Head-to-Head Comparison between Peptide-Based Radiopharmaceutical for PET and SPECT in the Evaluation of Neuroendocrine Tumors: A Systematic Review

We compared head-to-head the most used radiolabeled peptides for single photon computed emission tomography (SPECT) and positron emission tomography (PET) imaging of neuroendocrine tumors (NETs). A comprehensive literature search was performed in PubMed, Web of Science, and Scopus databases. The following words, coupled two by two, were used: 68Ga-DOTATOC; 68Ga-DOTATATE; 68Ga-DOTANOC; 99mTc-EDDA/HYNIC-TOC; 64Cu-DOTATATE; and 111In-DTPA-octreotide. Moreover, a second-step search strategy was adopted by using the following combined terms: “Somatostatin receptor imaging,”; “Somatostatin receptor imaging” and “Functional,”; “Somatostatin receptor imaging” and “SPECT,”; and “Somatostatin receptor imaging” and “PET”. Eligible criteria were: (1) original articles focusing on the clinical application of the radiopharmaceutical agents in NETs; (2) original articles in the English language; (3) comparative studies (head-to-head comparative or matched-paired studies). Editorials, letters to the editor, reviews, pictorial essays, clinical cases, or opinions were excluded. A total of 1077 articles were found in the three electronic databases. The full texts of 104 articles were assessed for eligibility. Nineteen articles were finally included. Most articles focused on the comparison between 111In-DTPA-Octreotide and 68Ga-DOTATOC/TATE. Few papers compared 64Cu-DOTATATE and 68Ga-DOTATOC/TATE, or SPECT tracers. The rates of true positivity were 63.7%, 58.5%, 78.4% and 82.4%, respectively, for 111In-DTPA-Octreotide, 99mTc-EDDA/HYNIC-TOC, 68Ga-DOTATATE/TOC and 64Cu-DOTATATE. In conclusion, as highly expected, PET tracers are more suitable for the in vivo identification of NETs. Indeed, in comparative studies, they demonstrated a higher true positive rate than SPECT agents.


Introduction
Neuroendocrine tumors (NETs) are a heterogeneous family of relatively uncommon neoplasms originating from endocrine cells. They can originate from the lung, thymus, gastrointestinal tract, or pancreas, sharing some morphological and immunohistochemical characteristics. In recent years the incidence of NETs has increased, although it is still considered a rare neoplasm [1,2].
More than 80-90% of NETs express somatostatin receptors (SSTR), which are integral membrane glycoproteins that can be physiologically found in different tissues throughout the body, such as the spleen, kidneys, liver, pituitary, thyroid, and adrenal glands. Five different types of SSTR (sst1-sst5) have been identified with different tissue distributions [1,2]. The somatostatin receptor type 2 (sst2) is the one expressed more frequently by NETs, but also, sst3 and sst5 can be significantly found. The expression of SSTR by NETs offers a very specific target for diagnostic imaging and therapy. Thus, techniques such as single photon
Subsequently, the three databases were searched again with the following words: "Somatostatin receptor imaging", "Somatostatin receptor imaging" and "Functional", "Somatostatin receptor imaging" and "SPECT", "Somatostatin receptor imaging" and "PET". For this second round, we filtered only papers based on comparative studies.
Among the collected papers were selected the ones that meet these criteria: (1) original articles in the English language; (2) clinical application of the radiopharmaceutical agents in NETs, and (3) head-to-head comparative studies of SPECT and/or PET radiotracers in NETs imaging. Conversely, editorials, letters to the editor, reviews, pictorial essays, clinical cases, or opinions were excluded.
After the recovery of the PDF files, a new search across the reference lists in the selected studies was conducted by G.P. and L.E.
The quality of clinical papers was assessed with a modified version of the Critical Appraisal Skills Program (CASP) checklist for diagnostic studies [https://casp-uk.b-cdn.net/ wp-content/uploads/2018/03/CASP-Diagnostic-Checklist-2018_fillable_form.pdf] (access on 26 July 2022). This critical appraisal was done by two reviewers (G.P. and L.E.), and any divergence in opinion was resolved by discussion with a third author (D.C.).

Literature Search Analysis
A total of 1077 articles were found. All duplicates were removed, leaving 558 records. Then, all reviews and all articles not entirely consistent with the inclusion criteria were excluded. The full texts of 104 articles were assessed for eligibility, and a further three articles emerged upon checking the reference lists. Finally, 19 articles were included ( Figure 1). The quality of the selected articles, based on the CAPS for diagnostic studies, is reported in Supplementary Table S1.
