Diagnostic Efficacy And Influencing Factors Of SSTR-PET/CT For Endoscopy-Suspected GI-NEN

In recent years,due to the popularization of gastrointestinal endoscopic screening,the incidence of gastrointestinal neuroendocrine neoplasm（GI-NEN）is gradually increasing.However,due to the high risk and lack of diagnostic accuracy of biopsy in subepithelial lesion（SEL）,as well as the inability of endoscopy for staging,there is an urgent need to find a non-invasive,highly accurate,whole-body diagnostic modality for endoscopic suspected GI-NEN.Somatostatin receptor（SSTR）is a G protein-coupled receptor specifically expressed on the neuroendocrine neoplasm（NEN）.SSTR-positron emission tomography（PET）imaging,as a molecular imaging modality,can provide qualitative and quantitative information on the SSTR of NEN with a high pooled sensitivity and specificity of 91%and 94%,respectively.SSTR expression is closely related to the differentiation of NEN.Hence,guidelines only recommend SSTR-PET for well-differentiated neuroendocrine tumor（NET）.However,clinical studies referenced to clarify this indication are based on NEN of diverse origin,large tumor size and multiple staging status,while most GI-NEN detected by endoscopic screening are characterized by low grade,small size and early staging.Therefore,the diagnostic efficacy and indications of SSTR-PET for endoscopy-suspected GI-NEN remain to be further defined.Objectives:1.To evaluate the diagnostic efficacy of SSTR-PET for endoscopy-suspected GI-NEN.2.To retrospectively screen and prospectively validate the factors that affect the qualitative diagnosis of SSTR-PET and to further clarify the indications of SSTR-PET for endoscopy-suspected GI-NEN.3.To investigate the macroscopic and microscopic factors affecting the quantitative diagnosis of SSTR-PET.Methods:1.Thirty-nine patients with endoscopy-suspected GI-NEN who underwent SSTR-PET/CT from July 2018 to October 2021 were analyzed as a retrospective analysis set.Twenty-two patients with endoscopy-suspected GI-NEN from November 2021 to December 2022 were recruited as a prospective validation set to perform SSTR-PET/CT.The diagnostic criteria for PET-positive tumor were the tracer uptake above the background,but couldn’t be explained by physiological uptake.All patients were pathologically confirmed after SSTR-PET/CT.2.The diagnostic sensitivity and specificity of SSTR-PET in endoscopy-suspected GI-NEN were calculated.3.Univariate and multivariate Logistic regression analyses were used to screen independent factors affecting the qualitative disgnosis of SSTR-PET in GI-NEN patients in the retrospective analysis set.Candidate factors included age,gender,endoscopy ultrasound（EUS）tumor size,EUS tumor invasion depth,pathological tumor size,pathological tumor invasion depth,grade,intratumor lymphocytes score（ITLS）,Ki-67,somatostatin receptor2（SSTR2）Her2/neu score,and microvascular density（MVD）.Independent factors were verified in the prospective validation set and in all GI-NEN patients equally using univariate and multivariate Logistic regression analyses.4.Receiver operator characteristic（ROC）curves were plotted to explore the cut-off values of key independent factors that distinguish SSTR-PET qualitative diagnosis of GI-NEN patients in the retrospective analysis set,the prospective validation set,and all GI-NEN patients respectively.The optimum threshold was determined by comparing the actual improvements in diagnostic efficacy with different cut-offs.5.Multiple Spearman correlation analyses were performed to explore factors that correlate with the SSTR-PET quantitative parameter[maximum standardized uptake value(SUV_max)],and the semi-quantitative parameter（Krenning score）.Results:1.Fifty-five of 61 patients with endoscopy-suspected GI-NEN were pathologically confirmed（35 in the retrospective analysis set and 20 in the prospective validation set）,of whom 94.5%（53/55）were G1/G2 NET,whereas the overall diagnostic sensitivity of SSTR-PET/CT was 61.8%（34/55）and the specificity was 83.3%（5/6）.In the retrospective analysis set,97.1%（34/35）GI-NEN were G1/G2 NET and 94.3%（33/35）were positive in SSTR2 immunohistochemical staining,whereas the diagnostic sensitivity of SSTR-PET/CT was 65.7%（23/35）.2.In univariate Logistic regression analyses of GI-NEN patients in the retrospective analysis set,age,EUS tumor size,EUS tumor invasion depth,pathological tumor size,and MVD were significantly different between the PET-positive and PET-negative group（P<0.2）.In multivariate Logistic regression analyses,EUS tumor size was the independent influencing factor for SSTR-PET qualitative diagnosis（P=0.036）.EUS tumor size was also the independent influencing factor for SSTR-PET qualitative diagnosis in all GI-NEN patients（P=0.002）.3.The area under the ROC curve（AUC）of EUS tumor size of GI-NEN patients in the retrospective analysis set was 0.817（P=0.002）with a cut-off value of 6.35 mm.The AUC of EUS tumor size of GI-NEN patients in the prospective validation set was 0.924（P=0.001）with a cut-off value of 6.5 mm.The AUC of EUS tumor size of all GI-NEN patients was 0.849（P<0.001）with a cut-off value of 6.5 mm.The EUS tumor size thresholds of6.35 mm and 6.5 mm both improved the diagnostic sensitivity of SSTR-PET from 61.8%to100%,and the diagnostic accuracy from 63.9%to 96.2%and 96.0%,respectively.4.EUS tumor size was strongly correlated with tumor SUV_max（ρ=0.61,P=0.002）and Krenning score（ρ=0.55,P=0.006）.MVD was moderately correlated with EUS tumor size（ρ=0.41,P=0.049）,SUV_max（ρ=0.48,P=0.021）and Krenning score（ρ=0.47,P=0.024）.Ki-67,although moderately correlated with SUV_max（ρ=0.43,P=0.043）,was not correlated with Krenning score（ρ=0.39,P=0.069）.Conclusion:1.Despite the majority（94.5%）of GI-NEN included in this study were well-differentiated G1/G2 NET for which SSTR-PET imaging is recommended by guidelines and the positive rate of SSTR2 immunohistochemical staining was 94.3%in the retrospective analysis set,the overall sensitivity of SSTR-PET（61.8%）was much lower than the pooled sensitivity（91%）previously reported.2.For endoscopic GI-NEN,tumor size was an independent factor for qualitative diagnosis of SSTR-PET,while SSTR2 expression was not.The larger the tumor,the higher the probability of positive result and the diagnostic sensitivity in SSTR-PET.Taking 6.35mm as the EUS tumor size threshold for performing SSTR-PET could improve the diagnostic sensitivity of SSTR-PET from 61.8%to 100%,and the diagnostic accuracy from63.9%to 96.2%.Therefore,tumor size larger than 6.35 mm can be used as an indication of SSTR-PET for endoscopy-suspected GI-NEN to avoid false negative results.3.Tumor size and MVD were both positively correlated with the SSTR-PET quantitative parameter(SUV_max)and the semi-quantitative parameter（Krenning score）.This suggests that for endoscopic GI-NEN,tracer uptake of SSTR-PET is mainly determined by tumor size and blood perfusion instead of SSTR2 expression.