Comparison of [¹⁸F]fluorodeoxyglucose and [⁶⁸Ga]Gallium DOTA-TATE in patients with active giant cell arteritis

Abstract

Purpose

[¹⁸F]Fluorodeoxyglucose (FDG) is widely used in PET/CT imaging to detect large vessel vasculitis in giant cell arteritis (GCA), but its performance is suboptimal in patients receiving glucocorticoids. We aimed to compare [⁶⁸Ga]Ga-HA-DOTA-TATE, a somatostatin 2-analogue tracer, to [¹⁸F]FDG in a pilot study of patients with GCA.

Methods

Eight patients with active GCA were prospectively, sequentially scanned with both [¹⁸F]FDG PET/CT and [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT imaging. Images were evaluated by 2 blinded nuclear medicine specialists. Tracer uptake was assessed in 8 vascular territories using SUVmax, and target-background ratios (TBR) were calculated using both right atrium (TBR_RA) and liver mean (TBR_liver). Mean SUVmax and TBR of individual vascular territories and index vessels were compared.

Results

The patient median age was 71.5 years (range 64–82), and 4 (50%) were women. Active vasculitis (≥ grade 2 visual uptake in large vessels) was present in 62.5% of [¹⁸F]FDG scans. [¹⁸F]FDG scans had higher RA background activity than [⁶⁸Ga]Ga-HA-DOTA-TATE (mean RA SUVmean 1.88 vs. 0.36, p < 0.001), while [⁶⁸Ga]Ga-HA-DOTA-TATE had a significantly higher liver uptake (mean liver SUVmean 7.54 vs. 2.39, p < 0.001). Vascular uptake (as measured by both SUVmax and TBR_liver) was significantly higher in [¹⁸F]FDG than [⁶⁸Ga]Ga-HA-DOTA-TATE scans in every vascular territory (p < = 0.05 for all comparisons), including index vessels (SUVmax 4.04 vs. 1.91, p = 0.01, TBR_liver 1.73 vs. 0.27, p < 0.001).

Conclusion

In this pilot study of patients with active GCA, the arterial uptake of [⁶⁸Ga]Ga-HA-DOTA-TATE was lower and less conspicuous compared to [¹⁸F]FDG. While further evaluation in larger cohorts is needed, a clear advantage of [⁶⁸Ga]Ga-HA-DOTA-TATE over [¹⁸F]FDG for detecting vascular inflammation in GCA was not identified.

Trial registration, ClinicalTrials.gov

NCT 03812302, registered 2019-01-18, URL: https://clinicaltrials.gov/search?cond=dotatate%20%26;term=giant%20cell%20arteritis.

Introduction

Giant cell arteritis (GCA) is a granulomatous vasculitis of older persons, that damages the medium and large arteries (Weyand and Goronzy 2003). Due to early risks of permanent blindness and stroke, patients with suspected GCA are frequently treated with glucocorticoids while diagnostic testing is pursued.

Nuclear medicine permits visualization of biologic processes within the arterial wall, potentially allowing for an earlier diagnosis of vasculitis. [¹⁸F]fluorodeoxyglucose (FDG) PET/CT has a high reported sensitivity (90%) and specificity (98%) for the diagnosis of GCA (Soussan et al. 2015), and there are published acquisition and reporting standards (Slart and Writing 2018). However, glucocorticoids appear to inhibit [¹⁸F]FDG uptake by inflammatory cells and increase hepatic uptake, thus reducing the sensitivity of [¹⁸F]FDG imaging once therapy is initiated (Nielsen et al. 2018). Radiolabelled ⁶⁸Ga-high affinity-DOTA-TATE ([⁶⁸Ga]Ga-HA-DOTA-TATE), originally developed for the imaging of neuroendocrine tumours (Sanli et al. 2018), binds specifically to the somatostatin receptor subtype-2 (SSTR₂) upregulated in M1 proinflammatory macrophages (Helgebostad et al. 2022) and also the SSTR₅, expressed in activated T cells (Talme et al. 2001; Schottelius et al. 2014). Additionally, [⁶⁸Ga]Ga-DOTA-TATE shows minimal uptake in the myocardium and un-inflamed blood vessels, making it of interest for imaging vasculitis (Sanli et al. 2018). To date, the performance of [⁶⁸Ga]Ga-DOTA-TATE PET imaging has been evaluated in small studies of atherosclerosis (Tarkin et al. 2017) and in a few patients with large vessel vasculitis (LVV) with promising results (Corovic et al. 2023). However, so far there has been no direct comparison of [⁶⁸Ga]Ga-HA-DOTA-TATE with [¹⁸F]FDG for LVV detection.

