Selection of endometrial carcinomas for p53 immunohistochemistry based on nuclear features

POLE-mutated Endometrial "Carcinosarcoma"

The molecular subtype classification of endometrial carcinomas has conceptually changed our approach to this disease. However, open questions remain about how to integrate certain histotype diagnoses with the molecular subtype. We report 2 cases with morphologic suspicion for endometrial carcinosarcoma, that still fell short of the essential criteria for diagnosing carcinosarcoma. On subsequent molecular testing pathogenic POLE mutations were detected and a descriptive diagnosis of endometrial endometrioid carcinomas, low-grade with a homologous sarcoma component was rendered. This challenges the existence of POLE-mutated "carcinosarcoma."

TP53 as a Diagnostic Aid in the Distinction of Ovarian Mucinous Borderline Tumors From Mucinous Carcinoma

Ovarian mucinous borderline tumors (MBTs) are clinically managed as benign neoplasms while the management of ovarian mucinous carcinomas (MC) is dependent on tumor stage. Despite the standardization of sampling of ovarian mucinous neoplasms, limited interobserver reproducibility between MBT and MC persists. Based on our recent finding that abnormal TP53 expression is associated with unfavorable outcome in MBT, we hypothesized that TP53 status might improve the reproducible distinction of MBT from MC. A virtual slide set of 85 consecutive ovarian mucinous neoplasms received at a single institution, with each case represented by 3 full sections, were reviewed by 3 pathologists in 2 iterations. The initial assessment was based solely on morphologic review, while the second iteration was performed with knowledge of TP53 status. The reproducibility of a trinary categorization (MBT, MBT with intraepithelial carcinoma [IEC], MC) significantly improved from a κ of 0.60 based on the initial morphologic assessment to a κ of 0.76 (t-test, P =0.0042) after consideration of TP53 immunohistochemistry (IHC) results. Six out of 85 patients died of disease, and in 2 of them, at least 1 pathologist assessed MBT with IEC and not MC even after integration of TP53 IHC. With the integration of TP53 IHC, substantial interobserver agreement for MBT and MC can be reached, particularly in cases with an uncertain degree of confluent growth. TP53 IHC can also be used to highlight and support the presence of IEC in MBT, however, discordances remained in 2 cases with adverse outcome.

Dedifferentiated and Undifferentiated Ovarian Carcinoma: An Aggressive and Molecularly Distinct Ovarian Tumor Characterized by Frequent SWI/SNF Complex Inactivation

Dedifferentiated and undifferentiated ovarian carcinomas (DDOC/UDOC) are rare neoplasms defined by the presence of an undifferentiated carcinoma. In this study, we detailed the clinical, pathological, immunohistochemical, and molecular features of a series of DDOC/UDOC. We collected a multi-institutional cohort of 23 DDOC/UDOC and performed immunohistochemistry for core switch/sucrose nonfermentable (SWI/SNF) complex proteins (ARID1A, ARID1B, SMARCA4, and SMARCB1), mismatch repair (MMR) proteins, and p53. Array-based genome-wide DNA methylation and copy number variation analyses were performed on a subset of cases with comparison made to a previously reported cohort of undifferentiated endometrial carcinoma (UDEC), small cell carcinoma of the ovary, hypercalcemic type (SCCOHT), and tubo-ovarian high-grade serous carcinoma (HGSC). The age of all 23 patients with DDOC/UDOC ranged between 22 and 71 years (with an average age of 50 years), and a majority of them presented with extraovarian disease (16/23). Clinical follow-up was available for 19 patients. Except for 2 patients, the remaining 17 patients died from disease, with rapid disease progression resulting in mortality within a year in stage II-IV settings (median disease-specific survival of 3 months). Eighteen of 22 cases with interpretable immunohistochemistry results showed loss of expression of core SWI/SNF protein(s) that are expected to result in SWI/SNF complex inactivation as 10 exhibited coloss of ARID1A and ARID1B, 7 loss of SMARCA4, and 1 loss of SMARCB1. Six of 23 cases were MMR-deficient. Two of 20 cases exhibited mutation-type p53 immunoreactivity. Methylation profiles showed coclustering of DDOC/UDOC with UDEC, which collectively were distinct from SCCOHT and HGSC. However, DDOC/UDOC showed an intermediate degree of copy number variation, which was slightly greater, compared with SCCOHT but much less compared with HGSC. Overall, DDOC/UDOC, like its endometrial counterpart, is highly aggressive and is characterized by frequent inactivation of core SWI/SNF complex proteins and MMR deficiency. Its molecular profile overlaps with UDEC while being distinct from SCCOHT and HGSC.

