Inguinal draining-lymph node in 18F-FDG PET/CT images could be a new indicator for the diagnosis of fracture-related infection in the lower extremities

Purpose The imaging diagnosis of fracture-related infection is often challenging. The aim of this study was to evaluate the value of 18F-FDG PET/CT for the diagnosis of fracture-related infection (FRI) with internal fixation after orthopedic surgery in lower extremities. Methods A total of 254 consecutive patients who underwent 18F-FDG PET/CT scans with suspected FRI with internal fixation in lower extremities were retrospectively investigated 18F-FDG PET/CT images were semiquantitatively evaluated with multiple metabolic parameters. Additionally, morphological information of the inguinal draining lymph nodes (DLN) with the highest SUV value was also collected and analyzed. Results Patients were divided into two groups according to final diagnosis: the infected (N=197) and the non-infected group (N=57). The differences in the inguinal DLN-related parameters, including the long diameter, short diameter, maximum cross-sectional area, maximum standardized uptake value (SUVmax), metabolic volume (MV) 60%, MV70%, MV80%, total lesional glycolysis (TLG) 60%, TLG70%, TLG80%, and the infection suspected area related parameters, including SUVmax, MV25%, MV30%, MV35%, MV40%, MV50%, and TLG70%, between the two groups were statistically significant. We then compared the highest area under the curves (AUCs) among the morphological parameters of DLN, metabolic parameters of DLN, and metabolic parameters of the suspected infection area. The result demonstrated that SUVmax of the inguinal DLN showed the best diagnostic performance with an AUC of 0.939 (P<0.05). Conclusion Semiquantitative analysis (especially SUVmax) of the inguinal DLN in 18F-FDG PET/CT images could be a promising method for the diagnosis of suspected FRI with internal fixation after orthopedic surgery in lower extremities.


Introduction
Fracture-related infection (FRI) is one of the most prevalent and challenging complications followed by trauma surgery, especially in patients receiving orthopedic implants (1)(2)(3)(4).The most common microorganisms causing implant-associated infections are Staphylococcus aureus and Coagulase-negative staphylococci (5).Infection is among the main causes of persistent non-union fractures and thus osteosynthesis failure, which requires invasive treatment and is a major burden for healthcare systems (6).What is more, the growing amount of placed fracture-related implants and patients with inducing conditions (e.g., diabetes mellitus and peripheral vascular diseases) may increase the incidence of FRI (7,8).
The timely identification of infection after orthopedic surgery is essential for effective treatment (9).A consensus on how to diagnose FRI was made by the AO Foundation and European Bone and Joint Infection Society (EBJIS) recently (10), which compromised a set of confirmatory criteria (infection definitely present: for instance, fistula, sinus, purulent drainage, and the presence of pus) and suggestive criteria (infection possibly present: for instance, related clinical manifestations and radiological imaging).The evaluation of suspected FRI can be characterized by a multimodality workup containing physical and imaging examination and laboratory parameters (i.e., an elevated erythrocyte sedimentation rate and C-reactive protein).However, FRI does not always announce itself with evident symptoms.Therefore, a definite diagnosis of FRI could sometimes be quite challenging (11).
Conventional imaging methods such as plain X-rays and computed tomography (CT) can help little for the diagnosis of FRI due to nonspecific imaging features (12).Magnetic resonance imaging (MRI) can distinguish osteomyelitis with high certainty (13).However, the diagnostic value of MRI is weakened for FRI because sterile inflammation and reparative fibrovascular scar tissue usually show similar signal intensities such as reactive marrow edema (14-16).Furthermore, MRI images can be hampered by the presence of internal fixation material (17).Considering its pathophysiological processes, multiple nuclear imaging modalities such as three-phase bone scintigraphy (BS), leukocyte scintigraphy (namely, white blood cell [WBC] scintigraphy), and 18 Ffluorodeoxyglucose ( 18 F-FDG) positron emission tomography (PET)/computed tomography (CT) have been investigated to evaluate the diagnostic efficacy of orthopedic surgery related infections including periprosthetic joint infection and FRI (18).
In a previously published meta-analysis, 18 F-FDG PET/CT scans have presented satisfactory accuracy for the diagnosis of osteomyelitis compared with bone scintigraphy and leucocyte scintigraphy (19).Institutions using 18 F-FDG PET/CT to evaluate suspected FRI are steadily growing worldwide, and this nuclear medicine imaging modality showed encouraging diagnostic accuracy based on qualitative assessments (20)(21)(22)(23)(24). Uptake of 18 F-FDG reflects cellular glucose metabolism and can be semiquantitatively expressed as SUV, MV, TLG, etc. Increased glucose consumption induces increased 18 F-FDG-uptake in the areas with enhanced metabolic activity of the tissue involved or the invasion of inflammatory cells (25)(26)(27).In addition, studies also evaluated the value of semiquantitative measurements in 18 F-FDG PET/CT for the diagnosis of suspected infections for nonunion fractures, and they achieved satisfactory diagnostic accuracy (28,29).Results from previous studies indicated that 18 F-FDG PET/CT could play an essential role in the diagnosis of FRI.
In clinical practice, increased uptake of 18 F-FDG in inguinal lymph nodes can be frequently found during PET/CT scans in patients with suspected FRI in lower extremities.Enlargement of draining lymph nodes (DLNs) is related to infection, sterile inflammation, autoimmune disorder, cancer, and other diseases where the immune system is involved.The immune response in the DLNs can be etiology and pathogenesis specific (30)(31)(32)(33)(34).Thus, we supposed that the 18 F-FDG uptake pattern in the inguinal DLNs could be different in fracture-related infections and aseptic inflammation in the lower extremities.To the best of our knowledge, the diagnostics value of DLNs in 18 F-FDG PET/CT images for FRI has not been reported.The aim of our study, therefore, was to further evaluate semiquantitative measurements of 18 F-FDG PET/CT for the diagnosis of suspected FRI with internal fixation in lower extremities and to focus not only on the suspected infection area but also the DLNs in the inguinal region.

