PURPOSE To investigate the correlation between multiple detector computed tomography (MDCT) features of pancreatic neuroendocrine tumors (pNETs) and histopathologic grade and find valuable imaging criteria for grade prediction. MATERIAL AND METHODS MDCT of 61 patients with 65 masses, which pNETs were approved histopathologically, underwent revision retrospectively. Each MDCT was evaluated for various radiologic characteristics. Absolute and relative (R: tumor/pancreas, D: tumor-pancreas) tumor enhancements were calculated in multiple post contrast phases. RESULTS 61 patients [mean age = 50.70 ± 14.28 y/o and 30(49.2%) were male] were evaluated and classified into 2 groups histopathologically: G1: 32 (49.2%) and G2,3: 33 (50.8%). Significant relationships were observed between histopathologic tumor grade regarding age (p = 0.006), the longest tumor size (p = 0.006), presence of heterogeneity (p < 0.0001), hypodense foci in delayed phase (p = 0.004), lobulation (p = 0.002), vascular encasement (p < 0.0001), adjacent organ invasion (p = 0.01), presence (p < 0.0001) and number (0.02) of liver metastases, presence of lymphadenopathy with short axis of more than 10 mm (LAP) (p = 0.008), pathologic lymph node size (p = 0.004), relative (R and D) (p = 0.05 and 0.02, respectively), and percentage of arterial hyper-enhancing area (p = <0.0001). Tumor grades, however, had no significant relationship with gender, tumor location, tumor outline, calcification, cystic change, or pancreatic (PD) or biliary duct (BD) dilation (p = 0.21, 0.60, 0.05, 0.05 1, 0.10, and 0.51, respectively). Then, we suggested a novel imaging criteria consisting of six parameters (tumor size > 33 mm, relative (R) tumor enhancement in arterial phase ≤ 1.33, relative (D) tumor enhancement in arterial phase ≤ 16.5, percentage of arterial hyper-enhancing area ≤ 75%, vascular encasement, and lobulation), which specificity and accuracy of combination of all findings (6/6) for predicting G2,3 were 100% and 70.1%, respectively. The highest accuracy (84.21%) was seen in combinations of at least 4 of 6 findings, with 80.00% sensitivity, 87.5% specificity, 83.33% PPV, and 84.85% NPV. CONCLUSION We suggested reliable imaging criteria with high specificity and accuracy for predicting the histopathologic grade of pNETs.

中文翻译：

---

胰腺神经内分泌肿瘤 (pNETs)：MDCT 特征在组织病理学分级分化中的预测价值。

目的 研究胰腺神经内分泌肿瘤 (pNET) 的多检测器计算机断层扫描 (MDCT) 特征与组织病理学分级之间的相关性，并为分级预测找到有价值的影像学标准。材料与方法 对 61 例 65 个肿块的 MDCT 进行了回顾性修正，pNETs 经组织病理学批准。对每个 MDCT 的各种放射学特征进行评估。在多个对比后阶段计算绝对和相对（R：肿瘤/胰腺，D：肿瘤-胰腺）肿瘤增强。结果 对 61 名患者 [平均年龄 = 50.70 ± 14.28 岁/岁和 30 (49.2%) 名男性] 进行了评估，并在组织病理学上分为 2 组：G1：32 (49.2%) 和 G2,3：33 (50.8%)。在组织病理学肿瘤分级与年龄之间观察到显着关系（p = 0.006），最长的肿瘤大小 (p = 0.006)、异质性的存在 (p < 0.0001)、延迟期低密度病灶 (p = 0.004)、分叶 (p = 0.002)、血管包裹 (p < 0.0001)、邻近器官侵犯 (p) = 0.01)，肝转移的存在 (p < 0.0001) 和数量 (0.02)，存在短轴超过 10 毫米的淋巴结病 (LAP) (p = 0.008)，病理淋巴结大小 (p = 0.004)，相对（R 和 D）（分别为 p = 0.05 和 0.02），以及动脉过度增强区域的百分比（p = <0.0001）。然而，肿瘤分级与性别、肿瘤位置、肿瘤轮廓、钙化、囊性改变或胰腺 (PD) 或胆管 (BD) 扩张没有显着关系（p = 0.21、0.60、0.05、0.05 1、0.10 和0.51，分别）。然后，我们提出了一种新的成像标准，包括六个参数（肿瘤大小 > 33 mm，动脉期相对（R）肿瘤增强≤1.33，动脉期相对（D）肿瘤增强≤16.5，动脉高强化区百分比≤75%，血管包裹和分叶），其特异性和准确性所有结果 (6/6) 的组合预测 G2,3 分别为 100% 和 70.1%。最高准确度 (84.21%) 出现在 6 个结果中的至少 4 个的组合中，敏感性为 80.00%，特异性为 87.5%，PPV 为 83.33%，NPV 为 84.85%。结论 我们建议使用具有高特异性和准确性的可靠影像标准来预测 pNETs 的组织病理学分级。所有结果 (6/6) 的组合预测 G2,3 的特异性和准确性分别为 100% 和 70.1%。最高准确度 (84.21%) 出现在 6 个结果中的至少 4 个的组合中，敏感性为 80.00%，特异性为 87.5%，PPV 为 83.33%，NPV 为 84.85%。结论 我们建议使用具有高特异性和准确性的可靠影像标准来预测 pNETs 的组织病理学分级。所有结果 (6/6) 的组合预测 G2,3 的特异性和准确性分别为 100% 和 70.1%。最高准确度 (84.21%) 出现在 6 个结果中的至少 4 个的组合中，敏感性为 80.00%，特异性为 87.5%，PPV 为 83.33%，NPV 为 84.85%。结论 我们建议使用具有高特异性和准确性的可靠影像标准来预测 pNETs 的组织病理学分级。