European Heart Journal - Cardiovascular Imaging (2022) 23, 1260–1271 ORIGINAL PAPER https://doi.org/10.1093/ehjci/jeab279 The valve uptake index: improving assessment of prosthetic valve endocarditis and updating [18F]FDG PET/CT(A) imaging criteria Albert Roque 1,2,3,4,*, Marıa N. Pizzi 3,4,5,6, Nuria Fernández-Hidalgo 3,4,7, Guillermo Romero-Farina 4,5, Gemma Burcet1,2,3,4, Jose  Luis Reyes-Juarez1,2,3,4, Carina Espinet2,3,8, Joan Castell-Conesa2,3,8, Manuel Escobar1,2, Ignacio Ferreira-González3,5,9, Santiago Aguad e-Bruix3,4,6,8, and Hug Cuellar-Calabria1,2,3,4 1 Department of Radiology, Hospital Universitari Vall d’Hebron, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; 2IDI (Institut de Diagnòstic per la Imatge), Barcelona, Spain; 3 Departament de Medicina, Universitat Autònoma de Barcelona, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; 4Vall d’Hebron Institut de Recerca (VHIR), Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; 5Department of Cardiology, Hospital Universitari Vall d’Hebron, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; 6Centro de Investigación Biomedica en Red en Enfermedades Cardiovasculares (CIBERCV), Madrid, Spain; 7Department of Infectious Diseases, Hospital Universitari Vall d’Hebron, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; 8Department of Nuclear Medicine, Hospital Universitari Vall d’Hebron, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain; and 9CIBER de Epidemiologıa y Salud Pública (CIBERESP), Madrid, Spain Received 23 September 2021; editorial decision 10 December 2021; accepted 13 December 2021; online publish-ahead-of-print 6 January 2022 Aims Diagnosis of prosthetic valve endocarditis (PVE) by positron emission computed tomography angiography (PET/CTA) is based on visual and quantitative morpho-metabolic features. However, the fluorodeoxyglucose (FDG) uptake pat- tern can be sometimes visually unclear and susceptible to subjectivity. This study aimed to validate a new parameter, the valve uptake index [VUI, maximum standardized uptake value (SUVmax)-mean standardized uptake value (SUVmean)/SUVmax], designed to provide a more objective indication of the distribution of metabolic activity. Secondly, to re-evaluate the utility of traditionally used PVE imaging criteria and determine the potential value of add- ing the VUI in the diagnostic algorithm of PVE. ................................................................................................................................................................................................... Methods Retrospective analysis of 122 patients (135 prosthetic valves) admitted for suspicion of endocarditis, with a conclu- and results sive diagnosis of definite (N = 57) or rejected (N = 65) PVE, and who had undergone a cardiac PET/CTA scan as part of the diagnostic evaluation. We measured the VUI and recorded the SUVmax, SUVratio, uptake pattern, and the presence of endocarditis-related anatomic lesions. The VUI, SUVmax, and SUVratio values were 0.54 ± 0.1 vs. 0.36 ± 0.08, 7.68 ± 3.07 vs. 3.72 ± 1.11, and 4.28 ± 1.93 vs. 2.16 ± 0.95 in the ‘definite’ PVE group vs. the ‘rejected’ group, respectively (mean ± SD; P < 0.001). A cut-off value of VUI > 0.45 showed a sensitivity, specificity, and diag- nostic accuracy for PVE of 85%, 88%, and 86.7% and increased diagnostic ability for confirming endocarditis when combined with the standard diagnostic criteria. ................................................................................................................................................................................................... Conclusions The VUI demonstrated good diagnostic accuracy for PVE, even increasing the diagnostic power of the traditionally used morphometabolic parameters, which also confirmed their own diagnostic performance. More research is needed to assess whether the integration of the VUI into the PVE diagnostic algorithm may clarify doubtful cases and thus improve the diagnostic yield of PET/CTA. 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 *Corresponding author. Tel: þ34 932746029; Fax: þ34 932746764. E-mail: albert_roque@hotmail.com C The Author(s) 2022. Published by Oxford University Press on behalf of the European Society of Cardiology. V This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com The valve uptake index: a new PVE parameter by PET/CTA 1261 Graphical Abstract ........................................................................................................................................................................................... Keywords prosthetic heart valves • positron emission tomography • infective endocarditis • nuclear imaging .. Introduction .. .. reference parameter providing a more objective indication of the dis- tribution of metabolic activity, which is not well reflected by standar- .. Positron emission computed tomography with 18F-fluorodeoxyglu- .. dized uptake values (SUVs).9 We recently developed a new measure, cose ([18F]FDG PET/CT) associated with ECG-gated cardiac CT .. the ‘valve uptake index’ (VUI), first evaluated in a cohort of patients .. angiography (PET/CTA) has significantly improved the diagnostic .. without suspected infection who underwent serial cardiac PET/CTA yield of prosthetic valve endocarditis (PVE) when combined with the .. examinations after PV implantation surgery.5 The VUI provided infor- .. Duke criteria.1,2 Hence, PET/CT findings have been incorporated in .. mation about the degree of homogeneity of the metabolic activity, as the 2015 ESC and in the 2020 ACC/AHA guidelines for the diagnosis .. it attenuates the effect of high SUV values when uptake is diffuse. The .. of infective endocarditis (IE).3,4 It is known that FDG uptake can be .. main aim of this study