From morphology to methylome: epigenetic studies of Müllerian mesonephric-like adenocarcinoma reveal similarities to cervical mesonephric adenocarcinoma†.

Mesonephric adenocarcinomas (MAs) and mesonephric-like adenocarcinomas (MLAs) are rare, aggressive neoplasms that arise in the gynecologic tract and show overlapping morphologic, immunohistochemical, and molecular features. While MAs occur in the cervix and are thought to arise from mesonephric remnants, MLAs occur in the endometrium and ovary and are believed to originate from transdifferentiation of Müllerian lesions. Both MAs and MLAs show a variety of architectural patterns, exhibit frequent expression of GATA3 by immunohistochemistry, and harbor KRAS mutations. In a recent article published in The Journal of Pathology, Kommoss and colleagues used DNA methylation profiling to extend these similarities and showed that MLAs and MAs cluster together based on their epigenetic signatures and are epigenetically distinct from other Müllerian adenocarcinomas. They also showed that MLAs and MAs harbor a high number of global copy number alterations. This study provides evidence that MLAs more closely resemble MAs than Müllerian carcinomas on an epigenetic level. As a result, the authors argue that MLA should be renamed 'mesonephric-type adenocarcinoma.' Further research is needed to establish the relationship between these two entities, their etiology, and pathogenesis. © 2024 The Pathological Society of Great Britain and Ireland.

Ultrasound Surveillance in Melanoma Management: Bridging Diagnostic Promise with Real-World Adherence: A Systematic Review and Meta-Analysis.

BACKGROUND: Ultrasound surveillance has become the new standard of care in stage III melanoma after the 2017 Multicenter Selective Lymphadenectomy Trial II (MSLT-II) demonstrated non-inferior 3-year survival compared with complete lymph node dissection. OBJECTIVE: We aimed to quantify diagnostic performance and adherence rates of ultrasound surveillance for melanoma locoregional metastasis, offering insights into real-world applicability. METHODS: Conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines, we systematically searched the Medline, Embase, Cochrane Library, CINAHL, Scopus, and Web of Science databases from inception until 11 October 2023. All primary studies that reported data on the diagnostic performance or adherence rates to ultrasound surveillance in melanoma were included. R statistical software was used for data synthesis and analysis. Sensitivity and specificity were aggregated across studies using the meta-analytic method for diagnostic tests outlined by Rutter and Gatsonis. Adherence rates were calculated as the ratio of patients fully compliant to planned follow-up to those who were not. RESULTS: A total of 36 studies including 18,273 patients were analysed, with a mean age of 56.6 years and a male-to-female ratio of 1:1.11. The median follow-up duration and frequency was 36 and 4 months, respectively. The pooled sensitivity of ultrasound examination was 0.879 (95% confidence interval [CI] 0.878-0.879) and specificity was 0.969 (95% CI 0.968-0.970), representing a diagnostic odds ratio of 224.5 (95% CI 223.1-225.9). Ultrasound examination demonstrated a substantial improvement in absolute sensitivity over clinical examination alone, with a number needed to screen (NNS) of 2.95. The overall adherence rate was 77.0% (95% CI 76.0-78.1%), with significantly lower rates in the United States [US] (p <  0.001) and retrospective studies (p <  0.001). CONCLUSION: Ultrasound is a powerful diagnostic tool for locoregional melanoma metastasis. However, the real applicability to surveillance programmes is limited by low adherence rates, especially in the US. Further studies should seek to address this adherence gap.

Characterizing the limitations of using diagnosis codes in the context of machine learning for healthcare.

BACKGROUND: Diagnostic codes are commonly used as inputs for clinical prediction models, to create labels for prediction tasks, and to identify cohorts for multicenter network studies. However, the coverage rates of diagnostic codes and their variability across institutions are underexplored. The primary objective was to describe lab- and diagnosis-based labels for 7 selected outcomes at three institutions. Secondary objectives were to describe agreement, sensitivity, and specificity of diagnosis-based labels against lab-based labels. METHODS: This study included three cohorts: SickKids from The Hospital for Sick Children, and StanfordPeds and StanfordAdults from Stanford Medicine. We included seven clinical outcomes with lab-based definitions: acute kidney injury, hyperkalemia, hypoglycemia, hyponatremia, anemia, neutropenia and thrombocytopenia. For each outcome, we created four lab-based labels (abnormal, mild, moderate and severe) based on test result and one diagnosis-based label. Proportion of admissions with a positive label were presented for each outcome stratified by cohort. Using lab-based labels as the gold standard, agreement using Cohen's Kappa, sensitivity and specificity were calculated for each lab-based severity level. RESULTS: The number of admissions included were: SickKids (n = 59,298), StanfordPeds (n = 24,639) and StanfordAdults (n = 159,985). The proportion of admissions with a positive diagnosis-based label was significantly higher for StanfordPeds compared to SickKids across all outcomes, with odds ratio (99.9% confidence interval) for abnormal diagnosis-based label ranging from 2.2 (1.7-2.7) for neutropenia to 18.4 (10.1-33.4) for hyperkalemia. Lab-based labels were more similar by institution. When using lab-based labels as the gold standard, Cohen's Kappa and sensitivity were lower at SickKids for all severity levels compared to StanfordPeds. CONCLUSIONS: Across multiple outcomes, diagnosis codes were consistently different between the two pediatric institutions. This difference was not explained by differences in test results. These results may have implications for machine learning model development and deployment.

