Background: Nomograms and grading/staging systems are fundamental in guiding oncology disease management. Histological grading has efficacy in the management of gastroenteropancreatic neuroendocrine tumors (GEP-NET) but tumor heterogeneity renders outcome unpredictable. Clinical parameters alone have limited prognostic ability. A previously described NET nomogram calculated risk for overall survival (OS). To augment accuracy, we developed a circulating NET multigene expression blood test. The NETest exhibits high level metrics as a molecular prognostic for progression free survival (PFS). We evaluated whether combining clinical information with the NETest would provide accurate prognostic information for PRRT treatment.

Aim: Create a combined clinical and gene expression nomogram to predict OS and PFS in PRRT-treated GEP-NETs.

Methods: 177Lu-based-PRRT-treated GEP-NETs (n=57; median age 65 yrs (31-83); 34M:23F; small bowel (47%), pancreas (35%), predominantly G2: 65%) followed for a median 15 months (range 1-33). The clinical nomogram comprised 10 variables (including: age, gender, grade, Ki67, stage symptoms, liver metastases, somatostatin analog use, surgery extent, Chromogranin A level) with a score range: 25-200. This data was interpolated with pre-therapy NET transcript analysis results (range 0-100) into a hybrid molecular nomogram (range: 25-300). Outcomes were OS and PFS (RECIST criteria). Receiver operating characteristics (ROC), Kaplan-Meier survival and non-parametric (Mann-Whitney) analyses were performed.

Results: Median OS was not reached (8 deaths); mPFS was 19 months (20/57 disease progression). Neither grade nor stage alone predicted a difference in OS or PFS. Clinical nomogram scores for non-survivors was significantly higher (139±20 vs. 87±6, p<0.001) than survivors. Pre-treatment NETest values were higher in non-survivors (56±10 vs. 36±3, p<0.04). Integration of clinical and molecular nomogram scores into a hybrid nomogram significantly amplified the prognostic efficacy. Non-survivors had higher scores than survivors (189±14 vs. 113±9, p<0.001). A cut-off score of >130 exhibited an AUC for OS prediction of 0.84±0.05, (p=0.002) with a Hazard ratio (HR) of 8.7 (2.1-35.4, p<0.002). A score of <130 predicted 100% survival. PFS prediction using the hybrid nomogram cut off of 130 had an AUC of 0.68±0.07, p=0.02. The Gehan-Breslow-Wilcoxon test value was 7.2 (p<0.008) demonstrating that high scores were prognostic of early progression (treatment failure). The HR was 3.1 (1.2-7.4, p<0.02) and the mPFS was 17 months vs. not reached.

Conclusions: Hybrid molecular nomogram data derived from a circulating gene expression signature combined with a clinical NET nomogram accurately predicts OS in PRRT-treated GEP-NETs. The nomogram also predicted PFS particularly individuals with elevated transcript values who will not respond to PRRT.

Citation Format: Mark Kidd, Lisa Bodei, Stefano Severi, Ignat Drozdov, Sylvia Nicolini, Giovanni Paganelli, Irvin M. Modlin. A gene expression-based nomogram predicts overall and progression free survival in PRRT-treated gastroenteropancreatic neuroendocrine tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1789. doi:10.1158/1538-7445.AM2017-1789