Detection Rate of Culprit Tumors Causing Osteomalacia Using Somatostatin Receptor PET/CT: Systematic Review and Meta-Analysis

Background: Tumor-induced or oncogenic osteomalacia (TIO) is a rare paraneoplastic syndrome in which osteomalacia is a consequence of fibroblast growth factor 23 (FGF23) secretion by a mesenchymal tumor. The localization of the culprit lesion in patients with TIO is often challenging. Several studies have evaluated the detection rate (DR) of these tumors using somatostatin receptor positron emission tomography (SSTR-PET/CT). We aimed to summarize literature findings on this topic providing pooled estimates of DR. Methods: A comprehensive literature search by screening PubMed, Embase and Cochrane library electronic databases through August 2019 was performed. The pooled DR of culprit tumors using SSTR-PET/CT in patients with TIO was calculated using a random-effects statistical model. Results: Fourteen studies on the use of SSTR-PET/CT in detecting the culprit tumor in patients with TIO were included in the qualitative analysis. The pooled DR of SSTR-PET/CT on a per-patient-based analysis calculated using eleven studies (166 patients) was 87.6% (95% confidence interval (95% CI) 80.2–95.1%). Statistical heterogeneity among studies was detected (I-square = 63%), likely due to the use of different radiolabeled somatostatin analogues, as demonstrated by a subgroup analysis. Conclusions: Despite limited literature data due to the rarity of the disease, SSTR-PET/CT demonstrated a very high DR of culprit tumors in patients with TIO and it could be used as first-line imaging method for this indication.


Search Strategy
Three co-authors (M.M, M.N.L. and G.T.) independently performed a comprehensive computer literature search of the PubMed/MEDLINE, Cochrane library and Embase databases to find relevant published articles on the DR of culprit tumors in patients with TIO using SSTR-PET/CT. This search string-based on a combination of key words, Boolean operators and truncations (*) -was created and used: (A) "DOTA*' OR 'somatostatin' AND (B) 'PET' OR 'positron*' AND (C) 'osteomalacia' OR 'culprit' OR 'mesench*'. No beginning date limit nor language restrictions were used. The literature search was updated until 31st August 2019. To expand the literature search, the references of the retrieved articles were also screened for possible additional records.

Study Selection
Studies assessing the DR of culprit tumors in patients with TIO using SSTR-PET/CT were eligible for inclusion in the qualitative analysis (systematic review).
The exclusion criteria for the systematic review were: (a) articles not within the field of interest; (b) reviews, editorials, letters, comments, conference proceedings; (c) case reports and small case series (less than 5 patients included).
All the studies included in the systematic review were included in the meta-analysis, except those with possible patient data overlap. If studies with possible patient data overlap were found, only the article with more complete information was included in the meta-analysis.
Three co-authors (M.M, M.N.L. and G.T.) independently screened the abstracts of the retrieved articles, applying the predefined inclusion and exclusion criteria. Subsequently, the researchers independently reviewed the full text of the selected articles to assess their eligibility for inclusion in the systematic review. Any disagreement was solved through a consensus meeting among the researchers performed in September 2019 at the Department of Nuclear Medicine and Molecular Imaging of the University Hospital of Lausanne, Switzerland.

Data Extraction
For each selected article, information was collected on basic study characteristics (authors, year of publication, country, study design), patient characteristics (type and number of patients evaluated, age and sex ratio, FGF23 serum levels), technical details (type of hybrid imaging used, radiolabeled somatostatin analogues used, injected activity, time between radiopharmaceutical injection and PET/CT image acquisition, image analysis and other functional imaging methods performed for comparison), data on DR, and the site of the culprit tumors detected by SSTR-PET/CT, including the number and type of tumors proven by histopathology.

Quality Assessment
The quality of the studies included in this systematic review was critically appraised using the revised "Quality Assessment of Diagnostic Accuracy Studies" tool (QUADAS-2) [13].

Statistical Analysis
The DR of culprit tumors using SSTR-PET/CT was obtained from individual studies on a per-patient-based analysis. A random-effects model was used for the statistical pooling of DR. Pooled data were presented with 95% confidence intervals (95% CI) and displayed using forest plots. Heterogeneity was estimated using the I-square index (I 2 ); a statistical heterogeneity among studies was present if I 2 was higher than 50% [14]. If significant heterogeneity was found, subgroup analyses taking into account the type of radiolabeled somatostatin analogues used (DOTATATE, DOTATOC, DOTANOC) were performed. Publication bias was assessed through the visual evaluation of a funnel plot and the Egger's test [15].

