Head-to-Head Comparison Between 18 F-FDG PET and Leukocyte Scintigraphy to Monitor Treatment Responses in Necrotizing Otitis Externa

Abstract

Background

Necrotizing otitis externa (NOE) is a rare disease associated with high morbidity and mortality, and there is currently no available accurate biomarker to assess treatment responses. The aim of the current study was to evaluate and directly compare the diagnostic performances of 18-Fluoro-deoxyglucose positron emission tomography (¹⁸F-FDG PET) and labeled leukocyte scintigraphy (LS) to monitor treatment responses in NOE.

Methods

Consecutive patients with NOE who underwent ¹⁸F-FDG PET at the end of antibiotic therapy and planar as well as single photon emission computed tomography-labeled leukocyte scintigraphy after completing the initial antibiotic treatment were retrospectively included. Semiquantitative analyses were performed to determine the ratios of affected/nonaffected sides for PET and 4 hour and 24 hour LS acquisitions as well as the kinetic PET ratios (at diagnosis and post-treatment) and LS (4 and 24 hours). The final treatment responses were assessed by 2 experienced ENT physicians based on clinical, otoscopic, and biological data and subsequent 3-month follow-up.

Results

Seventeen patients (74.0 ± 10.6 years old, 5 women) were included. The best diagnostic performances were obtained with the PET maximum standardized uptake value (SUV_max)-lesion-to-background ratio and the tomographic LS lesion-to-background ratio at the 4-hour acquisition timepoint (thresholds of 4.1 and 1.19, yielding accuracies of 100% and 88%, respectively). In the multivariate analysis, the PET SUV_max-lesion-to-background ratio was the only predictive factor of recovery when associated with all clinical parameters (P < .001).

Conclusion

¹⁸F-FDG PET is the first-line imaging modality for evaluating NOE treatment responses, with excellent diagnostic performances. LS with only 4-hour acquisitions appeared to suffice to evaluate NOE treatment responses. Both biomarkers constitute early prognostic biomarkers for predicting antibiotic treatment response in patients with NOE.

Trial registration

The institutional ethics committee (Comité d’Ethique du CHRU de Nancy) approved the evaluation of retrospective patient data, and the trial was registered at ClinicalTrials.gov (n°2023PI003-404).

Graphical abstract

Background

Necrotizing otitis externa (NOE), formerly known as malignant otitis externa, is a rare but severe pathology. NOE is an osteitis that predominantly affects diabetic, elderly, or immunocompromized patients and may progress to osteomyelitis. Its severity predominantly stems from complications: nerve palsy, meningitis, venous sinus thrombosis and extensive basicranial infection, ¹ with the latest studies reporting a mortality rate of approximately 8%. ² Cohen and Friedman proposed the first diagnosis criteria for NOE in 1987. ³ These included clinical signs such as otalgia, otorrhea, granulation tissue, or edema and radiobiological signs such as a positive technetium-99 m bone scan or signs suggestive of a localized Pseudomonas aeruginosa infection. The optimal treatment modality remains unclear in terms of the type, route (oral or intravenous), and duration of antibiotherapy. Most of the recent studies opt for a 6-week, combining oral and intravenous treatments^4,5 to obtain a quick auditory canal recovery. Biological signs of infection also promptly return to normal, but osteitis may take longer to resolve. Due to its well-known radio-clinical delay, the computed tomography (CT) bone scan is a flawed choice to establish complete healing. Signs of osteolysis persist weeks after osteitis. ⁶ Consequently, clinical and biological markers should not be used in isolation to evaluate full recovery and need follow-up in clinical practice to ascertain full recovery. There is currently a lack of accurate biomarkers to identify good treatment responders early, limiting the prescription of noneffective prolonged antibiotic therapies and selecting patients at risk of nontreatment response, requiring a close follow-up.

Nuclear medicine techniques such as technetium-99 m bone scan and gallium-67 or, more recently, leukocyte scintigraphy (LS) have been proposed as imaging biomarkers.^7-10 LS has shown good results in terms of sensitivity and specificity for osteitis and is now regularly used for bone and prosthesis infections. ¹¹

Interestingly, positron emission tomography with a 2-deoxy-2fluoro-D-glucose CT ( ¹⁸ F-FDG-PET/CT) has been proposed as a reliable imaging modality for treatment cessation in NOE. ^{12 18} F-FDG-PET/CT has the advantage of being more accessible and less expensive than LS. However, to the best of our knowledge, no study has to date evaluated and directly compared the performances of LS and ¹⁸F-FDG-PET/CT for providing the rationale of when to stop antibiotic treatment.

