Predicting distant recurrence of endometrial cancer (EC) is crucial for personalized adjuvant treatment. The current gold standard of combined pathological and molecular profiling is costly,... Show morePredicting distant recurrence of endometrial cancer (EC) is crucial for personalized adjuvant treatment. The current gold standard of combined pathological and molecular profiling is costly, hampering implementation. Here we developed HECTOR (histopathology-based endometrial cancer tailored outcome risk), a multimodal deep learning prognostic model using hematoxylin and eosin-stained, whole-slide images and tumor stage as input, on 2,072 patients from eight EC cohorts including the PORTEC-1/-2/-3 randomized trials. HECTOR demonstrated C-indices in internal (n = 353) and two external (n = 160 and n = 151) test sets of 0.789, 0.828 and 0.815, respectively, outperforming the current gold standard, and identified patients with markedly different outcomes (10-year distant recurrence-free probabilities of 97.0%, 77.7% and 58.1% for HECTOR low-, intermediate- and high-risk groups, respectively, by Kaplan–Meier analysis). HECTOR also predicted adjuvant chemotherapy benefit better than current methods. Morphological and genomic feature extraction identified correlates of HECTOR risk groups, some with therapeutic potential. HECTOR improves on the current gold standard and may help delivery of personalized treatment in EC. Show less
Background Endometrial cancer can be molecularly classified into POLEmut , mismatch repair deficient (MMRd), p53 abnormal (p53abn), and no specific molecular profile (NSMP) subgroups. We aimed to... Show moreBackground Endometrial cancer can be molecularly classified into POLEmut , mismatch repair deficient (MMRd), p53 abnormal (p53abn), and no specific molecular profile (NSMP) subgroups. We aimed to develop an interpretable deep learning pipeline for whole-slide-image-based prediction of the four molecular classes in endometrial cancer (im4MEC), to identify morpho-molecular correlates, and to refine prognostication. Methods This combined analysis included diagnostic haematoxylin and eosin-stained slides and molecular and clinicopathological data from 2028 patients with intermediate-to-high-risk endometrial cancer from the PORTEC-1 (n=466), PORTEC-2 (n=375), and PORTEC-3 (n=393) randomised trials and the TransPORTEC pilot study (n=110), the Medisch Spectrum Twente cohort (n=242), a case series of patients with POLEmut endometrial cancer in the Leiden Endometrial Cancer Repository (n=47), and The Cancer Genome Atlas-Uterine Corpus Endometrial Carcinoma cohort (n=395). PORTEC-3 was held out as an independent test set and a four-fold cross validation was performed. Performance was measured with the macro and class-wise area under the receiver operating characteristic curve (AUROC). Whole-slide images were segmented into tiles of 360 & mu;m resized to 224 x 224 pixels. im4MEC was trained to learn tile-level morphological features with self-supervised learning and to molecularly classify whole-slide images with an attention mechanism. The top 20 tiles with the highest attention scores were reviewed to identify morpho-molecular correlates. Predictions of a nuclear classification deep learning model serve to derive interpretable morphological features. We analysed 5-year recurrence-free survival and explored prognostic refinement by molecular class using the Kaplan-Meier method. Findings im4MEC attained macro-average AUROCs of 0 & BULL;874 (95% CI 0 & BULL;856-0 & BULL;893) on four-fold cross-validation and 0 & BULL;876 on the independent test set. The class-wise AUROCs were 0 & BULL;849 for POLEmut (n=51), 0 & BULL;844 for MMRd (n=134), 0 & BULL;883 for NSMP (n=120), and 0 & BULL;928 for p53abn (n=88). POLEmut and MMRd tiles had a high density of lymphocytes, p53abn tiles had strong nuclear atypia, and the morphology of POLEmut and MMRd endometrial cancer overlapped. im4MEC highlighted a low tumour-to-stroma ratio as a potentially novel characteristic feature of the NSMP class. 5-year recurrence-free survival was significantly different between im4MEC predicted molecular classes in PORTEC-3 (log-rank p<0 & BULL;0001). The ten patients with aggressive p53abn endometrial cancer that was predicted as MMRd showed inflammatory morphology and appeared to have a better prognosis than patients with correctly predicted p53abn endometrial cancer (p=0 & BULL;30). The four patients with NSMP endometrial cancer that was predicted as p53abn showed higher nuclear atypia and appeared to have a worse prognosis than patients with correctly predicted NSMP (p=0 & BULL;13). Patients with MMRd endometrial cancer predicted as POLEmut had an excellent prognosis, as do those with true POLEmut endometrial cancer. Interpretation We present the first interpretable deep learning model, im4MEC, for haematoxylin and eosin-based prediction of molecular endometrial cancer classification. im4MEC robustly identified morpho-molecular correlates and could enable further prognostic refinement of patients with endometrial cancer. Copyright & COPY; 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Show less