Subclonal p53 immunostaining in the diagnosis of endometrial carcinoma molecular subtype

AIMS

The significance of subclonal expression of p53 (abrupt transition from wild-type to mutant-pattern staining) is not well understood, and the arbitrary diagnostic cut-off of 10% between NSMP and p53abn molecular subtypes of endometrial carcinoma (EC) has not been critically assessed. Our aim was to characterise subclonal p53 and discrepant p53 expression/TP53 sequencing results in EC and assess their clinical significance.

METHODS AND RESULTS

Subclonal p53 immuostaining on whole sections from 957 ECs was recorded. Agreement between TP53 mutational assessment and p53 immunostaining was evaluated. Subclonal p53 IHC staining was seen in 4.0% (38 of 957) of cases, with 23 of 957 (2.4%) showing mutant-pattern p53 staining in ≥10% of tumour cells. It was most commonly seen in POLEmut (nine of 65, 14%) and MMRd (13 of 274, 4.7%) EC ('multiple classifier' ECs), where subclonal p53 staining does not impact the molecular subtype diagnosis. Excluding POLEmut and MMRd EC, 11 of 957 (1.1%) showed ≥10% subclonal p53 from which four patients died of disease, while there were no deaths due to disease in the five patients with <10% mutant-pattern p53 staining. Agreement between p53 immunostaining and TP53 sequencing was 92.6%; most of the discrepant results were in the ultramutated POLEmut or hypermutated MMRd ECs. In NSMP and p53abn EC the agreement between IHC and sequencing was 95.8%.

CONCLUSIONS

Subclonal p53 staining ≥10% is present in only 1.1% of EC after excluding 'multiple classifier' ECs. The cut-off of ≥10% subclonal p53 staining identified patients at increased risk of dying from EC, supporting its use to diagnose p53abn molecular subtype.

Clinical Behavior and Molecular Landscape of Stage I p53-Abnormal Low-Grade Endometrioid Endometrial Carcinomas

PURPOSE

The clinical significance of the p53-abnormal (p53abn) molecular subtype in stage I low-grade endometrioid endometrial carcinoma (EEC) is debated. We aimed to review pathologic and molecular characteristics, and outcomes of stage I low-grade p53abn EEC in a large international cohort.

EXPERIMENTAL DESIGN

Previously diagnosed stage I p53abn EC (POLE-wild-type, mismatch repair-proficient) low-grade EEC from Canadian retrospective cohorts and PORTEC-1&2 trials were included. Pathology review was performed by six expert gynecologic pathologists blinded to p53 status. IHC profiling, next-generation sequencing, and shallow whole-genome sequencing was performed. Kaplan-Meier method was used for survival analysis.

RESULTS

We identified 55 stage I p53abn low-grade EEC among 3,387 cases (2.5%). On pathology review, 17 cases (31%) were not diagnosed as low-grade EEC by any pathologists, whereas 26 cases (47%) were diagnosed as low-grade EEC by at least three pathologists. The IHC and molecular profile of the latter cases were consistent with low-grade EEC morphology (ER/PR positivity, patchy p16 expression, PIK3CA and PTEN mutations) but they also showed features of p53abn EC (TP53 mutations, many copy-number alterations). These cases had a clinically relevant risk of disease recurrence (5-year recurrence-free survival 77%), with pelvic and/or distant recurrences observed in 12% of the patients.

CONCLUSIONS

A subset of p53abn EC is morphologically low-grade EEC and exhibit genomic instability. Even for stage I disease, p53abn low-grade EEC are at substantial risk of disease recurrence. These findings highlight the clinical relevance of universal p53-testing, even in low-grade EEC, to identify women at increased risk of recurrence.