The current study therefore aimed to evaluate and directly compare the diagnostic performances of ¹⁸F-FDG-PET/CT and LS for monitoring NOE treatment responses.

Methods

Study Population

All patients with NOE referred to our department between December 2020 and March 2023 for planar and SPECT/CT ^99mTc-HMPAO-LS as well as an 18-Fluoro-deoxyglucose positron emission tomography ( ¹⁸ F-FDG PET) scan to monitor antibiotic therapy responses within 1 month of completing treatment were retrospectively included in the study. ¹⁸F-FDG PET scans were performed at least 1 week after the end of antibiotic therapy, as recommended.^13-15 The institutional ethics committee (Comité d’Ethique du CHRU de Nancy) approved the evaluation of retrospective patient data, and the trial was registered at ClinicalTrials.gov (n°2023PI003-404). This research complied with the principles of the Declaration of Helsinki. Informed consent was obtained from all individuals included in the study.

ENT Consultations

Patients attended at least 3 ENT specialist consultations, the first at the initial diagnosis (V1), the second at the end of the antibiotic treatment (V2) (6 weeks), and the third at 3 months from the end of the antibiotic therapy (V3). Patients requiring external ear calibration had additional consultations for otologic care. The initial consultation considered comorbidities such as diabetes, kidney failure, and immunosuppressive therapy. Standard inflammatory diagnostics included blood count, C-reactive protein, renal function, and tests to confirm that diabetes was well controlled. NOE was suspected based on clinical (fever, otalgia, cranial nerve palsy) and otoscopic criteria (otorrhea, inflammation of the external auditory canal, osseous exposition, granulation tissue). The NOE diagnosis was confirmed by the detection of osteitis, with/without external ear edema or middle-ear opacities, on the CT scan. MRI is not systematic for NOE diagnosis in our practice and has been realized only in 3 patients: in 1 case to confirm osteitis, due to the absence of clear osseous erasure on CT scan and in 2 cases because of neurological impairments associated with NOE at the diagnosis with extended osteitis, an association that makes malignancy more credible as an alternative diagnosis. Bacteriological ear canal samples were systematically analyzed. Biopsies were only taken for the pathological examination of granulation tissue as a potential differential diagnosis for tumors. After the definitive NOE diagnosis, an antibiotic treatment combining intravenous ceftazidime and oral ciprofloxacin was prescribed for a duration of 6 weeks. The second and third consultations were scheduled at the end of the antibiotic therapy (V2) and at 3 months from the end of this therapy (V3) to search for the resolution of the infection (based constantly on clinical and otoscopic parameters and biological parameters at V2). Patients were considered to be still infected if any one of the previously defined clinical or otoscopic signs persisted. The recovery of each individual patient was assessed by the combination of clinical, otoscopic, and biological recovery at V2 associated with the persistence of recovery during the V3 consultation. Biological recovery was assessed as the normalization of C-reactive protein (CRP) at the end of the treatment if it was high at diagnosis, or the persistence of CRP’s negativity. At these points, the criteria for scoring a recovery were solely based on the clinical and otoscopic signs previously described, and ENT specialists were blinded to the imaging results, including morphological imaging such as CT or MRI scans.

^99mTc-HMPAO leukocyte scintigraphy

A ^99mTc-HMPAO LS was planned for each patient at the end of the antibiotic therapy as for the ¹⁸F-FDG PET scan. As recommended for LS, autologous leukocytes were isolated from peripheral blood and radiolabeled ex vivo and then reinjected back into the same patient for imaging. Leukocytes were labeled with Technetium-99 m-hexamethylpropyle-neamineoxime (^99mTc-HMPAO) as recommended by the European guidelines for the labeling of leukocytes with ^99mTc-HMPAO. ^{16 99m}Tc-HMPAO-labeled leukocytes (200-500 Mbq) were injected into the patient intravenously.

