The association of intra-therapeutic heterogeneity of somatostatin receptor expression with morphological treatment response in patients undergoing PRRT with [177Lu]-DOTATATE

Aim Purpose of this study was to evaluate the association of the spatial heterogeneity (asphericity, ASP) in intra-therapeutic SPECT/ CT imaging of somatostatin receptor (SSR) positive metastatic gastroenteropancreatic neuroendocrine neoplasms (GEP-NEN) for morphological treatment response to peptide receptor radionuclide therapy (PRRT). Secondly, we correlated ASP derived form a pre-therapeutic OctreoScan (ASP[In]) and an intra-therapeutic [177Lu]-SPECT/CT (ASP[Lu]). Materials and methods Data from first therapy cycle [177Lu-DOTA0-Tyr3]octreotate ([177Lu]-DOTATATE)-PRRT was retrospectively analyzed in 33 patients (m = 20; w = 13; median age, 72 [46–88] years). The evaluation of response to PRRT was performed according to RECIST 1.1 in responding lesions [RL (SD, PR, CR), n = 104] and non-responding lesions [NRL (PD), n = 27]. The association of SSR tumor heterogeneity with morphological response was evaluated by Kruskal-Wallis test and receiver operating characteristic curve (ROC). The optimal threshold for separation (RL vs. NRL) was calculated using the Youden-index. Relationship between pre- and intra-therapeutic ASP was determined with Spearman’s rank correlation coefficient (ρ) and Bland-Altman plots. Results A total of 131 lesions (liver: n = 59, lymph nodes: n = 48, bone: n = 19, pancreas: n = 5) were analyzed. Lesions with higher ASP values showed a significantly poorer response to PRRT (PD, median: 11.3, IQR: 8.5–15.5; SD, median: 3.4, IQR: 2.1–4.5; PR, median 1.7, IQR: 0.9–2.8; CR, median: 0.5, IQR: 0.0–1.3); Kruskal-Wallis, p<0.001). ROC analyses revealed a significant separation between RL and NRL for ASP after 4 months (AUC 0.85, p<0.001) and after 12 months (AUC 0.94, p<0.001). The optimal threshold for ASP was >5.45% (sensitivity 96% and specificity 82%). The correlation coefficient of pre- and intra-therapeutic ASP revealed ρ = 0.72 (p <0.01). The mean absolute difference between ASP[In] and ASP[Lu] was -0.04 (95% Limits of Agreement, -6.1–6.0). Conclusion Pre- and intra-therapeutic ASP shows a strong correlation and might be an useful tool for therapy monitoring.