Socioeconomic Disparities in Research Participation: Bias in Plastic Surgery Residency Match.

Background: Integrated plastic surgery residency applicants have increased at a rate disproportionate to available positions. Research productivity has become a surrogate marker for competitiveness, and many applicants pursue it to distinguish themselves. To date, no study has investigated socioeconomic disparities in extended research experience (ERE) participation. Methods: A 35-question cross-sectional survey was distributed to applicants to United States-based integrated plastic surgery residency programs during the 2019-2022 application cycles. Summary tables, student t test, and chi-square tests were used for statistical analysis. Results: A total of 161 responses (response rate: 20.9%) were recorded. Fifty-nine (40.7%) respondents participated in an ERE. The most common reason for ERE participation was strengthening one's application. The most common reason against participation was avoiding delays in career progression. A greater percentage of respondents from Northeastern medical schools participated in EREs (P = 0.019). There were no significant differences in debt burden between those who did or did not participate in an ERE. A greater percentage of applicants whose parents had advanced degrees participated in EREs (P = 0.053). Conclusions: There may be geographic and socioeconomic biases present in access to ERE for students interested in plastic surgery. The growing popularity of EREs may have unintended consequences for applicant diversity. As most plastic surgeons ultimately practice in nonacademic settings, applicants and plastic surgeons may consider the financial hardships and possible socioeconomic disparities in research opportunities before participating in or recommending them.

Increased PI3K pathway activity is associated with recurrent breast cancer in patients with low and intermediate 21-gene recurrence score.

AIMS: We investigated key signalling pathways' activity and mutational status of early-stage breast carcinomas with low and intermediate 21-gene recurrence score (RS) to identify molecular features that may predict recurrence. METHODS: This is a retrospective case-control study of 18 patients with recurrent breast carcinoma with low and intermediate 21-gene RS (<25) and control group of 15 non-recurrent breast cancer patients. DNA and mRNA were extracted from tumour tissue. mRNA expression of genes involved in oestrogen receptor (ER), androgen receptor (AR), PI3K and MAPK signalling pathways was measured by real-time quantitative reverse transcription-qPCR (OncoSIGNal G4 test, InnoSIGN). Tumour mutational landscape was assessed by targeted DNA sequencing (Oncomine Precision Assay). RESULTS: There were no statistical differences between the groups' demographic and clinicopathological characteristics. PI3K pathway showed significantly higher activity in cases compared with controls (p=0.0014). Receiver operating characteristic curve analysis showed an area under the curve of 0.79 for PI3K pathway activity in the prediction of recurrent disease in low and intermediate 21-gene RS breast cancer. There was no difference in ER, AR and MAPK pathway activity. PIK3CA alterations were the most common driver mutations, but no difference was found between the groups (p=0.46) and no association with PI3K pathway activity (p=0.86). Higher Ki67 gene expression was associated with recurrences (p=0.042) CONCLUSION: Increased PI3K pathway activity, independent of PIK3CA mutations, may play a role in the recurrence of early-stage breast cancer with low and intermediate 21-gene RS. Pathway analysis can help to identify high-risk patients in this setting.

Intermediate-Term Outcomes of Hip Arthroscopy for Femoroacetabular Impingement Syndrome in Patients With Global Versus Isolated Lateral Acetabular Overcoverage.