Literature Search
Literature search results are summarized in Figure 1 and briefly described below.

Literature Search
Literature search results are summarized in Figure 1 and briefly described below. Overall, 56 records were identified through the comprehensive computer literature search of the PubMed/MEDLINE, Cochrane library and Embase databases. Screening 56 abstracts, 36 records were excluded: 11 as they were not in the field of interest, four as reviews/editorials/letters, 21 as case reports or small case series (less than five patients included). Twenty articles were selected and their full text was retrieved. No additional records were found screening the references of these articles, whereas six articles were excluded after the analysis of the full text. Therefore, 14 articles were included in the qualitative analysis (systematic review) [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. Three articles were excluded from the meta-analysis for possible patient data overlap [18,25,26]. Overall, 11 articles (166 patients with TIO) were included in the quantitative analysis (meta-analysis) [16,17,[19][20][21][22][23][24][27][28][29]. The characteristics of the 14 studies included in the systematic review are presented in Tables 1 and 2. The diagnostic accuracy data from these articles are shown in Table 3. The quality appraisal of studies included in the systematic review is reported in Figure 2. Overall, 56 records were identified through the comprehensive computer literature search of the PubMed/MEDLINE, Cochrane library and Embase databases. Screening 56 abstracts, 36 records were excluded: 11 as they were not in the field of interest, four as reviews/editorials/letters, 21 as case reports or small case series (less than five patients included). Twenty articles were selected and their full text was retrieved. No additional records were found screening the references of these articles, whereas six articles were excluded after the analysis of the full text. Therefore, 14 articles were included in the qualitative analysis (systematic review) [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. Three articles were excluded from the meta-analysis for possible patient data overlap [18,25,26]. Overall, 11 articles (166 patients with TIO) were included in the quantitative analysis (meta-analysis) [16,17,[19][20][21][22][23][24][27][28][29]. The characteristics of the 14 studies included in the systematic review are presented in Tables 1 and 2. The diagnostic accuracy data from these articles are shown in Table 3. The quality appraisal of studies included in the systematic review is reported in Figure 2.

Basic Study and Patient Characteristics
Screening the selected databases, 14 articles evaluating the DR of culprit tumors in patients with TIO using SSTR-PET/CT were selected (Table 1) [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. Most of the selected articles were retrospective (93%) or single-center (86%) studies. All the selected articles were published in the last 6 years (from 2013 to 2019) by research groups of different continents (Asia, Europe, America and Oceania), but studies from Asia were the most represented (71%). The patients included in the selected articles have a clinical and biochemical diagnosis of TIO or a suspected TIO. The mean age of the patients included in these studies ranged from 36 to 53 years and the percentage of male patients (sex ratio) largely ranged from 17% to 80%. The majority of patients included in the selected studies had symptomatic osteomalacia with hypophosphatemia and evidence of increased serum levels of FGF23. The most common presenting symptoms were bone pain and muscle weakness [16][17][18][19][20][21][22][23][24][25][26][27][28][29].

Basic Study and Patient Characteristics
Screening the selected databases, 14 articles evaluating the DR of culprit tumors in patients with TIO using SSTR-PET/CT were selected (Table 1) [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. Most of the selected articles were retrospective (93%) or single-center (86%) studies. All the selected articles were published in the last 6 years (from 2013 to 2019) by research groups of different continents (Asia, Europe, America and Oceania), but studies from Asia were the most represented (71%). The patients included in the selected articles have a clinical and biochemical diagnosis of TIO or a suspected TIO. The mean age of the patients included in these studies ranged from 36 to 53 years and the percentage of male patients (sex ratio) largely ranged from 17% to 80%. The majority of patients included in the selected studies had symptomatic osteomalacia with hypophosphatemia and evidence of increased serum levels of FGF23. The most common presenting symptoms were bone pain and muscle weakness [16][17][18][19][20][21][22][23][24][25][26][27][28][29].