Targeted Molecular Testing in Endometrial Carcinoma: Validation of a Clinically Driven Selective ProMisE Testing Protocol

Incorporation of molecular classification into clinicopathologic assessment of endometrial carcinoma (EC) improves risk stratification. Four EC molecular subtypes, as identified by The Cancer Genome Atlas, can be diagnosed through a validated algorithm Pro active M olecular R is k Classifier for E ndometrial Cancer (ProMisE) using p53 and mismatch repair (MMR) protein immunohistochemistry (IHC), and DNA polymerase epsilon ( POLE) mutational testing. Cost and access are major barriers to universal testing, particularly POLE analysis. We assessed a selective ProMisE algorithm (ProMisE-S): p53 and MMR IHC on all EC's with POLE testing restricted to those with abnormal MMR or p53 IHC (to identify POLEmut EC with secondary abnormalities in MMR and/or p53) and those with high-grade or non-endometrioid morphology, stage >IA or presence of lymphovascular space invasion (so as to avoid testing on the lowest risk tumors). We retrospectively compared the known ProMisE molecular classification to ProMisE-S in 912 EC. We defined a group of "very low-risk" EC (G1/G2, endometrioid, MMR-proficient, p53 wild-type, stage IA, no lymphovascular space invasion) in whom POLE testing will not impact on patient care; using ProMisE-S, POLE testing would not be required in 55% of biopsies and 38% of all EC's, after evaluation of the hysterectomy specimen, in a population-based cohort. "Very low-risk" endometrioid EC with unknown POLE status showed excellent clinical outcomes. Fifteen of 166 (9%) of all p53abn EC showed G1/G2 endometrioid morphology, supporting the potential value of universal p53 IHC. The addition of molecular testing changed the risk category in 89/896 (10%) EC's. In routine practice, POLE testing could be further restricted to only those patients in whom this would alter adjuvant therapy recommendations.

p53 and ovarian carcinoma survival: an Ovarian Tumor Tissue Analysis consortium study

Our objective was to test whether p53 expression status is associated with survival for women diagnosed with the most common ovarian carcinoma histotypes (high-grade serous carcinoma [HGSC], endometrioid carcinoma [EC], and clear cell carcinoma [CCC]) using a large multi-institutional cohort from the Ovarian Tumor Tissue Analysis (OTTA) consortium. p53 expression was assessed on 6,678 cases represented on tissue microarrays from 25 participating OTTA study sites using a previously validated immunohistochemical (IHC) assay as a surrogate for the presence and functional effect of TP53 mutations. Three abnormal expression patterns (overexpression, complete absence, and cytoplasmic) and the normal (wild type) pattern were recorded. Survival analyses were performed by histotype. The frequency of abnormal p53 expression was 93.4% (4,630/4,957) in HGSC compared to 11.9% (116/973) in EC and 11.5% (86/748) in CCC. In HGSC, there were no differences in overall survival across the abnormal p53 expression patterns. However, in EC and CCC, abnormal p53 expression was associated with an increased risk of death for women diagnosed with EC in multivariate analysis compared to normal p53 as the reference (hazard ratio [HR] = 2.18, 95% confidence interval [CI] 1.36-3.47, p = 0.0011) and with CCC (HR = 1.57, 95% CI 1.11-2.22, p = 0.012). Abnormal p53 was also associated with shorter overall survival in The International Federation of Gynecology and Obstetrics stage I/II EC and CCC. Our study provides further evidence that functional groups of TP53 mutations assessed by abnormal surrogate p53 IHC patterns are not associated with survival in HGSC. In contrast, we validate that abnormal p53 IHC is a strong independent prognostic marker for EC and demonstrate for the first time an independent prognostic association of abnormal p53 IHC with overall survival in patients with CCC.

Interpretable deep learning model to predict the molecular classification of endometrial cancer from haematoxylin and eosin-stained whole-slide images: a combined analysis of the PORTEC randomised trials and clinical cohorts

BACKGROUND

Endometrial cancer can be molecularly classified into POLE^mut, 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 POLE^mut 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 μm resized to 224 × 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 0·874 (95% CI 0·856-0·893) on four-fold cross-validation and 0·876 on the independent test set. The class-wise AUROCs were 0·849 for POLE^mut (n=51), 0·844 for MMRd (n=134), 0·883 for NSMP (n=120), and 0·928 for p53abn (n=88). POLE^mut and MMRd tiles had a high density of lymphocytes, p53abn tiles had strong nuclear atypia, and the morphology of POLE^mut 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<0·0001). 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=0·30). 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=0·13). Patients with MMRd endometrial cancer predicted as POLE^mut had an excellent prognosis, as do those with true POLE^mut 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.