BACKGROUND: Previous studies evaluating the outcomes of hip arthroscopy for patients with global acetabular overcoverage and focal superolateral acetabular overcoverage suffer from short-term follow-up and inconsistent radiographic criteria when defining these subpopulations of patients with femoroacetabular impingement syndrome (FAIS). PURPOSE: To evaluate the intermediate-term postoperative outcomes for patients with FAIS in the setting of global acetabular overcoverage, lateral acetabular overcoverage, and normal acetabular coverage. STUDY DESIGN: Cohort study; Level of evidence, 3. METHODS: Patients undergoing hip arthroscopy for FAIS were enrolled in a prospective cohort study, and those with a minimum follow-up of 5 years were included in this analysis. Patients were grouped based on type of acetabular coverage: global overcoverage (lateral center-edge angle [LCEA] ≥40°, with coxa profunda), lateral overcoverage (LCEA ≥40°, without coxa profunda), and no overcoverage (LCEA <40°). Functional outcomes (modified Harris Hip Score and Nonarthritic Hip Score) and failure of primary hip arthroscopy were compared between groups. RESULTS: In total, 94 patients (mean age, 41.9 ± 14.2 years) were included with a mean follow-up duration of 6.1 ± 0.9 years. Of these patients, 40.4% had no acetabular overcoverage, 36.2% had lateral overcoverage, and 23.4% had global overcoverage. There was no difference between groups with respect to percentage of patients who underwent reoperation for either revision arthroscopy or conversion to total hip arthroplasty (28.9% for the normal acetabular coverage group, 29.4% for the lateral overcoverage group, and 31.8% for the global overcoverage group; P = .971). Among patients for whom primary hip arthroscopy did not fail, there was no difference in 5-year functional outcomes between groups. Postoperative LCEA >40° (ß = -13.3; 95% CI, -24.1 to -2.6; P = .016), female sex (ß = -14.5; 95% CI, -22.7 to -6.2; P = .001), and higher body mass index (ß = -1.9; 95% CI, -2.8 to -1.0; P < .001) were associated with worse intermediate-term hip function in terms of modified Harris Hip Score. CONCLUSION: There was no difference in functional outcomes or rate of reoperation at a minimum of 5 years postoperatively between those with global acetabular overcoverage, those with regional lateral overcoverage, and those with normal acetabular coverage. Provided that an appropriate acetabuloplasty is performed, there is no evidence to suggest that global acetabular overcoverage portends a worse prognosis than other FAIS subtypes.

Decreased Hip Labral Width Measured on Preoperative Magnetic Resonance Imaging Is Associated With Greater Revision Rate After Primary Arthroscopic Labral Repair for Femoroacetabular Impingement Syndrome at 5-Year Follow-Up.

PURPOSE: To examine the associations between hip labral width and patient-reported outcomes, clinical threshold achievement rates, and rate of reoperation among patients with femoroacetabular impingement syndrome (FAIS) who underwent hip arthroscopy and labral repair at minimum 5-year follow-up. METHODS: Patients were identified from a prospective database who underwent primary hip arthroscopy for treatment of labral tears and FAIS. Modified Harris Hip Score (mHHS) and Nonarthritic Hip Score (NAHS) were recorded preoperatively and at 5-year follow-up. Achievement of the minimal clinically important difference (MCID), substantial clinical benefit (SCB), and patient-acceptable symptom state (PASS) was determined using previously established values. Labral width magnetic resonance imaging measurements were performed by 2 independent readers at standardized "clockface" locations. Patients were stratified into 3 groups at each position: lower-width (<½ SD below mean), middle-width (within ½ SD of mean), and upper-width (>½ SD above mean). Multivariable regression was used to evaluate associations of labral width with patient-reported outcomes and reoperation rate. RESULTS: Seventy-three patients (age: 41.0 ± 12.0 years; 68.5% female) were included. Inter-rater reliability for labral width measurements was high at all positions (intraclass correlation coefficient 0.94-0.96). There were no significant intergroup differences in mHHS/NAHS improvement (P > .05) or in achievement rates of MCID/SCB/PASS at each clockface position (P > .05). Eleven patients (15.1%) underwent arthroscopic revision and 4 patients (5.5%) converted to total hip arthroplasty. Multivariable analysis found lower-width groups at 11:30 (odds ratio 1.75, P = .02) and 3:00 (odds ratio 1.59, P = .04) positions to have increased odds of revision within 5 years; however, labral width was not associated with 5-year improvement in mHHS/NAHS, achievement of MCID/PASS/SCB, or conversion to total hip arthroplasty (P > .05). CONCLUSIONS: Hip labral width <½ SD below the mean measured on preoperative magnetic resonance imaging at 11:30- and 3:00-clockface positions was associated with increased odds of reoperation after arthroscopic labral repair and treatment of FAIS. Labral width was not associated with 5-year improvement of mHHS, NAHS, achievement of clinical thresholds, or conversion to arthroplasty. LEVEL OF EVIDENCE: Level IV, case series.