LS acquisitions included both planar and single photon emission computed tomography (SPECT) associated with a CT scan. A 5-minute planar acquisition was performed 30 minutes after the injection of radiolabeled leukocytes for quality control purposes to determine the lung, liver, and spleen uptake and calculate the liver-to-spleen ratio (normal range greater than 0.5). Planar and SPECT/CT images centered on the head and neck were subsequently obtained 4 and 24 hours after the injection of ^99mTc-HMPAO-labeled leukocytes using a gamma camera (Symbia Siemens®, Germany, MUNICH) with low-energy high-resolution collimators (centered on the peak of 140 keV using a 15% window). Fixed time and fixed duration planar and SPECT/CT acquisitions were performed using a 128 × 128 matrix. Planar images were acquired for 240 seconds at 4 hours post-IV and 1200 seconds at 24 hours post-IV. SPECT images consisted of 32 projections over 360° for 30 seconds per projection at 4 hours post-IV and 45 seconds per projection at 24 hours postinjection. Three-dimensional (3D) images were reconstructed using the ordered subset expectation maximization algorithm. CT scans were performed immediately after SPECT. Scanning parameters were 130 kV, 90 mAs, and 6 × 1 mm collimation. For CT scans, bone-weighted (B70, Medium Sharp) reconstructions were performed to obtain 1.25 mm slices.

¹⁸ F-FDG PET

¹⁸ F-FDG PET scans were performed at the time of the initial diagnosis and at the end of the antibiotic therapy. All ¹⁸F-FDG-PET/CT scans were performed at least 1 week after the completion of the antibiotic therapy to limit the risk of false-negative results.^13-15 PET acquisitions were performed on a digital camera (Vereos, Philips®, Amsterdam) after the intravenous injection of 3 MBq/kg of ¹⁸F-FDG. All participants had fasted at least 6 hours before the injection and had a fasting blood glucose level less than 11 mmol/L, as recommended. ¹⁷ PET scans extended from the vertex to the liver with the patient’s hands down. PET images were obtained from 90-second acquisitions per bed position. PET images were reconstructed using an iterative algorithm (4 iterations, 8 subsets, Gaussian postfilter 5 mm full width at half maximum [FWHM] and 9.6 mm slices), corrected for scattering, random coincidences, and CT-based attenuation and resulted in 2 mm slices displayed in a 512 × 512 matrix. The CT scanning parameters were 120 kV, 100 mAs with mA modulation, and 0.625 mm collimation.

Image Analysis

The semiquantitative LS analysis was performed by a specialist nuclear medicine physician using an 80 mm region of interest (ROI) in the anterior planar projection and a 15 mm volume of interest (VOI) on SPECT/CT images of both affected and nonaffected sides. For SPECT/CT images, the VOIs were applied on 4 different anatomical locations: the external auditory canal, the temporo-mandibular joint, the mastoid, and the base of the skull, as previously reported. ¹² Semiquantitative analyses were performed by calculating the ratios of the number of counts for planar images and the concentrations in kBq/mL for SPECT images of the affected/nonaffected sides within the most severely affected anatomical location and its respective contralateral region at 4 and 24 hours acquisitions. Changes in these ratios between the 4 and 24 hours acquisitions were also determined.

The semiquantitative analysis of the ¹⁸F-FDG PET scans was performed by a specialist nuclear physician using a 15 mm VOI applied to the same anatomical location as the one most affected in LS (checking that it was also the anatomical location most affected in the PET scan). Standardized uptake values (SUVs), defined as the ratio of the radioactive concentration in the respective VOIs and the product of the injected dose and the body volume, were determined from the maximal and mean values of the VOIs on the affected side and the mean values of the VOIs on the unaffected side. Semiquantitative analyses were performed based on the ratios of affected/nonaffected sides for both ¹⁸F-FDG PET at the initial diagnosis and at the end of treatment. Changes between these ratios from the 2 PET scans were also calculated.

Statistical Analysis

Categorical variables are expressed as percentages, and continuous variables are expressed as medians (first and third quartiles). Receiver operating curves of LS and PET semiquantitative parameters were determined to quantify the rate of recovery. Decision cutoffs were selected when the product of sensitivity and specificity reached its maximum. Multivariate binary logistic regression analyses were performed to test for associations between recovery and LS and PET semiquantitative parameters as well as clinical parameters such as age, sex, comorbidities, treatment, type, and duration of symptoms as well as pathological microorganisms involved as confounding factors. A P value <.05 was considered significant. Statistical analyses were performed using IBM, USA, New York SPSS Statistics (Version 25).

