Ulrika Joneborg

Machine learning-enhanced gas sensor technology identifies ovarian and endometrial cancer of all stages through plasma volatile organic compound patterns

Ovarian cancer presents with non-specific symptoms that make early diagnosis challenging and the prognosis poor. Ovarian and endometrial cancers exhibit similar genomic mutations and biomarker profiles. Endogenous volatile organic compounds (VOCs) are products of metabolic activity. In cancer, metabolites increase due to tumour necrosis, leading to cancer-specific VOC patterns. The aim of this study was to evaluate VOC analyses in plasma as diagnostic tests for early diagnosis of ovarian and endometrial cancer. Preoperative plasma from 133 women with ovarian cancer (stages I-IV or borderline ovarian tumors) and 41 women with endometrial cancer (stages I-IV) was compared to 115 healthy controls with highly sensitive gas sensors. Data analyses were performed using feature extraction from 32 gas sensors per sample. 85 features were extracted from each signal (including statistical, time-domain, and frequency-domain features), and training and test datasets were formed. The features underwent an iterative redundancy removal process for stepwise optimization of models. Model robustness was assessed using a train/test split scheme with unique datasets, leading to a model optimized for diagnostic performance. By implementing sequential binary classification boosting-based models, it was possible to determine not only the presence or not of cancer, but also to distinguish between ovarian- and endometrial cancer, and the stage of the cancer. The VOC analysis, powered by a 5-fold cross-validated ensemble classifier, achieved exceptional diagnostic performance. It correctly identified all 133 patients with ovarian cancer and borderline ovarian tumors, all 41 cases of endometrial cancer, and all 115 healthy controls. For staging, the model accurately classified 172 out of 174 (98.9%) cancer cases as stage I vs. II-IV. On validation data, the classifier yielded an accuracy of 96.63% (95% CI: 96.56%-96.70%), sensitivity of 96.42% (95% CI: 96.29%-96.54%), and specificity of 96.88% (95% CI: 96.76%-97.01%). These metrics were robustly replicated on the independent test set, with an accuracy of 97.13% (95% CI: 96.80%-97.45%), sensitivity of 96.92% (95% CI: 96.49%-97.35%), and specificity of 97.37% (95% CI: 96.97%-97.77%). Complementing this, four additional classifiers (each with accuracy exceeding 90%) were developed and integrated into a cascaded algorithm to enable multi-class discrimination (ovarian cancer and endometrial cancer vs. healthy controls), cancer typing (ovarian vs. endometrial), and staging (stage I vs. later stages). The analysis of VOCs correctly identified 133 out of 133 patients with ovarian cancer and borderline ovarian tumour. All 41 cases of endometrial cancer were correctly identified, as were all the 115 healthy controls. In 172 out of 174 (98.9%) cancer cases the model correctly classified stage I vs. II-IV. VOC analysis emitted to gas-phase from plasma demonstrates high sensitivity and specificity for diagnosing ovarian cancer, including borderline ovarian tumors and endometrial cancers, compared to healthy controls. VOC analyses accurately differentiated between early and advanced stages of both cancer types. Future external validation needs to be performed. The Strategic Innovation Programs Swelife and MedTech4Health, a joint venture of Vinnova, Formas and the Energy Agency (grant No. 2022-03464 and grant No. 2023-03874) and within the Convergence Accelerator Program (Track L - Real-World Chemical Sensing Applications), funded by the Swedish Research Council, Vetenskapsrådet (grant No. 2023-07219), and Sweden's Innovation Agency, Vinnova (grant No. 2023-04186), in collaboration with the US National Science Foundation (NSF). The computations, data handling, and machine learning model training and testing were conducted in the MATLAB environment and enabled by resources provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS), partially funded by the Swedish Research Council through grant agreement no. 2022-06725. This work received funding from the European Union's Horizon Europe Research and Innovation Programme under Grant Agreement No 101214318 (DISARM). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the Health and Digital Executive Agency (HaDEA). Neither the European Union nor HaDEA can be held responsible for them.

