This study underscores the value of SSTR-directed PET/CT with [68Ga]Ga-DOTATOC as a crucial diagnostic tool in localizing PMT in patients with suspected TIO and offers new insights into the relationships between imaging and biochemical parameters.

The results demonstrate a detection rate of PMT using SSTR-directed PET/CT with [68Ga]Ga-DOTATOC of 60%, which is consistent with previous studies. El–Maouche et al. showed a detection rate of 55% for [68Ga]Ga-DOTATATE [9], Kato et al. 57% for [68Ga]Ga-DOTATOC [12] and Paquet et al. 60% for [68Ga]Ga-DOTATOC, respectively. In eight patients of our cohort, no PMT was detected. One patient eventually turned out to have FGF23-independent phosphate wasting. Considering only patients with confirmed diagnosis, SSTR-directed PET/CT shows a detection rate of 63%.

In one case, no immunoreactivity was observed in SSTR2A IHC due to suboptimal tissue fixation, which may be partly caused by the peculiar matrix produced by these tumors. Consequently, the result is interpreted as false negative. All other tumors showed SSTR2A expression at different levels in IHC, supporting the findings outlined by Houang et al., analyzing the expression of SSTR2A in 15 PMTs with positive staining in all tumors [11]. However, it cannot be ruled out that fixation artifacts may have affected the accuracy of SSTR2A staining and the resulting immunoreactive score (IRS) in other PMTs in this series as well, albeit to a much lesser extent.

We couldn’t find previous data analyzing the relation between PET-derived TV and TLU and SSTR2A expression in PMT and to the best of our knowledge, no IRS was calculated for PMT before. We found a fair to good correlation of SSTR2A IRS with SUVmax and SUVmean. Consistently, Boy et al. analyzed the relation of SUVmax in [68Ga]Ga-DOTATOC-PET/CT with the mRNA expression of the SSTR subtypes in human tissues and found a correlation especially for SUVmax and SSTR2 [20]. Our assessment revealed a negative correlation of SSTR2A IRS and TV. This is consistent with results of Kim et al., showing a negative correlation between histological tumor size and SSTR2 expression in rectal neuroendocrine tumors [21].

While PMT express several markers by IHC, none of these shows a high sensitivity and specificity [22]. The negative correlation between SSTR2A and TV may show the growing impact of other characteristics of PMT with growing size, including other SSTR subtypes, which is also reflected in the wide range of the IRS. However, this hypothesis needs further investigation.

Furthermore, we found no significant differences in SUVmax and SUVpeak between PMTs of soft tissue or bone origin. Again, we couldn`t find any other study which had evaluated these differences, so far.

To the best of our knowledge, so far no study evaluated the association of TRP and TmP/GFR with the PET parameter TLU. We found a significant negative correlation between PET derived TLU and TmP/GFR, potentially indicating an association between tumor size and/or receptor density and renal phosphate wasting. In that regard, one might hypothesize that TLU reflects the endocrinological activity of the tumor tissue or the amount of active tumor cells in a way that higher TLU values correlate with more pronounced impairment of TmP/GFR. The positive correlation between IRS and PET intensity confirms that immunohistochemical receptor expression corresponds to tracer uptake. Furthermore, the negative correlation between TV and IRS supports the hypothesis that SSTR subtypes other than SSTR2A exert a predominant influence, consistent with findings in other tumor types [23, 24]. So TLU might reflect the tumors functional or endocrine behavior, but the role of SSTR2A in PMT remains unclear and needs further investigation. Conversely, the correlation of TLU and TmP/GFR would suggest that lower TmP/GFR is potentially associated with a higher chance of detecting PMT on SSTR-directed PET/CT, due to a larger size and more intense SSTR expression. We found no significant correlation between other PET derived parameters and laboratory markers. As this analysis was intended for hypothesis generation rather than confirmatory inference, no adjustment for multiple comparisons was applied. The findings should therefore be regarded as preliminary and interpreted accordingly. Consistently, Paquet et al. found no significant correlations between SUVmax and TV with phosphorus, calcium, FGF23 and PTH serum levels [25]. Concordant to Kato et al., we found no differences in FGF23 and calcium levels between patients with and without detectable PMT [12], which is most likely due to a high level of inter- and intraindividual variability of values depending on current treatment, i.e. (id est) dosing of phosphate and active Vitamin D and sampling time in relation to the last dose.

The association of TLU of PMT in SSTR-directed PET/CT with laboratory indicators of phosphate wasting might add diagnostic accuracy and guidance when to order the PET/CT scan. Our findings are in line with the comprehensive review by Minisola et al., the expert recommendations by Brandi et al. and Dahir et al. as well as the global guidance by de Beur et al., which emphasize the role of biochemical markers in TIO diagnosis [1, 2, 7, 26]. However, given the small sample sizes, most results provide initial evidence and validation in further studies is required.

Our study demonstrated significant difference in SUVmax and SUVpeak values between PMTs and bone fractures in [68Ga]Ga-DOTATOC PET/CT. For generating new hypothesis, we determined a SUVmax threshold of 7.6 for distinguishing between PMT and fractures and the threshold shows both a high sensitivity and specificity. This could be of clinical utility in reducing false-positive findings and increasing diagnostic accuracy, considering the fact that patients frequently show tracer uptake in fractures besides PMT [27]. Kato et al. reported that it could be difficult to differentiate between true PMT and fractures [12]. Parghane and Basu reported a case of false-positive osseous findings in [68Ga]Ga-DOTATATE PET/CT with SUVmax values of 5.42, 3.84 and 2.3, respectively [28]. Based on our proposed SUVmax threshold of 7.6, these lesions could have been identified as non-PMT tracer accumulation. However, this cutoff value is derived from a limited cohort, and validation in a larger study is warranted.

Improvement of clinical symptoms was an early and reliable marker for successful therapy. Follow-up of patients after surgical resection showed a marked improvement in phosphate levels, confirming the effectiveness of surgical intervention which is inevitably dependent on the success of functional imaging. For those with biochemically confirmed TIO without detectable PMT, directly blocking FGF23 with burosumab provides a sound and well-tolerated therapeutic option for managing these patients. However, we also suggest to re-evaluate the correct execution of the PET/CT, especially with regard to the use of a SSTR-directed tracer and the complete coverage particularly of the peripheral limbs. In case of insufficient imaging and in line with consented recommendations, we encourage a follow-up imaging after one year to identify previously missed PMT. Even though we would not expect this imaging to be obscured by concomitant treatment, data in that regard is lacking.

Despite the promising results, our study has several limitations. Given the retrospective study design, CT acquisition protocols were not uniform, leading to minor heterogeneity in imaging analysis. Tumor volume was delineated using a 40% isocontour threshold. This method was applied given its established reproducibility; however, partial volume effects may lead to SUV underestimation in small or heterogeneous lesions, which should be considered when interpreting the volumetric data. Nerver the less, the threshold was chosen to provide a practical method that can be easily adopted in clinical practice. Regarding FGF23, assay heterogeneity limits comparability of absolute values. Although relative deviations from the upper limit of normal were calculated to partially address this limitation, inter-assay variability may have obscured a potential associations between FGF23 and PET derived parameters. The retrospective nature and relatively small sample size limit the generalizability of the findings. However, these limitations are common in TIO research due to the rarity of the condition. Future studies ideally with larger, prospective cohorts should re-evaluate our observed correlations and the suggested SUVmax threshold and implement standardized follow-up protocols in case of negative findings.