Introduction
Invasive fungal infections (IFIs) cause significant morbidity and mortality, especially in the immunocompromised host [1]. IFIs can affect any part of the human body. Imaging plays an essential role in the management of IFIs [2]. Fluorine‐18 fluorodeoxyglucose positron emission tomography integrated with computed tomography (FDG PET/CT) has been found useful in the management of IFIs [3, 4]. FDG PET/CT, as a whole‐body hybrid imaging technique, allows functional data from positron emission tomography (PET) to be correlated with anatomic data from the computed tomography (CT) from different regions of the body in one imaging session [5]. Anatomic‐based imaging procedures are more commonly used in imaging IFIs during the management of the infection. These modalities include chest X‐ray, high resolution computed tomography of the chest (HR CT), CT of the chest and other parts of the body, magnetic resonance imaging (MRI), and ultrasound imaging (US) [2,6‐8]. The acquisition of these anatomic‐based procedures is usually limited to a particular region of the body and frequently lacks the functional information provided by PET. Recent data suggest that FDG PET/CT may add extra information to the data supplied by anatomical based studies to help the management of IFIs.
In this study, we assessed the added value of FDG PET/CT to anatomic‐based imaging modalities that are used in the management of IFIs.
Material and Methods
The scans of patients diagnosed with IFIs who also underwent FDG PET/CT as part of the management of IFIs were retrospectively reviewed. We included patients who had proven, probable or possible IFIs according to the revised European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) classification [9]. The study was conducted in two Medical Centers in the Netherlands, The University Medical Center Groningen (UMCG) and The University Medical Center Utrecht (UMCU). All chest X‐rays, HR CT, CT scans, MRI, and ultrasound imaging that was performed within two weeks of the FDG PET/CT were retrieved, and results documented. The electronic patient files of all included patients were retrieved and scanned for relevant clinical data. Board‐certified nuclear medicine physicians and radiologists reported the images as part of routine clinical practice.
The images used in the study were acquired between October 2009 and March 2018. Given the retrospective nature and use of routine care data, no formal ethical approval or informed consent was required (UMCG 201600073). Data were handled according to general data protection regulation (GDPR) requirements. The FDG PET/CT was acquired according to the EANM guidelines [10]. The CT component of the FDG PET/CT was non‐contrast enhanced. The anatomic‐based imaging was conducted according to the departmental protocols. The FDG PET/CT study was considered to be concordant to the anatomical‐based study if abnormal lesions due to IFIs were detected in both studies or if no lesion could be attributed to IFI on both. We also recorded FDG PET/CT studies in which abnormal IFI lesions were present at sites that were not imaged by the other imaging studies done within two weeks of the PET/CT.
8
A retrospective study evaluating the value of PET in the management of IFIs
We defined added value as the ability of FDG PET/CT to detect lesions outside the region scanned by the other studies, which could potentially alter management. The FDG PET/CT was also considered to add value when the metabolic information provided by the study allowed the follow‐up of the activity of IFI lesions. Finally, FDG PET/CT was considered to add value when the scan was done to assess the metabolic activity of previously treated IFIs lesions to make therapeutic decisions.
Results
Finally, we included 155 FDG PET/CT studies from 73 patients who were diagnosed with IFIs. Twenty‐
four (33%) of the patients were females, and the mean age of the patients was 44.24 ± 21.81 years.
Fifty‐nine (81%) patients were from UMCG, and fourteen (19%) were from UMCU. Forty‐one patients had a single FDG PET/CT in the assessment of the IFIs and, 32 patients had repeated studies. The IFIs were classified according to the revised EORTC/MSG criteria and results displayed in Table 1. In 42 (57.5%) patients, there was a proven IFI, in 19 (26%%), there was a probable IFI and in 12 (16.4%) there was a possible IFI. Of the 32 patients who had repeated studies, 18 had a proven IFI, five had a probable IFI and nine had a possible IFI. In 31 patients with either a probable of possible IFI, 14 out of 31 (45.1%) had repeated FDG‐PET to monitor the treatment of the IFI.
