸FDG PET/CT compared with CT alone, the diagnostic yield was 32% [26].

A more recent study by Dibble et al [30] reported that ¹⁸F-FDG PET/CT had a 90% sensitivity for 42 hospitalized patients evaluated for fever and suspected infections, in contrast to the diagnostic yield of 67% for labeled leukocyte scans. Patients evaluated using ¹⁸F-FDG PET/CT were also more likely to have successful localization of an infection or cancer than those with a noninfectious inflammatory disorder. Furthermore, patients with negative ¹⁸F-FDG PET/CT results had an average chance of spontaneous remission that was approximately 6 times higher (relative risk = 5.6; 95% confidence interval [CI], 3.4–9.2; P < .001) than patients with positive PET/CT results. This suggests that, in patients who have undiagnosed classic FUO after a number of unrevealing investigations, a negative PET/CT result can be a reliable indicator of a favorable prognosis [26].

Radiation exposure in adults with whole-body ¹⁸F-FDG PET/CT (20 mCi dose) is estimated to be ∼2 rem (1.4 rem from ¹⁸F-FDG and 0.5 rem from the low-dose CT) [10], and comparable to radiation exposures from a conventional contrast-enhanced chest-abdomen-pelvis CT but without the potential risks of CT contrast [1, 26]. This amounts to 3–4 years of natural background radiation exposure per person in the United States. [31]. In children, ¹⁸F-FDG dose is weight based (0.10–0.14 mCi/kg) with an estimated radiation exposure of 0.21 rem/mCi (eg for a 20-kg child, the radiation exposure from ¹⁸F-FDG would be 0.42 rem). The accompanying whole-body CT would deliver another 0.5 rem, although child-friendly protocols can significantly reduce the CT radiation exposures [32].

Appropriate Use Criteria

The appropriate use criteria issued by the SNMMI provide guidance on nuclear medicine imaging in both adults and children with FUO [10]. Because SNMMI has been designated by the Centers for Medicare and Medicaid Services as a qualified provider-led entity, the evidence used for this guideline to synthesize the scientific literature and expert opinion was the RAND/University of California, Los Angeles Appropriateness method with a modified Delphi process to achieve expert consensus [10, 33]. According to the RAND/University of California, Los Angeles Appropriateness method [33], a panel of 11 physicians (that also include S. K., R. B., B. L. S., C. J. P., and S. K. J. from this viewpoint) assessed data from a systematic search of the literature to develop these guidelines [10]. The panel then ranked the appropriateness of clinical scenarios on the diagnostic accuracy for gallium scintigraphy (ie, ⁶⁷Ga-citrate), labeled leukocyte scans (ie, either technetium-99 m or indium-111 labeled), and ¹⁸F-FDG PET/CT on a 9-point scale, in which a score in the range of 1–3 was considered “rarely appropriate,” 4–6 “may be appropriate,” and 7–9 “appropriate.” The panelists then pooled these scores to generate a median appropriateness score, which supported the early use of ¹⁸F-FDG PET/CT for use in FUO for both adults and children (Table 2).

Table 2.

SNMMI Appropriate Use Criteria

Imaging Test	Adults		Children
	Score	Diagnostic Yield (95% CI)	Score	Diagnostic Yield (95% CI)
Labeled leukocyte scintigraphy	4	20.0% (14.0%–28.0%)	3	ND
67Ga-citrate	5	35.0% (25.0%–46.0%)	3	ND
18F-FDG PET and 18F-FDG PET/CT	8	44.0% (31.0%–58.0%) and 58.0% (51.0%–64.0%)	8	ND

Adapted with permission from Palestro et al [8]. SNMMI 2023.

Abbreviations: ¹⁸F-FDG PET, 2-deoxy-2-¹⁸F-fluoro-D-glucose positron emission tomography; ⁶⁷Ga-citrate, gallium-67 citrate; CI, confidence interval; ND, no data; SNMMI, Society of Nuclear Medicine and Molecular Imaging.