Background: The endometrial cancer molecular classification has been integrated into the 2020 World Health Organization (WHO) diagnostic classification and European treatment guidelines, and... Show moreBackground: The endometrial cancer molecular classification has been integrated into the 2020 World Health Organization (WHO) diagnostic classification and European treatment guidelines, and provides direction towards more effective and less toxic adjuvant treatment strategies for women with endometrial cancer. Primary objective(s): The RAINBO program of clinical trials will investigate four molecular class-directed adjuvant treatment strategies following surgical resection to either increase cure rates through the addition of novel targeted therapies or safely reduce toxicity and improve quality of life through treatment de-escalation. Study hypothesis: Molecular-directed adjuvant treatment strategies will improve clinical outcomes and reduce toxicity of unwarranted therapies in women with endometrial cancer. The overarching and translational research RAINBO program will advance knowledge of predictive and prognostic (bio)markers that will improve prognostication and treatment allocation. Trial design: The RAINBO program is a platform of four international clinical trials and an overarching research program. The randomized phase III p53abn-RED trial for women with invasive stage I-III p53abn endometrial cancer compares adjuvant chemoradiation followed by olaparib for 2 years with adjuvant chemoradiation alone. The randomized phase III MMRd-GREEN trial for women with stage II (with lymphovascular space invasion (LVSI)) or stage III mismatch repair-deficient (MMRd) endometrial cancer compares adjuvant radiotherapy with concurrent and adjuvant durvalumab for 1 year to radiotherapy alone. The randomized phase III NSMP-ORANGE trial is a treatment de-escalation trial for women with estrogen receptor positive stage II (with LVSI) or stage III no specific molecular profile (NSMP) endometrial cancer comparing radiotherapy followed by progestin for 2 years to adjuvant chemoradiation. The POLEmut-BLUE trial is a phase II trial in which the safety of de-escalation of adjuvant therapy is investigated for women with stage I-III POLEmut endometrial cancer: no adjuvant therapy for lower-risk disease and no adjuvant therapy or radiotherapy alone for higher-risk disease. The overarching RAINBO program will combine data and tumor material of all participants to perform translational research and evaluate molecular class-based adjuvant therapy in terms of efficacy, toxicity, quality of life, and cost-utility. Major inclusion/exclusion criteria: Inclusion criteria include a histologically confirmed diagnosis of endometrial cancer treated by hysterectomy and bilateral salpingo-oophorectomy with or without lymphadenectomy or sentinel lymph node biopsy, with no macroscopic residual disease after surgery and no distant metastases, and molecular classification according to the WHO 2020 algorithm. Primary endpoint(s): Recurrence-free survival at 3 years in the p53abn-RED, MMRd-GREEN, and NSMP-ORANGE trials and pelvic recurrence at 3 years in the POLEmut-BLUE trial. Sample size: The p53abn-RED trial will include 554 patients, the MMRd-GREEN trial 316, the NSMP-ORANGE trial 600, and the POLEmut-BLUE trial 145 (120 for lower-risk disease and approximately 25 for higher-risk disease). The overarching research program will pool the four sub-trials resulting in a total sample size of around 1600. Estimated dates for completing accrual and presenting results: The four clinical trials will have different completion dates; main results are expected from 2028. Show less
Hummelink, K.; Noort, V. van der; Muller, M.; Schouten, R.D.; Lalezari, F.; Peters, D.; ... ; Thommen, D.S. 2022
