PURPOSE: Circulating tumor DNA (ctDNA) detection is a minimally invasive technique that offers dynamic molecular snapshots of genomic alterations in cancer. Although ctDNA markers can be used for early detection of cancers or for monitoring treatment efficacy, the value of ctDNA in guiding treatment decisions in solid cancers is controversial. Here, we monitored ctDNA to detect clinically actionable alterations during treatment of high-grade serous ovarian cancer, the most common and aggressive form of epithelial ovarian cancer with a 5-year survival rate of 43%. PATIENTS AND METHODS: We implemented a clinical ctDNA workflow to detect clinically actionable alterations in more than 500 cancer-related genes. We applied the workflow to a prospective cohort consisting of 78 ctDNA samples from 12 patients with high-grade serous ovarian cancer before, during, and after treatment. These longitudinal data sets were analyzed using our open-access ctDNA-tailored bioinformatics analysis pipeline and in-house Translational Oncology Knowledgebase to detect clinically actionable genomic alterations. The alterations were ranked according to the European Society for Medical Oncology scale for clinical actionability of molecular targets. RESULTS: Our results show good concordance of mutations and copy number alterations in ctDNA and tumor samples, and alterations associated with clinically available drugs were detected in seven patients (58%). Treatment of one chemoresistant patient was changed on the basis of detection of ERBB2 amplification, and this ctDNA-guided decision was followed by significant tumor shrinkage and complete normalization of the cancer antigen 125 tumor marker. CONCLUSION: Our results demonstrate a proof of concept for using ctDNA to guide clinical decisions. Furthermore, our results show that longitudinal ctDNA samples can be used to identify poor-responding patients after first cycles of chemotherapy. We provide what we believe to be the first comprehensive, open-source ctDNA workflow for detecting clinically actionable alterations in solid cancers.

Cancer antigen 125 (CA125) is generally considered the gold standard of biomarkers in the diagnosis and monitoring of high grade serous ovarian carcinoma (HGSC). We recently reported, that two CA125 glycoforms (CA125-STn and CA125-MGL) have a high specificity to HGSC and further hypothesized, that these cancer specific glycoforms are feasible candidates as biomarkers in HGSC treatment and follow up. Our cohort consisted of 122 patients diagnosed with HGSC. Serum samples were collected longitudinally at the time of diagnosis, during treatment and follow up. Serum levels of CA125, CA125-STn and CA125-MGL were determined and compared or correlated with different end points (tumor load assessed intraoperatively, residual disease, treatment response, progression free survival). Serum CA125-STn levels at diagnosis differentiated patients with low tumor load and high tumor load (p = 0,030), indicating a favorable detection of tumor volume. Similarly, the CA125-STn levels at diagnosis were significantly lower in patients with subsequent complete cytoreduction than in patients with suboptimal cytoreduction (p = 0,025). Conventional CA125 did not differentiate these patients (p = 0,363 and p = 0,154). The CA125-STn nadir value predicted the progression free survival of patients. The detection of disease relapse was improved with CA125-STn, which presented higher fold increase in 80,0% of patients and earlier increase in 37,0% of patients. CA125-STn showed promise as a useful biomarker in the monitoring and follow up of patients with HGSC utilizing a robust and affordable technique. Our findings are topical as a suitable indicator of tumor load facilitates patient selection in an era of new targeted therapies.

PURPOSE: We have created a cloud-based machine learning system (CLOBNET) that is an open-source, lean infrastructure for electronic health record (EHR) data integration and is capable of extract, transform, and load (ETL) processing. CLOBNET enables comprehensive analysis and visualization of structured EHR data. We demonstrate the utility of CLOBNET by predicting primary therapy outcomes of patients with high-grade serous ovarian cancer (HGSOC) on the basis of EHR data.MATERIALS AND METHODS: CLOBNET is built using open-source software to make data preprocessing, analysis, and model training user friendly. The source code of CLOBNET is available in GitHub. The HGSOC data set was based on a prospective cohort of 208 patients with HGSOC who were treated at Turku University Hospital, Finland, from 2009 to 2019 for whom comprehensive clinical and EHR data were available.RESULTS: We trained machine learning (ML) models using clinical data, including a herein developed dissemination score that quantifies the disease burden at the time of diagnosis, to identify patients with progressive disease (PD) or a complete response (CR) on the basis of RECIST (version 1.1). The best performance was achieved with a logistic regression model, which resulted in an area under receiver operating characteristic curve (AUROC) of 0.86, with a specificity of 73% and a sensitivity of 89%, when it classified between patients who experienced PD and CR.CONCLUSION: We have developed an open-source computational infrastructure, CLOBNET, that enables effective and rapid analysis of EHR and other clinical data. Our results demonstrate that CLOBNET allows predictions to be made on the basis of EHR data to address clinically relevant questions.