Evidence for ¹⁸F-FDG was based on 5 systematic reviews in adults (n = 5474 total subjects) reporting diagnostic yields of 76%–83% for ¹⁸F-FDG PET and 84%–98% for ¹⁸F-FDG PET/CT, which was rated as appropriate (score of 8) [10].The overall diagnostic yield was 44.0% (95% CI, 31.0–58.0) for ¹⁸F-FDG PET and 58.0% (95% CI, 51.0–64.0) for ¹⁸F-FDG PET/CT [10]. For those situations in which ¹⁸F-FDG PET/CT is not available, ⁶⁷Ga-citrate scintigraphy was preferred over labeled leukocyte scintigraphy [10]. The evidence for this recommendation was based on 2 systematic reviews (n = 550 subjects) in which the source of fever was correctly localized and had a diagnostic yield of 35% (95% CI, 25–46) in cases for ⁶⁷Ga-citrate scintigraphy and 20% (95% CI, 14–28) for labeled leukocyte scintigraphy [10]. Finally, for those situations in which there was a high index of suspicion for occult infection (ie, not readily apparent) as the cause of the fever, and ¹⁸F-FDG PET/CT is not available, labeled leukocyte scintigraphy could be used [10].

There were limited data available for children and the evidence for ¹⁸F-FDG was based on 3 retrospective investigations (n = 210 subjects) reporting diagnostic yields of 33%–62% of cases [10]. For children with FUO requiring intensive care support, 1 study (n = 19 subjects) reported that ¹⁸F-FDG PET/CT accurately localized the source of fever in 14 of 15 (93.3%) [10]. Among immunocompromised children, 2 studies (n = 17 subjects) reported that 58%–60% of ¹⁸F-FDG studies yielded a final diagnosis [10].

Both ⁶⁷Ga-citrate and labeled leukocyte scintigraphy were deemed rarely appropriate for children [10]. For ⁶⁷Ga-citrate, the recommendation was based on 2 investigations (n = 34 subjects) in which only 5 (14.7%) studies were positive, including 3 (8.8%) that localized infections that had not been identified with other imaging modalities [10]. The recommendation for labeled leukocyte scintigraphy was based on 3 studies (n = 56 subjects) reporting diagnostic yields that ranged from 45% to 94% [10].

Implications for Practice

The heterogeneity of FUO, the lack of multicenter, high-quality studies, and the extensive differential mean that clinical judgment remains an essential component of care. From a practical perspective, the goals of FUO evaluations include diagnostic stewardship ultimately reducing unnecessary, invasive, and expensive diagnostic investigations, achieving an earlier final diagnosis, and reducing the number of undiagnosed cases. ¹⁸F-FDG PET/CT appears to offer clinical benefit, particularly in the more difficult FUO cases.

There is also evidence suggesting that ¹⁸F-FDG PET/CT imaging should be performed earlier (ie, once patients complete the minimal standardized set of investigative tests that serves as the foundation for the qualitative criteria [3] or once patients complete the length of evaluation [7 or 3 days] required by the quantitative criteria), rather than later (ie, after other conventional imaging studies such as stand alone CT or magnetic resonance imaging), in the diagnostic evaluation of patients with FUO [1, 7, 18, 22, 33, 34]. This can reduce prolonged hospital stays associated with unnecessary diagnostic investigations and thus save costs. For example, in a retrospective study of 20 patients, Becerra Nakayo et al [34] studied both the theoretical and real cost-effective benefits of ¹⁸F-FDG PET/CT in FUO investigations. In the absence of ¹⁸FDG PET/CT, they estimated an actual cost of $12 064 for a mean hospital evaluation of 28 days and 1 outpatient consultation. If ¹⁸F-FDG PET/CT had been performed by the second week of investigations, $5910 per patient would have been saved. In another study, Chen et al [32] evaluated 326 patients with FUO who underwent ¹⁸F-FDG PET/CT were compared with 415 patients who did not. Although ¹⁸F-FDG PET/CT was used primarily in critically ill and hard-to-diagnose patients, the mean length of hospitalization and medical costs before diagnosis were significantly lower in patients who underwent ¹⁸F-FDG PET/CT within 1 week after hospital admission [35].

Several additional studies have suggested that ¹⁸FDG PET/CT, primarily when used early in FUO investigations, has helped establish the correct diagnosis sooner, reduce hospital length of stay, and avoid unnecessary tests, particularly when scans are used among patients with unintentional weight loss of at least 5% over the previous 6 months and anemia (eg, hemoglobin < 10.7 g/L) [1, 18, 23, 25–30, 34, 35]. These findings suggest that ¹⁸FDG PET/CT imaging should be included in the diagnostic algorithm and performed earlier in evaluations, rather than later, for diagnosing the patient with FUO [1, 7, 18]. Diagnostic considerations for which this technology may be most helpful include consideration of localized abscesses, osteomyelitis, sinusitis, sarcoidosis, vasculitis, adult-onset Still disease, Crohn disease, and subacute thyroiditis [1].