Extensive cytoreductive surgery in stage IV ovarian cancer may not be a justified treatment strategy

Complete macroscopic surgical cytoreduction improves survival in advanced ovarian cancer. However, our previous results suggest that the survival benefits of aggressive surgery to achieve complete macroscopic resection may vary across different subgroups. This study aimed to identify a potential poor-prognosis subgroup among patients with complete macroscopic resection. This cohort study included all patients with advanced ovarian cancer who underwent complete macroscopic resection in a public and centralized health care system setting. Two 3-year cohorts were compared: pre-implementation (cohort 1, n = 101) and post-implementation (cohort 2, n = 172) of a very high surgical proficiency setting. Overall survival was compared by relative risks (RR) stratified by stage and surgical extent. Cox regression estimated hazard ratios (HR) adjusted for relevant covariates. In patients with stage IV, median survival was 33 (cohort 2) vs 47 months (cohort 1), 7-year RR 1.01 (95 % CI 0.78 to 1.31). When extensive surgery was required in stage IV median survival was 27 (cohort 2) vs 58 months (cohort 1), 5-year RR 0.91 (95 % CI 0.81 to 1.01). By contrast, in stage III, median survival was 93 (cohort 2) vs 59 (cohort 1) months, 7-year RR 0.76 (95 % CI 0.59 to 0.98), with the greatest difference when less extensive surgery was required; Not reached vs 65 months, 7-year RR 0.62 (95 % CI, 0.43 to 0.88). Adjusted analysis confirmed improved survival in cohort 2 only after less extensive surgery, HR 0.56 (95 % CI, 0.33-0.95). The implementation of a very high surgical proficiency setting was not associated with improved survival in patients with stage IV despite complete macroscopic resection, as opposed to patients with stage III.

Incidence of and survival after surgical intervention for bowel obstruction in women with advanced ovarian cancer

AbstractIntroductionWomen with advanced ovarian cancer commonly present with peritoneal disease both at primary diagnosis and relapse, with risk of subsequent bowel obstruction. The aims of this study were to assess the cumulative incidence of and survival after intervention for bowel obstruction in women with advanced ovarian cancer, to identify factors predictive of survival and the extent to which the intended outcome of the intervention was achieved.Material and methodsWomen diagnosed with advanced ovarian cancer stages III and IV in 2009–2011 and 2014–2016 in the Stockholm‐Gotland Region in Sweden were identified in the Swedish Quality Registry for Gynecologic Cancer. Through hospital records, types of intended and executed interventions for bowel obstruction were assessed, and as well as when in the course of oncologic treatment, the intervention was performed. Time from first intervention to death was analyzed with survival methodology and proportional hazard regression was used.ResultsOf 751 identified women, 108 had an intervention for bowel obstruction. Laparotomy was the most prevalent intervention and was used in 87% (94/108) of all women, with a success rate of 87% (82/94). An intervention for bowel obstruction was performed before or during first line treatment in 32% (35/108) with a cumulative incidence in the whole cohort of 14% (108/751, 95% confidence interval [CI] 11–16). Median survival after intervention for bowel obstruction was 4 months (95% CI 3–6). The hazard of death increased when the intervention was performed after completion of primary treatment (HR 4.46, 95% CI 1.61–12.29, P < 0.01), with a median survival of 3 months. In women subjected to radical surgery during primary treatment, the hazard of death after intervention for bowel obstruction decreased (hazard ratio [HR] 0.54, 95% CI 0.32–0.91, P = 0.02).ConclusionsWomen with advanced ovarian cancer undergoing intervention for bowel obstruction have a dismal prognosis, regardless of which line of oncologic treatment the intervention was performed. In the majority of women an intervention for bowel obstruction was performed in a relapse situation with an even worse survival. Our findings emphasize the importance of a holistic approach in the decision‐making before an intervention for bowel obstruction in women with advanced ovarian cancer.

