Strong Prognostic Value Research Articles

Multistain deep learning for prediction of prognosis and therapy response in colorectal cancer.

Although it has long been known that the immune cell composition has a strong prognostic and predictive value in colorectal cancer (CRC), scoring systems such as the immunoscore (IS) or quantification of intraepithelial lymphocytes are only slowly being adopted into clinical routine use and have their limitations. To address this we established and evaluated a multistain deep learning model (MSDLM) utilizing artificial intelligence (AI) to determine the AImmunoscore (AIS) in more than 1,000 patients with CRC. Our model had high prognostic capabilities and outperformed other clinical, molecular and immune cell-based parameters. It could also be used to predict the response to neoadjuvant therapy in patients with rectal cancer. Using an explainable AI approach, we confirmed that the MSDLM's decisions were based on established cellular patterns of anti-tumor immunity. Hence, the AIS could provide clinicians with a valuable decision-making tool based on the tumor immune microenvironment.

Prognostic and predictive value of Immunoscore and its correlation with ctDNA in stage II colorectal cancer

ABSTRACT This study aimed to validate the prognostic value of Immunoscore (IS) in stage II colorectal cancer (CRC), and explore the roles of IS and circulating tumor DNA (ctDNA) in the adjuvant treatment for early-stage CRC. Resected tumor samples from stage II CRC patients were collected from the Sun Yat-sen University Cancer Center. The densities of CD3+ and CD8+ lymphocytes were quantified and converted to IS and classified into Low, Intermediate (Int), and High groups according to predefined cutoffs. A total of 113 patients were included in the study. Patients with IS-High, Int, and Low were 43 (38%), 62 (55%), and 8 (7%), respectively. Patients with IS-High had an excellent clinical outcome, with none recurring during a median follow-up of 3 years, including 15 (35%) clinical high-risk patients. The 3-year disease-free survival (DFS) was 100% for IS-High, 76% for IS-Int, and 47% for IS-Low (P < .001). In the multivariate Cox analysis, IS was the only significant parameter associated with DFS. IS-Int and IS-Low patients with adjuvant chemotherapy had improved DFS compared to those who did not receive adjuvant chemotherapy (HR = 0.3; 95% CI 0.1–0.92; P = .026). Among the 49 patients with postoperative ctDNA data, IS-High patients had the lowest ctDNA positivity rate, suggesting that they were most eligible for chemotherapy-free treatment. IS had a strong prognostic value in Chinese patients with stage II CRC and demonstrates its clinical utility. IS and ctDNA will jointly optimize the adjuvant treatment strategies for early-stage CRC.

Pretreatment albumin is a prognostic and predictive biomarker for response to atezolizumab across solid tumors.

Reliable predictive biomarkers for response to immune checkpoint inhibition (ICI) are lacking. Pretreatment serum albumin, a known prognostic and predictive factor in ICI-treated patients, has been proposed as a potential pharmacokinetic surrogate marker for anti-PD1/PD-L1 antibodies, as it shares a homeostatic pathway with IgG. However, this hypothesis is currently based on theoretical considerations and limited evidence from retrospective data. Therefore, we comprehensively investigated the prognostic and predictive value of pretreatment albumin and its relationship with anti-PD-L1 IgG levels. We analysed pretreatment albumin and atezolizumab serum levels and clinical response in four trials (IMvigor210, IMvigor211, IMmotion151 and OAK) of patients with metastatic lung-, renal- or urothelial cancer who received atezolizumab alone or in combination. A total of 3391 patients were analysed. Correlation between serum albumin and atezolizumab levels was weak (Pearson's coefficient 0.23). We found a strong prognostic value for pretreatment serum albumin across all trials. Both atezolizumab serum levels and serum albumin were independently correlated with overall survival. Importantly, in the three randomised phase III clinical trials, the survival benefit for immunotherapy compared with the active comparator arm was limited to patients with pretreatment serum albumin > 35 g L-1. Our data do not support the hypothesis that albumin serves as a surrogate for atezolizumab pharmacokinetics. However, we show that albumin on its own exerts strong prognostic value for patients treated with immunotherapy. As benefit from immunotherapy was limited to patients with normal/elevated serum albumin levels, baseline albumin could potentially be used as a predictive marker for immune checkpoint inhibition.