FUO experts have recently proposed rational approaches to the diagnostic investigation of these patients that incorporate ¹⁸F-FDG PET/CT earlier in the process (Figure 1) [7, 36]. However, practical issues in the United States still limit the utility of ¹⁸F-FDG PET/CT because not all hospital centers have this technology available or accessible, and certain health care insurers such the U.S. Centers for Medicare and Medicaid Services, do not universally reimburse ¹⁸F-FDG PET/CT use for the FUO, creating disparities across different settings [37]. Additional challenges lie in the need for several hours of fasting before imaging, determining the appropriate use of this combined technology to the clinical setting (ie, inpatient vs outpatient); reducing wait times resulting from obtaining preauthorization for the procedures and appropriately staffing of ¹⁸F-FDG PET/CT centers. This demonstrates the need for a standardized approach to FUO with the goal of increasing healthcare equity.

Figure 1.

Three-step diagnostic evaluation for fever of unknown origin. Minimal obligatory studies: CBC including differential, CMP including calcium and liver function tests (LFT), ESR, CRP, ferritin, TSH, RF, and ANA. Microbiology studies include HIV-1/2 serology, urinalysis with microscopy, Mycobacterium tuberculosis skin test or whole-blood interferon-γ releasing assay, and blood cultures (3 sets). Imaging studies must consist of both abdomen ultrasonography and posteroanterior-lateral view chest plain-film or chest/abdominal/pelvic CT; secondary investigative studies: infectious diseases molecular tests, serologies or other blood testing (eg, rheumatologic panel), specialized imaging (TTE/TEE, MRI, or ¹⁸F-FDG PET/CT scans) or appropriate biopsy methods for culture and histologic analysis; repeat comprehensive history and physical examination should include periodic reevaluation of the patient's history and physical examination for new potential diagnostic clues (PDC). For any new PDCs, use proper invasive and/or noninvasive diagnostic testing considering any further information. If fevers continue and PDC-driven evaluation has been exhausted, consider ¹⁸F-FDG PET/CT, if not already performed. If the patient remains ill without a diagnosis other than FUO, discuss the risks and advantages with the patient of empiric use of nonsteroidal anti-inflammatory drugs, immunosuppressive medications such as corticosteroids, or anti-infectives (eg, therapy for tuberculosis, which should be considered as an empiric treatment before patients become severely ill) may be considered. Abbreviations: ¹⁸F-FDG PET, 2-deoxy-2-¹⁸F-fluoro-D-glucose positron emission tomography; ANA, antinuclear antibodies; CBC, complete blood count; CMP, comprehensive metabolic panel; CRP, C-reactive protein; CT, computed tomography; ESR, erythrocyte sedimentation rate; HIV, human immunodeficiency virus; MRI, magnetic resonance imaging; RF, rheumatoid factor; TEE, transesophageal echocardiogram; TSH, thyroid-stimulating hormone; TTE, transthoracic echocardiogram. Adapted with permission from Wright et al [9]. Am J Med. 2022.

CONCLUSION

FUO remains a challenging clinical condition and use of ¹⁸F-FDG PET/CT offers clinical benefit, particularly among patients lacking potential diagnostic clues. The recent SNMMI guidelines provide an equitable framework for the use of nuclear medicine imaging in the management of FUO, supporting the early use of ¹⁸F-FDG PET/CT in these patients. Clinicians and insurers should consider ¹⁸F-FDG PET/CT when evaluating patients with FUO, particularly when other clinical clues and preliminary studies are unrevealing. Promising PET-based, pathogen-specific imaging technologies are also emerging [38–43], which could further enhance the clinical value of PET-based approaches for FUO.

Notes

Author Contributions . Concept and design: W. F. W., S. K. J., C. J. P. Drafting of the primary manuscript: W. F. W., S. K. J. Critical revision of the manuscript for important intellectual content: all authors Administrative, technical, or material support: all authors.

Financial support . W. F. W. has been supported by the Johns Hopkins Institute for Clinical and Translational Research (ICTR), which is funded in part by Grant Number UL1 TR003098 from the National Center for Advancing Translational Sciences (NCATS), a component of the National Institutes of Health (NIH), and the NIH Roadmap for Medical Research. S. K. J. is supported by NIH grants R01-AI145435-A1, R01-AI153349, R01-AI161829-A1, and R01-EB025985.