Application of Machine Learning for the Prediction of Etiological Types of Classic Fever of Unknown Origin

Background: The etiology of fever of unknown origin (FUO) is complex and remains a major challenge for clinicians. This study aims to investigate the distribution of the etiology of classic FUO and the differences in clinical indicators in patients with different etiologies of classic FUO and to establish a machine learning (ML) model based on clinical data.Methods: The clinical data and final diagnosis results of 527 patients with classic FUO admitted to 7 medical institutions in Chongqing from January 2012 to August 2021 and who met the classic FUO diagnostic criteria were collected. Three hundred seventy-three patients with final diagnosis were divided into 4 groups according to 4 different etiological types of classical FUO, and statistical analysis was carried out to screen out the indicators with statistical differences under different etiological types. On the basis of these indicators, five kinds of ML models, i.e., random forest (RF), support vector machine (SVM), Light Gradient Boosting Machine (LightGBM), artificial neural network (ANN), and naive Bayes (NB) models, were used to evaluate all datasets using 5-fold cross-validation, and the performance of the models were evaluated using micro-F1 scores.Results: The 373 patients were divided into the infectious disease group (n = 277), non-infectious inflammatory disease group (n = 51), neoplastic disease group (n = 31), and other diseases group (n = 14) according to 4 different etiological types. Another 154 patients were classified as undetermined group because the cause of fever was still unclear at discharge. There were significant differences in gender, age, and 18 other indicators among the four groups of patients with classic FUO with different etiological types (P < 0.05). The micro-F1 score for LightGBM was 75.8%, which was higher than that for the other four ML models, and the LightGBM prediction model had the best performance.Conclusions: Infectious diseases are still the main etiological type of classic FUO. Based on 18 statistically significant clinical indicators such as gender and age, we constructed and evaluated five ML models. LightGBM model has a good effect on predicting the etiological type of classic FUO, which will play a good auxiliary decision-making function.

Download Full-text

Patients with fever of unknown origin (FUO): diagnosis by nuclear medicine imaging

Nuklearmedizin ◽

10.1055/s-0038-1623872 ◽

2001 ◽

Vol 40 (03) ◽

pp. 59-70 ◽

Cited By ~ 13

Author(s):

W. Becker ◽

J. Meiler

Keyword(s):

Nuclear Medicine ◽

Fever Of Unknown Origin ◽

Tumor Imaging ◽

White Blood Cells ◽

Unknown Origin ◽

Final Diagnosis ◽

Recurrent Fever ◽

Nuclear Medicine Imaging

SummaryFever of unknown origin (FUO) in immunocompetent and non neutropenic patients is defined as recurrent fever of 38,3° C or greater, lasting 2-3 weeks or longer, and undiagnosed after 1 week of appropriate evaluation. The underlying diseases of FUO are numerous and infection accounts for only 20-40% of them. The majority of FUO-patients have autoimmunity and collagen vascular disease and neoplasm, which are responsible for about 50-60% of all cases. In this respect FOU in its classical definition is clearly separated from postoperative and neutropenic fever where inflammation and infection are more common. Although methods that use in-vitro or in-vivo labeled white blood cells (WBCs) have a high diagnostic accuracy in the detection and exclusion of granulocytic pathology, they are only of limited value in FUO-patients in establishing the final diagnosis due to the low prevalence of purulent processes in this collective. WBCs are more suited in evaluation of the focus in occult sepsis. Ga-67 citrate is the only commercially available gamma emitter which images acute, chronic, granulomatous and autoimmune inflammation and also various malignant diseases. Therefore Ga-67 citrate is currently considered to be the tracer of choice in the diagnostic work-up of FUO. The number of Ga-67-scans contributing to the final diagnosis was found to be higher outside Germany than it has been reported for labeled WBCs. F-l 8-2’-deoxy-2-fluoro-D-glucose (FDG) has been used extensively for tumor imaging with PET. Inflammatory processes accumulate the tracer by similar mechanisms. First results of FDG imaging demonstrated, that FDG may be superior to other nuclear medicine imaging modalities which may be explained by the preferable tracer kinetics of the small F-l 8-FDG molecule and by a better spatial resolution of coincidence imaging in comparison to a conventional gamma camera.