As illustrated in the supplementary material, in many cases, the studies have not included a standard of reference, or different types of analyses were used (i.e., lesion-based, region-based, or patient-based), thus rendering difficult the comparison between or among the radiopharmaceutical agents. Moreover, in many cases, the impact of the imaging on the selected population was not clearly stated. region-based, or patient-based), thus rendering difficult the comparison between among the radiopharmaceutical agents. Moreover, in many cases, the impact of the ima ing on the selected population was not clearly stated.      68 Ga-DOTATOC None Visual analysis by three observers 68 Ga-DOTATOC outperformed 111 In-DTPA-Octretide detecting more lesions 4 Kowalski et al. [12] 2003 Germany 4 111 In-DTPA-Octreotide vs. 68 Ga-DOTATOC

None
Visual analysis Both 64 Cu-DOTATATE and 68 Ga Three papers aimed to assess the comparison between 111 In-DTPA-Octreotide and 99m Tc-EDDA/HYNIC-TOC [9,10,13]. In the study by Bangard et al. [9], the authors compared nine patients who underwent a scintigraph examination with both tracers and described different biokinetics between them. The uptake of 99m Tc-EDDA-HYNIC-TOC was lower in the spleen and kidney than 111 In-DTPA-Octreotide. However, lesion-based analysis, 111 In-DTPA-Octreotide, detected more lesions, mainly in the liver, while 99m Tc-EDDA-HYNIC-TOC identified more abdominal lesions. Conversely, in a head-to-head comparison, Decristoforo et al. [10] demonstrated, in 10 patients, that 99m Tc-EDDA-HYNIC-TOC was simpler to produce and more detectable of lesions than 111 In-DTPA-Octreotide, thus, opening the way for new alternative SPECT agents for the NET detection. Three years later, Gabriel et al. [13] concluded that 99m Tc-EDDA-HYNIC-TOC scintigraphy is more performant than 111 In-DTPA-Octreotide, mainly if the acquisition is made by an early and late acquisition (after 1-2 h from the tracer injection), in order to improve the tumor/background ratio.
Eleven out of 19 papers, including 498 patients, focused on the comparison between 111 In-DTPA-Octreotide SPECT or SPECT/CT and 68 Ga-DOTATOC PET or PET/CT. [11,12,[14][15][16][17][18][19][20]22,26] Hofmann et al. [11] enrolled eight patients with metastatic carcinoid who underwent 68 Ga-DOTATOC SPECT, CT, MRI, and 111 In-DTPA-Octreotide, showing the power of PET tracer in detecting the lesions (100% vs. 85%, respectively for 68   In-DTPA-Octreotide). A similar and limited experience was reported by Kowalski et al. in four patients [12]. In this small patient population, the researchers found that PET was able to better detect small lesions with low-density SSTR expression. The studies by Buchmann et al. [14], Gabriel et al. [15], Srirajaskanthan et al. [17], Krausz et al. [18], Hofman et al. [19] demonstrated that 68 Ga-DOTATOC/TATE PET/CT was able to detect more NET lesions than 111 In-DTPA-Octreotide SPECT or SPECT/CT during patient-based and lesion-based analysis. In particular, 68 Ga-DOTATOC/TATE PET/CT was able to better define the extension of metastatic disease in the liver, skeleton, and thoracic/abdominal lymph nodes. Moreover, based on the study by Krausz et al. [18], primary NET in the pancreas was more often detected by 68 Ga-DOTATOC/TATE PET/CT than 111 In-DTPA-Octreotide SPECT/CT, thus increasing its performance also in primary tumors and not only for metastatic disease. Additionally, in the papers by Buchmann et al. [14], Gabriel et al. [15], Srirajaskanthan et al. [17], and Hofman et al. [19], 68 Ga-DOTATOC/TATE PET/CT was able to improve the clinical management in comparison with 111 In-DTPA-Octreotide SPECT/CT. Indeed, based on the study by Buchmann et al. [14], surgical intervention was extended in seven patients owing to PET findings. Similarly, PET was able to change the therapeutic approach from a surgical to a systemic one after identifying more distant NET lesions (12/51 patients; 24%), in accordance with Gabriel et al. [15]. In the study by Srirajaskanthan et al. [17], the change of management with PET imaging was reported in 36/51 (70.6%) patients, mainly by providing the opportunity to undergo PRRT with 90 Y/ 177 Lu-DOTATATE/TOC. Finally, Hofman et al. [19] reported that PET imaging had a high management impact in 28/58 (47%) patients. Indeed, 68 Ga-DOTATOC/TATE PET/CT increased the number of lesions detected; thus, many patients received systemic therapy rather than undergoing surgery.
