ISSN 1470-3947 (print) | ISSN 1479-6848 (online)

Endocrine Abstracts (2017) 52 OC1 | DOI: 10.1530/endoabs.52.OC1

A predictive quotient index, comprising neuroendocrine gene cluster analysis in blood and tissue grading is specifically predicts PRRT efficacy

Lisa Bodei1, Mark Kidd2, Wouter van der Zwaan3, Aviral Singh4, Stefano Severi5, Ignat Drozdov2, Dik Kwekkeboom3, Jaroslaw Cwikla6, Agnieska Kolasinska-Cwikla7, Richard P Baum4, Giovanni Paganelli5, Eric Krenning3 & Irvin M Modlin8

1Memorial Sloan Kettering Cancer Center, New York, USA; 2Wren Laboratories, Branford, Connecticut, USA; 3Erasmus Medical Center, Rotterdam, Netherlands; 4Zentralklinik, Bad Berka, Germany; 5ISRSCT, Meldola, Italy; 6University of Warmia and Mazury, Olztyn, Poland; 7Insitute of Oncology, Warsaw, Poland; 8Yale University, New Haven, Connecticut, USA.

Background: The efficacy of PRRT is based upon NET over expression of somatostatin receptor (SSR) to deliver targeted isotope therapy. SSR expression (Krenning scale) compared to Predictive Quotient Index (PQI) (circulating NET transcript analysis mathematically integrated with grade) indicates the latter is more accurate for predicting PRRT efficacy. We evaluated whether PQI was specifically predictive or was prognostic for PRRT compared to other therapeutic strategies.

Methods: We evaluated three treatment cohorts. 177Lu-PRRT-treatment (n=146 (Rotterdam: Meldola; Bad Berka) and two Comparator cohorts. These comprised GEPNETs (n=106, in a watch-and-wait program) and somatostatin analog (SSA)-treated GEP-NETs (n=28). Blood prospectively collected. Baseline evaluations: Grade (Ki67) and NETest (qRT-PCR - multianalyte algorithmic analyses). All samples were blinded. The PQI (NETest genes regulating two ‘omic’ clusters metabolism and growth factor signaling) integrated with the Ki67 index. The PQI has two prediction outputs: ‘PRRT-responder’ (R) vs ‘PRRT-non-responder’ (NR). Disease control was by RECIST criteria (R (stable, partial and complete response) vs NR (disease progression)) Statistics: Kaplan-Meier survival analysis.

Results: PRRT cohort (n=146). Median follow-up: 14–16 months. Cohort Meldola: mPFS for patients identified as ‘PRRT-responders’ was not reached versus predicted ‘non-responders’ mPFS 17 months (χ2=38, P<0.0001). Cohort Bad Berka: Not reached vs. 17 months (χ2=27.4, P<0.0001). Cohort Rotterdam: Not reached vs. 10.4 months (χ2=34.9, P<0.0001). The PQI accurately predicted response in 94-97% of PRRT-treated individuals. SSA cohort (n=28). Median follow-up 11 months (9–15). No significant difference in mPFS was noted between ‘Responders’ and ‘Non-responders’. The PQI does not predict SSA response. Comparator cohort (n=106). Median follow-up 19 months (1–36). No significant differences in median survival between ‘Responder’ vs Non-Responder (both mPFS: 24 months). The PQI was neither predictive nor prognostic in the comparator groups.

Conclusion: An integrated measurement (PQI) of ‘omic’ NET gene analysis with grading in an individual tumor is a specific predictive marker for PRRT therapeutic efficacy in neuroendocrine tumors.