Download Full-text

Big Data for Health Care Analytics using Extreme Machine Learning Based on Map Reduce

International Journal of Engineering and Advanced Technology - Regular Issue ◽

10.35940/ijeat.c5808.029320 ◽

2020 ◽

Vol 9 (3) ◽

pp. 2758-2762

Keyword(s):

Machine Learning ◽

Big Data ◽

Data Storage ◽

Clinical Data ◽

Disease Risk ◽

Learning Algorithm ◽

Information Storage ◽

Support Vector ◽

Machine Learning Algorithm ◽

Data Set

A large volume of datasets is available in various fields that are stored to be somewhere which is called big data. Big Data healthcare has clinical data set of every patient records in huge amount and they are maintained by Electronic Health Records (EHR). More than 80 % of clinical data is the unstructured format and reposit in hundreds of forms. The challenges and demand for data storage, analysis is to handling large datasets in terms of efficiency and scalability. Hadoop Map reduces framework uses big data to store and operate any kinds of data speedily. It is not solely meant for storage system however conjointly a platform for information storage moreover as processing. It is scalable and fault-tolerant to the systems. Also, the prediction of the data sets is handled by machine learning algorithm. This work focuses on the Extreme Machine Learning algorithm (ELM) that can utilize the optimized way of finding a solution to find disease risk prediction by combining ELM with Cuckoo Search optimization-based Support Vector Machine (CS-SVM). The proposed work also considers the scalability and accuracy of big data models, thus the proposed algorithm greatly achieves the computing work and got good results in performance of both veracity and efficiency.

Download Full-text

Patients with Fever of Unknown Origin and Splenomegaly: Diagnostic Value of Splenectomy and Preoperative Risk Factors Suggestive of Underlying Lymphomas

Acta Haematologica ◽

10.1159/000473859 ◽

2017 ◽

Vol 137 (4) ◽

pp. 240-246

Author(s):

Lu Zhang ◽

Wei Zhang ◽

Huacong Cai ◽

Xinxin Cao ◽

Miao Chen ◽

...

Keyword(s):

Risk Factors ◽

Lymph Nodes ◽

Hodgkin Lymphoma ◽

Fever Of Unknown Origin ◽

Characteristic Curve ◽

Unknown Origin ◽

Final Diagnosis ◽

Diagnostic Value ◽

Discrimination Power ◽

Non Hodgkin Lymphoma

Background: We reviewed patients with fever of unknown origin (FUO) and splenomegaly and assessed the diagnostic value of splenectomy and measured risk factors suggestive of an underlying lymphoma. Methods: FUO patients (n = 83) who had splenomegaly and underwent splenectomy were enrolled into this retrospective single-center study. Clinical presentations were documented and risk factors suggestive of an underlying lymphoma were tested. Results: Seventy-four patients (89.2%) had a diagnosis of lymphoma or not after splenectomy and follow-up. Of those (55.4%) diagnosed with lymphoma, 29 had B-cell non-Hodgkin lymphoma and 12 had T-cell non-Hodgkin lymphoma. The remaining 33 (44.6%) had diseases other than lymphoma. Using multivariate logistic analysis, the following 3 independent risk factors were found to be related to a final diagnosis of lymphoma: age (continuous) (HR 1.086; 95% CI 1.033-1.141; p = 0.001), massively enlarged spleen (HR 7.797; 95% CI 1.267-47.959; p = 0.027), and enlarged intra-abdominal lymph nodes (HR 63.925; 95% CI 7.962-513.219; p < 0.001). The calibration of the model was satisfactory (p = 0.248 using the Hosmer-Lemeshow test), and the discrimination power was good (area under the receiver operating characteristic curve 0.925; 95% CI 0.863-0.987). Conclusions: Splenectomy is an effective diagnostic procedure for patients with FUO and splenomegaly and lymphoma is a common cause. Older age, a massively enlarged spleen, and enlarged intra-abdominal lymph nodes are risk factors suggesting an underlying lymphoma, and surgery for high-risk patients should be considered.

Download Full-text

A machine learning based prediction model of anti-PD-1 therapy response using noninvasive clinical information and blood markers of lung cancer patients.

Journal of Clinical Oncology ◽

10.1200/jco.2019.37.15_suppl.e14138 ◽

2019 ◽

Vol 37 (15_suppl) ◽

pp. e14138-e14138

Author(s):

Beung-Chul AHN ◽

Kyoung Ho Pyo ◽

Dongmin Jung ◽

Chun-Feng Xin ◽

Chang Gon Kim ◽

...