Patients
We retrospectively enrolled 254 consecutive patients of suspected FRI with internal fixation in lower extremities who had undergone 18 F-FDG PET/CT scan in Shanghai Sixth People's Hospital between Dec 2015 and July 2021.The surgery following was performed within 1 month.The time interval between the latest fracture surgery and the PET/CT scan was more than 3 months in all patients.The final diagnosis of FRI was confirmed according to the consensus (10) made by the AO Foundation and EBJIS.Of the 254 patients, 197 were diagnosed as having an infection (89 with positive bacterial culture; 108 with negative bacterial culture), and 57 were diagnosed as having aseptic inflammation.

F-FDG PET/CT acquisition protocol
Before 18 F-FDG PET/CT scans, patients were required to fast for at least 6 hours with a maximum blood glucose level of 11 mmol/L.Scans were performed 40-60mins after intravenous administration of 3.7MBq/Kg dose of 18 F-FDG by using an integrated PET/CT scanner (Discovery VCT, GE, USA).PET/CT scans of the lower extremities were performed from the anterior superior spine to toes with the following settings: CT scan, 120V and 80mA, 64 slices, 3.75mm slide thickness; PET scan, 3D model, 2.5 min/bed.PET images were then reconstructed iteratively by using ordered subset expectation maximization (OSEM), and attenuation correction was applied according to CT for all scans.
Image analysis 18 F-FDG PET/CT images were interpreted semiquantitatively by two experienced senior nuclear medicine physicians.For the semiquantitative analysis, the volume of interest (VOI) was set in the target area with suspected infection and the corresponding inguinal draining lymph node with the highest SUVmax value.The semiquantitative parameters, including the maximum standardized uptake value (SUVmax), metabolic volume (MV)25%, MV30%, MV35%, MV40%, MV50%, MV60%, MV70% of suspected infection area, and SUVmax, MV60%, MV70%, MV80% of the inguinal lymph node within the VOIs, were detected.The corresponding total lesional glycolysis (TLG) value (TLG25%, TLG30%, TLG35%, TLG40%, TLG50%, TLG60%, and TLG70%) was calculated according to the following formula: TLG = SUVmean × MV (with corresponding SUV threshold).In addition to the above metabolic parameters, morphological information in CT images, including long diameter, short diameter, and maximum cross-sectional area of the inguinal DLN (with the highest SUVmax value), was also collected and analyzed.