was to validate the utility of this index in often present in non-infected prosthetic valves (PVs), especially in .. .. patients with suspicion of PVE and establish cut-off values to deter- the recent postoperative period, with the potential risk of misinter- .. mine when a PV shows normal uptake (low probability of infection) .. preting reactive inflammation as infection.5 Nevertheless, FDG up- .. or abnormal uptake (high probability of PVE). In addition, we re-eval- take in non-infected PVs is usually diffuse, homogeneous and of mild .. uated the traditionally used PVE imaging findings to establish the PET/ .. intensity, and differs from the focal/heterogeneous uptake normally .. CTA imaging criteria for infection. seen in PVE. The combination of a considered-normal uptake pattern .. .. with the absence of anatomic lesions on CTA (the two proposed .. ‘major criteria of normality’) are the cornerstones of PV evaluation .. Methods .. by PET/CTA, and usually allows postoperative inflammation to be .. .. We performed a retrospective analysis of 135 patients with suspicion of distinguished from active infection.5–8 Thus, if PET/CT is not com- .. .. PVE admitted to our IE-referral hospital between February 2013 and bined with cardiac CTA, the uptake pattern would remain as the .. February 2020, and who underwent cardiac PET/CTA as part of the diag- main criterion to classify a PET/CT scan as positive or negative for IE. .. nostic work-up. In order to determine how the VUI correlated with the Since assessment of the FDG distribution is basically visual, it may be .. .. visual pattern, only patients with a clearly classifiable uptake pattern were open to subjectivity and occasionally unclear, hampering its classifica- .. included. Besides, only patients with a final conclusive diagnosis of ‘defin- tion as homogeneous or heterogeneous pattern. Therefore, in add- .. ite’ (N = 57) or ‘rejected’ (N = 65) PVE by the Expert Team after a 6- .. ition to qualitative assessment, it would be desirable to have a . month follow-up were analysed, while inconclusive cases (‘possible’ PVE, 1262 A. Roque et al. .. 13) were excluded. In patients with double mitroaortic PV replacement .. classified as: ‘absent’, ‘focal/multifocal’, or ‘diffuse’, with the latter (n = 17), the analysis was performed in both prostheses in the ‘rejected’ .. being further sub-classified into ‘homogeneous’ or ‘heterogeneous’, as .. previously defined.9,10 Focal/multifocal and diffuse-heterogeneous dis- group (10/17), in both prostheses in the ‘definite’ group with mitroaortic .. endocarditis (3/17), but only in the involved PV in the remaining cases (4/ .. tributions are considered ‘abnormal patterns’ suggestive of infection, 17). Finally, a total of 75 rejected and 60 definite PVs were analysed (en- .. against absent and diffuse-homogenous uptakes that are considered .. rolment flowchart is shown in Figure 1). .. ‘normal patterns’. CTA images were systematically evaluated to de- The study protocol was approved by the hospital ethics committee, .. tect IE-related anatomic lesions (vegetations, abscesses, pseudoaneur- .. and written informed consent was obtained from all participants. .. ysms, and fistulae). .. .. PET/CTA technical considerations and .. Quantitative analysis: uptake intensity and VUI .. Quantification was made on a circular region of interest (ROI) manually image interpretation .. 18 Patients underwent specific cardiac [ F]FDG PET/CTA including myo- .. adjusted to encompass the PV, and drawn on a 10 mm thick maximum in- .. tensity projection (MIP) transverse view of the valve plane in the cardial FDG suppression measures and the combination of both metabol- .. morpho-metabolic fusion images. Two sets of metabolic acquisitions ic (PET) and anatomic (CTA) ECG-gated cardiac acquisitions, as .. previously described.10–12 Fused PET/CTA images were evaluated by .. were used for the fusion: the standard AC images commonly used in .. PET/CT for other purposes, and a localized ECG-gated cardiac bed (C- two experienced cardiac imaging specialists, a cardiac radiologist and a .. nuclear cardiologist. Double-oblique planes were used to obtain trans- .. Bed) to improve evaluation of cardiac structures. The standard maximum .. and mean standardized uptake values (SUVmax and SUVmean) were col- verse views of the PVs. .. lected. The target-to-background ratio (SUVratio) was calculated as the .. .. ratio between the SUVmax of the valve and the SUVmean of the blood Visual (non-quantitative) analysis: metabolic pattern and .. pool in the left atrium. Finally, we obtained the valve uptake index (VUI), .. anatomic lesions .. first described in our previous control group5 and calculated as follows: The prosthetic/periprosthetic [18F]FDG uptake pattern was visually .. (SUVmax-SUVmean)/SUVmax (Figure 2). As the images most commonly analysed in attenuation correction (AC), localized ECG-gated cardiac .. used worldwide are those of AC, the statistical analysis was performed .. (C-Bed), and non-attenuation correction (NAC) images, and was . with the values obtained in these reconstructions, with the aim of making Figure 1 Study flowchart. *From our tertiary referral hospital for infective endocarditis with an estimated volume of 60 definite endocarditis per year, of which approximately 25 definite PVE per year. PV, prosthetic heart valve; PVE, prosthetic valve endocarditis. The valve uptake index: a new PVE parameter by PET/CTA 1263 Figure 2 PET/CTA systematic evaluation. Visual (non-quantitative) and quantitative analysis of an aortic mechanical prosthetic valve. (A, top and bottom) Multiplanar reformation of PET/CTA fusion images to obtain a transverse view of the prosthesis. A diffuse and homogeneous FDG uptake pattern is seen in both fusion (B) and attenuation-corrected