Introduction
Neuroendocrine neoplasms of the gastroenteropancreatic system (GEP-NEN) represents a rare and heterogeneous tumor entity. However, the incidence of GEP-NEN increased recently, mainly due to improved functional imaging and device-specific sensitivity [1]. Beyond commonly "cold" agents, such as somatostatin analogs, as a first-line antiproliferative drug, peptide receptor radionuclide therapy (PRRT) with [ 177 Lu-DOTA 0 -Tyr 3 ]octreotate ([ 177 Lu]DOTA-TATE) has emerged as a highly effective treatment in metastatic well-differentiated GEP-NEN of low and intermediate grade G 1 and G 2 . [2,3]. Several retrospective studies could also show the superiority of [ 177 Lu]DOTATATE-PRRT in advanced inoperable GEP-NEN compared to other treatment modalities [4][5][6]. The recently published results of the first prospective randomized study in patients with progressive metastatic midgut NEN, NETTER-1, found a rate of 65.2% progression-free survival at month 20 after PRRT. In particular, the overall survival (OS) has not yet been reached in the study cohort [7]. In contrast, the current published ENETS guidelines recommend the use of PRRT in intestinal (midgut) NEN with distance and/ or loco regional metastases as a second-to third-line therapy after progression under somatostatin-analogues (SSA) irrespective of the abovementioned NETTER-1 results. According to guidelines in pancreatic NEN with advanced locoregional disease PRRT should be even used as a third line therapy after failure of SSA, everolimus and/ or cytotoxic chemotherapy [8].
Meanwhile, application of PRRT can be only repeated to a limited extent, imposed by critical dose exposure in different organs at risk, e.g. in bone marrow (<2Gy) [9,10] in 70% of patients after treatment with [ 177 Lu]DOTATATE and in kidney (<40 Gy) [11]. Histological diversities in metastatic SSR-expression are common findings, and may deliver objective prediction of therapy efficacy [12,13]. However, not all patients benefit from PRRT. Therefore, better stratification criteria are highly desirable to identify patients with high probability for tumor response.
Until now, diagnosis and staging of NEN are merely based on analysis of tumor spread and morphology in imaging, irrespective of SSR expression and density.
Promising results have been published considering features from the non-invasive SSRimaging that can be used separately or in combination to characterize more accurately biological behavior of NEN, thus leading to a better understanding of tumor/therapy response. The Krenning score, a well-established qualitative measure [3,14], the metastases to liver ratio (M/ L ratio), a first scanner-independent quantitative surrogate in [ 68 Ga-DOTA 0 -Phe 1 -Tyr 3 ] octreotide (DOTATOC)-positron emission tomography (PET) combined with computed tomography (CT) [15], as well as the intratumoral SSR-heterogeneity in [ 68 Ga]DOTATATE--PET/CT [16] represent the most common methodologies in prediction of response in PRRT. The assessment of asphericity (ASP), characterizing lesion's spatial SSR-heterogeneity e.g. by [ 111 In-DTPA 0 ]octreotide ([ 111 In]octreotide, OctreoScan) scintigraphy [13], demonstrated an further promising parameter to predict response in PRRT planning [17]. Whereas, translation and verification of this novel image based methodology to intra-therapeutic workflow is missing.
The aim of the present study was to evaluate the predictive capability of tumoral SSR-receptor heterogeneity based on intra-therapeutic imaging with [ 177 Lu]DOTATATE. Secondly, we correlated ASP derived form a pre-therapeutic OctreoScan (ASP [In] )and an intra-therapeutic [ 177 Lu]DOTATATE-SPECT/CT imaging (ASP [Lu] ).
All patients provided written informed consent on the evaluation of their data, and approval from the institutional ethics committee of the Otto-von-Guericke-University (reference ID: RAD279; vote, 07/16) was obtained.

SPECT/CT imaging
Intra-therapeutic imaging was performed with a dual head SPECT/CT (Discovery NM/ CT670, GE, Haifa, Israel) according to standard protocols. Planar imaging and SPECT with low-dose CT (X-ray tube current = 40 mAs) of thorax and abdomen for morphological correlation was carried out 24 hours p.i.. SPECT data were reconstructed by iterative algorithms with CT-based correction for attenuation as previously described [13]. Imaging data was evaluated with a dedicated workstation (Xeleris-Workstation, GE Healthcare, Waukesha, USA) at standard clinical settings. Diagnostic CE-CT as well as contrast enhanced MRI of the liver was performed according to a standard protocol, published elsewhere [13,19].

Image analysis
Intra-therapeutic SPECT data sets were evaluated regarding SSR-heterogeneity in accumulation pattern as previously described [13]. In brief, SPECT data measured 24h p.i. at the first cycle [ 177 Lu]-DOTATATE-PRRT was analyzed by a dedicated software (ROVER version 2.1.20, ABX, Radeberg, Germany). The functional tumor volume, a measure of the SSR uptake of tumor tissue, was delineated by a semi-automatic algorithm based on adaptive threshold taking the local background into account [20,21]. For the resulting volumes of interest (VOI) the ASP was computed using the following equation: where S and V are the mean surface and volume of the lesions' functional active part, respectively. The ASP described the deviation of an activity accumulation from spherical geometry. A small ASP (Range 1-5%) represents a more spherical accumulation while increasing ASP represents considerably deviating pattern (e.g. elliptical). A more detailed description has been given in recent studies [13,22].
For lesion based analysis the heterogeneity of lymph-node-, bone-, liver-as well as pancreatic-metastases were assessed to ensure a systemic and not organ-based approach. According to the response evaluation criteria in solid tumors (RECIST 1.1) only two lesions per organ and not more than five per patient were included. In the case of more than two metastases per organ ASP was derived from two individual lesions with the most representative SSR-uptakes on visual assessment.
The association between ASP and results from RECIST-evaluation was evaluated for three different time points (pre PRRT: mean = 4 weeks, range: 1-9 weeks), 4 months post PRRT, and 12 months post PRRT). In generally, morphological assessment was performed by CE-CT. Alternatively, CE-MRI was used for morphological evaluation (contrast-enhanced, CE-MRI; n = 23/33 70%) if available. Analysis of the CE-CT (CE-MRI) data was performed using dedicated radiologic workstation INFINITT (INFINITT Healthcare Co., Ltd., Seoul, South Korea). In order to permit comparison with CE-CT, generally the short axis of lymph-nodes and long diameters in the transverse plane for liver, pancreatic and bone metastases were drawn. Lesions with a short axis of <15 mm (lymph nodes) and <10 mm (pancreas, liver lesions, bone lesions) were not included into analysis to avoid partial volume effects. The morphological changes in diameter were classified according to the RECIST 1.

