Clinical Judgment, Ultrasonography and Organoid Technology to improve Diagnostic and Therapeutic Strategies for Endometrial Pathology – An Endometrial Elegy

Endometrial cancer (EC)  annually affects 382,000 new cases worldwide and is thereby the third cause of death from women's cancers in industrialized countries. Its incidence is steadily increasing and is estimated to do so even more in the following decades, thereby exposing an increasing number of women to disease- and therapy-related risks. Most endometrial cancers are diagnosed early owing to the presence of abnormal uterine bleeding (AUB). This early warning sign is highly sensitive, but has a low specificity. Current literature remains inconclusive regarding the impact of its intensity, duration and frequency. In this thesis, we analyzed exactly which signs and symptoms should raise awareness for underlying endometrial malignancy. We performed a systematic review that highlighted the current lack of knowledge regarding the diagnostic accuracy of several signs and symptoms for uterine cancer. Further research is needed to establish the rule-in or rule-out value of specific clinical signs. On the other hand, the individual value of various anthropometric or lifestyle factors might be unduly overestimated. Assessing the right factors and constructing a robust referral strategy is beneficial to patients and  caregivers in both primary and secondary settings. Adequate referrals facilitate timely diagnosis and proper management without delay, whereas inappropriate referrals could lead to unnecessary use of healthcare resources. In this thesis, we construct an endometrial cancer risk model in women with AUB based on patient history, anthropometric and lifestyle factors. We found that Age, BMI and parity suffice for endometrial cancer risk estimation in women with AUB. Other clinical patient information offers limited added value. As a next step, transvaginal ultrasound with color/power Doppler imaging allows to assess the endometrium. In some cases, diagnostic accuracy may be optimized by instilling fluid into the uterine cavity. Fluid acts as an anechogenic sonographic contrast and pushes aside both endometrial layers. This procedure might, however, exert shear stress on the endometrium, causing exfoliation, flushing of endometrial tissue through the fallopian tubes and eventually intra-abdominal deposition of endometrial cells. We performed an in vitro randomized trial comparing gel to water and found that gel is associated with less tubal flow and therefore could be a safer diagnostic test compared to saline infusion sonography or hysteroscopy. We used International Endometrial Tumor Analysis (IETA) guidelines to describe endometrial findings of women with and without AUB. Using these variables, we validated previous and constructed new binary and multiclass prediction models that can support clinical management. We found that these models significantly improved diagnostic capacity in these patients. For patients with histologically proven EC, studies have linked grayscale and Doppler patterns to endometrial cancers of a different size, grade, invasion pattern and stage. Based on both sonographic patterns and biopsy results, the risk of lymph node metastases can be estimated. This has an impact on surgical planning, disease recurrence and progression, and disease-specific survival. Based on survival data of an observational cohort, we now combined demographic, sonographic and histologic findings to predict the risk of recurrence of or death due to endometrial cancer. Our models were able to accurately predict which patients would experience higher burden of disease. To further understand how the high-risk population relates to those with favorable outcomes, it is important to understand the key drivers of the disease. Despite large efforts to understand the pathogenesis of EC, cellular and molecular mechanisms remain largely unknown, and therapy efficiency and overall survival rate (particularly for the most advanced stages) have not substantially improved in the last decade. This is mainly due to the lack of reliable preclinical models. The current models (such as tumor-derived cell lines and their xenografts) do not reliably mimic the disease, or do not cover the EC spectrum. A novel, promising and powerful in vitro tool to model human diseases is provided by the organoid technology. Organoids are three-dimensional self-forming in vitro reconstructions of the normal or diseased organ, developing from the tissue (stem/tumor stem) cells under specific Wingless/ints (WNT)-activated culture conditions. We provide an overview of the latest progress on the use of organoids as a preclinical model to study the human female reproductive tract and established and characterized an organoid model for endometrial polyps, endometrial hyperplasia and endometrial cancer. The organoids highly resembled the original organ or its diseased state (histologically, cellular composition, genomically, transcriptomically). They could be maintained and expanded in culture for a long period, as well as cryopreserved, without losing pheno- and genotypical characteristics. Using them as a drug screening platform resulted in comparable drug sensitivities and observed in vivo. An organoid biobank was established that can fuel further studies and serves as a valuable platform to test (novel) drugs affecting new molecular targets and to facilitate personalized medicine.