Integrative network analysis reveals subtype-specific long non-coding RNA regulatory mechanisms in head and neck squamous cell carcinoma

Head and neck squamous cell carcinoma (HNSC) is one of most common malignancies with high mortality worldwide. Importantly, the molecular heterogeneity of HNSC complicates the clinical diagnosis and treatment, leading to poor overall survival outcomes. To dissect the complex heterogeneity, recent studies have reported multiple molecular subtyping systems. For instance, HNSC can be subdivided to four distinct molecular subtypes: atypical, basal, classical, and mesenchymal, of which the mesenchymal subtype is characterized by upregulated epithelial-mesenchymal transition (EMT) and associated with poorer survival outcomes. Despite a wealth of studies into the complex molecular heterogeneity, the regulatory mechanism specific to this aggressive subtype remain largely unclear. Herein, we developed a network-based bioinformatics framework that integrates lncRNA and mRNA expression profiles to elucidate the subtype-specific regulatory mechanisms. Applying the framework to HNSC, we identified a clinically relevant lncRNA LNCOG as a key master regulator mediating EMT underlying the mesenchymal subtype. Five genes with strong prognostic values, namely ANXA5, ITGA5, CCBE1, P4HA2, and EPHX3, were predicted to be the putative targets of LNCOG and subsequently validated in other independent datasets. By integrative analysis of the miRNA expression profiles, we found that LNCOG may act as a ceRNA to sponge miR-148a-3p thereby upregulating ITGA5 to promote HNSC progression. Furthermore, our drug sensitivity analysis demonstrated that the five putative targets of LNCOG were also predictive of the sensitivities of multiple FDA-approved drugs. In summary, our bioinformatics framework facilitates the dissection of cancer subtype-specific lncRNA regulatory mechanisms, providing potential novel biomarkers for more optimized treatment of HNSC.

1027 CONGENITAL CORONARY ARTERY ABNORMALITIES: THE DIAGNOSTIC AND PROGNOSTIC VALUE OF CARDIAC COMPUTED TOMOGRAPHY

Abstract Introduction To the best of our knowledge, there are no previous univocal data on how CCT can accurately describe morphological variants and subtypes of congenital coronary artery abnormalities. Taking into consideration the extensive application of CCT in clinical practice in the last decades, it is not uncommon to have a congenital coronary artery anomaly detected at CCT. Thus, the study aim is to evaluate the diagnostic and prognostic impact of specific cardiac CT parameters in subjects with a diagnosis of congenital coronary artery anomalies. Material and Methods This is an intermediate evaluation of a prospective clinical registry on a population of subjects who underwent a cardiac computed tomography imaging evaluation in the period between January 2007 and October 2015 and were diagnosed to have a congenital coronary abnormality. For the present preliminary analysis only coronary anomalies of origin have been considered and 92 patients have been included. Follow-up have been collected either through a structured telephonic interview or through the evaluation of clinical records of subsequent hospitalization or ambulatory visits. Results The population enrolled in the present study has an average age of 63,0 ± 12,8 years (range 34 to 85 years), with a male prevalence of 69,6%. The left main artery is missing in 46 individuals (50%), with separate origin of LAD and LCX. Only two duplications (2.2%), one for the left coronary artery and one for the right coronary artery, have been discovered. Six participants (6.5%) had a non-coronary artery origin abnormalities and pulmonary artery was the site of origin in the majority of cases. 51 subjects (55,4%) have an anomalous origin of the coronary artery from a different coronary sinus and 50 subjects had also an anomalous course with the retro-aortic being the most common (30,4%). 11 participants (12,0%) displayed also an intramural segment and 16 (17,4%) had abnormal ostial morphology. The high take-off of the vessel was uncommon (3,3%). Age appears to be the only relevant characteristic; indeed, it displays a significant correlation both to MACE (OR 95% CI 1.03, 1.01-1.07; p=0.0349) and to all-cause of death (OR 95% CI 1.14, 1.02-1.27; p=0.0151). None of the traditional cardiovascular risk factors were found to be significantly linked with adverse outcomes in this study sample of coronary anomalies. Conclusions The result of this intermediate evaluation is that cardiac CT can be successfully used to define the anatomy and features of CAA. However, it demonstrated that in middle-aged patients, the tomographic finding of an abnormality of coronary origin might not have a meaningful, strong negative prognostic value in terms of major cardiovascular events and all-cause of death. This means that probably, in the majority of the cases, once the diagnosis is performed later on in life, no further systematic assessment is needed but a personalized approach should be suggested.

Prognostic Relevance of Minimal Residual Disease in Therapy Related and Secondary Acute Myeloid Leukemia Receiving CPX-351 or Fludarabine-Based Induction