FUNDING

The Hanarth Foundation, the Promedica Foundation, and the Swiss Federal Institutes of Technology.

The evolving role of morphology in endometrial cancer diagnostics: From histopathology and molecular testing towards integrative data analysis by deep learning

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 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.

Improving the Diagnosis of Endometrial Hyperplasia Using Computerized Analysis and Immunohistochemical Biomarkers

Endometrial hyperplasia (EH) is a precursor lesion to endometrial carcinoma (EC). Risks for EC include genetic, hormonal and metabolic factors most notably those associated with obesity: rates are rising and there is concern that cases in pre-menopausal women may remain undetected. Making an accurate distinction between benign and pre-malignant disease is both a challenge for the pathologist and important to the gynecologist who wants to deliver the most appropriate care to meet the needs of the patient. Premalignant change may be recognized by histological changes of endometrial hyperplasia (which may occur with or without atypia) and endometrial intraepithelial neoplasia (EIN). In this study we created a tissue resource of EH samples diagnosed between 2004 and 2009 (n = 125) and used this to address key questions: 1. Are the EIN/WHO2014 diagnostic criteria able to consistently identify premalignant endometrium? 2. Can computer aided image analysis inform identification of EIN? 3. Can we improve diagnosis by incorporating analysis of protein expression using immunohistochemistry. Our findings confirmed the inclusion of EIN in diagnostic criteria resulted in a better agreement between expert pathologists compared with the previous WHO94 criteria used for the original diagnosis of our sample set. A computer model based on assessment of stromal:epithelial ratio appeared most accurate in classification of areas of tissue without EIN. From an extensive panel of putative endometrial protein tissue biomarkers a score based on assessment of HAND2, PTEN, and PAX2 was able to identify four clusters one of which appeared to be more likely to be benign. In summary, our study has highlighted new opportunities to improve diagnosis of pre-malignant disease in endometrium and provide a platform for further research on this important topic.

The Evolution of Ovarian Carcinoma Subclassification

Simple Summary

Historically, cancers presenting with their main tumor mass in the ovary have been classified as ovarian carcinomas (a concise term for epithelial ovarian cancer) and treated with a one-size-fits-all approach. Over the last two decades, a growing molecular understanding established that ovarian carcinomas consist of several distinct histologic types, which practically represent different diseases. Further research is now delineating several molecular subtypes within each histotype. This histotype/molecular subtype subclassification provides a framework of grouping tumors based on molecular similarities for research, clinical trial inclusion and future patient management.

Abstract

The phenotypically informed histotype classification remains the mainstay of ovarian carcinoma subclassification. Histotypes of ovarian epithelial neoplasms have evolved with each edition of the WHO Classification of Female Genital Tumours. The current fifth edition (2020) lists five principal histotypes: high-grade serous carcinoma (HGSC), low-grade serous carcinoma (LGSC), mucinous carcinoma (MC), endometrioid carcinoma (EC) and clear cell carcinoma (CCC). Since histotypes arise from different cells of origin, cell lineage-specific diagnostic immunohistochemical markers and histotype-specific oncogenic alterations can confirm the morphological diagnosis. A four-marker immunohistochemical panel (WT1/p53/napsin A/PR) can distinguish the five principal histotypes with high accuracy, and additional immunohistochemical markers can be used depending on the diagnostic considerations. Histotypes are further stratified into molecular subtypes and assessed with predictive biomarker tests. HGSCs have recently been subclassified based on mechanisms of chromosomal instability, mRNA expression profiles or individual candidate biomarkers. ECs are composed of the same molecular subtypes (POLE-mutated/mismatch repair-deficient/no specific molecular profile/p53-abnormal) with the same prognostic stratification as their endometrial counterparts. Although methylation analyses and gene expression and sequencing showed at least two clusters, the molecular subtypes of CCCs remain largely elusive to date. Mutational and immunohistochemical data on LGSC have suggested five molecular subtypes with prognostic differences. While our understanding of the molecular composition of ovarian carcinomas has significantly advanced and continues to evolve, the need for treatment options suitable for these alterations is becoming more obvious. Further preclinical studies using histotype-defined and molecular subtype-characterized model systems are needed to expand the therapeutic spectrum for women diagnosed with ovarian carcinomas.