PET images (D). (C) Cardiac CTA shows no associated anatomic lesions. (B) A circular re- gion of interest drawn on the valve plane allows the SUVmax and SUVmean to be obtained, from which the VUI is subsequently calculated (formula). FDG, fluorodeoxyglucose; PET/CTA, positron emission computed tomography angiography; SUVmax and SUVmean, maximum and mean standar- dized uptake value; VUI, valve uptake index. .. them more exportable. The intra-class correlation coefficient agreement .. Results between the two observers for VUI measurements was computed. .. .. Details on quantification procedure are supplied in Supplementary data .. Study population and patient online, Material I and Videos. .. .. characteristics .. A total of 122 patients under suspicion of PVE who underwent a car- Statistical analysis .. All continuous data are expressed as mean (standard deviation), and all .. diac PET/CTA, with a classifiable FDG uptake pattern and a final con- .. non-continuous variables are expressed as percentages. Continuous vari- .. clusive diagnosis of ‘confirmed or rejected’ IE were included in the ables were compared using the t-test for unpaired samples. Differences .. analysis. Of the 122 patients under suspicion of PVE, 60 prostheses .. between proportions were compared using the X2 test. .. were analysed in the ‘definite’ group (aortic-mitral/biological–mech- First, we evaluated the diagnostic accuracy of a VUI > 0.45 for PVE, a .. anical: 43-17/41-19) and 75 in the ‘rejected’ group (51-24/33-42). .. value that we previously described to be the mean þ 1SD VUI (0.35 ± 0.1) in non-infected prosthetic heart valves.5 ... The most frequent pathogens identified were coagulase-negative .. staphylococci (17/57, 29.8%, mainly Staphylococcus epidermidis), and Then we analysed both the qualitative and quantitative variables of the .. PET/CTA images routinely used for the diagnosis of IE. Multiple logistic .. Enterococcus faecalis (14/57, 24.6%), followed by Streptococcus spp. .. (10/57, 17.5%). The included patients, PV type and location are regression models were analysed to predict PVE. Forward stepwise se- .. lection techniques (FSTEP) were used to identify which variables were in- .. shown in Figure 1 and are described in Table 1. .. dependently associated with IE. The threshold for variable entry into all .. multivariable models was a P-value <0.05 and that for variable removal .. Visual (non-quantitative) analysis: was a P-value >0.10. All models were adjusted for age and PV type, and .. .. metabolic pattern and anatomic lesions the area under the curve (AUC) was calculated. Models were considered .. reasonable when the AUC was >0.7 and strong when it exceeded 0.8. .. Prosthetic/periprosthetic FDG uptake was visually detected in all 60 .. (100%) ‘definite’ PVE, but in only 26/75 (35%) of the ‘rejected’ group. A value of P < 0.05 was considered as indicative of statistical signifi- .. cance. All statistical tests were two-sided. The data were analysed using .. The FDG uptake pattern was ‘focal/multifocal’ in 34/60 (57%), ‘dif- Stata 15 (College Station, TX, USA). .. fuse-heterogeneous’ in 24/60 (40%) and ‘diffuse-homogeneous’ in 1264 A. Roque et al. Table 1 Patient characteristics, study parameters, and morpho-metabolic findings According to final diagnosis (n 5 122 patients) Definite PVE (n 5 57) Rejected PVE (n 5 65) .................................................................................................................................................................................................................... Clinical and study characteristics Age (years), mean ± SD 73 ± 5 69 ± 14 Men, n (%) 34 (60) 41 (63) Symptoms highly suggestive of PVE (2015-ESC diagnostic criteriaa), n (%) 57 (100) 55 (85) Positive microbiological findings at completion of follow-up,b n (%) 57 (100) 39 (60) Time from symptom onset-PET/CTA (days), median (IQR)c 19 (10–41) 19 (11–37) Days of antibiotics-PET/CTA, mean ± SD 11 ± 12 9±8 [18F]FDG myocardial suppression, adequate,d n (%) 56 (98) 50 (77) .................................................................................................................................................................................................................... According to final diagnosis (n 5 135 prosthesis) (n 5 60) (n 5 75) .................................................................................................................................................................................................................... Position and type of prosthesis Aortic (n = 94) 43 51 Mitral (n = 41) 17 24 Biological (n = 74) 41 33 Mechanical (n = 61) 19 42 Time from surgery to PET/CTA (months), median (IQR) 35 (8–81) 49 (5–134) <1 year, n (%) 23 (38) 27 (36) >_1 year, n (%) 37 (62) 48 (64) .................................................................................................................................................................................................................... Visual (Non-quantitative) analysis .................................................................................................................................................................................................................... Metabolic pattern ([18F]FDG distribution), n (%) Any visually detected activity 60 (100) 26 (35) Absence of uptake 0 (0) 49 (65) Focal or multifocal uptake 34 (57) 2 (3) Diffuse heterogeneous uptake 24 (40) 1 (1) Diffuse homogeneous uptake 2 (3) 23 (31) Anatomic lesions, n (%) Any IE-related lesion 39 (65) 9 (12) Images consistent with vegetations 19/39 (49) 2/9 (22) Abscesses 26/39 (67) 0 (0) Pseudoaneurysms 13/39 (33) 3/9 (33) Fistulae 3/39 (8) 4/9 (44) Number of lesions per patients (when present) 1 lesion 19 (49) 7 (9) 2 lesions 