Background: The high prognostic value of minimal residual disease (MRD) assessment in Acute Myeloid Leukemia patients (AML) undergoing intensive induction therapy has been confirmed by several studies. The most widely adopted MRD techniques include multicolor flow cytometry (MFC), RT-PCR for recurrent genetic lesion and, for patients lacking specific markers, RT-PCR for the pan-leukemic marker WT1. However, mostly younger patients receiving a conventional 3+7 induction are included in clinical studies evaluating MRD in AML. As a consequence, secondary AML, evolving from a previous myelodisplastic syndrome (s-AML) or therapy-related AML (t-AML), which usually affect elderly AML patients are usually under-represented in those trials. Recently, CPX-351 was approved for the treatment of s-AML and t-AML, following the positive results of the Phase III trial by Lancet et al. However, as MRD was not assessed in this trial, only a limited ammount of data are available on the kinetics and the prognostic value of MRD in elderly s- AML and t-AML patients receiving CPX-351 induction. Aims: In order to assess the prognostic value and to define the best time points for MRD assessment, MFC-based and WT1 based MRD was evaluated in a cohort of elderly s-AML or t-AML patients receiving CPX-351 induction. Timepoints for assessment were after induction (TP1), after first consolidation (TP2) and at the end of treatment (TP3). As a comparison, we also evaluated MRD in a cohort of elderly s-AML and t-AML patients who received an age-adjusted fludarabine-containing regimen (FLAI3). MRD analysis in this cohort was performed with the same methodics and at the same timepoints. Methods: A total of 151 elderly (>60 year, median age 68, range 60-77) patients were analyzed in this study. Patients were treated between January 2005 and January 2020 in our Center, either with CPX-351 (n=50) or fludarabine- high dose cytarabine-idarubicin (FLAI), with (n=72) or without (n =29) gemtuzumab-ozogamicin (GO). MRD was analyzed in all patients achieving hematological complete remission (CR) with both MFC and WT1 expression levels. All patients were affected by s-AML or t-AML, defined according to the WHO 2016 criteria. Results: After first induction cycle, 95 patients (59%) achieved CR. Patients treated with CPX-351 had a significantly higher CR rate, if compared to patients receiving FLAI (40/50, 80% and 55/101, 54.5%, respectively, p<0.05). Adding GO to FLAI did not increase CR rate. Overall, in the two cohorts, an MRD negative CR was achieved in 51 (53.7%) and 53 responding patients (55.8%) assessed with MFC or WT1-based methods, respectively. The probability of achieving MFC MRD negativity was significantly higher among patients treated with CPX-351, when compared to patients receiving FLAI with or without GO (MFC MRD negativity rate of 26/40, 65% and 25/55, 45%, respectively, p<0.05). In both cohort, the most informative timepoint in terms of survival prediciton was after first cycle (TP1). WT1-based MRD assessment led to superimposable results. In multivariate analysis, MRD was significantly correlated with an higher probability of longer Overall Survival (OS) in all treatment group (2-year OS of 34% and 77% in patients with or without residual MFC MRD after induction, respectively, p<0.05). Allogeneic stem cell transplantation consolidation (HSCT) was also significantly correlated with longer OS in both treatment groups (p<0.05). Notably, 12/40 (30%) CR patients treated with CPX 351 underwent HSCT, whereas only 8 CR patients proceeded to HSCT in the FLAI group (15%) Conclusions: MRD assessment has a strong prognostic value also in the context of elderly s-AML and t-AML patients. WT1-based and MFC-based MRD assessment lead to similar conclusions thus allowing to obtain MRD data from virtually all AML patients treated in the selected time period. Overall, CPX-351 resulted in higher CR rate with deeper MRD responses, if compared to FLAI3. This increased efficacy translated in a significantly higher number of elderly AML patients proceeding to HSCT in CR, which is the only curative approach in this particular AML setting.

Abstract 14106: Prevalence and Respective Predictive Value of Late Gadolinium Enhancement Phenotypes in Patients Referred for Suspected Infiltrative Cardiomyopathy

Introduction: The respective value of different late gadolinium enhancement (LGE) phenotypes in patients with suspected infiltrative cardiomyopathy remains uncertain. Methods: In a large cohort of patients referred for suspected infiltrative cardiomyopathy, we evaluated respective associations between the dominant LGE phenotype and the composite endpoint of heart failure hospitalization, survived cardiac arrest or death. Kaplan-Meier survival curves were executed for six pre-defined LGE phenotypes. Multivariable analysis was performed to assess their independent association with the composite outcome, adjusted for relevant confounders. Final diagnosis was adjudicated using standardized disease-specific criteria. Results: 588 patients (68% male, median age 70 years, mean LVEF 54.7±12.7%) were studied. Of these patients, 139 (24%) were confirmed to have cardiac amyloidosis (CA) (81 [58%] transthyretin, 54 [39%] light chain and 4 [3%] undifferentiated) by objective non-CMR diagnostic criteria. Any LGE was present in 394 patients (67%) with the dominant phenotype being diffuse in 137 (23%), mid-wall patchy in 118 (20%), ischemic in 55 (9.4%), subepicardial in 18 (3.1%), mid-wall striae in 8 (1.4%), and RV insertion site in 58 (9.9%). The sensitivity and specificity of diffuse LGE for CA was 86% and 96%, respectively. Over a median 855 (IQR 372, 1542) days, 239 (41%) experienced the primary outcome. Unadjusted survival curves for each LGE phenotype are shown in the Figure. Following adjustment, diffuse (HR 2.30, p&lt;0.001), mid-wall patchy (HR 1.52, p=0.040) and ischemic LGE (HR 1.78, p=0.023) remained independently associated (Figure). Conclusions: In patients referred for suspected infiltrative cardiomyopathy, diffuse, mid-wall patchy and ischemic patterns all showed independent prognostic value. We confirm that diffuse LGE pattern reliably identifies CA patients with high sensitivity and specificity while delivering strong prognostic value.