Sadowski et al. [22] assessed the comparison between 68 Ga-DOTATATE PET/CT and 111 In-DTPA-Octreotide SPECT/CT in a cohort of patients affected by MEN1, demonstrating that 68 Ga-DOTATATE PET/CT is more sensitive in detecting MEN1 lesions than SPECT imaging and it could also alter the management; therefore the authors strongly recommend to introduce this imaging modality in the diagnostic flow-chart of patients affected by MEN1 syndrome.
Finally, in the study by Van Binnebeek et al. [20] and Hope et al. [26] appeared, the term "tumor burden" relative to the extension of SSTR-positive disease. In both the studies, 68 Ga-DOTATOC or 68 Ga-DOTATATE was superior to SPECT imaging with 111 In in assessing the tumor burden, mainly for the identification of small lesions detected by the PET scanner rather than by the SPECT one.
The articles focused on the comparison between 68 Ga-DOTATOC/ 68   Tc-EDDA/HYNIC-TOC were 2 [23,25]. In the study by Madrzak et al. [23], 24 patients underwent both images with PET/CT and SPECT/CT. The authors found that 68 Ga-DOTATOC PET/CT altered the treatment procedures in only 8.4% of patients (2 persons). However, due to the limited patient enrolment, the authors suggested additional studies to confirm this assumption. Therefore, one year later, Kunikowska et al. [25] enrolled 68 patients showing the advantages of 68 Ga-DOTATATE PET/CT over 99m Tc-EDDA/HYNIC-TOC in detecting NET lesions and underlined that 68 Ga-DOTATATE PET/CT was able to change the clinical decision in one-third of patients.
Two articles (n = 64 patients) [24,27] assessed the comparison between 64 Cu-DOTATATE and 68 Ga-DOTATOC/TATE. Johnbeck et al. [24] enrolled 59 patients who underwent PET imaging with both tracers within 1 week. Through an intra-patient analysis, it emerged that PET images were concordant in 37 patients and discordant in 22 patients. Among this later subset of patients, most additional lesions were found by 64 Cu-DOTATATE vs. 68 Ga-DOTATOC/TATE (14 vs. eight patients, respectively). Although 64 Cu-DOTATATE seems more performant in this study, in a very recent pilot analysis performed on five patients, Jha et al. [27] concluded that the data currently available was not conclusive about the superiority of one over the other.
Finally, the residual paper was a comparative analysis between 111 In-DTPA-Octreotide and 64 Cu-DOTATATE [21]. This was a large experience in 112 patients demonstrating that, similar to 68 Ga radiolabeled peptides, 64 Cu-DOTATATE PET is superior to 111 In-DTPA-Octreotide SPECT.
Indeed, as is visible from Table 3, the sensitivity was, as expected, higher for PET radiopharmaceuticals either with 68 Ga and 64 Cu than for SPECT agents. However, by comparing 111 In-DTPA-Octreotide and 99m Tc-EDDA/HYNIC-TOC, the latter has a higher sensitivity, although a lower specificity. Until now, only limited data are available on the comparison between 68 Ga-DOTATOC/TATE and 64 Cu-DOTATATE. The study by Johnbeck et al. [24] found a slightly higher sensitivity and specificity for 64 Cu-DOTATATE when compared to 68 Ga-DOTATOC. However, the data are still limited for final evidence.
For lesion-based and site-based analyses, radiopharmaceuticals for PET imaging were more performant than SPECT agents in identifying the number and the presence of lesions in the musculoskeletal system, bone, liver, and lymph nodes (mainly in the abdominal region).
The paper by Mussig et al. [16] analyzed the association between the expression of sst2 and the uptake of 68   In-DTPA-Octreotide. The authors found that a positive scan with both the tracers was associated with a high expression of sst2; however, tumors without immunohistochemical sst2 expression could show 68 Ga-DOTATOC tracer uptake, probably due to the expression of sst3 or sst5, or simply because of the tumor heterogeneity.

The Theragnostic Role of Radiopharmaceuticals for NET
The theragnostic role of the abovementioned radiopharmaceuticals in the different settings of disease (detection, staging, status of SSTR, and follow-up) has scarcely been reported in comparative studies.
From a careful analysis of the available data, 56 (7.8%) patients were enrolled for the detection of NET, 120 (16.9%) for staging, and 535 (75.3%) for the assessment of SSTR expression and follow-up. As expected, follow-ups for the evaluation of SSTR expression were the most common indication in many selected studies. Higher diagnostic performances have been reported for the assessment of SSTR status in the follow-up settings for PET tracers as compared to SPECT tracers, for patient-based analysis but also regional-and lesion-based ones. After different previous treatments, the assessment of SSTR expression is essential in planning PRRT, and the PET tracer results were more accurate in these settings at any level of analysis. However, no information has been found about PET and SPECT tracers in monitoring the response to PRRT in NETs, although it would be an interesting and important topic from a diversified point of view.