Purpose: Durable clinical benefit to PD-1 blockade in non- small cell lung cancer (NSCLC) is currently limited to a small fraction of patients, underlining the need for predictive biomar-kers. We... Show morePurpose: Durable clinical benefit to PD-1 blockade in non- small cell lung cancer (NSCLC) is currently limited to a small fraction of patients, underlining the need for predictive biomar-kers. We recently identified a tumor-reactive tumor-infiltrating T lymphocyte (TIL) pool, termed PD-1T TILs, with predictive potential in NSCLC. Here, we examined PD-1T TILs as biomarker in NSCLC. Experimental Design: PD-1T TILs were digitally quantified in 120 baseline samples from advanced NSCLC patients treated with PD-1 blockade. Primary outcome was disease control (DC) at 6 months. Secondary outcomes were DC at 12 months and survival. Exploratory analyses addressed the impact of lesion-specific responses, tissue sample properties, and combination with other biomarkers on the predictive value of PD-1T TILs. Results: PD-1T TILs as a biomarker reached 77% sensitivity and 67% specificity at 6 months, and 93% and 65% at 12 months,respectively. Particularly, a patient group without clinical benefit was reliably identified, indicated by a high negative predictive value (NPV) (88% at 6 months, 98% at 12 months). High PD-1T TILs related to significantly longer progression-free (HR 0.39, 95% CI, 0.24-0.63, P < 0.0001) and overall survival (HR 0.46, 95% CI, 0.28-0.76, P < 0.01). Predictive performance was increased when lesion-specific responses and samples obtained immediately before treatment were assessed. Notably, the pre-dictive performance of PD-1T TILs was superior to PD-L1 and tertiary lymphoid structures in the same cohort. Conclusions: This study established PD-1T TILs as predictive biomarker for clinical benefit to PD-1 blockade in patients with advanced NSCLC. Most importantly, the high NPV demon-strates an accurate identification of a patient group without benefit. Show less
Randomized controlled trials (RCTs) are considered the gold standard for testing causal hypotheses in the clinical domain; however, the investigation of prognostic variables of patient outcome in a... Show moreRandomized controlled trials (RCTs) are considered the gold standard for testing causal hypotheses in the clinical domain; however, the investigation of prognostic variables of patient outcome in a hypothesized cause-effect route is not feasible using standard statistical methods. Here we propose a new automated causal inference method (AutoCI) built on the invariant causal prediction (ICP) framework for the causal reinterpretation of clinical trial data. Compared with existing methods, we show that the proposed AutoCI allows one to clearly determine the causal variables of two real-world RCTs of patients with endometrial cancer with mature outcome and extensive clinicopathological and molecular data. This is achieved via suppressing the causal probability of non-causal variables by a wide margin. In ablation studies, we further demonstrate that the assignment of causal probabilities by AutoCI remains consistent in the presence of confounders. In conclusion, these results confirm the robustness and feasibility of AutoCI for future applications in real-world clinical analysis.The invariant causal prediction (ICP) framework tries to determine the causal variables given an outcome variable, but considerable effort is needed to adapt existing ICP methods to the clinical domain. The authors propose an automated causal inference method that could potentially address the challenges of applying the ICP framework to complex clinical datasets. Show less
Horeweg, N.; Workel, H.H.; Loiero, D.; Church, D.N.; Vermij, L.; Leon-Castillo, A.; ... ; Bruyn, M. de 2022
B-cells play a key role in cancer suppression, particularly when aggregated in tertiary lymphoid structures (TLS). Here, we investigate the role of B-cells and TLS in endometrial cancer (EC).... Show moreB-cells play a key role in cancer suppression, particularly when aggregated in tertiary lymphoid structures (TLS). Here, we investigate the role of B-cells and TLS in endometrial cancer (EC). Single cell RNA-sequencing of B-cells shows presence of naive B-cells, cycling/germinal center B-cells and antibody-secreting cells. Differential gene expression analysis shows association of TLS with L1CAM overexpression. Immunohistochemistry and co-immunofluorescence show L1CAM expression in mature TLS, independent of L1CAM expression in the tumor. Using L1CAM as a marker, 378 of the 411 molecularly classified ECs from the PORTEC-3 biobank are evaluated, TLS are found in 19%. L1CAM expressing TLS are most common in mismatch-repair deficient (29/127, 23%) and polymerase-epsilon mutant EC (24/47, 51%). Multivariable Cox regression analysis shows strong favorable prognostic impact of TLS, independent of clinicopathological and molecular factors. Our data suggests a pivotal role of TLS in outcome of EC patients, and establishes L1CAM as a simple biomarker.Tertiary lymphoid structures (TLS) are associated with a reduced risk of cancer recurrence and improved response to immune checkpoint blockade in several tumor types. Here the authors identify L1CAM as a marker for mature TLS and show that the presence of TLS is associated with favorable prognosis in patients with endometrial cancer from the PORTEC-3 trial. Show less