The influence of atrial fibrillation on the prognostic value of brain natriuretic peptide (BNP) in risk assesment of paroxysmal ventricular tachycardia

Introduction. Brain natriuretic peptide (BNP) as marker of higher left ventricular myocardial mechanical stress has a strong prognostic value in patients with heart failure (HF). Elevated BNP levels are associated with the of malignant ventricular arrhythmias. Atrial fibrillation (AF) is known to futher elevate BNP levels and this can influence on BNP prognostic value. The aim of the study was to assess the predictive value of BNP in risk assessment of paroxysmal ventricular tachycardia (PVT) in patients with HF and AF. Materials and methods. Totally, 92 patients with sustained or persistent AF, HF with left ventricular ejection fraction &lt; 50 %. All patients underwent transthoratic echocardiography; 24-hour ECG monitoring to assess the ventricular arrhythmia burden. BNP determination by enzyme immunoassay in venous blood serum. Results. Patients with HF, AF and PVTcompared with patients with HF and AF without PVT have higher levels of BNP (298,8 [149; 500,6] pg/ml versus 152 [145,7; 335,4] pg/ml, р = 0,02). There is a significant positive correlation between the level of BNP and the amount of PVT per 24 hours (r = 0,26; р &lt; 0,05). Increased BNP levels are associated with an increased risk of VPT, OR = 3,71 [95 % CI 1,2 to 13,1]. Discussion. BNP is a novel biomarker to improve risk stratification of sudden cardiac death not only for patients in sinus rhythm but for patients with AF. Conclusions. BNP has a predictive value in risk assessment of PVT in patients with HF and AF.

Strong prognostic value of SLAMF7 protein expression in patients with lymph node-positive breast cancer.

Breast cancer (BC) in women is the second most commonly diagnosed type of cancer worldwide, and the leading cause of cancer-related mortality among women. To date, surgery is the main treatment option, often combined with other (neo)adjuvant therapeutic modalities to treat this malignancy and prevent relapse. Despite the invasive aspects of the majority of these therapeutic interventions and their associated side-effects, these treatments are still unable to effectively cure this disease and prevent relapse. Thus, there is an urgent need for the identification of more relevant biomarkers for more effective theranostics. Signaling lymphocytic activation molecule F7 (SLAMF7) is a glycosylated cell surface protein that plays a critical role in immune cell functions under both healthy conditions and in cancer. It is specifically targeted by elotuzumab for the treatment of multiple myeloma. The present retrospective study aimed to investigate the expression patterns of SLAMF7 and evaluate its associations with clinicopathological features, including the survival outcomes of patients with BC. The protein expression of SLAMF7 in BC was investigated in 278 lymph node-positive formalin-fixed and paraffin-embedded (FFPE) tissue blocks using tissue microarray and immunohistochemistry techniques. The results revealed a significant association between cytoplasmic SLAMF7 protein expression and several clinicopathological parameters, particularly age at diagnosis (P<0.007), tumor invasion (P<0.008) and vascular invasion (P=0.05). Kaplan-Meier analysis revealed that the overexpression of SLAMF7 was a strong positive prognosticator of both disease-free and disease-specific survival in the patients with BC (log-rank P=0.001 and P=0.008, respectively). This suggests that patients with SLAMF7 protein expression have a higher survival rate and a lower recurrence rate. On the whole, the present study demonstrated that a weak or no SLAMF7 expression was a powerful prognosticator of poor survival outcomes associated with both tumor and vascular invasion. Therefore, elotuzumab (as SLAMF7monoclonal antibody therapy) may be a promising option for targeted therapy worthy of clinical testing in patients with BC.

In This Issue

In This Issue

272TiP HARMONIA SOLTI-2101 / AFT-58: A head-to-head phase III study comparing ribociclib (RIB) and palbociclib (PAL) in patients (pts) with hormone receptor-positive/HER2-negative/HER2-enriched (HR+/HER2-/HER2-E) advanced breast cancer (ABC)

RIB and PAL are CDK4/6 inhibitors with efficacy and tolerability in pts with HR+/HER2- ABC when administered in combination with endocrine therapy (ET). RNA-based intrinsic subtyping has strong prognostic and predictive value in advanced HR+/HER2- tumors treated with ET. Non-Luminal subtypes, such as HER2-E and basal-like, are relatively endocrine-resistant and have poorer prognosis than luminal. tumors can switch subtypes over time to a more aggressive and less endocrine-responsive biology, which may also be reversed by effective treatment.