Epithelial ovarian cancer (EOC) generally responds well to oncological treatments, but the eventual development of a refractory disease is a major clinical problem. Presently, there are no prognostic blood-based biomarkers for the stratification of EOC patients at the time of diagnosis. We set out to assess and validate the prognostic utility of a novel two-lipid signature, as the lipidome is known to be markedly aberrant in EOC patients. The study consisted of 499 women with histologically confirmed EOC that were prospectively recruited at the university hospitals in Turku (Finland) and Charité (Berlin, Germany). Lipidomic screening by tandem liquid chromatography–mass spectrometry (LC-MS/MS) was performed for all baseline serum samples of these patients, and additionally for 20 patients of the Turku cohort at various timepoints. A two-lipid signature, based on the ratio of the ceramide Cer(d18:1/18:0) and phosphatidylcholine PC(O-38:4), showed consistent prognostic performance in all investigated study cohorts. In the Turku cohort, the unadjusted hazard ratios (HRs) per standard deviation (SD) (95% confidence interval) were 1.79 (1.40, 2.29) for overall and 1.40 (1.14, 1.71) for progression-free survival. In a Charité cohort incorporating only stage III completely resected patients, the corresponding HRs were 1.59 (1.08, 2.35) and 1.53 (1.02, 2.30). In linear-mixed models predicting progression of the disease, the two-lipid signature showed higher performance (beta per SD increase 1.99 (1.38, 2.97)) than cancer antigen 125 (CA-125, 1.78 (1.13, 2.87)). The two-lipid signature was able to identify EOC patients with an especially poor prognosis at the time of diagnosis, and also showed promise for the detection of disease relapse.

Background: The cancer antigen 125 (CA125) immunoassay (IA) does not distinguish epithelial ovarian cancer (EOC) from benign disease with the sensitivity needed in clinical practice. In recent studies, glycoforms of CA125 have shown potential as biomarkers in EOC. Here, we assessed the diagnostic abilities of two recently developed CA125 glycoform assays for patients with a pelvic mass. Detailed analysis was further conducted for postmenopausal patients with marginally elevated conventionally measured CA125 levels, as this subgroup presents a diagnostic challenge in the clinical setting. Methods: Our study population contained 549 patients diagnosed with EOC, benign ovarian tumors, and endometriosis. Of these, 288 patients were postmenopausal, and 98 of them presented with marginally elevated serum levels of conventionally measured CA125 at diagnosis. Preoperative serum levels of conventionally measured CA125 and its glycoforms (CA125-MGL and CA125-STn) were determined. Results: The CA125-STn assay identified EOC significantly better than the conventional CA125-IA in postmenopausal patients (85% vs. 74% sensitivity at a fixed specificity of 90%, P = 0.0009). Further, both glycoform assays had superior AUCs compared to the conventional CA125-IA in postmenopausal patients with marginally elevated CA125. Importantly, the glycoform assays reduced the false positive rate of the conventional CA125-IA. Conclusions: The results indicate that the CA125 glycoform assays markedly improve the performance of the conventional CA125-IA in the differential diagnosis of pelvic masses. This result is especially valuable when CA125 is marginally elevated.

Human epididymis protein 4 (HE4) is a validated, complementary biomarker to cancer antigen 125 (CA125) for high grade serous ovarian carcinoma (HGSC). Currently, there are insufficient data on the ...