Keyword(s):

Machine Learning ◽

Flow Cytometry ◽

Supervised Learning ◽

Clinical Data ◽

Ridge Regression ◽

Predictive Score ◽

Support Vector ◽

Data Set ◽

Test Set ◽

Flow Cytometry Data

e14138 Background: Immune checkpoint inhibitors have become breakthrough therapy for various types of cancers. However, regarding their total response rate around 20% based on clinical trials, predicting accurate aPD-1 response for individual patient is unestablished. The presence of PD-L1 expression or tumor infiltrating lymphocyte may be used as indicators of response but are limited. We developed models using machine learning methods to predict the aPD-1 response. Methods: A total of 126 advanced NSCLC patients treated with the aPD-1 were enrolled. Their clinical characteristics, treatment outcomes, and adverse events were collected. Total clinical data (n = 126) consist of 15 variables were divided into two subsets, discovery set (n = 63) and test set (n = 63). Thirteen supervised learning algorithms including support vector machine and regularized regression (lasso, ridge, elastic net) were applied on discovery set for model development and on test set for validation. Each model were evaluated according to the ROC curve and cross-validation method. Same methods were used to the subset which had additional flow cytometry data (n = 40). Results: The median age was 64 and 69.8% were male. Adenocarcinoma was predominant (69.8%) and twenty patients (15.1%) were driver mutation positive. Clinical data set (n = 126) demonstrated that the Ridge regression (AUC: 0.79) was the best model for prediction. Of 15 clinical variables, tumor burden, age, ECOG PS and PD-L1, were most important based on the random forest algorithm. When we merged the clinical and flow cytometry data, the Ridge regression model (AUC:0.82) showed better performance compared to using clinical data only. Among 52 variables of merged set, the top most important immune markers were as follows: CD3+CD8+CD25+/Teff-CD28, CD3+CD8+CD25-/Teff-Ki-67, and CD3+CD8+CD25+/Teff-NY-ESO/Teff-PD-1, which indicate activated tumor specific T cell subset. Conclusions: Our machine learning based model has benefit for predicting aPD-1 responses. After further validation in independent patient cohort, the supervised learning based non-invasive predictive score can be established to predict aPD-1 response.

Download Full-text

Diagnostic Use of Serum Ferritin Levels to Differentiate Infectious and Noninfectious Diseases in Patients with Fever of Unknown Origin

Disease Markers ◽

10.1155/2013/915389 ◽

2013 ◽

Vol 34 (3) ◽

pp. 211-218 ◽

Cited By ~ 14

Author(s):

Seong Eun Kim ◽

Uh Jin Kim ◽

Mi Ok Jang ◽

Seung Ji Kang ◽

Hee Chang Jang ◽

...

Keyword(s):

Infectious Disease ◽

Infectious Diseases ◽

Serum Ferritin ◽

Fever Of Unknown Origin ◽

Inflammatory Diseases ◽

Unknown Origin ◽

University Hospital ◽

Disease Group ◽

Hematologic Disease ◽

Hematologic Diseases

INTRODUCTION: In this study, we determined whether serum ferritin levels could be used to differentiate between fever of unknown origin (FUO) caused by infectious and noninfectious diseases.METHODS: FUO patients were hospitalized at Chonnam National University Hospital between January, 2005 and December, 2011. According to the final diagnoses, five causes were identified, including infectious diseases, hematologic diseases, noninfectious inflammatory diseases, miscellaneous and undiagnosed.RESULTS: Of the 77 patients, 11 were caused by infectious diseases, 13 by hematologic diseases, 20 by noninfectious inflammatory diseases, 8 by miscellaneous diseases, and 25 were undiagnosed. The median serum ferritin levels in infectious diseases was lower than those in hematologic diseases and (median (interquartile range) of 282.4 (149.0–951.8) ng/mL for the infectious disease group, 1818.2 (485.4–4789.5) ng/mL for the hematologic disease group, and 563.7 (399.6–1927.2) ng/mL for the noninfectious inflammatory disease group,p= 0.048, Kruskal–Wallis test). By comparison using the Mann–Whitney test, statistically significant differences were found only between the infectious disease and hematologic disease groups (p= 0.049) and between the infectious disease and groups (p= 0.04).CONCLUSION: An optimal cutoff value of serum ferritin levels to predict FUO caused by a noninfectious disease (hematologic diseases, noninfectious inflammatory diseases) was established as 561 ng/mL.