A Potential Four-Gene Signature and Nomogram for Predicting the Overall Survival of Papillary Thyroid Cancer.

Background Although the prognosis of papillary thyroid cancer (PTC) is relatively good, some patients experience recurrence or distant metastasis after thyroidectomy and progress to radioactive iodine refractory stage. Therefore, accurate prediction of clinical outlook can aid to screen out the minority of patients with poorer prognosis and avoid excessive treatment in low-risk patients. Methods The RNA-seq and clinical data of PTC patients was downloaded from the Gene Expression Omnibus (GEO) and the Cancer Genome Atlas (TCGA) databases. Multivariate and Lasso Cox regression analyses were used to construct a prognostic nomogram to predict overall survival (OS). Thereafter, quantitative RT-PCR and Human Protein Atlas (HPA) database were employed to verify the expression of key genes. Results A four-gene risk score comprising ABI3BP, DPT, MRO, and TENM1 was exhibited strong prognostic value. Moreover, an integrated nomogram was established based on the risk score, age, AJCC (American Joint Commission on Cancer) stage, tumor size, extrathyroidal extension, and history of neoadjuvant treatment, which exhibited significantly better predictive performance than TNM stage system (P < 0.05). GSEA (Gene Set Enrichment Analysis) and GSVA (Gene Set Variation Analysis) revealed that the different tumor-associated hallmarks, biological processes, and pathways were substantially enriched in the poor-prognosis group. In addition, a ceRNA network was constructed by including the four genes (ABI3BP, DPT, MRO, and TENM1), 54 lncRNAs, and 10 miRNAs. Finally, both the relative mRNA and protein expression of ABI3BP, DPT, MRO, and TENM1 were validated. Conclusion The present study identified a four-gene risk signature and developed a novel nomogram, which could be regarded as a reliable prognostic model for PTC patients. The findings also revealed preliminary potential mechanisms that may influence the prognosis outcome. These results can be conducive to design personalized treatment and prognosis management in affected patients.

Machine learning-based screening of an epithelial-mesenchymal transition-related long non-coding RNA signature reveals lower-grade glioma prognosis and the tumor microenvironment and predicts antitumor therapy response.

Epithelial-mesenchymal transition (EMT) confers high invasive and migratory capacity to cancer cells, which limits the effectiveness of tumor therapy. Long non-coding RNAs (lncRNAs) can regulate the dynamic process of EMT at different levels through various complex regulatory networks. We aimed to comprehensively analyze and screen EMT-related lncRNAs to characterize lower-grade glioma (LGG) tumor biology and provide new ideas for current therapeutic approaches. We retrieved 1065 LGG samples from the Cancer Genome Atlas and Chinese Glioma Genome Atlas by machine learning algorithms, identified three hub lncRNAs including CRNDE, LINC00665, and NEAT1, and established an EMT-related lncRNA signature (EMTrLS). This novel signature had strong prognostic value and potential clinical significance. EMTrLS described LGG genomic alterations and clinical features including gene mutations, tumor mutational burden, World Health Organization (WHO) grade, IDH status, and 1p/19q status. Notably, stratified analysis revealed activation of malignancy-related and metabolic pathways in the EMTrLS-high cohort. Moreover, the population with increased EMTrLS scores had increased cells with immune killing function. However, this antitumor immune function may be suppressed by increased Tregs and macrophages. Meanwhile, the relatively high expression of immune checkpoints explained the immunosuppressive state of patients with high EMTrLS scores. Importantly, we validated this result by quantifying the course of antitumor immunity. In particular, EMTrLS stratification enabled assessment of the responsiveness of LGG to chemotherapeutic drug efficacy and PD1 blockade. In conclusion, our findings complement the foundation of molecular studies of LGG, provide valuable insight into our understanding of EMT-related lncRNAs, and offer new strategies for LGG therapy.

Molecular subtype stratified response to adjuvant therapy in endometrial cancer (086)

