Abstract

BACKGROUND:

The clinical course of pancreatic endocrine tumors (PET) depends on tumor size, the presence of invasion or metastasis, the Ki-67 index, mitoses per high power field, and mutational damage. Most of this information is not available before surgery for clinical decision making or prognostication.

OBJECTIVE:

To evaluate PET EUS-guided FNA (EUS-FNA) microsatellite loss analysis in the context of PET-related mortality.

DESIGN:

A single institution retrospective cohort.

PATIENTS:

Patients with PET diagnosed by EUS-FNA who underwent DNA microsatellite loss analysis and at least 1 year of follow-up or subsequent death.

INTERVENTION:

PET microsatellite loss analysis results and current clinical status were compared.

RESULTS:

Twenty-nine patients were included in the final analysis; the mean age of the patients was 57 years, and 10 were women (35%). The mean follow-up was 33.7 months (median 30 months, range 2-66 months). Twelve patients had disease progression, and 8 died, all from disease-specific causes. Malignant PET contained multiple microsatellite losses, with a median fractional allelic loss (FAL) of 0.37 (range 0.12-0.69, interquartile range [IQR] 0.23-0.42), significantly different from benign PET, median FAL 0 (range 0-0.18, IQR 0-0.08, P < .0001). Survival analysis revealed a significant difference in disease recurrence or progression at 2 years (P < .0001) and in the 5-year survival between patients with FAL </=0.2 compared with >0.2 (P < .0001). Logistic regression could not be performed because of the perfect association between an FAL >0.2 and disease status or mortality.

LIMITATIONS:

Retrospective design, referral bias, and DNA analysis availability.

CONCLUSIONS:

PET EUS-FNA microsatellite loss analysis provides preoperative prognostic information. An FAL >0.2 is not only associated with disease progression but also with mortality.