J. J. Baechle1, P. M. Smith2, M. Tan2, C. Bailey2, C. Solorzano2, A. G. Lopez-Aguiar3, M. Dillhoff4, E. W. Beal4, G. Poultsides5, E. Makris5, F. G. Rocha6, A. Crown6, C. Cho7, M. Beems7, E. R. Winslow8, V. R. Rendell8, B. A. Krasnick9, R. Fields9, S. K. Maithel3, K. Idrees2 1Meharry Medical College,School Of Medicine,Nashville, TN, USA 2Vanderbilt University Medical Center,Department Of Surgery,Nashville, TN, USA 3Emory University,Department Of Surgery,Atlanta, GA, USA 4Ohio State University,Comprehensive Cancer Center,Columbus, OH, USA 5Stanford University Medical Center,Palo Alto, CA, USA 6Virginia Mason Medical Center,Seattle, WA, USA 7University Of Michigan,Hepatopancreatobiliary And Advanced Gastrointestinal Surgery,Ann Arbor, MI, USA 8University Of Wisconsin,School Of Medicine And Public Health,Madison, WI, USA 9Washington University,School Of Medicine,St. Louis, MO, USA
Introduction: Pancreatic neuroendocrine tumors (PNETs) are often indolent, but rapidly progressing variants have been reported. To better inform prognosis and treatment decisions, improved understanding of patients at-risk for rapidly progressing PNETs is critical, particularly for patients with small PNETs who may be candidates for expectant management under current treatment guidelines. Specific growth rate (SGR) has been demonstrated in multiple malignancies to be predictive of overall and disease-free survival, but SGR has not been examined in PNETs. The aim of this study is to determine the predictive value of SGR on oncological outcomes in patients with PNETs.
Methods: A retrospective cohort study of adult patients who underwent surgical resection of PNET from 2000-2016 was performed utilizing the multi-institutional U.S. Neuroendocrine Study Group database. Patients with PNET and more than one pre-operative cross-sectional imaging study at least thirty days apart were included in our analysis. The tumor SGR (% growth/day) was calculated using the tumor diameters measured on initial (Di) and final (Df) pre-operative imaging utilizing the previously published equation: SGR = 3ln(Di-Df)/ΔT. Patients with a SGR above the ninetieth percentile were termed “high SGR” and the remaining patients were termed “low SGR”. Overall survival (OS) was analyzed by Kaplan-Meier method and log-rank test. Cox proportional hazard models were used to assess the impact of SGR on OS after adjusting for patient and tumor characteristics.
Results: Of the 1,247 PNET patients who underwent resection, 288 (23%) had two or more pre-operative cross-sectional imaging studies at least 30 days apart. High SGR was associated with higher T Stage at resection (p=0.01), shorter doubling time (p<0.01), and elevated HbA1c (p=0.01). Patients with high SGR also had significantly decreased 5-year OS and disease-specific survival (DSS) compared to those with low SGR (63 vs 80%, p=0.01, Figure 1a; 72 vs 86%, p=0.03, Figure 1b). In patients with small (≤2cm) tumors, high SGR predicted lower 5-year OS (85 vs 91%, p=0.01, Figure 1c). When examining all patients by multivariate analysis controlling for T, N, M stage and HbA1c, high SGR was independently associated with worse OS (Hazard Ratio 2.67, 95% Confidence Interval 1.05 – 6.84, p=0.04).
Conclusion: High SGR in PNETs, including small tumors (<2cm), is associated with worse survival. High SGR can potentially be utilized as a useful marker in the clinical decision process particularly when weighing close observation versus surgical resection in patients with small PNETs.