In this pilot study we aimed to prospectively compare [⁶⁸Ga]Ga-HA-DOTA-TATE to [¹⁸F]FDG PET/CT for detecting medium and large vessel inflammation in patients with active GCA receiving glucocorticoids.

Patients and methods

Patients

Consecutive patients with active GCA were prospectively approached for enrolment from rheumatology practices at the University of Alberta, and community rheumatology practices in Edmonton, Canada. For study inclusion, patients were identified by their treating rheumatologist and were required to have a diagnosis of active new or relapsing GCA (Unizony et al. 2013), as defined by the presence of all 3 of the following criteria: (1) age ≥ 50 years at disease onset, (2) the presence of unequivocal symptoms of GCA (defined as new headache, scalp or temporal artery tenderness, ischemia-related vision loss, or jaw/mouth claudication) or polymyalgia rheumatica (defined as bilateral shoulder and/or hip girdle pain associated with inflammatory stiffness), and (3) confirmation of vasculitis by either a positive temporal artery biopsy or imaging evidence of large-vessel vasculitis, confirmed on either magnetic resonance angiography (MRA), computed tomography angiography (CTA), or [¹⁸F]FDG positron emission tomography-computed tomography (PET-CT) by an imaging specialist. Active GCA was defined as the presence of active symptoms (of either GCA or polymyalgia rheumatica) at the time of enrollment requiring initiation of glucocorticoids (or increase in baseline glucocorticoid dose, in the case of relapses) as determined by the treating physician. All patients were required to have initiated glucocorticoids or have had a dose escalation within 2 weeks of study enrolment. Ethics approval was obtained from the University of Alberta, and all patients provided written informed consent for study participation. The study protocol was registered on Clinicaltrials.gov, # NCT03812302. Although the initial study protocol called for enrolment of 15 patients (to simply acquire preliminary results in a pilot trial), due to slow enrolment during the COVID 19 pandemic and absence of clear positive results with [⁶⁸Ga]Ga-HA-DOTA-TATE (see results section below) enrolment was stopped early after 8 patients had completed the study (see Fig. 1.)

[¹⁸F]FDG studies.

[¹⁸F]FDG PET/CT imaging was performed as soon as possible, median time 0.5 days (0–16), after enrolment. Patients received weight-based [¹⁸F]FDG intravenously (per institutional protocol) followed by CT and PET imaging 60 min later. Scans were performed using a GE MI Discovery PET/CT system (GE HealthCare, Chicago, Il, USA) from vertex to knees. Unless contraindications were present, bolus intravenous CT contrast (Omnipaque 350, weight-based volume per institutional standard) was administered with a saline flush, followed by a diagnostic multidetector CT angiographic acquisition through the same anatomic region (either arterial delay or with bolus tracking, per institutional practice at the time of the examinations).

[⁶⁸Ga]Ga-HA-DOTA-TATE studies.

[⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT imaging was performed in all patients as close as possible temporally to [¹⁸F]FDG PET/CT (per protocol within 4 weeks of [¹⁸F]FDG PET/CT scan), at a median of 11 days (3–23) from enrolment. Patients received an intravenous infusion of 3MBq/kg of [⁶⁸Ga]Ga-HA-DOTA-TATE (maximum of 200 MBq). After 60 min, a whole-body PET acquisition was obtained as described above, with low dose CT through the same anatomic region for attenuation correction.