Statistical analysis
The statistical analysis was performed using IBM SPSS version 26.0 and MedCalc version 19.2.6.Comparisons were assessed using the Student's T test (parametric data) or the Mann-Whitney U test (non-parametric data).Data were expressed as mean ± standard deviation (SD).Receiver operating characteristic (ROC) curves for parameters were plotted, and the corresponding area under the curve (AUC) values were computed.The diagnostic sensitivity and specificity of all parameters were calculated.Differences of AUCs among parameters were compared by MedCalc.A P value < 0.05 was taken as the threshold of statistical significance.
SUVmax is the easiest parameter obtained in clinical routines.We then compared the AUCs of SUVmax from DLNs and suspected infection areas.The result demonstrated that SUVmax from the inguinal DLNs showed better diagnostic efficiency than SUVmax from suspected infection area (0.939 vs 0.836, P=0.0006) (Figure 2).Representative cases are provided in Figures 3, 4.

Discussion
Results of the current study showed that 18 F-FDG PET/CT could be a promising diagnostic tool for FRI with internal orthopedic implants.Among all the parameters evaluated, SUVmax of the inguinal DLNs showed the best diagnostic accuracy with an AUC of 0.939, sensitivity of 86.8%, and  specificity of 93.0%.To the best of our knowledge, the current study is the first one in the literature that investigated the diagnostic value of inguinal DLNs in 18 F-FDG PET/CT images for FRI in lower extremities.

