Gyllensten Ulf, Hedlund-Lindberg Julia, Svensson Johanna, Manninen Johanna, Öst Torbjörn, Ramsell Jon, Åslin Matilda, Ivansson Emma, Lomnytska Marta, Lycke Maria, Axelsson Tomas, Liljedahl Ulrika, Nordlund Jessica, Edqvist Per-Henrik, Sjöblom Tobias, Uhlén Mathias, Stålberg Karin, Sundfeldt Karin, Åberg Mikael, Enroth Stefan
Department of Immunology, Genetics, and Pathology, Biomedical Center, SciLifeLab Uppsala, Uppsala University, SE-75108 Uppsala, Sweden.
Stellenbosch Institute for Advanced Study (STIAS), Marais Rd., Mostertsdrift, Stellenbosch 7600, South Africa.
Cancers (Basel). 2022 Mar 30;14(7):1757. doi: 10.3390/cancers14071757.
Ovarian cancer is the eighth most common cancer among women and has a 5-year survival of only 30-50%. The survival is close to 90% for patients in stage I but only 20% for patients in stage IV. The presently available biomarkers have insufficient sensitivity and specificity for early detection and there is an urgent need to identify novel biomarkers.
We employed the Explore PEA technology for high-precision analysis of 1463 plasma proteins and conducted a discovery and replication study using two clinical cohorts of previously untreated patients with benign or malignant ovarian tumours ( = 111 and = 37).
The discovery analysis identified 32 proteins that had significantly higher levels in malignant cases as compared to benign diagnoses, and for 28 of these, the association was replicated in the second cohort. Multivariate modelling identified three highly accurate models based on 4 to 7 proteins each for separating benign tumours from early-stage and/or late-stage ovarian cancers, all with AUCs above 0.96 in the replication cohort. We also developed a model for separating the early-stage from the late-stage achieving an AUC of 0.81 in the replication cohort. These models were based on eleven proteins in total (ALPP, CXCL8, DPY30, IL6, IL12, KRT19, PAEP, TSPAN1, SIGLEC5, VTCN1, and WFDC2), notably without MUCIN-16. The majority of the associated proteins have been connected to ovarian cancer but not identified as potential biomarkers.
The results show the ability of using high-precision proteomics for the identification of novel plasma protein biomarker candidates for the early detection of ovarian cancer.
卵巢癌是女性中第八大常见癌症，其5年生存率仅为30% - 50%。I期患者的生存率接近90%，而IV期患者仅为20%。目前可用的生物标志物在早期检测方面的敏感性和特异性不足，迫切需要识别新的生物标志物。
我们采用探索性蛋白质组学分析（Explore PEA）技术对1463种血浆蛋白进行高精度分析，并使用两个临床队列（分别为111例和37例）对未经治疗的良性或恶性卵巢肿瘤患者进行了发现和验证研究。
发现分析确定了32种蛋白质，与良性诊断相比，这些蛋白质在恶性病例中的水平显著更高，其中28种蛋白质的关联在第二个队列中得到了验证。多变量建模基于每种4至7种蛋白质确定了三个高度准确的模型，用于将良性肿瘤与早期和/或晚期卵巢癌区分开来，在验证队列中所有模型的曲线下面积（AUC）均高于0.96。我们还开发了一个区分早期和晚期的模型，在验证队列中的AUC为0.81。这些模型总共基于11种蛋白质（碱性磷酸酶（ALPP）、趋化因子配体8（CXCL8）、二氢嘧啶酶样30（DPY30）、白细胞介素6（IL6）、白细胞介素12（IL12）、角蛋白19（KRT19）、妊娠相关血浆蛋白（PAEP）、四跨膜蛋白1（TSPAN1）、唾液酸结合免疫球蛋白样凝集素5（SIGLEC5）、V域免疫球蛋白超家族成员1（VTCN1）和乳脂肪球表皮生长因子8（WFDC2）），值得注意的是不包括粘蛋白16（MUCIN-16）。大多数相关蛋白质与卵巢癌有关，但未被确定为潜在的生物标志物。
结果表明，高精度蛋白质组学有能力识别用于早期检测卵巢癌的新型血浆蛋白生物标志物候选物。