Tracer interpretation. All images were reviewed using Segami Oasis workstations (Segami Corporation, Columbia, MD, USA). Tracer uptake in 8 vascular territories [thoracic aorta, abdominal aorta, coronary arteries (left anterior descending or LAD, right coronary or RCA and circumflex or LCx), bilateral temporal, carotid, vertebral, subclavian, and iliac/femoral arteries) was assessed both visually and semi-quantitatively by 2 experienced nuclear medicine physicians blinded to clinical data, and results were averaged. [¹⁸F]FDG visual uptake scores were graded 0–3, by comparing arterial uptake to that of the liver (0 = none, 1 < liver, 2 = liver, 3 > liver), as previously published (Slart and Writing 2018), and [⁶⁸Ga]Ga-HA-DOTA-TATE visual uptake scores were graded 0–2 as follows: 0 = none, 1 = low grade (barely visible tracer uptake above background), 2 = clearly positive (easily visible above background.). Scans with visual [¹⁸F]FDG uptake ≥ 2 were considered positive for LVV.

Semi-quantitative scoring was determined by measuring the maximum standardized uptake value (SUVmax) of each target vessel (or the hottest of each paired vessel or coronary artery). The index vessel was defined as the artery with the highest SUVmax for each individual patient. Target-to-background ratios (TBR) were then calculated for each vessel (or the hottest of each paired vessel or coronary artery) in 2 ways: (1) by dividing the target vessel SUVmax by the SUVmean of the R atrium (TBR_RA), and (2) by dividing the target vessel SUVmax by SUVmean of the liver (TBR_liver).

Follow up Imaging: The initial study protocol planned for repeat [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT imaging to be performed at 6 months from enrollment, however, the study plan was altered after the emergence of the COVID 19 pandemic. Due to a safety concerns related to requesting elderly and immunosuppressed patients come to the hospital for additional research imaging scans, the protocol was amended to allow 6 months follow-up imaging only if the baseline [⁶⁸Ga]Ga-HA-DOTA-TATE scans revealed clear-cut (grade 2) visual vascular uptake. Since no baseline scan met this requirement, no follow-up scans were perfomed.

Statistical analysis

Descriptive statistics (median and range, mean and standard deviation) were used to characterize enrolled patients. The mean liver and right atrium SUVmean (± SD) were reported for [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE scans. Mean (± SD) SUVmax, TBR_RA and TBR_liver values in each of 8 vascular territories, and index vessels were reported and compared between FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE scans using paired student’s t-tests. P values < 0.05 were considered statistically significant.

Results

Twenty-two consecutive patients were screened for study eligibility, of whom 8 patients with active GCA completed both [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT imaging. See Fig. 1 for flow diagram of enrolled study participants. Median patient age was 71.5 years (range 64–82), and 4 (50%) were women. As per the inclusion criteria, all included participants were receiving glucocorticoids, begun within 2 weeks of enrolment, at a median dose of 60 mg/day (range 25–60 mg/day). See Table 1 for the clinical characteristics of the enrolled participants.

Fig. 1

Flow diagram of enrolled participants. Of 22 patients with active GCA screened for study participation, ultimately 8 were enrolled and completed both baseline [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT scans. Clear-cut (> grade 2) visual vascular [⁶⁸Ga]Ga-HA-DOTA-TATE uptake was not appreciated on any baseline PET/CT therefore 6 month follow up [⁶⁸Ga]Ga-HA-DOTA-TATE scans were not performed in any participant. Abbrev: wks (weeks), mos (months)

Table 1 Baseline demographics and clinical characteristics of GCA patients

Qualitative results

Tracer uptake was assessed in a total of 120 individual vessels. On visual inspection, ≥grade 2 vascular uptake was observed in at least 1 vascular territory in 5 of the 8 [¹⁸F]FDG scans (62.5%). With respect to [⁶⁸Ga]Ga-HA-DOTA-TATE scans, low blood pool activity (almost negligible) and marked hepatic activity were noted on all scans. Visual [⁶⁸Ga]Ga-HA-DOTA-TATE vascular uptake was barely perceptible above background activity in all cases, with no scans demonstrating grade 2 visual uptake. Accordingly, 6 month follow-up [⁶⁸Ga]Ga-HA-DOTA-TATE scans were not performed in any participant. See examples of [⁶⁸Ga]Ga-HA-DOTA-TATE and corresponding [¹⁸F]FDG PET images in Figs. 2, 3 and 4.