15 (38) 1 (1) 3 lesions 5 (13) 0 Distant/peripheral lesions (embolic or metastatic lesions), n (%) 17/57 (30) 3/65 (5) .................................................................................................................................................................................................................... Quantitative analysis [18F] FDG uptake intensity and VUI, mean ± SD .................................................................................................................................................................................................................... SUVmax AC images 7.68 ± 3.07 3.72 ± 1.11 SUVmax C-Bede images 7.95 ± 3.4 3.64 ± 1.28 SUVmean AC images 3.29 ± 0.94 2.38 ± 0.7 SUVmean C-Bed images 3.5 ± 1.12 2.41 ± 0.82 SUVratiof AC images 4.28 ± 1.93 2.16 ± 0.95 SUVratio C-Bed images 4.73 ± 2.29 2.3 ± 1.03 VUIg AC images 0.54 ± 0.1 0.36 ± 0.08 VUI C-Bed images 0.53 ± 0.1 0.34 ± 0.07 .................................................................................................................................................................................................................... According to prosthesis position and type Aortic Mitral Aortic Mitral .................................................................................................................................................................................................................... SUVmax AC, mean ± SD Biological 8.19 ± 3.24 9.97 ± 2.42 3.7 ± 0.86 4.3 ± 0.25 Mechanical 6.11 ± 2.59 5.77 ± 1.37 3.57 ± 1.22 3.77 ± 1.24 SUVratio AC, mean ± SD Biological 4.39 ± 1.78 5.27 ± 2.22 2.11 ± 1.03 2.41 ± 0.53 Mechanical 3.56 ± 1.9 3.88 ± 2.18 2.36 ± 1.1 2.2 ± 0.86 Continued The valve uptake index: a new PVE parameter by PET/CTA 1265 .................................................................................................................................................................................................................... Table 1 Continued According to prosthesis position and type Aortic Mitral Aortic Mitral .................................................................................................................................................................................................................... VUI AC, mean ± SD Biological 0.54 ± 0.09 0.64 ± 0.1 0.31 ± 0.06 0.4 ± 0.07 Mechanical 0.50 ± 0.11 0.54 ± 0.08 0.35 ± 0.07 0.41 ± 0.07 a 2015-ESC diagnostic criteria,3 considered symptoms (temperature >38 C, heart failure, peripheral symptoms, and septic shock). b Positive blood cultures at admission or during clinical course or positive cultures of surgical material. c 25th–75th interquartile. d Adequate myocardial suppression was defined as complete suppression or partial suppression that did not interfere with PV evaluation. e Cardiac C-Bed. f Prosthesis-to-background (blood pool) SUV ratio. g AC, attenuation correction; IE, infective endocarditis; PET/CTA, positron emission computed tomography angiography; SUVmax, maximum standardized uptake value; [18F]FDG, 18F-fluorodeoxyglucose; VUI, valve uptake index. 2/60 (3%) in the definite group, and was absent in 49/75 (65%), ‘dif- .. specificity and 0.79 AUC (CI 0.72–0.85) for the diagnosis of PVE .. fuse-homogeneous’ in 23/75 (31%), ‘focal/multifocal’ in 2/75 (3%), .. (Figure 4A). Conversely, lesions were identified in only 8 out of 75 and ‘diffuse-heterogeneous’ in 1/75 (1%) in the ‘rejected’ group. All .. .. (11%) of the rejected cases: 7 patients (9%) showed 1 lesion and 1 pa- cases of positive FDG uptake visualized in the AC and C-Bed images .. tient (1%) showed 2 lesions, with fistula being the most frequent. were also identified in NAC images. .. .. A description of the FDG uptake patterns and type and frequency In the multiple logistic regression analysis (included variables: focal, .. of anatomic lesions for each group and according to valve position .. heterogeneous and homogeneous patterns, and anatomic lesions), focal .. and type are shown in Table 1. pattern (X2: 37.445; P < 0.001; S 56.7%, E 97.3%, AUC 0.79) was an in- .. .. dependent predictor for PVE followed by the heterogeneous pattern .. Quantitative analysis: uptake intensity (X2: 28.222; P < 0.001; S 40%, E 98.7%, AUC 0.73); adjusted by age and .. .. and VUI PV type. Although in clinical practice, a prosthesis can only display one .. pattern per episode (Supplementary data online, Figure S1), both the .. Mean ± SD uptake values (AC images) for the ‘definite’ PVE group .. and the ‘rejected’ group were: SUVmax 7.68 ± 3.07 vs. 3.72 ± 1.11, focal/multifocal and the heterogeneous patterns are ‘a priori’ ‘abnormal’ .. and, when evaluated as a group, its high diagnostic power (AUC 0.96) .. SUVmean 3.29 ± 0.94 vs. 2.38 ± 0.7, SUVratio 4.28 ± 1.93 vs. .. 2.16 ± 0.95, and VUI 0.54 ± 0.1 vs. 0.36 ± 0.08. These differences confirms the relevance of the visual criteria (Table 2 and Figure 3). .. Cardiac CTA images showed IE-related lesions in 39 out of 60 .. were all statistically significant (P < 0.001). Values calculated in the C- .. (65%) definite PVE cases, of which 19 (49%), 15 (38%), and 5 (13%) ... Bed images are summarized in Table 1. No differences were found in patients showed 1, 2, or 3 anatomic lesions, respectively, with peri- .. uptake values related to valve type (mechanical or biological; P = 0.3) prosthetic abscess being the most frequently identified lesion. The .. or position (aortic or mitral; P = 0.21) in either in the ‘definite’ or the .. presence of these anatomic lesions showed 65% sensitivity, 90.7% . ‘rejected’ group. There was an excellent absolute agreement Table 2 Results of diagnostic ability of each quantitative and qualitative parameter analysed and with the addition of the VUI Sensitivity (%,CI) Specificity (%,CI) Accuracy (%,CI) .................................................................................................................................................................................................................... Qualitative parameters Focal pattern 56.7 (44.1–68.4) 97.3 (90.8–99.3) 79.3 (71.7–85.2) Heterogeneous pattern 40 (28.6–52.6) 98.7 (92.8–99.8) 72.6 (64.5–79.4) Abnormal patterns (focal OR heterogeneous uptake) 96.7 (88.6–99.1) 96 (88.9– 98.6) 96.3 (91.6–98.4) Anatomic lesions 65 (52.4–75.8) 90.7 (81.9–95.4) 79.3 (71.6–85.2) Quantitative parameters (attenuation correction images, AC) SUVmax 83.3 (71.9–90.7) 88 (78.7–93.6) 85.9 (79.1–90.8) SUVmax >5.28 84.3 (74.5–90.9) 83.3 (74.2–88.9) 84.1 (77.6–88.6) SUVratio 80 (68.2–88.2) 90.7 (81.9–95.4) 85.9 (79.1–90.8) SUVratio >2.58 84 (72.2–90.4) 77.2 (67.6–84.1) 79.7 (72.6–84.9) VUI 83.3 (71.9–90.7) 93.3 (85.3–97.1) 88.9 (82.5–93.2) VUI >0.45 85 (73.9–91.9) 88 (78.7–93.6) 86.7 (79.9–91.4) Qualitative parameters 1 VUI >0.45 Focal pattern þ VUI >0.45 95 (86.3–98.3) 86.7 (77.2–92.6) 90.4 (84.2–94.3) Heterogeneous pattern þ VUI >0.45 86.7 (75.8–93.1) 87 (77.2–92.6) 86.7 (79.9–91.4) Anatomic lesions þ VUI >0.45 85 (73.9–91.9) 88 (78.7–93.6) 86.7 (80–91) All values adjusted by age and prosthesis type. SUVmax, maximum standardized uptake value; VUI, valve uptake index. 1266 A. Roque et al. Figure 3 Receiver–operator characteristics (ROC) curves showing the PVE diagnostic accuracy of the focal and the heterogeneous pattern indi- vidually (red and green curves, respectively) and the focal OR heterogeneous patterns considered as a group of abnormal FDG uptake (blue curve). FDG, fluorodeoxyglucose; PVE, prosthetic valve endocarditis. Figure 4 (A) Receiver–operator characteristics (ROC) curves showing the PVE diagnostic accuracy of the VUI and VUI > 0.45 cut-off value (upper row). Added value of the VUI to the qualitative parameters (bottom row). (B) Mean ± SD VUI values according to the FDG uptake pattern and distri- bution of rejected and definite cases. Focal and heterogeneous patterns correlate with significantly higher VUI values in comparison with absent and homogenous patterns. FDG, fluorodeoxyglucose; PVE, prosthetic valve endocarditis; VUI, valve uptake index. The valve uptake index: a new PVE parameter by PET/CTA 1267 .. between the observers, using the two-way random effect models .. those with >1 lesions, values were 7.72 ± 2.74/4.56 ± 1.82/ and ‘single rater’ unit, kappa = 0.99, P < 0.001. .. 0.56 ± 0.09 vs. 9.45 ± 3.23/5.32 ± 2.22/0.60 ± 0.08). .. In the multiple logistic regression analysis (included variables: .. There were no differences in SUVmax and VUI values among SUVmax, SUVratio, and VUI), VUI (X2: 33.411; P < 0.001) was an in- .. ‘rejected’ cases (n = 75) with (n = 7) or without (n = 68) anatomic .. dependent predictor for PVE adjusted by age and PV type, yielding a .. lesions (3.68 ± 1.1 vs. 3.73 ± 1.1; P = 0.758; and 0.37 ± 0.08 vs. diagnostic sensitivity, specificity and accuracy (95% CI) of 83% (71.9– .. 0.34 ± 0.07; P = 0.307, respectively). Fistula/leak was the most fre- .. 90.7%), 93.3% (85.3–97.1%), and 88.9% (82.5–93.2%), respectively. A .. quent lesion observed in the ‘rejected’ group. Surprisingly, most of cut-off value of VUI > 0.45 improved PVE diagnostic yield, with 85% .. them were found in PV without visually detectable FDG uptake (ab- .. (73.9–91.9%) sensitivity, 88% (78.7–93.6%) specificity, and 86.7% .. sent pattern), indicating that this association may be poorly suggestive .. of infection. (79.9–91.4%) diagnostic accuracy. SUVmax and SUVratio also .. showed good global diagnostic ability (complete results in Table 2, .. .. Supplementary data online, Material II and Figure 4A). .. .. Discussion .. Incremental diagnostic power of .. The results of this study show that patients with ‘definite’ PVE had sig- .. combined qualitative and quantitative .. nificantly higher ‘valve uptake index’ (VUI) values (0.54 ± 0.1) than features .. those with rejected PVE (0.36 ± 0.08) (P < 0.001). A VUI cut-off value .. Mean ± SD VUI values according to the FDG uptake pattern and final .. of >0.45 yielded a diagnostic accuracy of 86.7% to confirm PVE (85% .. sensitivity and 88% specificity). Currently used PET/CTA qualitative diagnosis are shown in Figure 4B. Only in a few cases the VUI slightly .. mismatched with the final diagnosis and are explained in .. (uptake pattern and anatomic lesions) and quantitative (SUVmax and .. Supplementary data online, Material III. The combination of the cur- .. SUVratio) parameters also confirmed their utility for the diagnosis of rently used PET/CTA diagnostic criteria for PVE with the new pro- .. infection. A higher PVE diagnostic accuracy was achieved by addition .. posed measurement (VUI > 0.45) significantly increased the .. of the VUI to the qualitative parameters. diagnostic ability of this technique for confirming endocarditis. The .. The VUI correlated with the visual metabolic pattern of FDG up- .. addition of a VUI > 0.45 to a focal pattern increased the diagnostic ac- .. take. Low VUI values were associated with no visually detectable .. metabolic activity or with a diffuse-homogeneous uptake (considered curacy (CI) from 79.3% (71.7–85.2%) to 90.4% (84.2–94.4%), and the .. addition of a VUI > 0.45 to a heterogeneous pattern increased the ac- .. normal patterns). A high VUI was associated with focal/multifocal or .. diffuse-heterogeneous uptake (highly suggestive of infection) curacy from 72.6% (64.5–79.4%) to 86.7% (79.9–91.4%). Finally, ana- .. tomic complications had a diagnostic accuracy for PVE of 79.3% .. (Figures 4B and 5). Therefore, the VUI constitutes a numerical repre- .. sentation of the distribution of metabolic activity, and can help to as- (71.7–85.2%), which increased