Development and validation of the SickKids Enterprise-wide Data in Azure Repository (SEDAR).

Objectives: To describe the processes developed by The Hospital for Sick Children (SickKids) to enable utilization of electronic health record (EHR) data by creating sequentially transformed schemas for use across multiple user types. Methods: We used Microsoft Azure as the cloud service provider and named this effort the SickKids Enterprise-wide Data in Azure Repository (SEDAR). Epic Clarity data from on-premises was copied to a virtual network in Microsoft Azure. Three sequential schemas were developed. The Filtered Schema added a filter to retain only SickKids and valid patients. The Curated Schema created a data structure that was easier to navigate and query. Each table contained a logical unit such as patients, hospital encounters or laboratory tests. Data validation of randomly sampled observations in the Curated Schema was performed. The SK-OMOP Schema was designed to facilitate research and machine learning. Two individuals mapped medical elements to standard Observational Medical Outcomes Partnership (OMOP) concepts. Results: A copy of Clarity data was transferred to Microsoft Azure and updated each night using log shipping. The Filtered Schema and Curated Schema were implemented as stored procedures and executed each night with incremental updates or full loads. Data validation required up to 16 iterations for each Curated Schema table. OMOP concept mapping achieved at least 80 % coverage for each SK-OMOP table. Conclusions: We described our experience in creating three sequential schemas to address different EHR data access requirements. Future work should consider replicating this approach at other institutions to determine whether approaches are generalizable.

Preoperative Symptom Severity Predicts 5-Year Hip Arthroscopy Outcomes.

PURPOSE: Thisstudy sought to assessthe prognostic effect of preoperative symptom severity on hip arthroscopy outcomes for femoroacetabular impingement syndrome (FAI). METHODS: Patients undergoing hip arthroscopy between September 2012 and July 2014 for FAI with a minimum of 5-year clinical outcomes were compiled. Patient reported outcomes (PROs) including modified Harris Hip Score (mHHS) and Nonarthritic Hip Score (NAHS) were collected. High and low preoperative function (PF) subgroups were created using baseline population median mHHS (43.3) as a threshold with PROs below the median score indicating low preoperative function and vice versa for scores above the median. Kaplan-Meier analysis, Cox proportional modeling, analysis of variance (ANOVA), and linear regressions were used for analysis. RESULTS: One hundred five of 131 eligible patients(80.2% inclusion; age: 42.6 ± 1.4 years; body mass index: 25.3 ± 0.4 kg/m2 ) met the study criteria. The 5-year survival-torevision rate (85% versus 61%, p = 0.013) and survivalto-arthroplasty rate (95% vs. 82%, p = 0.022) were greater in the high versus low PF group. ANOVA demonstrated the high versus low PF group had higher baseline (mHHS: 52.7 ± 1.4 vs. 36.1 ± 1.1, p < 0.001; NAHS: 57.4 ± 1.6 vs. 39.3 ± 1.2, p < 0.001) and 1-year (mHHS: 91.9 ± 1.8 vs. 79.5 ± 2.7, p < 0.001; NAHS: 91.7 ± 1.6 vs. 80.8 ± 2.5, p < 0.001) outcomes. High versus low PF achieved higher Minimal Clinically Important Difference (77% vs. 57%, p = 0.026) at 5-years. High versus low PF achieved higher Patient Acceptable Symptomatic State rates at 1 year (79% vs. 47%, p < 0.001) and 5 years (66% vs. 45%, p = 0.032). Linear regression demonstrated body mass index (mHHS: p = 0.002; NAHS: p = 0.005), pincer resection (mHHS: p = 0.046), and preoperative symptom severity (mHHS: p = 0.001; NAHS: p = 0.002) to be predictors of 5-year change in PROs. CONCLUSION: Preoperative symptom severity is a reliable prognostic indicator of clinical survival rates and PROs after hip arthroscopy for FAI. Subjects with high PF are likely to have increased longevity of the index procedure while maintaining excellent PASS and MCID rates mid-term as opposed to those with low PF.

Severe Contour Deformity of the Hip Following Corticosteroid Injection.