Table 1: Demographics of patients, type of IFIs and those with repeated studies
Number (%)
Total number of patients 73 (100%)
Female (n, %) 24 (33%)
Patients with repeated FDG PET/CT 32 (44%) Patients with proven IFIs 42 (26%) Patients with probable IFIs 19 (56%) Patients with possible IFIs 12 (16%) Patients with proven molds 13 (18%) Patients with proven yeasts 29 (40%)
Conditions underlying the IFIs
Acute leukemia was the most common underlying condition in 27 (37%) of IFI patients. Other hematologic disorders made up the second‐largest group 18 (24.7%). Other underlying conditions included solid organ transplantation, invasive procedures, immunosuppressive therapies, abnormal anatomy due to a previous disease, and in five patients, no underlying condition was identified. Table 2 depicts the underlying conditions for the IFI.
Table 2: Underlying conditions of IFIs
Condition Number (%)
Acute leukemia 27 (37%)
Hematologic conditions excluding acute leukemia 18 (25%)
Solid‐organ transplantation 10 (14%)
Invasive procedures 4 (4%)
High dose steroids 4 (5%)
Intense chemotherapy 3 (4%)
Lung cavitation 2 (3%)
No clear underlying condition 5 (7%)
Indications for the FDG PET/CT
134 135
8
Chapter Eight
The indications for the FDG PET/CT were classified under five categories. In thirty‐two patients (44%), FDG PET/CT was used to monitor response to antifungal therapy. In 18 (25%), it was used to stage the IFI. In 10 (14%), there was unexplained fever or increasing infective markers. In seven (10%) patients who were previously treated for IFIs were being assessed for allogeneic stem cell transplantation and in six (8%) to determine whether there was active disease in persistent anatomical lesions. Table 3 shows the various indications for the FDG PET/CT scans. Figure 1 demonstrates the initial chest X‐ray, HR CT and fused axial FDG PET/CT through the chest of a patient with a proven fungal infection and FDG PET/CT was initially requested to stage infection and was subsequently used to monitor treatment.
Table 3: Indication for the FDG PET/CT studies
Indication Number of patients (%)
Monitor response to antifungal therapy 32 (44%)
Stage infection 18 (25%)
Unexplained fever or increasing infective markers 10 (14%)
Evaluation for ASCT 7 (10%)
Active disease 6 (8%)
Figure 1: Chest X–ray, HR CT and FDG PET/CT of a 60‐
year‐old female with acute myeloid leukemia on chemotherapy. She had a proven fungal infection (Alternaria alternata)
A‐ Initial chest X‐ray, B‐CT scan done within two weeks of the FDG PET/CT. C‐ Fused axial image of FDG PET/CT of the chest. After 2 months of antifungal therapy. D‐ Chest X‐ray showing response with a residual lesion with E fused axial FDG PET/CT showing an almost complete metabolic response in the IFI lesion.
FDG PET/CT studies that had anatomic‐based imaging within two weeks
Of the 155 FDG PET/CT studies that were analyzed, 34 (22%) had no anatomic‐based imaging within two weeks of the FDG PET/CT study. Forty‐three (28%) of the FDG PET/CT studies had one anatomic‐
based study; sixty‐two (40%) had two other anatomic‐based studies within two weeks. Fifteen of the FDG PET/CT had three different anatomic‐based imaging studies, and one had four other anatomic‐
based imaging studies done within two weeks of the FDG PET/CT study. Table 4 outlines the number of anatomic‐based studies done with two weeks of the FDG PET/CT scan.