to 86.7% (79–91%) when adding a .. VUI > 0.45 (Figure 4A and Table 2). .. sess whether FDG uptake is homogeneous or heterogeneous, which .. is a key feature when evaluating the probability that PV metabolic ac- .. Metabolic findings according to the time .. tivity corresponds or not to infection. Visual interpretation of PET/ .. from surgery to PET/CTA .. CTA images is based on subjective evaluation, and some cases of dif- .. fuse uptake may be difficult to interpret. This becomes especially evi- Twenty-three (38%)/27 (36%) of the prostheses were implanted be- .. fore 1 year and 37 (62%)/48 (64%) after 1 year in the ‘definite’ and .. dent when uptake is very intense, which has been described as .. normal in some PV types,13 and metabolic activity may not be com- the ‘rejected’ groups, respectively. There were no significant differen- .. ces between any of the quantitative values before and 1 year after .. pletely expressed by the SUVmax or SUVratio, which express more .. the intensity of the uptake. The VUI attenuates the effect of high SUV implantation in either the ‘definite’ or ‘rejected’ cases (mean ± SD, .. .. values when uptake tends to be diffuse, reinforcing the visual inter- _1 year after implantation): SUVmax 8.21 ± 3.11/7.36 ± 3.04, .. pretation of homogeneous, and probably not pathological, metabolic SUVmean 3.31 ± 0.9/3.04 ± 1.01, SUVratio 4.23 ± 1.7/4.42 ± 2.46, and .. .. activity. In addition, an association was observed between the degree VUI 0.54 ± 0.08/0.54 ± 0.14 for the ‘definite’ group; SUVmax .. of FDG uptake and the presence of IE-related anatomic lesions. 3.94 ± 0.95/3.61 ± 1.18, SUVmean 2.52 ± 0.62/2.3 ± 0.72, SUVratio .. .. Higher VUI values correlated with a higher frequency of anatomic 2.41 ± 1.2/2.01 ± 0.72 and VUI 0.36 ± 0.07/0.35 ± 0.08 for the .. lesions in general, and with a higher frequency of periprosthetic ‘rejected’ group. Of note, FDG uptake was visually detectable in only .. .. complications. 33% (16/48) of ‘rejected’ cases when PV was implanted >_1 year. .. .. The VUI can help to finally determine the pattern of FDG uptake Nevertheless, the uptake pattern in these cases was diffuse/homoge- .. in doubtful cases, compensating for the effect of extreme values of neous in 81% (13/16). .. .. SUV—which can be subject to variability—on visual evaluation. In .. addition, the VUI is correlated with the presence of anatomical Correlation between metabolic activity .. .. lesions and periprosthetic complications. Therefore, ‘the VUI integra- and the presence of anatomic lesions .. tes both qualitative and quantitative information into a single numer- .. Patients with ‘definite’ PVE and no anatomic lesions (n = 21) had a .. ical value’ and would add consistency to the already established lower mean SUVmax of 5.96 ± 2.21 (P < 0.0001), SUVratio .. imaging criteria, increasing the diagnostic yield of PET/CTA. In add- .. 3.04 ± 0.72 (P < 0.001) and VUI 0.47 ± 0.09 (P < 0.001) than patients .. ition, given the well-known problem of the lack of standardization of with at least one lesion (n = 39): 8.61 ± 3.09, 4.95 ± 2.04, and .. semi-quantitative measures, the VUI could be a parameter less influ- .. 0.58 ± 0.08, respectively (if divided in patients with only 1 lesion and . enced by technical differences, equipment or protocols, which can 1268 A. Roque et al. Figure 5 FDG uptake distribution patterns (visual assessment) and their correlation with VUI values. FDG uptake in infected prostheses (PVE, left panel), compared with non-infected prostheses (non-PVE, right panel). PET/CTA fusion images of the valve plane (upper row), and their correspond- ing attenuation-corrected PET images (lower row), on which FDG distribution has been schematically represented. High VUI values correlate with focal/multifocal or diffuseþheterogeneous uptake patterns, which are highly suggestive of infection, whereas low VUI values reflect a diffu- seþhomogeneous pattern or absence of visually detectable FDG uptake, more characteristic of inflammation. FDG, fluorodeoxyglucose; PET/CTA, positron emission computed tomography angiography; PVE, prosthetic valve endocarditis; VUI, valve uptake index. affect the absolute values of the SUV. This would make the VUI a po- .. rejected group and in 93% of the control group. Likewise, the VUI, .. tentially exportable measure between different centres and equip- .. which in essence reflects FDG distribution, was superimposable, ment. However, further studies are needed to validate this potential .. being 0.35 ± 0.08 and 0.35 ± 0.10, respectively. A pathological uptake .. reproducibility. Finally, obtaining this parameter does not imply any .. pattern (focal/multifocal or diffuse-heterogeneous) was seen in 97% .. change in either image acquisition or reconstruction, since it is calcu- .. cases in the definite PVE group. lated directly from the available SUV values. .. .. No significant differences were seen in the VUI values according to .. valve type (mechanical or biological) or position (aortic or mitral) in .. .. Diagnostic algorithm and criteria either the definite or the rejected group, as previously reported in .. of PVE patients with no suspicion of IE.5 Prosthetic/periprosthetic FDG up- .. .. take was visually detectable in all cases (100%) of ‘definite’ PVE, but in .. The diagnostic approach by PET/CTA of a suspected PVE should in- only 35% of the ‘rejected’ group, a frequency markedly lower than .. clude, when possible, both local assessment of the infection at the .. that reported in prostheses without suspicion of infection (79.3%). In .. cardiac level (which will provide major diagnostic criteria), as well as .. addition, quantitative values of metabolic activity were lower in .. extracardiac evaluation to determine the presence of distant lesions rejected