Objectives: Molecular classification of endometrial cancer (EC) has been shown to have a strong prognostic value. Recent data support the predictive implications of molecular subtype assignment. Using combined EC cohorts that have undergone molecular classification, our objective was to assess the response to adjuvant therapy based on ESMO 2016 risk group criteria within the four molecular subtypes (MMRd, POLEmut, NSMP, and p53abn). Methods: Historical institutional (29%) and population-based cohorts of EC from 2016 (53%) and prior to 2016 (18%) that had undergone molecular classification were identified. We assessed outcomes (overall, disease-specific, or progression-free survival [OS, DSS, PFS]) based on the adjuvant treatment given, including none, vault brachytherapy (VB) only, external beam radiation (EBRT) ± VB, or chemotherapy (± any RT including EBRT or VB) across ESMO risk groups (low, intermediate, high-intermediate, high and advanced/metastatic) within the four molecular subtypes. Conclusions: Although retrospective data regarding where the treatment was given was not confined to clinical trial parameters and need to be interpreted with caution, these results are consistent with the PORTEC3 trial subgroup analysis showing chemotherapy does not appear to add benefit within MMRd EC but may improve outcomes for patients with p53abn EC. These findings support molecular subtype-directed trials, such as radiation alone ± investigational agent for MMRd, chemotherapy ± investigational agent for p53abn, and de-escalation of adjuvant therapy for POLEmut EC. Objectives: Molecular classification of endometrial cancer (EC) has been shown to have a strong prognostic value. Recent data support the predictive implications of molecular subtype assignment. Using combined EC cohorts that have undergone molecular classification, our objective was to assess the response to adjuvant therapy based on ESMO 2016 risk group criteria within the four molecular subtypes (MMRd, POLEmut, NSMP, and p53abn). Methods: Historical institutional (29%) and population-based cohorts of EC from 2016 (53%) and prior to 2016 (18%) that had undergone molecular classification were identified. We assessed outcomes (overall, disease-specific, or progression-free survival [OS, DSS, PFS]) based on the adjuvant treatment given, including none, vault brachytherapy (VB) only, external beam radiation (EBRT) ± VB, or chemotherapy (± any RT including EBRT or VB) across ESMO risk groups (low, intermediate, high-intermediate, high and advanced/metastatic) within the four molecular subtypes. Conclusions: Although retrospective data regarding where the treatment was given was not confined to clinical trial parameters and need to be interpreted with caution, these results are consistent with the PORTEC3 trial subgroup analysis showing chemotherapy does not appear to add benefit within MMRd EC but may improve outcomes for patients with p53abn EC. These findings support molecular subtype-directed trials, such as radiation alone ± investigational agent for MMRd, chemotherapy ± investigational agent for p53abn, and de-escalation of adjuvant therapy for POLEmut EC.

Mind the target: circulating tumour DNA in gastrointestinal malignancies.

Circulating tumour DNA (ctDNA) is an appealing minimally invasive tool with significant theranostic potential. In this review, we highlighted recent studies evaluating three major applications of ctDNA in gastrointestinal malignancies. ctDNA demonstrated a strong prognostic value in colorectal and gastroesophageal cancers in assessing minimal residual disease after radical surgery. ctDNA-guided interventional studies are ongoing.Tracking clonal dynamics with early identification of response and resistance to therapies is of particular interest in gastrointestinal cancers especially for established targeted therapies such as antiepidermal growth factor receptor (EGFR), BRAF inhibitors and immune checkpoint inhibitors.Early cancer detection via ctDNA approaches is encouraging and of particular relevance in gastrointestinal cancers in view of limited screening programmes and yet poor outcomes of metastatic patients. Although feasible and powerful tool, the clinical utility of ctDNA to guide patients' management is challenging to demonstrate and requires further investigations in large interventional clinical trials.

452 Quantitative Assessment Of AI-based CCTA Plaque Volume Compared With IVUS

Introduction: Intravascular ultrasound (IVUS) is the gold-standard for invasive assessment of plaque burden, which has strong prognostic value for patients with coronary artery disease. The purpose of this study is to compare IVUS with AI-based plaque volume derived from coronary computed tomography angiography (CCTA; HeartFlow Inc, Mountain View, CA).

O-134 Predictive value of seminal oxidation-reduction potential (ORP) and sperm DNA fragmentation (SDF) analysis for reproductive outcomes of intracytoplasmic sperm injection (ICSI) cycles

