Watrowski Rafał, Zeillinger Robert
Department of Gynecology and Obstetrics, St. Josefskrankenhaus, Teaching Hospital of the University of Freiburg, Sautierstr. 1, Freiburg, Germany.
Department of Obstetrics and Gynecology, Molecular Oncology Group, Medical University of Vienna, Währinger Gürtel 18-20, Vienna, Austria.
Tumour Biol. 2016 Apr;37(4):4343-9. doi: 10.1007/s13277-015-4280-7. Epub 2015 Oct 23.
We assessed the diagnostic accuracy of a newly developed laboratory score-based on CA125, platelet count (PLT), C-reactive protein (CRP), and fibrinogen levels-in the preoperative diagnosis of adnexal mass. In this retrospective single-center study, we analyzed records of 142 patients with 54 malignant (38 %) and 88 benign (62 %) ovarian tumors. Preoperative levels of CA125, PLT, CRP, and fibrinogen were dichotomized according to the common cutoff values (CA125, 35 U/ml; PLT, 350/nl; CRP, 5.0 mg/l; fibrinogen, 400 mg/dl), resulting in "1" for results above the cutoff and "0" for results within the normal ranges. The values (1 or 0) were summarized to a "low" (0-2) or "high" (3-4) score. Its diagnostic accuracy was compared to the "gold standard," CA125. All parameters differed significantly between malignant and benign cases. The score was false positive in 5/88 (5.7 %) and false negative in 13/54 (24 %) of cases. Conversely, CA125 was false positive in 18/88 (20.4 %) and false negative in 4/54 (7.4 %). The diagnostic accuracy of CA125 (>35 U/ml) was sensitivity 0.93, specificity 0.80, positive predictive value (PPV) 0.74, negative predictive value (NPV) 0.95, and positive likelihood ratio (weighted by prevalence) (+LH/p) 2.78. The diagnostic accuracy of the score was sensitivity 0.76, specificity 0.94, PPV 0.89, NPV 0.86, and +LH/p 8.2. In conclusion, the score is easy to use and generates no additional costs. It provides a better specificity, PPV, and +LH/p than CA125. The sensitivity and NPV are lower, but acceptable. A validation of the score in a large patient cohort is needed.
我们评估了一种新开发的基于CA125、血小板计数(PLT)、C反应蛋白(CRP)和纤维蛋白原水平的实验室评分在附件包块术前诊断中的诊断准确性。在这项回顾性单中心研究中,我们分析了142例患者的记录,其中有54例(38%)为恶性卵巢肿瘤,88例(62%)为良性卵巢肿瘤。根据常用的临界值(CA125,35 U/ml;PLT,350/nl;CRP,5.0 mg/l;纤维蛋白原,400 mg/dl)将术前CA125、PLT、CRP和纤维蛋白原水平进行二分法划分,结果高于临界值记为“1”,结果在正常范围内记为“0”。这些值(1或0)汇总为“低”(0 - 2)或“高”(3 - 4)分。将其诊断准确性与“金标准”CA125进行比较。恶性和良性病例之间所有参数均有显著差异。该评分在88例中的5例(5.7%)为假阳性,在54例中的13例(24%)为假阴性。相反,CA125在88例中的18例(20.4%)为假阳性,在54例中的4例(7.4%)为假阴性。CA125(>35 U/ml)的诊断准确性为:敏感性0.93,特异性0.80,阳性预测值(PPV)0.74,阴性预测值(NPV)0.95,阳性似然比(按患病率加权)(+LH/p)2.78。该评分的诊断准确性为:敏感性0.76,特异性0.94,PPV 0.89,NPV 0.86,+LH/p 8.2。总之,该评分易于使用且不产生额外费用。它比CA125具有更好的特异性(specificity)、PPV和+LH/p。敏感性和NPV较低,但可以接受。需要在大型患者队列中对该评分进行验证。