Download Full-text

OP0293 USE OF WHOLE-BODY MAGNETIC RESONANCE TO IDENTIFY POTENTIAL DIAGNOSTIC CLUES IN CHILDREN WITH FEVER OF UNKNOWN ORIGIN (FUO)

Annals of the Rheumatic Diseases ◽

10.1136/annrheumdis-2020-eular.2640 ◽

2020 ◽

Vol 79 (Suppl 1) ◽

pp. 182-183

Author(s):

S. Signa ◽

R. Caorsi ◽

G. Stagnaro ◽

F. Minoia ◽

P. Picco ◽

...

Keyword(s):

Magnetic Resonance ◽

Fever Of Unknown Origin ◽

Unknown Origin ◽

Final Diagnosis ◽

Whole Body ◽

Recurrent Fever ◽

Eosinophilic Fasciitis ◽

Entire Body ◽

The Mean ◽

Diagnostic Clue

Background:Whole-body magnetic resonance imaging (WBMRI) is a fast and accurate method to detect diseases throughout the entire body without exposure to ionizing radiation. Possible emerging applications for this technique include rheumatologic field and evaluation of fever of unknown origin (FUO).Objectives:To evaluate the ability of WBMRI to identify significant potential diagnostic clue (PDC) in patients presenting a non specific inflammatory clinical picture.Methods:We retrospectively collected cases of pediatric patients followed in a single pediatric rheumatology center who underwent WBMRI between January 2010 and December 2015 for the following indications: i) FUO (temperature greater than 38.3°C for more than three weeks or failure to reach diagnosis after one week of investigations), iii) recurrent fever (febrile episodes separated by periods of normal temperature), iii) Inflammation of unknown origin, IUO (an illness of at least 3 weeks’ duration, with raised inflammatory markers and fever below 38.3°C).WBMRI studies were acquired with coronal and sagittal planes (slice thickness 5mm) with acquisition of several image sets with automatic direct image realignment after acquisition creating a whole-body scan.Sequences include short τ inversion recovery (STIR) and T1-weighted. All studies have been evaluated twice, the second time according to a predefined checklist, defined by an experienced radiologist, considering systematically single /multifocal bone lesion, bone marrow, joint effusion, soft tissues, adenopathies, parenchymal and vessels looking for PDC. We considered as a Potential Diagnostic Clue each alteration of the examined district that can potentially guide the diagnosis. Each alteration found is a PDC. We retrospectively evaluated patients’ clinical history and final diagnosis and we classified the PDCs identified during both first evaluation and re-evaluation as: Not useful (the identified PDC did not guide the diagnosis and is not coherent with the final diagnosis), consistent (the identified PDC is congruent with the patient’s final diagnosis) or diagnostic (the identification of the considered PDC strongly orient the final diagnosis).Results:We collected 104 patients who underwent WBMRI; 24 (23%) of them presenting FUO, 28 (27%) presenting recurrent fever and 52 (50%) presenting IUO. The mean age of onset symptoms was 6 years and nine months (range: 2 weeks old- 17 years and 6 months). The mean age of execution of WBMRI was 9 years (range: 5 months old- 19 years). After the whole diagnostic work-out a final diagnosis was achieved in 44 patients (42%).PDCs were identified at the first evaluation in 78/104 cases (75%). In 22 cases (21%) the identified PDCs were consistent with the diagnosis, whereas in 9 cases (8.5%) the identified PDCs were considered diagnostic. Globally we can consider that at first evaluation PDCs were somehow contributory to the diagnosis in 31 cases (30%; 6 JIA, 7 systemic infections, 5 monogenic inflammatory diseases, 4 ALPS, 2 Goldbloom’s Syndrome,2 Vasculitis,1 eosinophilic fasciitis, 1 hystiocytosis, 3 oncologic diagnosis).Blind re-evaluation of WBMRI allowed the identification of additional PDCs in 52 patients (12 of them previously negative). In 10 cases the PDC found after re- evaluation were consistent with the final diagnosis (2 JIA, one infectious disease, one neuroblastoma, 3 ALPS, 1 monogenic inflammatory disease, 1Takayasu arteritis, 1 Goldbloom’s syndrome).Conclusion:WBMRI can be a powerful diagnostic tool in patients with FUO. A predefined checklist increases sensitivity of WBMRI in the identification of PDC.Disclosure of Interests:Sara Signa: None declared, Roberta Caorsi: None declared, Giorgio Stagnaro: None declared, Francesca Minoia: None declared, Paolo Picco: None declared, Angelo Ravelli: None declared, Maria Beatrice Damasio: None declared, Marco Gattorno Consultant of: Sobi, Novartis, Speakers bureau: Sobi, Novartis