Concordance of FDG PET/CT and all the anatomic‐based studies in the region of interest of the anatomical‐based study
One hundred and twenty‐one FDG PET/CT scans had at least one anatomic‐based imaging within two weeks. Forty‐two of 43 FDG PET/CT scans (98%) that had only one anatomic‐based study within two weeks were concordant with the only anatomic‐based study done. Fifty‐three of 62 FDG PET/CT procedures (86%) that had two anatomic‐based imaging studies were concordant with both anatomic‐
based studies. Thirteen of 15 FDG PET/CT scans (87%) that had three anatomic‐based studies were concordant with all three anatomic‐based studies. The only FDG PET/CT study that had four anatomic‐
based studies within two weeks was concordant with all four anatomic‐based studies. In total, 109, (90%) of the 121 FDG PET/CT were concordant with all anatomic‐based imaging done within two
8
A retrospective study evaluating the value of PET in the management of IFIs
weeks. Table 4 depicts all the numbers of FDG PET/CT scans that were concordant with all the anatomic‐based studies.
Specific anatomic‐based studies done within two weeks of FDG PET/CT study A total of 216 anatomic‐based imaging studies were performed within two weeks of the 155 FDG PET/CT studies. Eighty (37%) of these were chest radiographs, 62 (29%) were HR CT or CT of the chest, 27 (13%) were CT scans of extra thoracic regions, 14 (6%) were MRI scans, and 33 (15%) were ultrasound studies (Table 5).
Table 4: Number of anatomic‐based studies done within two weeks of each FDG PET/CT Anatomic‐based studies for
each FDG PET/CT scan done FDG PET/CT
scans done FDG PET/CT scans that were concordant with all the anatomic‐based studies done (%)
0 34 (21.9%) N/A
1 43 (27.7%) 42 (97.6%)
2 62 (40.0%) 53 (85.5%)
3 15 (9.7%) 13 (86.6%)
4 1 (0.1%) 1 (100%)
Total 155 (100%) 109 of 121 (90.1%)
Concordance of FDG PET/CT with specific anatomical‐based imaging
The results of 74 out of 80 (92.5%) FDG PET/CT that had a chest X‐ray within two weeks were concordant. In the six that were not concordant, abnormal lesions due to IFIs were found on the FDG PET/CT but not on the X‐ray. All of the 62 FDG PET/CT that had HR CT or thoracic CT within two weeks were concordant with the anatomic‐based studies. Again, all the 27 FDG PET/CT that had CT scans done for areas outside the thorax done within two weeks study were concordant with the extra thoracic CT scans. Thirty of the 33 (90.9%) FDG PET/CT scans that had ultrasound scans within two weeks were concordant with the ultrasound studies. In the two that were not concordant, FDG PET/CT determined IFI lesions not noted on ultrasound imaging. Eleven out of 14 (78.6%) FDG PET/CT that had MRI imaging within two weeks were concordant. In the three FDG PET/CT scans that were not concordant with MRI imaging, MRI detected lesions that were not detected on FDG PET/CT. The overall concordance of FDG PET/CT with anatomic‐based studies was 94.9%. Table 5 demonstrates the concordance of the specific anatomic based studies with FDG PET/CT.
Extra lesions detected only by FDG PET/CT because the lesion is not in the field of view of the anatomic‐based study
Of the 80 plain radiographs done, 37 (46%) had lesions on the FDG PET/CT outside the thorax. For the HR CT or thoracic CT, 22 out of 62 (35%) had lesions outside the thorax on the PET scans. For extra thoracic CT scans, 14 out of 27 (52%) had lesions outside the area that was imaged by the CT. Ten out of the 14 (71%) MRIs done had lesions outside the region of interest, and 22 out of 33 (67%) of the ultrasounds had lesions outside the area of interest that were detected on FDG PET/CT. In total, in 105 of the 216 (48.6%) anatomic‐based imaging studies that were done within two weeks of FDG PET/CT, the FDG PET/CT detected lesions outside the region of interest imaged by the anatomic‐based study.