Results

Population

A total of 17 patients with a median age of 72.0 [62.5-78] years were included in this longitudinal study. Patient characteristics are summarized in Table 1. The most prevalent NOE risk factors were diabetes (n = 10, 59%) and renal failure (n = 6, 35%), with 5 patients (29%) showing both diabetes and renal failure risk factors. Pseudomonas aeruginosa was the most prevalent etiological agent identified (n = 6, 35%). The median duration of symptoms prior to the first consultation in our tertiary center was 70 [28-108.5] days. At the initial ENT consultation, most patients presented symptoms such as otalgia (n = 12, 70%), otorrhea (n = 13, 76%), inflammation of the external auditory canal (n = 12, 70%), and cranial nerve paralysis (n = 5, 12%). The final ¹⁸F-FDG PET scans were performed at 9 [7;13] days after the end of the antibiotic therapy. LS was performed at 1 [1;5] day from the final ¹⁸F-FDG PET scans. Thirteen patients were considered to have recovered according to the clinical and otoscopic criteria and follow-ups defined. Four patients were deemed not to have recovered: 2 continued the combined ceftazidime + ciprofloxacin antibiotic treatment for at least an additional 6 weeks, 1 was prescribed voriconazole for fungal osteitis, and the last patient had a relapse shortly after discontinuing the initial antibiotic treatment. In our practice, patients are followed for at least 3 months after establishing recovery and then are considered healed. They are not asked to come back again for systematic control. But in practice, patients often ask for further controls, to be reassured. For this particular series, we have now data for a mean delay of 19 months [9-37]. As 100% of healed patients did not relapse in this delay, our data obtained nevertheless in a small series of patients seems to validate our clinical practice.

Table 1.

Patient Characteristics.

Age (years)	72.0 [62.5-78]
Sex (female)	5 (29%)
Risk factors
Diabetes	11 (65%)
Renal failure	6 (35%)
None	6 (35%)
Causative agents
Pseudomonas aeruginosa	6 (35%)
Staphylococcus aureus	2 (12%)
Others	7 (41%)
Not identified	3 (18%)
Symptoms
Otalgia	12 (70%)
Otorrhea	13 (76%)
Inflammation of the external auditory canal	12 (70%)
Cranial nerve paralysis	5 (12%)
Duration of symptoms before treatment (days)	70 [28-108.5]
Duration of antibiotic treatment (days)	45.5 [42-62.5]

Semiquantitative Analyses

For ¹⁸ F-FDG PET at initial diagnosis, the maximum standardized uptake value (SUV_max)-lesion-to-background ratio was 5.37 [3.44;5.87], and the mean standardized uptake value (SUV_mean)-lesion-to-background ratio was 3.20 [1.98;3.56]. These values did not correlate with age, sex, comorbidities, whether images were acquired under treatment, the type/duration of symptoms, or the type of microorganism detected (P > .15).

When comparing the initial PET with the final PET and the kinetics between the 2 time points, the ¹⁸F-FDG PET scans as well as planar or tomographic LS performed at 4 and 24 hours and the respective kinetics between these 2 time points, the SUV_max-lesion-to-background of the final ¹⁸F-FDG PET scan showed a maximal area under the curve (AUC) of 1.0 [1.0;1.0] to predict patient recovery. Other significant imaging parameters to predict patient recovery were the SUV_max-lesion-to-background of the initial ¹⁸F-FDG PET scan, the SUV_mean-lesion-to-background ratio values of the final ¹⁸F-FDG PET scan, and the 4-hour tomographic LS lesion-to-background ratios with respective AUCs of 0.85 [0.63;1.0], 0.96 [0.87;1.0] and 0.90 [0.76;1.0] (P < .05). All the AUC results are presented in Supplemental Table 1. When restricting the analysis to the 14 patients who had discontinued antibiotic therapy at least 1 week prior to LS (Supplemental Table 2), the SUV_max-lesion-to-background of the final ¹⁸F-FDG PET scan remained the best parameter, showing a maximal AUC of 1.0 [1.0;1.0] to predict patient recovery, and both 4 hours planar and tomographic LS lesion-to-background ratios showed significant AUCs to predict recovery.

When comparing the diagnostic performances of different cutoffs, as expected, an SUV_max-lesion-to-background ratio of 4.1 for ¹⁸F-FDG PET resulted in an accuracy of 100% in predicting patient recovery. The 4-hour tomographic LS lesion-to-background ratio of 1.19 presented an accuracy of 88% (sensitivity of 85%, specificity of 100%) to predict patient recovery. Box plots of SUV_max-lesion-to-background ratios in PET and 4-hour tomographic LS lesion-to-background ratios are depicted in Figure 1. Interestingly, an SUV_max-lesion-to-background ratio of the initial ¹⁸F-FDG PET scan of 4.73 was associated with a specificity of 100% for predicting patient recovery. The diagnostic performances of cutoff values for parameters showing significant AUCs are given in Table 2.