Technical Aspects
Technical details about SSTR-PET/CT from the included studies are summarized in Table 2. Hybrid PET/CT was performed in 100% of the studies, without contrast-enhanced CT in the majority of cases (93%). The injected radiolabeled somatostatin analogues were 68 Ga-DOTATATE (in 10 studies), 68 Ga-DOTANOC (in two studies) and 68 Ga-DOTATOC (in one study). The remaining study used both 68 Ga-DOTATATE and 68 Ga-DOTANOC. The radiopharmaceutical injected activity was quite different among the studies. The time interval between radiopharmaceutical injection and SSTR-PET/CT acquisition ranged from 20 min to 90 min. A whole-body PET/CT acquisition (from head to toes) was performed in all the studies. The analysis of SSTR-PET/CT images was performed using qualitative criteria (visual analysis) in all the studies and additional semi-quantitative parameters, as the maximal standardized uptake value (SUV max ), in 64% of cases. At visual analysis, the areas of focal increased radiopharmaceutical uptake greater than the surrounding tissue and not judged as physiological activity were considered abnormal.
Other functional imaging modalities were used for the comparison of SSTR-PET/CT findings in most of the articles; in particular SSTR scintigraphy or SPECT/CT (using 111 In-or 99m Tc-octreotide), fluorine-18 fluorodeoxyglucose ( 18 F-FDG) PET/CT, 99m Tc-sestamibi scintigraphy and bone scintigraphy.
Histopathological results (gold standard) and/or clinical/imaging/biochemical follow-up were used as the reference standard in the included studies.

Main Findings
As shown in Table 3, most of the selected studies showed a good DR of culprit tumors in patients with TIO using SSTR-PET/CT. The culprit lesions were usually small benign tumors located in the bones or soft tissues, presenting a high uptake of radiolabeled somatostatin analogues ( 68 Ga-DOTATATE, 68 Ga-DOTATOC or 68 Ga-DOTANOC) at SSTR-PET/CT [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. The most frequent site of culprit tumors were the lower limbs; other frequent sites were the cranio-facial region and the trunk, whereas the localization of culprit tumors in the upper limbs was less frequent. Most of the culprit tumors detected by SSTR-PET/CT were confirmed by histopathology and phosphaturic mesenchymal tumor was the most frequent histological type. Malignant or metastatic tumors detected by SSTR-PET/CT in patients with TIO were rare [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. A very high inter-observer concordance among PET/CT masked readers was reported by one study for the visual detection of culprit tumors by SSTR-PET/CT [20]. Interestingly, a significant correlation between SUVmax of the culprit tumor at SSTR-PET/CT and serum FGF23 levels was not reported [17,20]. Furthermore, there was no significant difference for any of the biochemical parameters and for the duration of the disease between SSTR-PET/CT-positive and SSTR-PET/CT-negative cases [20].
Fractures can be a common consequence of TIO, and they can also lead to increased radiopharmaceutical uptake at SSTR-PET/CT, potentially affecting the accuracy of this method in detecting the culprit tumors. However, it has been demonstrated that mild radiopharmaceutical uptake at the sites of fracture is not a major challenging factor in the interpretation of SSTR-PET/CT when both the intensity of the radiopharmaceutical uptake at PET and the morphology of CT are assessed. In fact, fractures show a characteristic morphology at CT, and they usually present a lower radiopharmaceutical uptake at SSTR-PET compared to culprit tumors in patients with TIO [19,21]. Inflammatory lesions, such as granulomatous lesions, may also cause false positive findings for culprit tumors in patients with TIO, due to the high expression of SSTRs by activated inflammatory cells [20,25].
When compared to 18 F-FDG PET/CT, SSTR-PET/CT had a higher DR of culprit tumors in patients with TIO. Additionally, 18 F-FDG PET/CT may lead to a higher number of false positive results compared to SSTR-PET/CT for this indication. Furthermore, in those TIO patients with culprit tumors positive at both 18 F-FDG PET/CT and SSTR-PET/CT, the lesion-to-background contrast was higher at SSTR-PET/CT compared to that of 18 F-FDG PET/CT, enabling a more confident diagnosis [16,17,20,22,23,26,28,29].
Bone scintigraphy was performed in some studies, but it showed a significantly lower DR of the culprit tumor in patients with TIO compared to SSTR-PET/CT. Furthermore, bone scintigraphy may show areas of focal radiopharmaceutical uptake at the site of fractures or scintigraphic signs suggestive of metabolic bone disease in patients with TIO [16,24,29].
Overall, allowing the detection of culprit tumors which remained occult after conventional work-up, SSTR-PET/CT induced a change of management in a significant percentage of patients with TIO. In particular, in most cases of culprit tumors detected by SSTR-PET/CT, patients with TIO were referred to surgery for the excision of the culprit tumor. The most frequent outcome after surgery was normalization of biochemical parameters and clinical remission of TIO [16][17][18][19][20][21][22][23][24][25][26][27][28][29].