136 137
8
Chapter Eightfor the FDG PET/CT was staging, FDG PET/CT detected IFI lesions in five patients that were not previously known, this was also considered to add value. In the 10 patients with the indication of number that FDG PET/CT detected IFI lesions outside the region of interest of the anatomical‐based study
Modality Done within two weeks of
FDG PET/CT (percentage of
Chest X‐ray 80 (37%) 74 (92.5%) 37 (42.3%)
HR CT and thoracic CT 62 (29%) 62 (100%) 22 (35.5%)
Extra thoracic CT scan 27 (13%) 27 (100%) 14 (51.8%)
MRI 14 (6%) 11 (78.6%) 10 (71.4%)
Ultrasound imaging 33 (15%) 30 (90.9%) 22 (66.7%)
Overall 216 (100%) 204 (94.4%) 105 (48.6%)
Table 6: Patients that FDG PET/CT was considered to add value
Value added Number (%)
Assess response to therapy 32 (44%)
Detected previously undiagnosed sites of IFIs 5 (7%) Unexplained fever leading to biopsy guided diagnosis of IFI 4 (5%)
Evaluation for ASCT 7 (10%)
Metabolic activity of residual IFI lesion on anatomic‐based imaging 6 (8%) Total number of patients FDG PET/CT added value 54 (74%)
Discussion
There are currently many imaging procedures the clinician has available to image IFIs. In this current study, we demonstrate that FDG PET/CT adds value to anatomic‐based studies. We also showed that
8
A retrospective study evaluating the value of PET in the management of IFIs
FDG PET/CT was frequently concordant with other imaging studies modalities that are commonly used in the assessment of pathology due to IFIs.
The diagnosis of fungal infections is very challenging. The direct isolation of fungi from sterile site is not always possible. Antifungal therapy is often initiated in patients without proven fungal infection.
This strategy helps reduce the morbidity and mortality associated with delay in initiating treatment for antifungal infection. However, this may lead to the unnecessary administration of antifungal treatment, which may be associated with adverse effects and development of resistance. Pre‐emptive therapy which is not diagnosis driven is often used. The EORTC/MSG has a revised classification for the diagnosis of IFIs. In this classification, patients may have proven, probable or possible IFIs. The patients with probable IFI had clinical features of IFI, indirect microbiologic evidence such as galactomannan or β‐1,3 D‐glucan and imaging features of IFI. The patients with possible IFI have clinical features and imaging features of IFI in the absence of indirect microbiologic evidence [9]. The identification and monitoring of antifungal treatment of IFI lesions in the probable and possible group would be of great value to clinicians. FDG PET/CT would identify IFI lesions and assess disease activity over time. This will help to monitor response to antifungal therapy, identify early if the therapy needs to be changed and may even identify alternative pathology by directing biopsy in unresponsive lesion [3, 11]. In our study, 45% of patient with either possible or probable IFIs had repeated FDG PET/CT scan to monitor therapy.
FDG PET/CT as a whole‐body imaging procedure enables the detection of IFI from different sites of the body in a single imaging session. The metabolic uptake from the PET study makes it very useful for monitoring disease activity over time [12]. In contrast, anatomic‐based imaging procedures are usually limited to one region per study. The plain chest radiograph is readily available, and often, one of the earliest tests requested as IFIs commonly affects the lungs. CT scan and MRI can acquire whole body images, however, this is not routine clinical practice and this study was done to determine the added value of FDG PET/CT to anatomic‐based imaging used in the routine clinical setting [13].
The plain chest radiograph, however, is less useful compared to other procedures such as CT of the chest in the early stages of IFI’s [14, 15]. In our study, FDG PET/CT detected IFI lesions in the chest that had a chest X‐ray reported as normal within two weeks.