A 50-year-old woman developed severe soft tissue atrophy of the hip following a triamincolone acetonide injection to the greater trochanteric bursa. Saline injection therapy was initially attempted without improvement and the defect was ultimately treated effectively with serial fat grafting. Adverse soft tissue reactions are rare but potentially devastating complications of corticosteroid injections, and the use of soluble steroid preparations and proper injection techniques can minimize the risk to surrounding tissue. Serial fat grafting represents a promising treatment option for severe cases of steroid-induced soft tissue atrophy.

Buccal Fat Pad in Primary and Secondary Cleft Palate Repair: A Systematic Review of the Literature.

OBJECTIVE: To review the evidence supporting the use of buccal fat pad (BFP) in primary and secondary cleft palate repair and its short- and long- term clinical outcomes. DESIGN: Systematic review conducted by 2 independent reviewers following PRISMA guidelines. SETTING: NONE PARTICIPANTS: Articles were identified from three databases (Pubmed/Medline, Embase and Web of Science). Search terms included "cleft palate", "palatoplasty", "palate repair", "buccal fat pad". INTERVENTIONS: Use of BFP in primary and secondary cleft palatoplasty. MAIN OUTCOME MEASURES: Primary outcomes were immediate postoperative complications, postoperative fistula, and maxillary growth. Secondary outcomes were palatal length, speech, and donor site morbidity. RESULTS: Ninety-one reports were retrieved after excluding duplicates. Twenty-three studies were included (13 case series and 10 comparative studies). Overall level of evidence was low. Randomized and non-randomized studies had a high risk of bias. In primary palatoplasty, BFP was more frequently used filling lateral relaxing incisions(57.4%), or in the hard-soft palate junction and covering mucosal defects(30.1%). In these patients, post operative fistula incidence was 2.8%. Two studies found wider transverse maxillary dimensions after BFP use. No higher incidence of bleeding, infection, dehiscence, or flap necrosis was reported. In secondary palatoplasty, no recurrent fistulas were reported for patients undergoing BFP for fistula repair. CONCLUSIONS: BFP appears to be associated with a favorable impact in fistula prevention and management, as well as in transverse maxillary growth. However, there is a high heterogeneity among studies, high risk of bias and overall low quality of evidence. More high-quality research with long-term follow-up is warranted.

Medical Specialty Interest Survey (MSIS): understanding career interest and specialty training in commencing Australian medical interns.

AIM: To assess specialisation interests in commencing interns and create a standardised survey to aid medical schools, supervisors and health services in quantifying, understanding and supporting medical career development to improve medical workforce planning. METHOD: The Medical Specialty Interest Survey (MSIS) cross-sectional study was used. Incoming interns at a multisite tertiary hospital network in Melbourne, Australia rated their desire to pursue each specialty as a career using a Likert scale (1-5). 47 Medical Board of Australia Medical Specialties were included in the survey. RESULTS: Completion rate was 123 of 124 (99.2%). The overall mean desirability was 2.62, suggesting on average more specialties were deemed less preferred. Critical care specialties were most popular, while surgical specialties had least interest. Gastroenterology and cardiology were most popular among internal medicine specialties. General practice had low correlation with other specialties (Pearson correlation mean R coefficient 0.106 compared with overall mean 0.208), suggesting interns interested in general practice exhibit less interest in other specialties, and interest in specialisation confers low interest in general practice. Psychiatry had the lowest mean R coefficient of 0.088. CONCLUSIONS: The MSIS quantifies relative interest in 47 medical specialties and specialty interest correlations among final-year medical students/incoming interns. The MSIS may be a tool for medical schools, healthcare services and government agencies to better understand the career interest among medical students and pre-vocational doctors and therefore improve doctor retention and well-being.

DNA Methylation Identifies Epigenetic Subtypes of Triple-Negative Breast Cancers With Distinct Clinicopathologic and Molecular Features.