Statistical analysis
Data analyses were performed using SPSS 22 (IBM Corp., Armonk, NY, USA). Descriptive values were expressed as median, interquartile range (IQR, 25th percentile-75th percentile), and range (minimum-maximum) and depicted as boxplots. According to histograms and quantile-quantile plots, the distribution of data-sets was assumed non-parametric. The ASP within the context of intra-therapeutic dosimetry was analyzed using the Kruskal-Wallis test and receiver operating characteristic curve (ROC). The optimal threshold for separation of RL and NRL was calculated using the Youden's index [24].
The cutoff values were assessed separately for prediction response at 4 and 12 months post PRRT. ASP values was binarized using ROC cutoffs and the associated sensitivity and specificity were determined. Statistical significance was assumed at at a P value < 0.05 and high significance at P value < 0.001. A subanalysis was performed in patients who underwent a pretherapeutic [ 111 In]octreotide SPECT/CT. The correlation of ASP [In] and ASP [Lu] was compared by Spearman's rank correlation coefficient rho (ρ) and the limits of agreement were determined in a Bland-Altman diagram. Concordance was assumed at a deviation of � 5%.
A total of 131 lesions (liver: n = 59, lymph nodes: n = 48, bone: n = 19, pancreas: n = 5) were analyzed. Of the 131 lesions, 104 lesions had been assigned to the RLs and 27 to the NRLs