Figure 2: 63‐year‐old female with acute myeloid leukemia with on treatment for pulmonary aspergillosis. FDG PET/CT was done because there was unexplained fever and the IFI lesions seen on HR CT (A) were similar to HR CT scan done a month and 2 months earlier. FDG PET/CT scan (B) done within two weeks of (A) showed multiple IFI lesions outside the thorax.
138 139
8
Chapter Eight
CT of the chest, especially HR CT, has been useful in the management of IFIs and has been incorporated in international guidelines for the management of IFIs [9, 15, 16]. Our study found an excellent concordance between the HR CT and the thoracic findings of FDG PET/CT. The lung is the most common site of IFIs. Some authors have recommended the combination of HR CT and FDG PET/CT when evaluating pulmonary lesions to increase sensitivity and specificity [17]. In our study, we demonstrated concordance of the CT of both thoracic and extra thoracic CT with FDG PET/CT. FDG PET/CT has CT as part of the data it acquires. The CT component was a non‐contrasted enhanced diagnostic CT and was able to detect all lesions due to IFI. In addition, our study found that FDG PET/CT detected lesions outside the field of view of CT in 36% and 52% of the thoracic and extra thoracic CT representing 40%
(36/89) of total number of CT scans performed within two weeks, to add value to the CT study.
MRI is another crucial imaging modality for the evaluation of IFIs. In the brain and sinuses, where physiological FDG uptake may compromise the detection of IFI foci by PET/CT, MRI is beneficial. The excellent soft‐tissue resolution of MRI allows better visualization of small thin structures such as the meninges and the different acquisition sequences useful in IFI diagnosis [2, 18]. The advantages of each imaging procedure must be carefully considered when imaging a patient with known or suspected IFIs taking into account the limitation of each imaging study. In our study, MRI detected abnormal brain IFI lesions, which were not detected by FDG PET/CT. FDG PET/CT, however, detected lesions IFI lesions outside the region of interest imaged by IFI (Table 5). FDG PET/CT may add value to MRI imaging by detecting sites of disease outside the region evaluated by the MRI.
Ultrasound (US) of the abdomen is also used in imaging IFI located in the liver, spleen, or kidneys in IFI’s like chronic disseminated candidiasis. Ultrasound imaging can be easily and quickly done at the bedside. However, it lacks sensitivity in patients with neutropenia and is operator dependent [6, 19].
FDG PET/CT, as shown in our study, was useful for the detection of IFI both at the site and beyond the region of interest of the ultrasound study (Table 5).
Occult lesions in IFIs could be potentially devastating if intense immunosuppressive procedures such as ASCT are considered as part of their management [20]. Moreover, anatomic lesions may persist in patients with IFIs despite the resolution of the IFI that may lead to undue prolongation in antifungal therapy. As metabolic changes precede anatomic changes, FDG PET/CT can provide information to enable clinical decisions on antifungal treatment to be done as quickly as possible [3, 21]. These changes will allow clinicians to switch treatment when therapy is ineffective and stop treatment early to avoid adverse effects and reduce the duration and cost of antifungal therapy.
The findings of our study compare favorably with other studies in literature. Douglas et al. compared the FDG PET/CT to conventional CT in detecting IFI and guiding management of the infection [22]. In their study, FDG PET/CT localized occult disease and detected IFI dissemination to other organs in 40%
and 38% of patients respectively. This indirectly compares with our findings where we detected IFI lesions outside the region imaged by the CT scan in 40% of the FDG PET/CT performed within two weeks of the CT. The study by Douglas et al. also found that FDG PET/CT detected significantly more sites of IFI dissemination compared to CT (35% VS. 5%, p < 0.001). The authors concluded that FDG
and 38% of patients respectively. This indirectly compares with our findings where we detected IFI lesions outside the region imaged by the CT scan in 40% of the FDG PET/CT performed within two weeks of the CT. The study by Douglas et al. also found that FDG PET/CT detected significantly more sites of IFI dissemination compared to CT (35% VS. 5%, p < 0.001). The authors concluded that FDG