Discussion
To the best of our knowledge, this is the first systematic review and meta-analysis which has evaluated the DR of culprit tumors in patients with TIO using SSTR-PET/CT. Several studies have been published on this topic, but these studies have a limited statistical power, as a small number of patients with TIO were enrolled due to the rarity of the disease. Therefore, we have pooled data

Discussion
To the best of our knowledge, this is the first systematic review and meta-analysis which has evaluated the DR of culprit tumors in patients with TIO using SSTR-PET/CT. Several studies have been published on this topic, but these studies have a limited statistical power, as a small number of patients with TIO were enrolled due to the rarity of the disease. Therefore, we have pooled data reported in the published studies, to obtain more robust estimates on the DR of SSTR-PET/CT in this setting.
Our systematic review and meta-analysis demonstrated a very good DR of culprit tumors in patients with TIO using SSTR-PET/CT-with a pooled value of about 90%-due to the overexpression of SSTRs in most of the culprit tumors causing osteomalacia. Even if false negative findings are possible, in about 10% of cases, it should be underlined that SSTR-PET/CT has allowed the detection of culprit tumors which remained occult with conventional imaging methods in most of the cases [16][17][18][19][20][21][22][23][24][25][26][27][28][29]. The main explanation for false negative results at conventional imaging methods is the reduced size of most culprit tumors causing osteomalacia [1,3,4]. False positive findings of SSTR-PET/CT in this setting are also described [20,25], in particular caused by inflammatory lesions due to the overexpression of SSTR by activated inflammatory cells [30,31]. Granulomatous lesions may be positive at SSTR-PET/CT in some cases and it could be difficult to differentiate them from tumors causing osteomalacia-or neuroendocrine tumors-using conventional imaging methods or SSTR-PET/CT. As culprit tumors detected by SSTR-PET/CT may be located everywhere throughout the body [16][17][18][19][20][21][22][23][24][25][26][27][28][29], it is important to perform a whole-body SSTR-PET/CT acquisition (from head to toes) to avoid missing lesions.
Semi-quantitative PET analysis (e.g., using SUVmax) can be used as an adjunct tool to visual PET analysis for SSTR-PET/CT interpretation, in particular to make a differential diagnosis between culprit tumors in patients with TIO and TIO-related fractures [19,21].
Contrast enhancement could further improve the DR of culprit tumors using SSTR-PET/CT in patients with TIO [16], but data on this regard are very limited and this should be better evaluated in further studies. Another topic that would need to be evaluated in further studies is the diagnostic performance of SSTR-PET/MRI compared to SSTR-PET/CT in this setting [32], because data on this regard are still lacking.
Some limitations and biases of our systematic review and meta-analysis should be considered. First of all, a quite limited number of studies and patients were available for the systematic review and the meta-analysis, but this is justified by the rarity of TIO. Moreover, as a composite reference standard was used in some studies, a possible verification bias could not be excluded in some cases; nevertheless, most of the culprit tumors detected by SSTR-PET/CT were verified by histopathology. Heterogeneity among studies (i.e., due to differences in patient characteristics, methodological aspects and study quality) may represent a bias in a meta-analysis. We have detected a statistical heterogeneity among the included studies in our meta-analysis, but we have explored this heterogeneity performing a subgroup analysis based on the radiopharmaceutical used for SSTR-PET/CT, thus demonstrating that the different PET radiopharmaceutical used may be cause of moderate heterogeneity even if the DR of culprit tumors were similar using different radiopharmaceuticals for SSTR-PET/CT. Lastly, we found a publication bias as demonstrated by the funnel plot and the Egger's test; therefore, the outcome of the studies has influenced the decision whether to publish or not the articles. We have tried to limit the publication bias, excluding from the analysis those studies including less than five patients with TIO.

Conclusions
Despite limited literature data due to the rarity of the disease, SSTR-PET/CT demonstrated a very high DR of culprit tumors in patients with TIO, and it could be used as first-line imaging method for this indication. Further prospective and multicenter studies, and in particular cost-effectiveness analyses, could strengthen the role of SSTR-PET/CT in this setting.