cases than in those reported in postsurgical cases without .. (minor criterion).4,14 In addition, PET/CTA allows the identification .. suspicion of PVE (SUVmax 3.72 ± 1.11 vs. 4.46 ± 1.50, SUVmean .. of the source of infection, or an alternative diagnosis if PVE is ruled- 2.38 ± 0.7 vs. 2.80 ± 0.62, and SUVratio 2.16 ± 0.95 vs. 2.28 ± 0.91). .. out, both highly useful for patient management. Nevertheless, the .. These differences could be explained because PV controls in our pre- .. main challenge remains the diagnosis of PV infection, often in a scen- vious study were scanned within the first year after implantation, .. ario in which echocardiographic evaluation is doubtful or negative, .. while the median time from surgery to PET/CTA was 49 months in .. probably leading the definitive diagnosis of endocarditis to be estab- .. rejected PVE patients. The known metabolic activity due to postop- .. lished by PET/CTA.3,4 erative inflammation may be higher in the recent post-implantation .. PET/CTA evaluation is based on metabolic and anatomical imaging .. period and, although not well established, may gradually decrease .. criteria: the visual assessment of the location and distribution pattern over time. Irrespective of this consideration, a normal uptake pattern .. of FDG uptake (visual analysis), the intensity of the uptake values .. (absent or diffuse-homogeneous) was seen in 96% of patients in the . (quantitative analysis), the presence or absence of anatomical lesions The valve uptake index: a new PVE parameter by PET/CTA 1269 Figure 6 Proposed diagnostic algorithm for PVE assessment by cardiac PET/CTA, with integration of the VUI to the evaluation of morphological and anatomical imaging criteria. IE, infective endocarditis; PET/CTA, positron emission computed tomography angiography; PVE, prosthetic valve endocarditis; VUI, valve uptake index; SUV, standardized uptake value. related to IE in the CTA, and the correlation between the metabolic .. correlation with the uptake distribution, the VUI could contribute to .. and anatomic images. Integrating the results of this study and the pre- .. a correct interpretation of these difficult patterns, reinforcing visual viously available knowledge, we have designed an improved diagnos- .. interpretation. Moreover, the addition of the VUI to the uptake pat- .. tic algorithm for PVE (Figure 6). .. tern has shown a significant increase in the diagnostic ability for PVE. .. The key step in the diagnostic algorithm will be to define if the pat- .. In this regard, it may provide an important benefit in patients with tern of FDG uptake is ‘normal’ or ‘abnormal’. Absent uptake or a dif- .. multiple prostheses, or in situations where the uptake pattern may .. fuse and homogeneous uptake should be considered a ‘normal’ .. be subject to changes over time due to separate episodes of sus- pattern. In contrast, a focal uptake or a diffuse and heterogeneous up- .. pected IE or in follow-up scans (Figure 7 and Supplementary data on- .. take will be highly indicative of infection (‘abnormal’ pattern). While .. line, Figure S1). In addition, the FDG uptake pattern, the traditionally focal FDG uptake should offer few diagnostic doubts, the challenge in .. used quantitative measures (SUVmax and SUV ratio) and the pres- .. classification would lie between the diffuse-homogeneous and .. ence or absence of anatomic lesions have confirmed their usefulness diffuse-heterogeneous patterns. Due to its demonstrated good .. in clinical practice. 1270 A. Roque et al. Figure 7 Added value of the VUI in a patient with multiple PVs. A 45-year-old woman who had undergone aortic and mitral valve replacement and tricuspid annuloplasty 4 years previously was admitted due to fever and positive blood cultures for Aggregatibacter aphrophilus (HACEK group). Transoesophageal echocardiography was negative. (A) fused PET/CTA view of the mitral plane showed focal and intense FDG uptake (SUVmax 8.86) in the inferoseptal aspect of the prosthetic ring-suture, with a VUI of 0.76, findings consistent with mitral PVE. (B) Aortic PV displays diffuse uptake of moderate intensity (SUVmax 5.19), but with a low VUI of 0.40, indicating a homogeneous distribution, suggesting reactive inflammation and not sup- porting infective involvement. (C) Maximum intensity projection oblique plane including all cardiac PVs in a single view allows global assessment of the loco-regional distribution of metabolic activity. Focality is seen only over the mitral valve in contrast with the homogeneous distribution along the fi- brous ring of the heart. Ao, aortic prosthetic valve; FDG, fluorodeoxyglucose; Mi, mitral prosthetic valve; PET/CTA, positron emission computed tom- ography angiography; PVE, prosthetic valve endocarditis; SUVmax, maximum standardized uptake value; Tri, tricuspid annuloplasty; VUI, valve uptake index. .. Based on the results obtained, we propose a series of metabolic .. an unclear visual pattern hence improving diagnosis. (iii) The pro- and anatomic ‘diagnostic criteria’ for PVE diagnosis by PET/CTA: (i) .. posed diagnostic algorithm includes searching for potential IE- .. abnormal FDG uptake pattern: focal/multifocal or diffuse/heteroge- .. related anatomic lesions, which can only be assessed if an associ- neous uptake; (ii) SUVmax > 5.28; SUVratio > 2.58; (iii) VUI > 0.45; .. ated ECG-gated cardiac CTA is performed, highly recommended .. and (iv) Presence of IE-related anatomic lesions, especially those that .. whenever possible. indicate periprosthetic extension of infection. The addition of extrac- .. In conclusion, the VUI has demonstrated to be a valid and helpful .. ardiac lesions to the main imaging criteria will support the diagnosis .. parameter for PVE diagnosis. Because of its simplicity and potential