Figure 1.

Box plots of the SUV_max-lesion-to-background of the final ¹⁸F-FDG PET scan and the 4-hour tomographic LS lesion-to-background ratios for predicting no recovery (left panel) or recovery (right panel) in patients with NOE according to the clinical follow-up. ¹⁸F-FDG PET, 18-Fluoro-deoxyglucose positron emission tomography; LS, leukocyte scintigraphy; NOE, necrotizing otitis externa; SUV_max, maximum standardized uptake value.

Table 2.

Diagnostic Performances of Cutoff Values for Parameters Showing Significant AUCs.

PET and scintigraphy parameters	Threshold	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)
Initial 18F-FDG PET SUVmax lesion	4.73	69	100	76	50	100
Final 18F-FDG PET SUVmax lesion	4.10	100	100	100	100	100
Final 18F-FDG PET SUVmean lesion	2.10	85	100	88	67	100
Tomographic LS 4 h	1.19	85	100	88	67	100

Abbreviations: ¹⁸F-FDG PET, 18-Fluoro-deoxyglucose positron emission tomography; AUC, area under the curve; LS, Leukocyte scintigraphy; NPV, negative predictive value; PPV, positive predictive value; SUV_max, maximum standardized uptake value; SUV_mean, mean standardized uptake value.

The multivariate analysis only identified the SUV_max-lesion-to-background ratio of the ¹⁸F-FDG PET as a predictive factor of recovery when combined with age, sex, comorbidities, the type/duration of symptoms, and the type of microorganism involved (P < .001).

Figure 2 shows representative ¹⁸F-FDG PET and LS images of patients with NOE who had and had not recovered.

Figure 2.

Axial PET/CT (a, c) and LS SPECT/CT 4 hour image (b, d) slices. (a, b) High uptake (red arrows) in the right ear 6 weeks after treatment in a 64-year-old female patient: persistent infection. (c, d) No significant uptake (blue arrows) in the left ear 6 weeks after treatment in a 68-year-old male patient: nonpersistent infection. CT, computed tomography; PET, positron emission tomography; LS, leukocyte scintigraphy; SPECT, single photon emission computed tomography.

Discussion

Our current study demonstrates the outstanding diagnostic performances of ¹⁸F-FDG PET and, to a lesser extent, LS for monitoring antibiotic therapy in NOE based on the threshold ratios evaluated and focalizing on the 4 hour postinjection acquisitions for LS. Nuclear medicine could thus be pivotal in monitoring NOE therapy by providing compelling prognostic biomarkers of patient recovery. These imaging biomarkers would allow us to limit the prescription of prolonged noneffective antibiotic therapy in patients or to identify poor treatment responders who require close clinical follow-ups early.

After 6 to 8 weeks of antibiotic treatment, ¹⁸F-FDG PET showed very high diagnostic performances for predicting recovery from NOE, with an accuracy of 100% based on an SUV_max-lesion-to-background ratio threshold of 4.1 (Table 2, Supplemental Table 1, Figure 1). Using a high ¹⁸F-FDG PET threshold value presumably guards against obtaining false-positive results, particularly due to inflammation.^18-20 Inflammatory processes, which are frequently associated with NOE, indeed lead to high glycolytic avidity.^18-20 Importantly, ¹⁸F-FDG PET was the only predictive factor of recovery when combined with all clinical factors, such as age, sex, comorbidities, the type/duration of symptoms, and the type of microorganism involved. It is therefore an invaluable prognostic biomarker for monitoring NOE during treatment. Our results are consistent with the potential role of ¹⁸F-FDG PET-CT in a series of 8 patients who underwent ¹⁸F-FDG PET for monitoring antibiotic treatment of NOE. ¹² Compared with LS, ¹⁸F-FDG PET offers better spatial resolution, lower costs, and greater availability. Moreover, ¹⁸F-FDG PET is well accepted by patients with NOE, who are often frail and elderly, with only 1 intravenous injection and a total examination time of less than 2 hours. Interestingly, an ¹⁸F-FDG PET scan performed at the initial NOE diagnosis provides 100% specificity by using a threshold of 4.73, which is of informative value (Table 2). Even if the observed PET uptake was highly variable between patients and did not correlate with clinical factors such as age, sex, comorbidities, treatment, type/duration of symptoms, and type of microorganism involved, this high specificity could identify poor responders to antibiotic treatment early. This predictive value needs to be further explored but is in line with the recommendations proposed for other infections, such as peripheral bone, ²¹ prosthetic joint, ²² and vascular graft ¹³ infections.