Discussion
In this retrospective study, we investigated the spatial heterogeneity (ASP) in SSR-positive/ functional metastatic GEP-NEN, using intra-therapeutic SPECT/CT data derived from the first cycle PRRT. Using this methodology we showed a strong association of ASP with morphological treatment response in patients undergoing PRRT with [177Lu]-DOTATATE. Even in diagnostic and in therapeutic imaging using [ 111 In]-OctreoScan or [ 177 Lu]-PRRT we demonstrated a strong correlation, resulting in high sensitivity and specificity.
A number of studies have already dealt with different texture analyses aimed at differentiation between benign and malignant lesions [25]. The tumor heterogeneity in lung carcinoma was evaluated by CT texture analysis. Considering a quite good contrast in lung tissue the quantification within lung tumors can be easily assessed. However, this is not the case within liver or other organ lesions. Here CT delination would be more challenging until semi-automatic segmentation of tumor delineation is limited due to worse tissue contrast in CT images. Lesions have to be evaluated in a manually, volumetric analysis-slice by slice. MRI delivers a better soft tissue contrast and the potential for accurate segmentation of metastases in the liver and other organs, respectively. Besides, several other studies worked with texture oriented analysis of morphologic data from CT or MRI in various entities (e.g. lung carcinoma [26] or  [16]. Heterogeneity was the favored parameter in distinction between responder and nonresponder. In particular, heterogeneity parameters (such as homogeneity, correlation, size variation and short zone emphasis and "entropy" with the highest AUC of 0.70) outperformed conventional parameters like standardized uptake value (SUV) and total receptor expression. In line with these findings, heterogeneity analysis of thyroid cancer showed only significant results in a patient-based setting. Despite our results, in a lesion based analysis only the parameter "entropy" demonstrated with an AUC of 0.73 and a definitely lower sensitivity of 67% and a specificity of 75% response prediction [29]. For adequate therapy-monitoring lesion-based control is always preferable compared to patient-based follow-up. In NEN the mitotic count (MC) and Ki-67 are the most common and reliable biomarkers for cell-proliferation and thus, indicators for biological phenotypes [30,31]. Although MC was used as the biomarker for GEP-NEN during the last decades, finally Khan et al. [32] showed the disparity of MC and Ki-67 and favored Ki-67 as a prognostic marker in grading. Actual Ki-67 plays the major role in dividing GEP-NEN into different grades of disease (G 1-3 NEN/ G 3 -NEC) which directly affects therapy management [33]. Unfortunately, the Ki-67, as a prognostic marker, delivers some pitfalls, physicians have to be aware of. Assessment of Ki-67 depends on expertise of the reporting pathologist [34] and core biopsies represent only a small tumor sample, which impedes accurate heterogeneity assessment, especially of intermediate G 2 -lesion [35]. Lastly, Ki-67 index is known to fluctuate in some patients, not only with the choice of therapy but also during the time of treatment [36,37]. Considering these findings, non-invasive and reliable whole-body assessment of inter-lesional heterogeneity may reflect the whole tumor burden in a more representative way-taking the divergent tumor areas into account. Even the well-established urine and serum-based parameters for treatment monitoring and follow-up, such as 5-hydroxyindole acetic acid (5-HIAA) [37] and Chromogranin A (CgA) [38], failed as prognostic biomarker. Since 18-33% of all patients with metastatic GEP-NEN show PD after PRRT, ASP might deliver a reliable surrogate parameter in selecting potential non-responder [39,40]. Especially, those patients can benefit from an early therapy-switch. Applying the ASP in pretherapeutic octreotide-scintigraphy, we could recently demonstrate that low ASP predicts response to PRRT [13]. However, when dealing with elaborative predictive parameters, it is advantageous to use comparable DOTA-conjugates, such as intra-therapeutic [ 177 Lu-DOTA 0 , Tyr 3 ]octreotate, with equal SSR-affinity to ensure robustness of the test and to minimize varieties.
Nevertheless, subgroup analysis presented a high correlation of ASP [In] and ASP [Lu] and thus, validated the method. The marginal better area under the curve for ASP, obtained in pretherapeutic SPECT/CT imaging, might has been affected by a smaller study population bias.
More recently, functional imaging of NEN with 68 Ga-labelled DOTA analogues (e.g., TATE/ TOC/ NOC) is the preferred SSR-imaging in follow-up examinations, due to better sensitivity, spatial resolution and less radiation [41]. Although there is still limited data available regarding PET-based follow-up, the ENETS consensus guidelines recommend MRI/ CT every 3-12 months and a receptor imaging in G 1-2 P-NEN each 12-24 months after treatment. If there is progression suspected, even anytime earlier [42]. Moreover, the post-therapeutic scan is performed during therapeutic procedure (e.g. for dosimetry) immediately after treatment (24 h p.i.). Providing an early surrogate for tumor response from analyzing. Our results suggest that ASP may deliver a strong approach in intra-therapeutic risk monitoring, indicating that follow-up intervals, e.g. CT every 3-12 months, [ 68 Ga]DOTANOC-PET/CT every 12-24 months, can be even prolonged. An encouraging aspect is the synergistic character of the ASP evaluation. ASP analysis uses functional data from standard diagnostics in patients with GEP-NEN or from dosimetry (in treatment validation) providing benefits without an additional test.
Different limitations may arise: Contrary to other qualitative visual scores, the ASP may deliver a semi-automatic parameter, which ensures reproducibility and robustness in prediction response. Nevertheless, the ASP is not a scanner independent surrogate parameter. It can be hypothesized that ASP is in small lesion affected by partial volume effect. Mainly due to remarkable progress in functional imaging and resolution, lesions smaller than 2.5ml [22], which used to appear with symmetrical sphere shape, can be depicted these days in less than a few millimeters in size. Thus, modern SPECT-devices, such as cadmium-zinc-telluride (CZT) detectors, exhibit a higher spatial resolution [43]. This might avoid adulteration of the ASP in smaller lesions. The limitations of our study are the small sample size, no comparison group and retrospective design. Nevertheless, considering GEP-NEN are commonly rare tumor entity [1], we were able to provide valuable findings and demonstrate significant effects that show some support for our hypotheses. However, further validation of ASP as a potential predictive marker is still required.

Conclusion
Pre-and intra-therapeutic ASP shows a strong correlation with morphological treatment response and might represent a potential predictive marker for PRRT. Further assessment of its predictive value is worthwhile and might improve therapeutic decision making in NEN.