Triple-negative breast cancers (TNBC) include diverse carcinomas with heterogeneous clinical behavior. DNA methylation is a useful tool in classifying a variety of cancers. In this study, we analyzed TNBC using DNA methylation profiling and compared the results to those of mutational analysis. DNA methylation profiling (Infinium MethylationEPIC array, Illumina) and 50-gene panel-targeted DNA sequencing were performed in 44 treatment-naïve TNBC. We identified 3 distinct DNA methylation clusters with specific clinicopathologic and molecular features. Cluster 1 (phosphoinositide 3-kinase/protein kinase B-enriched cluster; n = 9) patients were significantly older (mean age, 71 years; P = .008) with tumors that were more likely to exhibit apocrine differentiation (78%; P < .001), a lower grade (44% were grade 2), a lower proliferation index (median Ki-67, 15%; P = .002), and lower tumor-infiltrating lymphocyte fractions (median, 15%; P = .0142). Tumors carried recurrent PIK3CA and AKT1 mutations and a higher percentage of low HER-2 expression (89%; P = .033). Cluster 3 (chromosomal instability cluster; n = 28) patients were significantly younger (median age, 57 years). Tumors were of higher grade (grade 3, 93%), had a higher proliferation index (median Ki-67, 75%), and were with a high fraction of tumor-infiltrating lymphocytes (median, 30%). Ninety-one percent of the germline BRCA1/2 mutation carriers were in cluster 3, and these tumors showed the highest level of copy number alterations. Cluster 2 represented cases with intermediate clinicopathologic characteristics and no specific molecular profile (no specific molecular profile cluster; n = 7). There were no differences in relation to stage, recurrence, and survival. In conclusion, DNA methylation profiling is a promising tool to classify patients with TNBC into biologically relevant groups, which may result in better disease characterization and reveal potential targets for emerging therapies.

Self-supervised machine learning using adult inpatient data produces effective models for pediatric clinical prediction tasks.

OBJECTIVE: Development of electronic health records (EHR)-based machine learning models for pediatric inpatients is challenged by limited training data. Self-supervised learning using adult data may be a promising approach to creating robust pediatric prediction models. The primary objective was to determine whether a self-supervised model trained in adult inpatients was noninferior to logistic regression models trained in pediatric inpatients, for pediatric inpatient clinical prediction tasks. MATERIALS AND METHODS: This retrospective cohort study used EHR data and included patients with at least one admission to an inpatient unit. One admission per patient was randomly selected. Adult inpatients were 18 years or older while pediatric inpatients were more than 28 days and less than 18 years. Admissions were temporally split into training (January 1, 2008 to December 31, 2019), validation (January 1, 2020 to December 31, 2020), and test (January 1, 2021 to August 1, 2022) sets. Primary comparison was a self-supervised model trained in adult inpatients versus count-based logistic regression models trained in pediatric inpatients. Primary outcome was mean area-under-the-receiver-operating-characteristic-curve (AUROC) for 11 distinct clinical outcomes. Models were evaluated in pediatric inpatients. RESULTS: When evaluated in pediatric inpatients, mean AUROC of self-supervised model trained in adult inpatients (0.902) was noninferior to count-based logistic regression models trained in pediatric inpatients (0.868) (mean difference = 0.034, 95% CI=0.014-0.057; P < .001 for noninferiority and P = .006 for superiority). CONCLUSIONS: Self-supervised learning in adult inpatients was noninferior to logistic regression models trained in pediatric inpatients. This finding suggests transferability of self-supervised models trained in adult patients to pediatric patients, without requiring costly model retraining.

DNA Methylation Signature of Synchronous Endometrioid Endometrial and Ovarian Carcinomas.

Next-generation sequencing (NGS) studies have demonstrated that co-occurring sporadic endometrioid endometrial carcinoma (EEC) and endometrioid ovarian carcinoma (EOC) are clonally related, suggesting that they originate from a single primary tumor. Despite clonality, synchronous EEC and EOC when diagnosed at early stage behave indolently, similar to isolated primary EEC or isolated primary EOC. In the present study, we compared the DNA methylation signatures of co-occurring EEC and EOC with those of isolated primary EEC and isolated primary EOC. We also performed targeted NGS to assess the clonal relatedness of 7 co-occurring EEC and EOC (4 synchronous EEC and EOC and 3 metastatic EEC based on pathologic criteria). NGS confirmed a clonal relationship in all co-occurring EEC and EOC. DNA methylation profiling showed distinct epigenetic signatures of isolated primary EEC and isolated primary EOC. Endometrial tumors from co-occurring EEC and EOC clustered with isolated primary EEC while their ovarian counterparts clustered with isolated primary EOC. Three co-occurring EEC and EOC cases with peritoneal lesions showed a closer epigenetic signature and copy number variation profile between the peritoneal lesion and EOC than EEC. In conclusion, synchronous sporadic EEC and EOC are clonally related but demonstrate a shift in DNA methylation signatures between ovarian and endometrial tumors as well as epigenetic overlap between ovarian and peritoneal tumors. Our results suggest that tumor microenvironment in the ovary may play a role in epigenetic modulation of metastatic EEC.

Clinical Outcome and Morphology-Based Analysis of p53 Aberrant and Mismatch Repair Protein-Deficient Ovarian Clear Cell Carcinoma and Their Association With p16, HER2, and PD-L1 Expression.