Fig. 2

Comparison of [⁶⁸Ga]Ga-HA-DOTA-TATE and [¹⁸F]FDG PET/CT images. Maximal Intensity Projection (MIP) [⁶⁸Ga]Ga-HA-DOTA-TATE (A) and [¹⁸F]FDG (B) PET images in patient N.6. (A) Trace [⁶⁸Ga]Ga-HA-DOTA-TATE uptake in large arteries with intense liver and gastrointestinal uptake. (B) Grade 3 uptake of [¹⁸F]FDG in both subclavian arteries and the ascending aorta

Fig. 3

Additional [⁶⁸Ga]Ga-HA-DOTA-TATE and corresponding [¹⁸F]FDG PET/CT images. Panels A-D show [⁶⁸Ga]Ga-HA-DOTA-TATE images. A. MIP of total body uptake. B: Axial transmission image. Both A and B show minimal vascular uptake of [⁶⁸Ga]Ga-HA-DOTA-TATE with high liver uptake. C: Axial thoracic CT image showing the ascending aorta, main trunk and right pulmonary artery. D: [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT fusion images again showing barely visible vascular uptake of [⁶⁸Ga]Ga-HA-DOTA-TATE. Panels E-H are corresponding images of the same patient obtained with [¹⁸F]FDG. This tracer demonstrates intense vascular uptake and much lower liver uptake

Fig. 4

[⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT fusion images and corresponding [¹⁸F]FDG images. Panels A-A show [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT fusion images. A. Sagittal fusion image, C: coronal fusion image, C: axial fusion image. A-C show intense liver uptake and minimal vascular uptake. Panels D-F show corresponding [¹⁸F]FDG PET/CT fusion images in same patient. [¹⁸F]FDG demonstrates intense vascular uptake and also florid interspinous bursitis in the lumbar spine in keeping with concomitant active polymyalgia rheumatica

Semi-quantitative results

Mean SUV_max values were statistically significantly higher in every vascular territory and index vessel in [¹⁸F]FDG as compared to [⁶⁸Ga]Ga-HA-DOTA-TATE scans (see Fig. 5A and B and Supplemental Table 1 for p-values.) As expected, mean liver SUV_mean was significantly lower in [¹⁸F]FDG scans than [⁶⁸Ga]Ga-HA-DOTA-TATE scans [2.39 (0.48) vs. 7.54 (1.49), p < 0.001]. Correspondingly, mean TBR_liver values were also statistically significantly higher in every vascular territory (see Fig. 5C and Supplemental Table 2) and index vessels [1.73 (0.74) vs. 0.27 (0.11), p < 0.001, Fig. 5D] using [¹⁸F]FDG compared to [⁶⁸Ga]Ga-HA-DOTA-TATE.

Fig. 5

Mean tracer SUVmax and TBR_liver values for individual vessels and index vessels in [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE scans. (A) Mean SUVmax uptake in individual vascular territories (or hottest of paired territories). For each comparison, values were statistically significantly higher with [¹⁸F]FDG than [⁶⁸Ga]Ga-HA-DOTA-TATE (*=p < 0.05, see Supplemental data for individual p values). (B) Mean SUVmax of Index Vessels in [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE scans. (C) Mean TBR_liver values in [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE scans. For each comparison, values were statistically significantly higher with [¹⁸F]FDG than [⁶⁸Ga]Ga-HA-DOTA-TATE (*=p < 0.05, see Supplemental data for individual p-values), and (D) Mean TBR_liver of Index Vessels in [¹⁸F]FDG and [⁶⁸Ga]Ga-HA-DOTA-TATE tracers. Abbreviations: DOTA ([⁶⁸Ga]Ga-HA-DOTA-TATE), FDG([¹⁸F]FDG), Ao (aorta)

The mean RA SUV_mean was significantly higher in [¹⁸F]FDG scans than [⁶⁸Ga]Ga-HA-DOTA-TATE scans [1.88 (0.48) vs. 0.36 (0.17), p < 0.001], where this value often approached zero. As a result, the calculated mean TBR_RA ratios were significantly higher in several territories using [⁶⁸Ga]Ga-HA-DOTA-TATE as compared to [¹⁸F]FDG (see Supplemental Table 3). However, the mean TBR_RA of index vessels was not significantly different [[⁶⁸Ga]Ga-HA-DOTA-TATE 6.71 (5.31) vs. [¹⁸F]FDG 2.24 (0.90), p = 0.07].

