Casper Reijnen

Implementation of a Personalized Risk Model for Lymph Node Metastasis in Endometrial Carcinoma: Healthcare Providers' Perspectives on Use, Barriers, and Facilitators

ABSTRACT Background The ENDORISK model estimates the risk of lymph node metastases (LNM) in endometrial carcinoma (EC) patients using preoperative clinical variables and biomarkers. This qualitative study investigated healthcare providers' (HCP) perspectives on the use of the model and barriers and facilitators for clinical implementation. Methods Eight focus group interviews were performed among HCPs. A semi‐structured interview guide was used based on the Grol and Wensing implementation model. Results Focus groups included gynecologists, residents of gynecology, pathologists, radiation oncologists, and a nurse specialist ( n  = 41). ENDORISK was deemed supportive for counseling of patients and shared decision‐making for optimal surgical and adjuvant treatment. Barriers for implementation were difficulty in explaining the model and risk percentages to patients, differences in preoperative diagnostic tools used per hospital, and use of the model with the sentinel node procedure. Facilitators were a clear guideline for using the model with a predefined risk cutoff and making the model easily understandable for patients. A 10% risk cutoff was considered clinically relevant for lymph node assessment. Conclusion HCP found ENDORISK use in clinical practice supportive for patient counseling. Future implementation should focus on a user‐friendly interface, a cohesive guideline, and training to aid efficient use and counseling of patients.

Pure and mixed clear cell carcinoma of the endometrium: A molecular and immunohistochemical analysis study

AbstractBackgroundUterine clear cell carcinoma (CCC) consists of either pure clear cell histology but can also display other histological components (mixed uterine CCCs). In this study, the molecular and immunohistochemical background of pure and mixed uterine CCC was compared. Secondly, it was evaluated whether histological classification and molecular background affected clinical outcome.MethodsA retrospective multicenter study was performed comparing pure uterine CCCs (n = 22) and mixed uterine CCCs (n = 21). Targeted next‐generation sequencing using a 12‐gene targeted panel classified cases as polymerase‐ε (POLE) mutated, microsatellite instable (MSI), TP53 wildtype or TP53 mutated. Immunohistochemistry was performed for estrogen receptor, progesterone receptor, L1 cell adhesion molecule, MSH6, and PMS2.ResultsThe following molecular subgroups were identified for pure and mixed uterine CCCs, respectively: POLE mutated 0% (0/18) and 6% (1/18); MSI in 6% (1/18) and 50% (9/18); TP53 wildtype in 56% (10/18) and 22% (4/18); TP53 mutated in 39% (7/18) and 22% (4/18) (p = 0.013). Patients with mixed CCCs had improved outcome compared to patients with pure CCCs. Frequent TP53 mutations were found in pure CCCs and frequent MSI in mixed CCCs, associated with clinical outcome.ConclusionPure and mixed uterine CCCs are two entities with different clinical outcomes, which could be explained by different molecular backgrounds. These results underline the relevance of both morphological and molecular evaluation, and may assist in tailoring treatment.

L1CAM expression as a predictor of platinum response in high‐risk endometrial carcinoma

AbstractFor high‐risk endometrial cancer (EC) patients, adjuvant chemotherapy is recommended to improve outcome. Yet, predictive biomarkers for response to platinum‐based chemotherapy (Pt‐aCT) are currently lacking. We tested expression of L1 cell‐adhesion molecule (L1CAM), a well‐recognised marker of poor prognosis in EC, in tumour samples from high‐risk EC patients, to explore its role as a predictive marker of Pt‐aCT response. L1CAM expression was determined using RT‐qPCR and immunohistochemistry in a cohort of high‐risk EC patients treated with Pt‐aCT and validated in a multicentric independent cohort. The association between L1CAM and clinicopathologic features and L1CAM additive value in predicting platinum response were determined. The effect of L1CAM gene silencing on response to carboplatin was functionally tested on primary L1CAM‐expressing cells. Increased L1CAM expression at both genetic and protein level correlated with high‐grade, non‐endometrioid histology and poor response to platinum treatment. A predictive model adding L1CAM to prognostic clinical variables significantly improved platinum response prediction (C‐index 78.1%, P = .012). In multivariate survival analysis, L1CAM expression was significantly associated with poor outcome (HR: 2.03, P = .019), potentially through an indirect effect, mediated by its influence on response to chemotherapy. In vitro, inhibition of L1CAM significantly increased cell sensitivity to carboplatin, supporting a mechanistic link between L1CAM expression and response to platinum in EC cells. In conclusion, we have demonstrated the role of L1CAM in the prediction of response to Pt‐aCT in two independent cohorts of high‐risk EC patients. L1CAM is a promising candidate biomarker to optimise decision making in high‐risk patients who are eligible for Pt‐aCT.