OBJECTIVES: We studied the prevalence and prognostic significance of mismatch repair deficient (MMRD) and p53 aberrant ovarian clear cell carcinoma (CCO) and their association with other prognostic and theranostic biomarkers (p16, HER2, PD-L1). We also aimed to identify morphologic features to serve as screening tools for immunohistochemical testing for these biomarkers. METHODS: Tissue microarrays with 3-mm cores from 71 pure CCOs were immunostained with PMS2, MSH6, p53, p16, HER2, and PD-L1. Expression status was correlated with tumor recurrence/disease progression and survival. It was also correlated with morphologic features (tumor size, nuclear grade, tumor architecture, mitotic activity, presence of endometriosis, tumor budding, and tumor inflammation). RESULTS: p53 aberrant tumors were associated with shorter overall and recurrence-free survivals (P = .002 and P = .01, respectively). In multivariate analysis, p53 aberrant status and tumor stage were independently associated with recurrence/disease progression (hazard ratio [HR] = 3.31, P = .037 and HR = 1.465, P = .004, respectively). p53 aberrant status was associated with tumor budding (P = .037). MMRD, p16, HER2, and PD-L1 expression had no prognostic significance. HER2 and PD-L1 were expressed in 56% and 35% of tumors, respectively. MMRD was associated with tumor expression of PD-L1 (P > .05) but not with tumor inflammation. CONCLUSIONS: Aberrant p53 in CCO is infrequent but associated with poor prognosis independent of stage. Presence of tumor budding could be a screening tool for p53 testing. High prevalence of HER2 and PD-L1 expression indicates the eligibility of patients with CCO for ongoing clinical trials using these therapeutic targets.

HPV Cotesting of Unsatisfactory Papanicolaou Tests: Implications for Follow-up Intervals.

OBJECTIVES: The 2019 American Society of Colposcopy and Cervical Pathology management guidelines recommend that patients with an unsatisfactory Papanicolaou (Pap) test (UPT) and negative human papillomavirus (HPV) cotest undergo repeat age-based screening in 2 to 4 months. The rationale is that a negative HPV test in the setting of an UPT may reflect an inadequate sample and therefore should not be interpreted as truly "negative." For patients 25 years and older who are cotested, if HPV is positive for the 16 or 18 genotypes, direct referral for colposcopy is recommended. Our study aimed to determine if a negative HPV cotest result is predictive of the absence of a high-grade squamous intraepithelial lesion (HSIL) and whether these patients may be called back for repeat testing at an interval longer than 2 to 4 months. METHODS: Follow-up cervical cytology and biopsy results in women with UPT and HPV cotests from January 2017 to December 2021 were collected. Original UPT and HPV cotest results were correlated with the follow-up Pap and biopsy results. RESULTS: There were 1,496 (2.28%) UPT cases out of 65,641 total Pap tests. Among the 1,496 UPT cases, 1,010 (67.5%) had HPV cotesting; 676 (45.1%) were followed by repeat Pap or biopsy within 4 months and 850 (56.8%) within 12 months. The total follow-up rate was 81%, with a range of 3 days to 36 months. The HSIL rate in HPV-positive cases was 5.7% (3/53) vs 0.4% (2/539) (P = .006) in HPV-negative cases. In UPT, HPV cotesting showed negative predictive values for low-grade and high-grade squamous intraepithelial lesion detection of 98.5% and 99.6%, respectively, while positive predictive values were 19% and 5.7%. CONCLUSIONS: A negative HPV cotest in individuals with UPT predicted the lack of HSIL in our study. Compliance with the recommended follow-up time of 2 to 4 months for women with UPT was low (45.1%). Our study suggests that women with UPT and negative HPV cotest may be safely called back at an interval longer than 4 months.

EHR foundation models improve robustness in the presence of temporal distribution shift.