Discussion

In this pilot study, we compared [⁶⁸Ga]Ga-HA-DOTA-TATE to [¹⁸F]FDG for PET/CT diagnosis of large vessel vasculitis in patients with active GCA. Both scans were done prospectively in the same patient to permit direct comparison, and vascular uptake was described in the aorta, major branch vessels, and also temporal and coronary arteries for the first time. Despite the low blood pool background, we found that the vascular wall uptake with [⁶⁸Ga]Ga-HA-DOTA-TATE, while occasionally detectable on visual inspection, was very low in all cases. In contrast, on visual inspection of the [¹⁸F]FDG scans we showed that most patients (62.5%) had increased vascular uptake. The measured arterial uptake (SUVmax and TBR_liver) was overall significantly higher in [¹⁸F]FDG than [⁶⁸Ga]Ga-HA-DOTA-TATE scans in index vessels and every individual vascular territory.

Currently, there is no standardized reporting for vascular [⁶⁸Ga]Ga-HA-DOTA-TATE uptake. Since [⁶⁸Ga]Ga-HA-DOTA-TATE distribution differs from that of [¹⁸F]FDG, we calculated TBRs using both liver and RA as a surrogate for blood pool. Similar to SUVmax, TBR_liver values were significantly higher in every vascular territory and index vessels in [¹⁸F]FDG as compared to [⁶⁸Ga]Ga-HA-DOTA-TATE scans. In contrast, TBR_RA values were significantly higher with [⁶⁸Ga]Ga-HA-DOTA-TATE than [¹⁸F]FDG in several territories, due to the very low (near negligible) denominator used in the TBR_RA calculations.

Little has been published thus far regarding [⁶⁸Ga]Ga-DOTA-TATE imaging in patients with vasculitis. In Corovic et al., TBR measurements (using blood pool comparator only) were calculated in the thoracic aorta and aortic arch branch vessels in a small number of patients with GCA, Takayasu arteritis, and atherosclerosis (Corovic et al. 2023). [⁶⁸Ga]Ga-DOTA-TATE thoracic aorta TBR_max >1.8 was found to distinguish active from quiescent vasculitis with 82% sensitivity and 90% specificity, and [⁶⁸Ga]Ga-DOTA-TATE thoracic aorta TBR_max >1.6 distinguished aortitis from atherosclerosis (Corovic et al. 2023). For comparison, our blood pool-corrected TBRs (i.e.: [⁶⁸Ga]Ga-HA-DOTA-TATE TBR_RA) were much higher (mean [⁶⁸Ga]Ga-HA-DOTA-TATE TBR_RA 4.53 (2.94), and individual thoracic aorta TBR_RA values ranged from 1.95 to 11.5), supporting the clinical impression that our enrolled patients had clinically-active disease. But despite having numerically-higher TBR_RA measurements than what was reported previously, the visual uptake of [⁶⁸Ga]Ga-HA-DOTA-TATE in the large and medium vessels was overall extremely low.

The main limitation of this study is its small sample size. Additionally, although we aimed to perform the PET/CT scans with the 2 tracers as close together as possible, on average [⁶⁸Ga]Ga-HA-DOTA-TATE scans were performed 10 days after FDG scans. The administration of glucocorticoids to all patients before imaging may be considered a limitation, but may also be considered a strength, as results may reflect what is likely to occur in clinical practice. Additional study strengths include its novelty and the careful prospective comparison of both [⁶⁸Ga]Ga-HA-DOTA-TATE and [¹⁸F]FDG uptake in multiple vascular territories (including temporal and coronary arteries) in patients with active GCA. Given the absence of established standards, we reported [⁶⁸Ga]Ga-HA-DOTA-TATE TBRs using both liver and blood pool comparators.

Taken together, in this pilot study of patients with active GCA receiving glucocorticoids, [⁶⁸Ga]Ga-HA-DOTA-TATE PET/CT imaging was not found to be superior for identifying large vessel vasculitis compared to [¹⁸F]FDG PET/CT imaging. Further evaluation in larger cohorts of patients and in patients who have not commenced glucocorticoids is needed.