Time-interval to adjuvant chemotherapy and postoperative management after upper abdominal surgical procedures in advanced ovarian cancer

In advanced epithelial ovarian cancer (EOC), longer time-interval from surgery to initiation of adjuvant chemotherapy (TITC) is associated with decreased survival. Adding upper abdominal surgical procedures (UAP) increases rates of both complete gross resection and postoperative complications in EOC. Our objective was to investigate the association of UAP and TITC. Moreover, if specific postoperative monitoring after the most prevalent UAP increases early detection and management of complications. Women diagnosed with EOC 2014-2016 in the Stockholm/Gotland Region in Sweden were identified in the Swedish Quality Registry for Gynaecologic Cancer. The association between UAP and TITC was investigated by multivariable linear regression and adjusted for predefined confounders. The follow-up and detection of postoperative complications after diaphragm resection, splenectomy and cholecystectomy was examined. 240 women were selected for analysis. The TITC in women subjected to UAP was similar with a median of 30 days (p = 0.99). Moreover, despite a higher rate of postoperative and major complications (p < 0.001) and longer hospital stay (p < 0.001), in the adjusted analysis there was no association between UAP and prolonged TITC, with a mean difference of -2.27 days (95% Confidence Interval (CI), -5.99 to -1.45, p = 0.23). After the most prevalent UAP (diaphragm resection, splenectomy and cholecystectomy), eventual postoperative interventions were based on routine clinical management rather than procedure-specific postoperative surveillance. UAP does not prolong TITC despite an increased rate of postoperative complications and longer length of hospital stay. Clinical non-specific surveillance is sufficient to detect postoperative complications after the most prevalent UAP.

Treatment outcomes and the role of surgery in gestational trophoblastic neoplasia: a population-based cohort study

Background and purpose: Cure rates of gestational trophoblastic neoplasia (GTN) are excellent, however the surgical interventions in disease management are not well described. The primary aim of this study was to investigate the incidence and types of surgical procedures used for management of GTN and to report treatment outcomes in a population-based cohort. The secondary aim was to assess the impact of hysterectomy on time to human chorionic gonadotropin (hCG)-normalisation in low-risk GTN. Material and methods: Medical records of all patients treated for GTN at Karolinska University Hospital, Stockholm, Sweden between 1994 and 2020 were screened for treatment outcomes, types of surgical procedures and complications. Regression models were used to assess if hysterectomy affected time to complete remission in low-risk GTN. Results and interpretation: Over the 27-year study period, 185 patients with GTN were included. The primary complete remission rate was 98.4% and relapse rate 3.2%. Sixty-four patients (34.6%) underwent at least one surgical procedure; 39/154 (25.3%) of low-risk patients, 17/23 (73.9%) of high-risk patients and all (100%) patients with placental site or epithelioid trophoblastic tumour. No severe complications (Clavien-Dindo ≥3) were observed. Seven of 74 procedures (9.5%) were complicated by bleeding &gt;1,000 mL or surgical site infection. Therapeutic hysterectomy significantly shortened time to hCG-normalisation in the low-risk group (48 vs 74 days, p = 0.002). This population-based study confirms the excellent cure rates and low relapse rates for GTN. Surgery plays an important role in the management of GTN with low risk of complications. Hysterectomy shortens time to hCG normalisation.

Chemotherapy is not needed when complete evacuation of gestational choriocarcinoma leads to hCG normalization

The standard treatment for gestational choriocarcinoma is chemotherapy. To describe the risk of recurrence with expectant management of gestational choriocarcinoma that has reached a normal human chorionic gonadotropin level after tumor removal without adjuvant chemotherapy. A retrospective multicenter international cohort study was conducted from 1981 to 2017 involving 11 gestational trophoblastic disease reference centers with patient's follow-up extended until 2023. Clinical and biological data of included patients were extracted from each center's database. The inclusion criteria were i) histological diagnosis of gestational choriocarcinoma in any kind of placental tissue retrieved, ii) spontaneous normalization of human chorionic gonadotropin level following choriocarcinoma retrieval, iii) patient did not receive any oncological treatment for the choriocarcinoma, iv) and at least 6 months of follow-up after the first human chorionic gonadotropin level normalization. Among 80 patients with retrieved gestational choriocarcinoma and whose human chorionic gonadotropin level normalized without any other oncological therapy, none had a recurrence of choriocarcinoma after a median follow-up of 50 months. The median interval between choriocarcinoma excision and human chorionic gonadotropin level normalization was 48 days. The International Federation of Gynecology and Obstetrics/World Health Organization risk score was ≤6 in 93.7% of the cases. This multicenter international study reports that selected patients with gestational choriocarcinoma managed in gestational trophoblastic disease reference centers did not experience any relapse when the initial tumor evacuation is followed by human chorionic gonadotropin level normalization without any additional treatment. Expectant management may be a safe approach for highly selected patients.