Download Full-text

Clinical Predictive Models for COVID-19: Systematic Study (Preprint)

10.2196/preprints.21439 ◽

2020 ◽

Author(s):

Patrick Schwab ◽

August DuMont Schütte ◽

Benedikt Dietz ◽

Stefan Bauer

Keyword(s):

Machine Learning ◽

Health Care ◽

Intensive Care ◽

Clinical Data ◽

Predictive Models ◽

Health Care Systems ◽

Gradient Boosting ◽

Support Vector ◽

Care Systems ◽

To Receive

BACKGROUND COVID-19 is a rapidly emerging respiratory disease caused by SARS-CoV-2. Due to the rapid human-to-human transmission of SARS-CoV-2, many health care systems are at risk of exceeding their health care capacities, in particular in terms of SARS-CoV-2 tests, hospital and intensive care unit (ICU) beds, and mechanical ventilators. Predictive algorithms could potentially ease the strain on health care systems by identifying those who are most likely to receive a positive SARS-CoV-2 test, be hospitalized, or admitted to the ICU. OBJECTIVE The aim of this study is to develop, study, and evaluate clinical predictive models that estimate, using machine learning and based on routinely collected clinical data, which patients are likely to receive a positive SARS-CoV-2 test or require hospitalization or intensive care. METHODS Using a systematic approach to model development and optimization, we trained and compared various types of machine learning models, including logistic regression, neural networks, support vector machines, random forests, and gradient boosting. To evaluate the developed models, we performed a retrospective evaluation on demographic, clinical, and blood analysis data from a cohort of 5644 patients. In addition, we determined which clinical features were predictive to what degree for each of the aforementioned clinical tasks using causal explanations. RESULTS Our experimental results indicate that our predictive models identified patients that test positive for SARS-CoV-2 a priori at a sensitivity of 75% (95% CI 67%-81%) and a specificity of 49% (95% CI 46%-51%), patients who are SARS-CoV-2 positive that require hospitalization with 0.92 area under the receiver operator characteristic curve (AUC; 95% CI 0.81-0.98), and patients who are SARS-CoV-2 positive that require critical care with 0.98 AUC (95% CI 0.95-1.00). CONCLUSIONS Our results indicate that predictive models trained on routinely collected clinical data could be used to predict clinical pathways for COVID-19 and, therefore, help inform care and prioritize resources.

Download Full-text

The effectiveness of using Multi-prognostic index (MPI) domain scores, clinical data and machine learning to improve 12-month mortality risk prediction in older hospitalized patients (Preprint)

10.2196/preprints.26139 ◽

2020 ◽

Author(s):

Richard Woodman ◽

Kimberley Bryant ◽

Michael J Sorich ◽

Alberto Pilotto ◽

Arduino Aleksander Mangoni

Keyword(s):