Molecular profiling identifies synchronous endometrial and ovarian cancers as metastatic endometrial cancer with favorable clinical outcome

Synchronous primary endometrial and ovarian cancers (SEOs) represent 10% of all endometrial and ovarian cancers and are assumed to develop as independent entities. We investigated the clonal relationship between endometrial and ovarian carcinomas in a large cohort classified as SEOs or metastatic disease (MD). The molecular profiles were compared to The Cancer Genome Atlas (TCGA) data to explore primary origin. Subsequently, the molecular profiles were correlated with clinical outcome. To this extent, a retrospective multicenter study was performed comparing patients with SEOs (n = 50), endometrial cancer with synchronous ovarian metastasis (n = 19) and ovarian cancer with synchronous endometrial metastasis (n = 20). Targeted next‐generation sequencing was used, and a clonality index was calculated. Subsequently, cases were classified as POLE mutated, mismatch repair deficient (MMR‐D), TP53‐wild‐type or TP53‐mutated. In 92% of SEOs (46/50), the endometrial and concurrent ovarian carcinoma shared at least one somatic mutation, with a clonality index above 0.95, supporting a clonal origin. The SEO molecular profiles showed striking similarities with the TCGA endometrial carcinoma set. SEOs behaved distinctly different from metastatic disease, with a superior outcome compared to endometrial MD cases (p < 0.001) and ovarian MD cases (p < 0.001). Classification according to the TCGA identified four groups with different clinical outcomes. TP53 mutations and extra‐utero‐ovarian disease were independent predictors for poor clinical outcome. Concluding, SEOs were clonally related in an overwhelming majority of cases and showed a favorable prognosis. Their molecular profile implied a primary endometrial origin. TP53 mutation and extra‐utero‐ovarian disease were independent predictors for outcome, and may impact adjuvant systemic treatment planning.

Mutational analysis of cervical cytology improves diagnosis of endometrial cancer: A prospective multicentre cohort study

Endometrial carcinoma (EC) is traditionally diagnosed by a histopathological assessment of an endometrial biopsy, leaving up to 30% of patients undiagnosed due to technical failure or an inadequate amount of tissue. The aim of the current study is to assess whether mutational analysis of cervical cytology or pipelle endometrial biopsies improves the diagnostic accuracy of traditional histopathological diagnosis of EC. This prospective multicentre cohort study included patients surgically treated for EC or a benign gynaecological condition (control group). A Pap brush sample, cervicovaginal self‐sample, pipelle endometrial biopsy and surgical specimen of either the EC or normal endometrium were obtained. A targeted next‐generation sequencing panel was used to analyse these samples for mutations in eight genes. Sensitivity, specificity and predictive values were calculated. Fifty‐nine EC patients and 31 control patients were included. In these patients, traditional histopathological diagnosis by pipelle had a sensitivity of 79% and a specificity of 100%. For EC patients, 97% of surgical specimens contained at least one mutation. Mutational analysis of Pap brush samples, self‐samples and pipelle endometrial biopsies yielded a sensitivity of 78, 67 and 96% with a specificity of 97, 97 and 94%, respectively. Combining one of these three methods with histopathological pipelle endometrial biopsy evaluations yielded a sensitivity of 96, 93 and 96%, respectively. Our study has shown that mutational analysis of either cervical cytology or pipelle endometrial biopsies improves diagnosis of EC. Prospective validation will support implementation in clinical practice.

Predictive value of estrogen and progesterone receptors in endometrial hyperplasia and cancer

ENDORISK-2: A personalized Bayesian network for preoperative risk stratification in endometrial cancer, integrating molecular classification and preoperative myometrial invasion assessment

ENDORISK is a Bayesian network that can assist in preoperative risk estimation of lymph node metastasis (LNM) risk in endometrial cancer (EC) with consistent performance in external validations. To reflect state of the art care, ENDORISK was optimized by integrating molecular classification and preoperative assessment of myometrial invasion (MI). Variables for POLE, MSI, and preoperative assessment of MI, either by expert transvaginal ultrasound or pelvic magnetic resonance imaging (MRI), were added to develop ENDORISK-2. The p53 biomarker, part of the molecular classification, was already included in ENDORISK. External validation of ENDORISK-2 for LNM prediction was performed in two independent cohorts from: Brno (CZ), (n = 581) and Tübingen (DE), (n = 247). ENDORISK-2 yielded AUCs of 0·85 (95 % CI 0·80-0·90) (CZ) and 0·86 (95 % CI 0·77-0·96) (DE) for predicting LNM. In patients with low-grade histology, 83 % (CZ) and 89 % (DE) were estimated having less than 10 % risk of LNM, with false negative rates (FNR) of 4·3 % (CZ) and 2·2 % (DE). The previously defined set of minimally required variables, i.e.: preoperative tumor grade, three of the four immunohistochemical (IHC) markers, and one clinical marker, could be interchanged with the new variables, with comparable validation metrics, including AUC values of 0·79-0·87 for LNM prediction. Incorporation of molecular data and preoperative MI improved the flexibility of ENDORISK with comparable diagnostic accuracy for estimating LNM as when based on low-cost immunohistochemical biomarkers. In addition, the high diagnostic accuracy in patients with low-grade EC demonstrates how ENDORISK-2 could aid clinicians in identifying patients in whom surgical lymph node assessment may safely be omitted. These results underline its power for clinical use in both high and low resource countries.

Federated causal discovery with missing data in a multicentric study on endometrial cancer

Establishing causal dependencies is crucial in applied domains, such as medicine and healthcare, where decision-making must be explainable. In these settings, small sample sizes and missing data call for federated approaches to maximise the amount of information we can use. We propose a novel federated causal discovery algorithm capable of pooling information from multiple sources with heterogeneous missing data to learn a graph representing cause-effect relationships. In particular, we learn a causal graph on a centralised server while taking into account both prior knowledge and missingness mechanism specific to each client. We applied the proposed algorithm to synthetic data and real-world data from a multicentric study on endometrial cancer, validating the obtained causal graph through quantitative analyses and a clinical literature review. Our approach learns an accurate model despite data missing not-at-random.

Redefining the Position of Hormonal Therapy in Endometrial Cancer in the Era of Molecular Classification