LS showed lower prognostic performances when compared to ¹⁸F-FDG PET, with an accuracy of 88%, based on a 4 hours lesion-to-background ratio threshold of 1.19 (Table 2, Supplemental Table 1, Figure 1). Our results are consistent with a study that reported an 89% accuracy for the visual analysis based on planar and SPECT LS absorption kinetics between acquisitions at 4 and 24 hours after injection in a population of 27 patients with a relatively low rate of nonrecovery (only 15%). ¹⁴ Interestingly, we found better performances in predicting patient recovery with acquisitions performed at 4 hours postinjection than at 24 hours postinjection in planar and SPECT acquisitions. This likely reflects the rapid uptake kinetics of leukocytes in NOE, as has been previously reported in vascular graft infections. ¹³ Restricting acquisitions to only 4 hours postinjection would facilitate the LS procedure for frail and elderly patients with NOE. It is important to note that 3 patients underwent LS less than 8 days after stopping antibiotic treatment, which is not in accordance with the recommendations.^13-15 However, excluding these 3 patients had no impact on the diagnostic performances of LS, as shown in Supplemental Table 2.

To the best of our knowledge, our study is the first to directly compare the diagnostic performances of ¹⁸F-FDG PET and LS to monitor NOE treatments. The current study provides early prognostic biomarkers of antibiotic therapy responses in patients with NOE, which fills a gap in the current assessment of such patients.

Moreover, our study is the first to evaluate the prognostic role of ¹⁸F-FDG PET in a series of 17 patients compared with a previous study that examined 8 patients with NOE by 18F-FDG PET. ¹² Our outcome for the recovery from NOE was only based on clinical and otoscopic symptoms and by definition excluded any recourse to imaging techniques. All patients were nevertheless followed clinically with a mean delay of 19 months after completing the initial antibiotic therapy to ascertain that they were cured. All analyses were based on semiquantitative approaches to provide objective results. Our study was, however, limited by the small number of patients included, reflecting the rarity of this disease.

Conclusion

¹⁸ F-FDG PET is the first-line imaging modality for evaluating NOE treatment responses, with excellent diagnostic performances. LS with only 4 hour acquisitions appeared to suffice to evaluate NOE treatment responses. Both biomarkers constitute early prognostic biomarkers for predicting antibiotic treatment response in patients with NOE. Large multicenter trials using ¹⁸F-FDG PET are needed to validate these preliminary results.

Supplemental Material

sj-docx-1-ohn-10.1177_19160216241288810 – Supplemental material for Head-to-Head Comparison Between 18F-FDG PET and Leukocyte Scintigraphy to Monitor Treatment Responses in Necrotizing Otitis Externa

Supplemental material, sj-docx-1-ohn-10.1177_19160216241288810 for Head-to-Head Comparison Between 18F-FDG PET and Leukocyte Scintigraphy to Monitor Treatment Responses in Necrotizing Otitis Externa by Moïra Hurstel, Alice Vasseur, Saifeddine Melki, Nicolas Veran, Laetitia Imbert, Duc Trung Nguyen, Cécile Rumeau and Antoine Verger in Journal of Otolaryngology - Head & Neck Surgery

Supplemental Material

sj-docx-2-ohn-10.1177_19160216241288810 – Supplemental material for Head-to-Head Comparison Between 18F-FDG PET and Leukocyte Scintigraphy to Monitor Treatment Responses in Necrotizing Otitis Externa

Supplemental material, sj-docx-2-ohn-10.1177_19160216241288810 for Head-to-Head Comparison Between 18F-FDG PET and Leukocyte Scintigraphy to Monitor Treatment Responses in Necrotizing Otitis Externa by Moïra Hurstel, Alice Vasseur, Saifeddine Melki, Nicolas Veran, Laetitia Imbert, Duc Trung Nguyen, Cécile Rumeau and Antoine Verger in Journal of Otolaryngology - Head & Neck Surgery