Temporal distribution shift negatively impacts the performance of clinical prediction models over time. Pretraining foundation models using self-supervised learning on electronic health records (EHR) may be effective in acquiring informative global patterns that can improve the robustness of task-specific models. The objective was to evaluate the utility of EHR foundation models in improving the in-distribution (ID) and out-of-distribution (OOD) performance of clinical prediction models. Transformer- and gated recurrent unit-based foundation models were pretrained on EHR of up to 1.8 M patients (382 M coded events) collected within pre-determined year groups (e.g., 2009-2012) and were subsequently used to construct patient representations for patients admitted to inpatient units. These representations were used to train logistic regression models to predict hospital mortality, long length of stay, 30-day readmission, and ICU admission. We compared our EHR foundation models with baseline logistic regression models learned on count-based representations (count-LR) in ID and OOD year groups. Performance was measured using area-under-the-receiver-operating-characteristic curve (AUROC), area-under-the-precision-recall curve, and absolute calibration error. Both transformer and recurrent-based foundation models generally showed better ID and OOD discrimination relative to count-LR and often exhibited less decay in tasks where there is observable degradation of discrimination performance (average AUROC decay of 3% for transformer-based foundation model vs. 7% for count-LR after 5-9 years). In addition, the performance and robustness of transformer-based foundation models continued to improve as pretraining set size increased. These results suggest that pretraining EHR foundation models at scale is a useful approach for developing clinical prediction models that perform well in the presence of temporal distribution shift.

Evaluation of Feature Selection Methods for Preserving Machine Learning Performance in the Presence of Temporal Dataset Shift in Clinical Medicine.

BACKGROUND: Temporal dataset shift can cause degradation in model performance as discrepancies between training and deployment data grow over time. The primary objective was to determine whether parsimonious models produced by specific feature selection methods are more robust to temporal dataset shift as measured by out-of-distribution (OOD) performance, while maintaining in-distribution (ID) performance. METHODS: Our dataset consisted of intensive care unit patients from MIMIC-IV categorized by year groups (2008-2010, 2011-2013, 2014-2016, and 2017-2019). We trained baseline models using L2-regularized logistic regression on 2008-2010 to predict in-hospital mortality, long length of stay (LOS), sepsis, and invasive ventilation in all year groups. We evaluated three feature selection methods: L1-regularized logistic regression (L1), Remove and Retrain (ROAR), and causal feature selection. We assessed whether a feature selection method could maintain ID performance (2008-2010) and improve OOD performance (2017-2019). We also assessed whether parsimonious models retrained on OOD data performed as well as oracle models trained on all features in the OOD year group. RESULTS: The baseline model showed significantly worse OOD performance with the long LOS and sepsis tasks when compared with the ID performance. L1 and ROAR retained 3.7 to 12.6% of all features, whereas causal feature selection generally retained fewer features. Models produced by L1 and ROAR exhibited similar ID and OOD performance as the baseline models. The retraining of these models on 2017-2019 data using features selected from training on 2008-2010 data generally reached parity with oracle models trained directly on 2017-2019 data using all available features. Causal feature selection led to heterogeneous results with the superset maintaining ID performance while improving OOD calibration only on the long LOS task. CONCLUSIONS: While model retraining can mitigate the impact of temporal dataset shift on parsimonious models produced by L1 and ROAR, new methods are required to proactively improve temporal robustness.

Ovarian Clear Cell Carcinoma and Markers of Epithelial-Mesenchymal Transition (EMT): Immunohistochemical Characterization of Tumor Budding.

Tumor budding, largely considered a manifestation of epithelial-mesenchymal transition (EMT) is an established prognostic marker for several cancers. In a recent study, tumor budding was associated with poor clinical outcomes in early-stage ovarian clear cell carcinoma. Here, we evaluated the immune expression of 3 proteins shown to be associated with EMT (E-cadherin, ß-catenin, and glypican-3) in 72 primary tumors of ovarian clear cell carcinoma with median follow-up of 39.47 mo. E-cadherin and ß-catenin expression was further evaluated in tumor buds in 29 (40%) cases. In the tumor mass, diffuse membranous expression of E-cadherin and ß-catenin was seen in 83% (60/72) and 81% (58/72) cases, respectively. Nuclear accumulation of E-cadherin was seen in 7 (10%) cases, while none of the cases showed nuclear ß-catenin expression. Glypican-3 expression was diffuse in 33.3% (24/72), patchy in 29.2% (21/72), and absent in 37.5% (27/72) cases. Evaluation of tumor buds showed aberrant patterns of expression (complete loss/cytoplasmic accumulation/diminished, discontinuous incomplete membranous staining) of E-cadherin in 29/29 (100%) and of ß-catenin in 26/29 (90%) cases. E-cadherin, ß-catenin, and glypican-3 expression in the main tumor mass had no association with stage, lymph node status, recurrent/progressive disease, status at last follow-up, survival and histopathologic features ( P >0.05). Our finding of aberrant expression of both E-cadherin and ß-catenin in tumor buds indicates involvement of Wnt signaling pathway/EMT in tumor budding and outlines its significance as a prognostic marker especially for early-stage ovarian clear cell carcinoma.