Discussion and Conclusions
Scintigraphy with radiolabeled SSTR has gained widespread acceptance as the imaging method of choice in NET patients, showing high sensitivity and good specificity, as emerged from the previous study and from this systematic review. From planar imaging with 111 In-DTPA-Octreotide to SPECT with 99m Tc-HYNIC/EDDA-TOC, a gain in terms of detection has been obtained. However, due to the limited spatial resolution of SPECT imaging, PET tracers radiolabeled with either 64 Cu or 68 Ga have been introduced in clinical practice, thus increasing the detection of NET lesions.
From the present comparative review, it emerged that for patient-based analysis, the rate of true positive and the diagnostic performance is as expected, higher for PET tracers when compared to SPECT tracers, mainly when 111 In-DTPA-Octreotide was compared to 68 Ga-DOTATOC/TATE/NOC. Moreover, when analyzing the available data for lesionand region-based analyses, we found that PET radiopharmaceutical agents were more performant in detecting bone, lymph nodes, and liver metastases than SPECT agents. This advantage was mainly due to the PET scanner technology rather than the radiopharmaceutical itself. It would be interesting to understand if new technological achievements in SPECT technology, such as solid-state detectors and 360 • detector coverage, could fill this gap. Indeed, PET tomographic images can significantly improve the detection of deep lesions or visceral metastases when compared with planar or SPECT images.
By a comparative analysis between 64 Cu-DOTATATE and 68 Ga-DOTATOC emerged that the performances were quite similar; however, the number of true positive lesions was slightly higher for 64 Cu-DOTATATE than 68 Ga-DOTATOC (33 vs. 7), as reported by Johnbeck et al. [24] The affinity for SSTR was quite similar for all the imaging agents, particularly for those used in PET, as recently reported by some authors [28][29][30]. Nevertheless, the intrinsic physical characteristics of radioisotopes can have an important effect on lesion detectability. Indeed, 64 Cu has a shorter positron range than 68 Ga, thus possibly improving the detection rate of small lesions. Moreover, the radiation burden is different between 64 Cu and 68 Ga. Similarly, 99m Tc has the advantage of a lower radiation dose than 111 In. This latter physical characteristic can be translated into the advantageous use of 99m Tc for repeated investigations, for example, in monitoring the response to PRRT or in children. However, to date, no information about the cost-saving, other than the radioprotection information, is available for 99m Tc-EDDA/HYNIC-TOC SPECT in comparison to 68 Ga-DOTATOC/TATE PET. Conversely, Schreiter et al. [31] found that 68 Ga-DOTATOC PET/CT was considerably cheaper than 111 In-DTPA-octreotide with respect to both material and personnel costs. Therefore, additional cost analyses are welcome also for the other agents.
It should be noted, however, that comparisons between PET and SPECT agents were made considering different acquisition protocols. Table 4 reports some of the pros and cons of SPECT and PET imaging for detecting NETs.
The question that arises from the above considerations is, "Can the improved detection rate affect clinical management?" In NETs, changes in treatment strategy are nearly always based on clinical or imagingbased signs of progression. Thus, high performance in the detection of any new lesions is of great value in patients affected by these rare diseases. For example, the additional evidence of bone metastases can have either an important effect on the therapeutic intervention or a prognostic implication because unknown distant metastases are considered a negative prognostic factor, possibly requiring a more aggressive treatment regimen [32,33]. Based on the available data, the inclusion of PET imaging in clinical practice impacted the change of management from 3.7% to 70.6% (Table S2) of patients. Therefore, PET imaging should be preferred to SPECT imaging when available. From a careful analysis of the selected studies, no comparative data were available about the role of PET and SPECT imaging in monitoring the response to PRRT. The recent introduction of PRRT in clinical practice (Netter 1 trial) and the opportunity of monitoring the response to therapy, both in the interim and at the end, is essential for testing the efficacy of therapy. To date, some studies have been published about the role of 68 Ga-DOTATOC/TATE/NOC in monitoring the response to PRRT in comparison to morphological criteria without reporting conclusions [34,35]. Indeed, functional imaging is not yet accepted as a substitute for morphological imaging as a means to assess tumor response to treatment [36]. However, the opportunity to use both SPECT (especially with new scanners) and PET tracers during and after PRRT would also be an advantage in the case of retreatment or an early treatment interruption. Future studies should be conducted to test these hypotheses.
