Pancreatic neuroendocrine tumor: Added value of fusion of T2-weighted imaging and high b-value diffusion-weighted imaging for tumor detection

https://doi.org/10.1016/j.ejrad.2012.01.032Get rights and content

Abstract

Objectives

To investigate the added value of fusion of high b-value diffusion-weighted images (DWI) and T2-weighted (T2) MR images for the detection of pancreatic neuroendocrine tumors (PNT).

Methods

18 patients with 18 histologically proven PNT were included. Two radiologists independently and retrospectively reviewed four randomized images sets (T2+T1, DWI, T2+DWI, and DWI+T2 fusion). Lesion detection confidence level was assessed using a three grade score (no lesion; uncertain lesion and certain lesion); lesion size and signal intensity were recorded. Apparent diffusion coefficients (ADC) of tumor and adjacent pancreas were measured.

Results

Readers 1 and 2 respectively detected 14/18 and 16/18 lesions on T2+T1, 13/18 and 12/18 on DWI, 16/18 and 15/18 on T2+DWI and 17/18 and 16/18 on DWI+T2 fusion. Lesion median size was 16 mm (range: 7 mm–40 mm), 22% were hyperfunctioning (all insulinomas) and 72% were low-grade (Rindi 1). All tumors except one (with cystic component) showed lower ADC than adjacent pancreatic parenchyma. Fusion imaging had significantly better detection score by both authors (p < 0.005) and provided the higher inter-reader agreement (kappa 0.7). DWI alone had the worst score for both readers.

Conclusions

Fusion images improve the detection of PNT, especially in patients with small isointense lesions on conventional MR sequences.

Introduction

Pancreatic neuroendocrine tumor (PNT) corresponds to 1–2% of all pancreatic neoplasms [1] and typically presents as a hypervascular nodule [2]. Although considered as a rare pancreatic neoplasm, its incidence is increasing, which may be related to an increase of diagnostic sensitivity due to the development of imaging technologies [3].

Current magnetic resonance imaging (MRI) has a high sensitivity for PNT detection and should be recommended as the first imaging modality in patients with suspected PNT [4], allowing the investigation of the pancreas as well as the assessment of the disease extension into the liver. These tumors may present a broad spectrum of appearance. Nevertheless, their classic imaging features as arterial enhancement and hyperintensity on T2-weighted MR images may not be present, rendering them “invisible” on conventional MRI [2]. This is particularly true for small hyperfunctioning PNT [2], [5]. Diffusion-weighted imaging (DWI) may better depict those small PNT due to its greater image contrast [1]. However, as the b-value increases, anatomical location may become challenging because of the reduced signal-to-noise ratio and increased susceptibility artifacts. To overcome this problem, DWI can be fused to T2-weighted images (T2), improving anatomical localization of the findings identified on high b-value DWI [6].

There are promising published results on the use of fused DWI and T2 for the study of pelvic malignancy recurrence [7], pelvic lymph nodes [8], myometrial invasion in endometrial cancer [9], prostate cancer [6] and malignant tumor screening in general [10]. However, information about the use of DWI and T2 fusion to study PNT is scarce [5]. Our aim was to investigate the added value of b-1000 s/mm2 DWI and T2-weighted fused images, among sequences without the use of endovenous contrast, for the detection of pancreatic neuroendocrine neoplasms, using as reference the surgical and histological reports of the excised tumors.

Section snippets

Patients

The study complied with the recommendations of our Institutional Ethics Board and informed consent was waived.

We undertook a retrospective review of the medical records of 24 consecutive patients with proven PNT from 2007 to 2011. Six patients were excluded from the study because the baseline MR (previous to surgery) did not contain DWI sequences. Eighteen patients were finally included and all had histological confirmation of the PNT, 17 patients had a final confirmation by surgical excision.

Demographic characteristics and descriptive results

The median time between MR and surgery was 66 days (IQR 50 days). There were 12 men and 6 women, the median age was 61 years (IQR 14 years). The median tumors’ size was 16 mm (range: 7 mm–40 mm; IQR 10 mm) with 12/18 (67%) of the tumors measuring less than 20 mm. Most of the nodules were Rindi I (72%) or low-grade tumors (i.e. less than 2% of mitoses/10 HPF or less then 2% of Ki67 labeling). The remaining tumors were Rindi II (i.e. no more than 20% of KI 67 mitosis index). Nodules location, Rindi's

Discussion

Our sample was predominantly constituted of small and low-grade tumors (67% measured less than 20 mm – median size 16 mm – and 72% were Rindi I). Even so, MRI allowed for a correct diagnosis in all except one patient, whose nodule measured 6 mm and was located on the anterior periphery of the pancreatic body, simulating a peripancreatic lymphnode (Rindi I, T1Nx).

For both readers images’ interpretation was significantly improved when fusion imaging was included in their reading, and this was

Conclusions

The addition of post-processed fusion images of T2 and b-1000 DWI when reading diagnostic MR studies of pancreatic neuroendocrine tumors is helpful, especially in difficult cases of small lesions that might be isointense on conventional MRI. It significantly increases the agreement and confidence on MR diagnosis of this neoplasia, which is often a challenging diagnosis.

References (18)

There are more references available in the full text version of this article.

Cited by (0)

View full text