Location of non-union
During the immune response, immune cell activation, infiltration, and proliferation are the basic pathological changes in the DLNs.This process can be etiology and pathogenesis specific, and therefore the immune response in inguinal DLNs, due to chronic infection or sterile inflammation, resulting from fracture-related orthopedic surgery could be different.It is well known that immune cells increase glucose analogue 18 F-FDG uptake during activation, infiltration, and proliferation because of high energy consumption.So, it is reasonable to believe that the pathognomonic pattern of 18 F-FDG uptake in FRI could be different from aseptic chronic inflammation not only in suspected infection regions but also in corresponding draining lymph nodes.In this study, we found SUVmax values from suspected infection areas in the lower extremities in the infected group were significantly higher than that in the non-infected group.In addition, values of metabolic parameters, including SUVmax, TLG60%, TLG70%, and TLG80% derived from inguinal DLNs in the infected group, were all significantly higher than those in the non-infected group.Interestingly, the diagnostic efficiency of SUVmax from the inguinal DLN was better than SUVmax from the suspected infection areas in the lower extremities (0.939 vs 0.836, P=0.0006).
Vera Wenter et al. (35) investigated the diagnostic value of 18 F-FDG PET/CT for the detection of chronic osteomyelitis and implantassociated infection and reached a sensitivity of 70% and a specificity of 66% at the cut-off SUVmax value of 3.7.Furthermore, they found that the presence of an orthopedic implant did not reduce the diagnostic yield in patients undergoing PET/CT.Van Vliet et al. (29) used SUVmax to differentiate between aseptic and septic delayed union in the lower extremities.They found a sensitivity, specificity, and diagnostic accuracy of 65%, 77%, and 70%, respectively, at the cut-off SUVmax value of 4.0.Martina Sollini et al. (36) retrospectively tested the diagnostic performances of 18 F-FDG PET/CT in non-union fractures in 47 patients.The best diagnostic performance was observed when setting the cut-off SUVmax value to 5.92 with an AUC of 0.72.The cut-off SUVmax value identified in their study was higher than in our cohort.Besides differences in patient cohorts and heterogeneous patient populations, the fact that they used microbiological culture results for specimens collected during surgery as the only reference standard to define the final diagnosis could be a significant cause for a higher cut-off value.In real-world clinical practice, the proportion of patients who could not get positive  Comparison of receiver operating characteristics (ROC) curves between SUVmax from inguinal draining lymph node (DLN) and suspected infection area (AUC, 0.939 vs 0.836, P=0.0006).SUVmax, maximum standardized uptake value; Si, suspected infection.Case of a 50-year-old male patient with infected non-union.The thin arrow points to the swollen lymph node in the right groin with high metabolism.The thick arrow points to the infection area after internal fixation of fracture, and the metabolism of the infected area is increased.Case of a 33-year-old male patient with simple non-union.The thin arrow points to the increased metabolism after internal fixation of fracture.The thick arrow points to the multiple small lymph nodes in the right groin area without obvious abnormal metabolism.
bacterial culture results were not small relatively speaking, as low virulence bacteria such as the coagulasenegative Staphylococcus species may cause culture-negative infection (37).More recently, Lemans et al.
(22) reported in a large cohort of patients that the sensitivity and specificity were 80% and 72% at a cut-off SUVmax value of 4.2 for the diagnosis of FRI by using 18 F-FDG PET/CT.The cut-off value of SUVmax identified in their cohort was very similar to ours.Although 18 F-FDG PET/CT scans are shown to be promising in the diagnosis of FRI, the fact remains that 18 F-FDG is a non-specific tracer for infection.Therefore, specific radiopharmaceutical probes targeting microorganisms are warranted in the future.
There are several limitations in this study.Firstly, it is a retrospective study, so there is a potential risk of selection bias.
Retrospective studies analyze existing data and cannot control the data collection process, which may result in less accurate and reliable research findings.Secondly, some non-infected patients may not have had sufficient follow-up time, which could lead to false-negative results.Thirdly, some patients may have received antibiotics prior to PET imaging, which could affect the uptake of 18 F-FDG.Lastly, the number of non-infected patients is significantly smaller than the infected group, leading to data imbalance.This imbalance could result in statistical bias, affecting the reliability of the research results.Future research can improve the reliability of results by increasing the sample size, employing stricter research designs, and controlling for relevant factors.

Conclusion
In conclusion, our study suggests that semiquantitative analysis (especially SUVmax) of the inguinal draining-lymph node in 18 F-FDG PET/CT images shows promise as a diagnostic method for detecting FRI with internal fixation after orthopedic surgery in lower extremities.However, further research is necessary to validate and confirm these findings.

1
FIGURE 1 Receiver operating characteristics (ROC) curves of the parameters analyzed in this study regarding morphological and metabolic parameters of the inguinal draining lymph node (DLN) as well as metabolic parameters of the infection suspected area.(A) Morphological parameter of the inguinal DLN.The maximum cross-sectional area of DLN shows the highest area under curve (AUC) (0.889).The sensitivity and specificity are 81.2% and 82.5%, respectively, with a cut-off value of 117.27mm 2 .(B) Metabolic parameters of the inguinal DLN.SUVmax of DLN shows the highest AUC (0.939).The sensitivity and specificity are 86.8% and 93.0%, respectively, with a cut-off value of 1.55.(C) Metabolic parameters of the infection suspected area.SUVmax of suspected infection area shows the highest AUC (0.836).The sensitivity and specificity are 74.0%and 77.2%, respectively, with a cut-off value of 4.53.DLN, draining lymph node; Ln, lymph node; Sd, short diameter; Ld, long diameter; Cs, cross-sectional area; MV, metabolic volume; TLG, total lesion glycolysis; SUVmax, maximum standardized uptake value; Si, suspected infection.

TABLE 1
Clinical characteristics for patients in infected and noninfected groups.

TABLE 2
Data of the parameters analyzed with AUC > 0.8.