In this systematic review, we focused our attention on SPECT/PET radiotracers based on SSTR analogs, though theoretically, other PET tracers can be used for NETs imaging.
One alternative is represented by 11 C-hydroxytriptophan ( 11 C-5-HTP), a serotonin precursor that allows the evaluation of the serotonin pathway, which is one of the active metabolic pathways in NETs [37]. This tracer has a high sensitivity, especially for pancreatic NETS, but its use is limited by the half-life of the radionuclide (20 min), which requires the presence of an on-site cyclotron [3,37]. 18 F-DOPA ( 18 F-L-dihydroxyphenylalanine) is another PET tracer that finds a high application in NETs [37]. Indeed, NETs cells can often take up decarboxylate monoamine precursors, such as DOPA. 18 F-DOPA seems to be useful for imaging well-differentiated midgut tumors, though they often overexpress sst2 [38].
In a recent meta-analysis of head-to-head comparative studies emerged that at patientbased and region-based analysis, 68 Ga-DOTA-peptides performed better than 18 F-DOPA PET in detecting intestinal NETs, but at lesion-based analysis, 18 F-DOPA PET was more accurate [38].
Another alternative to peptide analogs is represented by 18 F-FDG ( 18 F-fluorodeoxyglucose), which is recently becoming the PET tracer of choice in many cancer forms. 18 F-FDG exploits cancer cells' preferential utilization of aerobic glycolysis. Similarly, to glucose, it enters cells via glucose transporters GLUT-1 and GLUT-3, but it doesn't follow the same metabolic pathway of glucose due to a lack of an oxygen atom in its C2 position. Thus, it accumulates in cells proportional to their glucose consumption. [37] However, for many years it was not used in NETs due to its low sensitivity in the detection of these tumors, but more recently, the utility of FDG-PET scans has been reassessed [1,3]. NETs with poor differentiation, a high grade, and rapid proliferation have a decreased expression of SSTR expression; thus, scans with peptide analogs may be negative, while 18 F-FDG imaging may be positive [1,39]. Liu et al. [39] analyzed 30 studies focused on 68 Ga-radiolabelled agonist SSTR and FDG PET/CT in NET patients. From the meta-analysis emerged that 18 F-FDG PET/CT has the lowest sensitivity in detecting NET lesions. However, it has a complementary role in the case of moderately or scarcely differentiated NETs.
The abovementioned radiopharmaceuticals have shown promising results, but these substances do not provide any theragnostic options, unlike somatostatin analogs.
Lastly, new interest is increasing in the use of SSTR antagonists. Compared to agonists, they showed better pharmacokinetics and image contrast, a higher tumor uptake, and a better residence time [37], [40]. Among them, 68 Ga-NOGADA-JR11 (or 68   and 68 Ga-DOTA-JR11 have also demonstrated advantages for potential theragnostic application [40][41][42]. In particular, the study by Zhu et al. [40] showed that in 12 patients undergoing imaging with both radiopharmaceutical agents on two consecutive days, 68 Ga-DOTA-JR11 outperformed better that 68 Ga-DOTATATE in detecting liver metastases, while 68 Ga-DOTATATE was better for the identification of bone lesions. However, to date, little clinical evidence is still available. The present systematic review has limitations. The limited number of studies comparing 64 Cu-DOTATATE vs. 68 Ga-DOTATATE. Moreover, in the study by Pfeifer et al. [21], 64 Cu-DOTATATE PET/CT was compared with 68 Ga-DOTATOC PET, therefore by using a hybrid vs. non-hybrid system, thus reducing the detection power of the second imaging modality. In the study by Buchman et al. [14], the authors reported that the region-based analysis could have overestimated the sensitivity of 111 In-DTPA-octreotide. Few data about the standard of reference, as also emerged by the CAPS evaluation; indeed, it missed 10/19 (53%) of papers, thus reducing the opportunity to perform an adequate comparison in terms of diagnostic performances.
In conclusion, PET imaging, as expected, is more suitable for the identification of NET. Indeed, they demonstrated a higher true positive rate than SPECT imaging. However, the availability of new SPECT scanners, more favorable radioprotection and synthetical characteristics (mainly for 99m Tc), and the consolidated experiences for conventional scintigraphy examination should be considered in the diagnostic and therapeutic path, also for health equity.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/xxx/s1, Table S1: Clinical studies assessment using Critical Appraisal Skill Progam (CASP) checklist for diagnostic studies; Table S2: Change of management with PET imaging on patients affected by the neuroendocrine tumor.