Abstract Study question Do seminal oxidation-reduction potential (ORP) and sperm DNA fragmentation (SDF) predict ICSI cycle outcomes? Summary answer Both SDF and seminal ORP have strong prognostic value in predicting good fertilization (≥80%), blastocyst development (≥60%), clinical pregnancy and live birth. What is known already Seminal oxidative stress (OS) has been found to play an important role in various aetiologies of male infertility and is induced by the imbalance between reactive oxygen species (ROS) and antioxidants. OS is a major cause of sperm DNA fragmentation (SDF). It is well established that high levels of seminal OS and SDF are negatively associated with ART outcomes. Previous studies have associated high levels of SDF with an increase risk of miscarriage and a decrease in live birth rate. Study design, size, duration This prospective study included 144 couples who had a fresh autologous ICSI cycle between June 2018 and December 2020. There was no restriction to patients with severe male factor infertility. All cryopreservation and third party assisted reproductive techniques were excluded. Participants/materials, setting, methods This multi-centre study was performed at three ART Clinics in Cape Town, South Africa. All couples included in the study had fresh autologous ICSI cycles with freshly ejaculated semen. Seminal ORP testing was performed using the MiOXSYS system and SDF by means of a fluorescence microscopic TUNEL assay on freshly ejaculated semen samples that were used for ICSI. The ART outcomes evaluated were fertilization rate, blastocyst development rate, clinical pregnancy rate and live birth rate. Main results and the role of chance The study shows that seminal ORP significantly negatively correlates with fertilization rate (r=-0.267; P = 0.0012), blastocyst development rate (r=-0.389; P &amp;lt; 0.0001), pregnancy (r=-0.296; P = 0.0003) and live birth (r=-0.353; P &amp;lt; 0.0001). ROC curve analysis shows significant predictive power for ORP for fertilization (≥80%; AUC=0.652; P = 0.0012), blastocyst development rate (≥60%; AUC=0.762; P &amp;lt; 0.0001), pregnancy (AUC=0.677; P = 0.0002) and live birth (AUC=0.729; P &amp;lt; 0.0001). Comparable results were obtained for SDF. An average cut-off value for ORP of 0.55 mV/106 sperm/mL was calculated. The AUCs between the SDF and ORP results for all reproductive outcome parameters did not differ. For all reproductive outcome parameters, normal sperm morphology shows the lowest predictive power. ORP with male age as confounding factor had significant effects on odds ratios (OR) of fertilization (OR = 0.1446; CI: 0.048-0.430; P = 0.0001) and blastocyst development (OR = 0.2468; CI: 0.095-0.635; P = 0.0029), while it was not significant for pregnancy and life birth. Multivariate logistic regression with fertilization, blastocyst development, pregnancy and live birth as dependent parameters and high/low seminal ORP, primary and secondary infertility, unexplained infertility, polycystic ovaries, endometriosis, tubal factor myomectomy, ovarian insufficiency and advanced maternal age as independent variables showed overall model fits (P &amp;lt; 0.0002) with a significant (P &amp;lt; 0.0084) influence of ORP on all reproductive outcome parameters. Limitations, reasons for caution Since only ICSI patients were included in this study, this might have contributed to the fact the cut-off value of ORP is lower than the published cut-off. A further limitation is the use of a fluorescence microscopic TUNEL assay which drastically limited the number of sperm evaluated. Wider implications of the findings This is the first study showing predictive value of seminal ORP for reproductive outcome including pregnancy and live birth. The method is easy and much quicker than current methods for the determination of SDF. Hence, it can be used a point of care method. Trial registration number Not applicable

Abstract 5168: Assessment and prognostic implication of 3-gene signature identified as potential biomarker in bladder cancer

Abstract The majority of bladder cancers present with non-muscle invasive disease (NIMBC) confined to the superficial layers of the bladder wall, however, 50-70% of patients will subsequently progress to muscle-invasive bladder cancer (MIBC) that will require more aggressive treatment. For patient stratification, the current prognostication relies on tumor stage and grade, which remains challenging due to the lack of appropriate biomarker(s) to monitor disease progress. Therefore, there is an urgent need for predictive biomarkers that can distinguish between progressive versus non-progressive disease for the identification of patients to better stratify their risk of progression. Here we aim to identify novel prognostic biomarker(s) associated with disease progression by exploring bladder cancer patient gene expression database. We explored and analyzed the transcriptomic gene expression of bladder cancer patient datasets consisting of NIMBC and MIBC subtype from the NCBI GEO (GSE154261, GSE57813, and GSE37317) database. These datasets were analyzed using GEO2R (an R-based web application) and Limma R packages. Pathway enrichment analysis of differentially expressed genes (DEGs) between the NMIBC and MIBC group was analyzed using the ingenuity pathway analysis (IPA), metascape web-based portal, and Cytoscape. The above finding was functionally validated in human bladder cancer cell lines; RT4 (transitional cell papilloma), J82 (transitional cell carcinoma), HT1197 (bladder carcinoma), and 253JB-V (metastatic phenotype) and compared with the relative expression ofUROtsa (benign) urothelial cell line. A total of 1516 DEGs were identified between non-muscle-invasive and muscle-invasive bladder cancer specimens. To identify genes of prognostic value, we performed Gene Ontology (GO) and Kyoto Gene and Genomic encyclopedia (KEGG)analysis. A total of seven genes including CDKN2A, CDC20, CTSV, FOXM1, MAGEA6, KRT23, and S100A9 were confirmed with strong prognostic values in bladder cancer and validated by qRT-PCR conducted in various human bladder cancer cells representing stage-specific disease progression. ULCAN, human protein atlas, and The Cancer Genome Atlas datasets were used to confirm the predictive value of these genes in bladder cancer progression. Moreover, Kaplan-Meier analysis and Cox-hazard ratio analysis were performed to determine the prognostic role of these genes. Univariate analysis performed on a validation set identified a 3-panel gene set viz. CDKN2A, CTSV and FOXM1 with 95.5% sensitivity and 100% specificity in predicting bladder cancer progression. Our study screened and confirmed a panel of a 3-genes biomarkers that could precisely predict the progression and prognosis of bladder cancer. Citation Format: Shiv Shankar Verma, Eswar Shankar, Spencer Lin, Vaibhav Singh, E. Ricky Chan, Shufen Cao, Gregory T MacLennan, Lee E. Ponsky, Sanjay Gupta. Assessment and prognostic implication of 3-gene signature identified as potential biomarker in bladder cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5168.

