Univeristy Hospital Center Zagreb, Department of Gastroenterology, Zagreb, Croatia.
University Hospital Center Zagreb, Zagreb, Croatia.
Endosc Ultrasound. 2014 Apr;3(Suppl 1):S4.
Endoscopic ultrasonography (EUS) quantitative elastography methods are developed for non-invasive differentiation of pancreatic masses.
First: To evaluate the diagnostic value of strain ratio (SR) and hue histogram (HH) in patients with pancreatic masses and to determine the cut-off value between pancreatic cancer and focal pancreatitis using a pancreatic tissue close to the mass as a reference area. Second: To calculate new variable HH ratio (HHR) in an attempt to improve sensitivity, specificity and accuracy of the method.
In a prospective single center study, 149 patients were examined: 105 with the pancreatic masses and 44 controls using Pentax EUS linear probes in combination with Hitachi platforms. SR and HH were automatically calculated by machine software. Finally, two groups were formed: Pancreatic cancer group (58 patients) and focal pancreatitis group (47 patients). All statistical analysis has been made in SPSS 14.0 (SPSS Inc., Chicago, USA).
Statistical analysis in our study showed that SR with a cut-off value of 7.59 reaches 100% sensitivity and 95% specificity with overall accuracy of 97% (confidence intervals [CI]: 92-97%) in detection of pancreatic cancer. Statistical analysis also showed that HH with a cut-off value of ≥86 reaches 100% sensitivity and just 45% specificity with overall accuracy of 66% (CI: 61-66%) in detection of pancreatic cancer. New variable HHR with cut-off ≥1.153 was slightly better with 98% sensitivity and 50% specificity, with overall accuracy of 69% (CI: 63-70%). SR showed significantly higher specificity compared with HH and HHR. More HH studies on Hitachi platforms are needed.
内镜超声检查(EUS)定量弹性成像方法是为了对胰腺肿块进行非侵入性鉴别而开发的。
首先,评估应变比(SR)和色调直方图(HH)在胰腺肿块患者中的诊断价值,并使用靠近肿块的胰腺组织作为参照区域来确定胰腺癌和局灶性胰腺炎之间的分界值。其次,计算新的变量 HH 比(HHR),试图提高该方法的敏感性、特异性和准确性。
在一项前瞻性单中心研究中,对 149 例患者进行了检查:105 例胰腺肿块患者和 44 例对照组,使用 Pentax EUS 线性探头与 Hitachi 平台相结合。SR 和 HH 由机器软件自动计算。最后,将两组分为:胰腺癌组(58 例)和局灶性胰腺炎组(47 例)。所有统计分析均在 SPSS 14.0(SPSS Inc.,芝加哥,美国)中进行。
我们的研究统计分析表明,SR 的截断值为 7.59 时,对胰腺癌的检测灵敏度为 100%,特异性为 95%,总准确率为 97%(置信区间[CI]:92-97%)。统计分析还表明,HH 的截断值≥86 时,对胰腺癌的检测灵敏度为 100%,特异性仅为 45%,总准确率为 66%(CI:61-66%)。新的变量 HHR 的截断值≥1.153 时略好,灵敏度为 98%,特异性为 50%,总准确率为 69%(CI:63-70%)。SR 与 HH 和 HHR 相比,特异性明显更高。需要在 Hitachi 平台上进行更多的 HH 研究。