Fremond, S.; Koelzer, V.H.; Horeweg, N.; Bosse, T. 2022
Endometrial cancer (EC) diagnostics is evolving into a system in which molecular aspects are increasingly important. The traditional histological subtype-driven classification has shifted to a... Show moreEndometrial cancer (EC) diagnostics is evolving into a system in which molecular aspects are increasingly important. The traditional histological subtype-driven classification has shifted to a molecular-based classification that stratifies EC into DNA polymerase epsilon mutated (POLEmut), mismatch repair deficient (MMRd), and p53 abnormal (p53abn), and the remaining EC as no specific molecular profile (NSMP). The molecular EC classification has been implemented in the World Health Organization 2020 classification and the 2021 European treatment guidelines, as it serves as a better basis for patient management. As a result, the integration of the molecular class with histopathological variables has become a critical focus of recent EC research. Pathologists have observed and described several morphological characteristics in association with specific genomic alterations, but these appear insufficient to accurately classify patients according to molecular subgroups. This requires pathologists to rely on molecular ancillary tests in routine workup. In this new era, it has become increasingly challenging to assign clinically relevant weights to histological and molecular features on an individual patient basis. Deep learning (DL) technology opens new options for the integrative analysis of multi-modal image and molecular datasets with clinical outcomes. Proof-of-concept studies in other cancers showed promising accuracy in predicting molecular alterations from H&E-stained tumor slide images. This suggests that some morphological characteristics that are associated with molecular alterations could be identified in EC, too, expanding the current understanding of the molecular-driven EC classification. Here in this review, we report the morphological characteristics of the molecular EC classification currently identified in the literature. Given the new challenges in EC diagnostics, this review discusses, therefore, the potential supportive role that DL could have, by providing an outlook on all relevant studies using DL on histopathology images in various cancer types with a focus on EC. Finally, we touch upon how DL might shape the management of future EC patients. Show less
Optimum risk stratification in early-stage endometrial cancer combines clinicopathologic factors and the molecular endometrial cancer classification defined by The Cancer Genome Atlas (TCGA). It is... Show moreOptimum risk stratification in early-stage endometrial cancer combines clinicopathologic factors and the molecular endometrial cancer classification defined by The Cancer Genome Atlas (TCGA). It is unclear whether analysis of intratumoral immune infiltrate improves this. We developed a machine-learning, image-based algorithm to quantify density of CD8(+) and CD103(+) immune cells in tumor epithelium and stroma in 695 stage I endometrioid endometrial cancers from the PORTEC-1 and -2 trials. The relationship between immune cell density and clinicopathologic/molecular factors was analyzed by hierarchical clustering and multiple regression. The prognostic value of immune infiltrate by cell type and location was analyzed by univariable and multivariable Cox regression, incorporating the molecular endometrial cancer classification. Tumor-infiltrating immune cell density varied substantially between cases, and more modestly by immune cell type and location. Clustering revealed three groups with high, intermediate, and low densities, with highly significant variation in the proportion of molecular endometrial cancer subgroups between them. Univariable analysis revealed intraepithelial CD8(+) cell density as the strongest predictor of endometrial cancer recurrence; multivariable analysis confirmed this was independent of pathologic factors and molecular subgroup. Exploratory analysis suggested this association was not uniform across molecular subgroups, but greatest in tumors with mutant p53 and absent in DNA mismatch repair-deficient cancers. Thus, this work identified that quantification of intraepithelial CD8thorn cells improved upon the prognostic utility of the molecular endometrial cancer classification in early-stage endometrial cancer. Show less