Identification of a novel subtype of endometrial cancer with unfavorable outcome using artificial intelligence-based histopathology image analysis.

5594 Background: Molecular subtyping of endometrial cancer (EC), unlike histopathological evaluation, offers an objective and reproducible classification system that has strong prognostic value and therapeutic implications. The Proactive Molecular risk classifier for Endometrial cancer (ProMisE) was developed by our team as a pragmatic, cost-effective, and clinically applicable molecular classifier for EC patients. ProMisE has four subtypes: (i) POLE mutant ( POLEmut), (ii) mismatch repair deficient (MMRd), (iii) p53 abnormal (p53abn) by immunohistochemistry, and (iv) NSMP (No Specific Molecular Profile), lacking any of the defining features of the other three subtypes. While ProMisE subtypes are associated with clinical outcomes, within each subtype, there are clinical/prognostic outliers. This is particularly true within the largest ProMisE subtype; NSMP (representing ̃50% of ECs), where a subset of patients experience a very aggressive disease course, comparable to what is observed in patients with p53abn ECs. Methods: We hypothesized that objective assessment of the digitized hematoxylin and eosin (H&amp;E)-stained histopathology slides of the largest and most diverse EC subset, NSMP, could potentially identify clinical outcome outliers. As such, we developed an artificial intelligence (AI)-based image analysis model to identify the NSMP cases that had similar histopathological features to the p53abn subtype, as assessed by H&amp;E stain. We used a discovery cohort of 182 and an external validation cohort of 195 NSMP ECs. Results: Our AI-based image analysis model, based on deep convolutional neural networks, identified 21 (11.5%) out of the 182 NSMP cases with similar histopathological features as p53abn cases. We refer to these cases as ‘p53abn-like’ NSMPs. Compared to the rest of the NSMP cases, these cases had markedly inferior disease-specific survival (DSS) (10-year DSS 58.9% vs. 93.1% ( p&lt;3.44e-8)) and progression-free survival (PFS) (10-year PFS 55.1% vs. 91.4% ( p&lt;3.76e-6)). These findings were confirmed in our validation cohort, with 10.7% of the 195 patients categorized as ‘p53abn-like’ tumors with 10-year DSS of 82% vs. 51.3% ( p&lt;5.28e-5) and PFS of 89.3% vs. 56.6% (p&lt;2.15e-4). Conclusions: Utilizing an AI-based approach for histopathology image analysis, we have discovered ‘p53abn-like’ NSMPs, a novel subtype of NSMP ECs with morphological features similar to p53abn cases. ‘p53abn-like’ NSMPs exhibit similar clinical behavior as p53abn, having noticeably inferior outcome compared to the rest of the NSMP cases in two independent cohorts. These findings warrant further molecular investigation of this novel subtype of EC to identify the biological underpinning and future therapeutic strategies.

Asymmetrical Lung Injury: Management and Outcome.

Among mechanically ventilated patients, asymmetrical lung injury is probably extremely frequent in the intensive care unit but the lack of standardized measurements does not allow to describe any prevalence among mechanically ventilated patients. Many past studies have focused only on unilateral injury and have mostly described the effect of lateral positioning. The good lung put downward might receive more perfusion while the sick lung placed upward receive more ventilation than supine. This usually results in better oxygenation but can also promote atelectasis in the healthy lung and no consensus has emerged on the clinical indication of this posture. Recently, electrical impedance tomography (EIT) has allowed for the first time to precisely describe the distribution of ventilation in each lung and to better study asymmetrical lung injury. At low positive-end-expiratory pressure (PEEP), a very heterogeneous ventilation exists between the two lungs and the initial increase in PEEP first helps to recruit the sick lung and protect the healthier lung. However, further increasing PEEP distends the less injured lung and must be avoided. The right level can be found using EIT and transpulmonary pressure. In addition, EIT can show that in the two lungs, airway closure is present but with very different airway opening pressures (AOPs) which cannot be identified on a global assessment. This may suggest a very different PEEP level than on a global assessment. Lastly, epidemiological studies suggest that in hypoxemic patients, the number of quadrants involved has a strong prognostic value. The number of quadrants is more important than the location of the unilateral or bilateral nature of the involvement for the prognosis, and hypoxemic patients with unilateral lung injury should probably be considered as requiring lung protective ventilation as classical acute respiratory distress syndrome.