Machine Learning ◽

Logistic Regression ◽

Diagnostic Accuracy ◽

Clinical Features ◽

Clinical Data ◽

Prediction Accuracy ◽

Prognostic Index ◽

Receiver Operating Curve ◽

Support Vector ◽

Feature Sets

BACKGROUND The Multidimensional Prognostic Index (MPI) is an aggregate comprehensive geriatric assessment scoring system derived from eight domains, that predicts adverse outcomes including 12-month mortality (12MM). However, prediction accuracy, using the 3 MPI categories (mild, moderate, severe risk) as per previous investigations was relatively poor in a recent study with older hospitalized Australian patients. Prediction modelling using the component domains of the MPI together with additional clinical features and Machine Learning (ML) algorithms might improve prediction accuracy OBJECTIVE To assess whether prediction accuracy for 12MM using logistic regression with maximum likelihood estimation (LR-MLE) with the 3-category MPI together with age and gender (feature-set 1) can be improved with the addition of 10 clinical features (sodium, haemoglobin, albumin, creatinine, urea, urea/creatinine ratio, estimated glomerular filtration rate, C-reactive protein, body mass index and anticholinergic risk score) (feature-set 2), and the replacement of the 3-category MPI in feature-sets 1 and 2 by the eight separate MPI domains (feature-sets 3 and 4 respectively). To also assess prediction accuracy of ML algorithms using the same feature-sets. METHODS MPI and clinical features were collected in patients aged ≥65 years admitted to either General Medical or Acute Care of the Elderly wards of a South Australian hospital between September 2015 and February 2017. The diagnostic accuracy of LR-MLE was assessed together with nine ML algorithms: decision-trees, random-forests, eXtreme gradient-boosting (XGBoost), support-vector-machines, naïve-bayes, k-nearest-neighbours, ridge regression, logistic regression without regularisation and neural-networks. A 70:30 Training:Test split of the data and a grid-search of hyper-parameters with 10-fold cross-validation was employed during model training of the ML algorithms. Area-under-curve (AUC) was used to assess prediction accuracy. RESULTS A total of 737 patients (F:M=50.2%/49.8%) with median (IQR) age 80 (72-86) years had complete MPI data recorded on admission and complete 12-month follow-up obtained. The area-under-the receiver-operating-curve (AUC) for LR-MLE was 0.632, 0.688, 0.738 and 0.757 for feature-sets 1 to 4 respectively. The best overall accuracy for the nine ML algorithms was obtained using the XGBoost algorithm (0.635, 0.706, 0.756 and 0.757 for feature-sets 1 to 4 respectively). CONCLUSIONS The use of MPI domains (feature-sets 3 and 4) with LR-MLE considerably improved prediction accuracy compared to that obtained using the traditional 3-category MPI. The XGBoost ML algorithm slightly improved accuracy compared to LR-MLE with feature-sets 1-3 but not with feature-set 4. Adding clinical data also provided small gains in accuracy for LR-MLE and some, but not all ML algorithms. Consideration should be given to using the underlying MPI domains of aggregate scoring systems, additional clinical data and ML algorithms when assessing the risk of 12MM. CLINICALTRIAL N/AMachine learning, Multidimensional Prognostic Index, mortality, diagnostic accuracy, XGBoost

Download Full-text

The value of 18F-FDG-PET/CT in identifying the cause of fever of unknown origin (FUO) and inflammation of unknown origin (IUO): data from a prospective study

Annals of the Rheumatic Diseases ◽

10.1136/annrheumdis-2017-211687 ◽

2017 ◽

Vol 77 (1) ◽

pp. 70-77 ◽

Cited By ~ 33

Author(s):

Verena Schönau ◽

Kristin Vogel ◽

Matthias Englbrecht ◽

Jochen Wacker ◽

Daniela Schmidt ◽

...

Keyword(s):

Prospective Study ◽

Fdg Pet ◽

Fever Of Unknown Origin ◽

Correct Diagnosis ◽

Unknown Origin ◽

Ct Scanning ◽

Final Diagnosis ◽

Intermittent Fever ◽

Pet Ct ◽

Fdg Pet Ct

BackgroundFever of unknown origin (FUO) and inflammation of unknown origin (IUO) are diagnostically challenging conditions. Diagnosis of underlying disease may be improved by 18F-fluorodesoxyglucose positron emission tomography (18F-FDG-PET).MethodsProspective study to test diagnostic utility of 18F-FDG-PET/CT in a large cohort of patients with FUO or IUO and to define parameters that increase the likelihood of diagnostic 18F-FDG-PET/CT. Patients with FUO or IUO received 18F-FDG-PET/CT scanning in addition to standard diagnostic work-up. 18F-FDG-PET/CT results were classified as helpful or non-helpful in establishing final diagnosis. Binary logistic regression was used to identify clinical parameters associated with a diagnostic 18F-FDG-PET/CT.Results240 patients were enrolled, 72 with FUO, 142 with IUO and 26 had FUO or IUO previously (exFUO/IUO). Diagnosis was established in 190 patients (79.2%). The leading diagnoses were adult-onset Still’s disease (15.3%) in the FUO group, large vessel vasculitis (21.1%) and polymyalgia rheumatica (18.3%) in the IUO group and IgG4-related disease (15.4%) in the exFUO/IUO group. In 136 patients (56.7% of all patients and 71.6% of patients with a diagnosis), 18F-FDG-PET/CT was positive and helpful in finding the diagnosis. Predictive markers for a diagnostic 18F-FDG-PET/CT were age over 50 years (p=0.019), C-reactive protein (CRP) level over 30 mg/L (p=0.002) and absence of fever (p=0.001).Conclusion18F-FDG-PET/CT scanning is helpful in ascertaining the correct diagnosis in more than 50% of the cases presenting with FUO and IUO. Absence of intermittent fever, higher age and elevated CRP level increase the likelihood for a diagnostic 18F-FDG-PET/CT.

Download Full-text