re- .. producibility, it could be incorporated as a complementary tool into of IE. .. Potential limitations of this study are: (i) although the VUI is a .. routine PET/CT evaluation, even when performed by less- .. experienced imaging readers. This study also confirms the utility of the relatively simple value to calculate, this new parameter was .. obtained in a single centre with extensive experience on cardiac .. most widely used parameters to date, both qualitative (uptake pattern .. and anatomic lesions) and quantitative (SUVmax and SUVratio) for the PET/CTA imaging and therefore its ultimate applicability must be .. verified through its use by other imaging readers. Likewise, as the .. diagnosis of infection. As the VUI provides both qualitative and quanti- .. VUI is obtained from a ratio between SUVmax and SUVmean, it .. tative information, its integration into the already established imaging .. criteria could improve the diagnostic yield of PET/CTA. could correct certain differences derived from the lack of stand- .. ardization in the acquisition and reconstruction of PET/CT images, .. .. as well as differences between equipment from different vendors. .. Supplementary data This possibility should be validated in further studies; (ii) although .. .. not a direct limitation of the present study, next step would be to .. Supplementary data are available at European Heart Journal - Cardiovascular evaluate the performance of the VUI in reclassifying patients with .. Imaging online. The valve uptake index: a new PVE parameter by PET/CTA 1271 .. Acknowledgements .. 5. Roque A, Pizzi MN, Fernández-Hidalgo N, Permanyer E, Cuellar-Calabria H, The authors thank the Hospital Universitari Vall d’Hebron .. Romero-Farina G et al. Morphometabolic post-surgical patterns of non-infected .. prosthetic heart valves by [18F]FDG PET/CTA: "normality" is a possible diagno- Endocarditis Team: Benito Almirante, MD, PhD, Laura Escolà-Verge, .. sis. Eur Heart J Cardiovasc Imaging 2020;21:24–33. .. 6. Pizzi MN, Roque A, Cuellar-Calabria H, Fernández-Hidalgo N, Ferreira-González MD, PhD, Ruben Fernández, MD, Nuria Fernández-Hidalgo, MD, .. I, González-Alujas MT et al. Infective versus inflammatory patterns in 18F-FDG- PhD, MSc, Maria Teresa González-Alujas, MD, Olga Maisterra, MD, .. .. PET/CTA of prosthetic cardiac valves and valve-tube grafts. JACC Cardiovasc PhD, Gerard Oristrell, MD, Marıa Nazarena Pizzi, MD, PhD, Pau .. Imaging 2016;9:1224–7. Rello, MD, Remedios Rıos, MD, PhD, Albert Roque, MD, Antonia .. 7. Mathieu C, Mikail N, Benali K, Iung B, Duval X, Nataf P et al. Characterization of .. 18F-fluorodeoxyglucose uptake pattern in noninfected prosthetic heart valves. Sambola, MD, PhD, and Toni Soriano, MD. .. Circ Cardiovasc Imaging 2017;10:e005585. .. 8. Swart LE, Gomes A, Scholtens A, Sinha B, Tanis W, Lam M et al. Improving the Conflict of interest: none declared. .. diagnostic performance of 18F-FDG PET/CT in prosthetic heart valve endocardi- .. .. tis. Circulation 2018;138:1412–27. References .. 9. Scholtens AM, Swart LE, Kolste H, Budde R, Lam M, Verberne H. Standardized .. uptake values in FDG PET/CT for prosthetic heart valve endocarditis: a call for 1. Pizzi MN, Roque A, Fernández-Hidalgo N, Cuellar-Calabria H, Ferreira-González I, .. standardization. J Nucl Cardiol 2018;25:2084–91. Gonzàlez-Alujas MT et al. Improving the diagnosis of infective endocarditis in pros- .. 10. Aguade-Bruix S, Roque A, Cuellar-Calabria H, Pizzi MN. Metodologıa de la 18F- thetic valves and intracardiac devices with 18F-FDG-PET/CT-angiography: initial .. FDG PET/CT cardiac para el diagnóstico de la endocarditis protesica y de dispo- results at an infective endocarditis referral center. Circulation 2015;132:1113–26. .. 2. Saby L, Laas O, Habib G, Cammilleri S, Mancini J, Tessonnier L et al. Positron .. sitivos intracardıacos. Rev Esp Med Nucl Imagen Mol 2018;37:163–71. .. 11. Roque A, Pizzi MN, Cuellar-Calàbria H, Aguade-Bruix S. 18F–FDG-PET/CT emission tomography/computed tomography for diagnosis of prosthetic valve .. angiography for the diagnosis of infective endocarditis. Curr Cardiol Rep 2017; endocarditis increased valvular 18F-fluorodeoxyglucose uptake as a novel major .. 19:15. criterion. J Am Coll Cardiol 2013;61:2374–82. .. 12. Erba PA, Lancellotti P, Vilacosta I, Gaemperli O, Rouzet F, Hacker M et al. 3. Habib G, Lancellotti P, Antunes MJ, Bongiorni MG, Casalta JP, Del Zotti F et al.; .. Recommendations on nuclear and multimodality imaging in IE and CIED infec- ESC Scientific Document Group. 2015 ESC Guidelines for the management of .. .. tions. Eur J Nucl Med Mol Imaging 2018;45:1795–815. infective endocarditis: The Task Force for the Management of Infective .. 13. Roque A, Pizzi MN, Fernández-Hidalgo N, González-Alujas MT, Rıos R, Castell- Endocarditis of the European Society of Cardiology (ESC). Endorsed by: .. Conesa J et al. Mosaic bioprostheses may mimic infective endocarditis by PET/ European Association for Cardio-Thoracic Surgery (EACTS), the European .. CTA: trust the uptake pattern to avoid misdiagnosis. JACC Cardiovasc Imaging Association of Nuclear Medicine (EANM). Eur Heart J 2015;36:3075–128. .. 2020;13:2239–44. 4. Otto CM, Nishimura RA, Bonow RO, Carabello BA, Erwin JP III, Gentile F et al. .. 14. Swart LE, Scholtens AM, Tanis W, Nieman K, Bogers A, Verzijlbergen F et al. 2020 ACC/AHA guideline for the management of patients with valvular heart dis- .. .. 18F-fluorodeoxyglucose positron emission/computed tomography and com- ease: report of the American College of Cardiology/American Heart Association .. puted tomography angiography in prosthetic heart valve endocarditis: from Joint Committee on Clinical Practice Guidelines. Circulation 2021;143:e72–227. . guidelines to clinical practice. Eur Heart J 2018;39:3739–49.