Molecular Subtype Classification Research Articles

Stromal Tumor-Infiltrating Lymphocytes Associated with Immunohistopathology and Molecular Subtypes of Breast Cancer in Vietnam.

Breast cancer is a heterogeneous disease with varied symptoms and pathogenesis, as well as variable prognosis and therapeutic outcomes. Stromal tumor-infiltrating lymphocytes, one of the tumor microenvironment factors, has been recognized as an important immunological biomarker that reflected the antitumor immune response in breast cancer. We analyzed 207 invasive breast cancer patients who had lumpectomy or mastectomy and have not received any pre-operative treatment. Clinicopathological characteristics, immunohistochemistry characteristics, molecular subtypes classification and stromal TILs evaluation were investigated. Stromal TILs correlated with well-established prognostic markers. Tumor grade showed significantly higher sTILs percentages in high-grade tumors than in low-grade tumors (p<0.001). There was a statistically significant association between intermediate and high levels of sTILs and a high Ki-67 index (p< 0.001). ER/PR negative was significantly related to high sTILs. Mean sTILs score was significantly higher in TNBC (40.1±31.6%) compared to others, statistically significant (p<0.001). In HER2-negative breast cancer, sTILs were significantly associated with histologic grade, ER status, PR status, and Ki67 index. sTILs played an important role, associated with unfavorable factors in breast cancer. Our findings support the use of stromal sTILs to identify a more aggressive phenotype of tumors.

MMDAE-HGSOC: A novel method for high-grade serous ovarian cancer molecular subtypes classification based on multi-modal deep autoencoder

High-grade serous ovarian cancer (HGSOC) is a type of ovarian cancer developed from serous tubal intraepithelial carcinoma. The intrinsic differences among molecular subtypes are closely associated with prognosis and pathological characteristics. At present, multi-omics data integration methods include early integration and late integration. Most existing HGSOC molecular subtypes classification methods are based on the early integration of multi-omics data. The mutual interference among multi-omics data is ignored, which affects the effectiveness of feature learning. High-dimensional multi-omics data contains genes unassociated with HGSOC molecular subtypes, resulting in redundant information, which is not conducive to model training. In this paper, we propose a multi-modal deep autoencoder learning method, MMDAE-HGSOC. MiRNA expression, DNA methylation, and copy number variation (CNV) are integrated with mRNA expression data to construct a multi-omics feature space. The multi-modal deep autoencoder network is used to learn the high-level feature representation of multi-omics data. The superposition LASSO (S-LASSO) regression algorithm is proposed to fully obtain the associated genes of HGSOC molecular subtypes. The experimental results show that MMDAE-HGSOC is superior to the existing classification methods. Finally, we analyze the enrichment gene ontology (GO) terms and biological pathways of these significant genes, which are discovered during the gene selection process.

Classification of colorectal cancer consensus molecular subtypes using attention-based multi-instance learning network on whole-slide images

Colorectal cancer (CRC) is the third most common and second most lethal cancer globally. It is highly heterogeneous with different clinical-pathological characteristics, prognostic status, and therapy responses. Thus, the precise diagnosis of CRC subtypes is of great significance for improving the prognosis and survival of CRC patients. Nowadays, the most commonly used molecular-level CRC classification system is the Consensus Molecular Subtypes (CMSs). In this study, we applied a weakly supervised deep learning method, named attention-based multi-instance learning (MIL), on formalin-fixed paraffin-embedded (FFPE) whole-slide images (WSIs) to distinguish CMS1 subtype from CMS2, CMS3, and CMS4 subtypes, as well as distinguish CMS4 from CMS1, CMS2, and CMS3 subtypes. The advantage of MIL is training a bag of the tiled instance with bag-level labels only. Our experiment was performed on 1218 WSIs obtained from The Cancer Genome Atlas (TCGA). We constructed three convolutional neural network-based structures for model training and evaluated the ability of the max-pooling operator and mean-pooling operator on aggregating bag-level scores. The results showed that the 3-layer model achieved the best performance in both comparison groups. When compared CMS1 with CMS234, max-pooling reached the ACC of 83.86 % and the mean-pooling operator reached the AUC of 0.731. While comparing CMS4 with CMS123, mean-pooling reached the ACC of 74.26 % and max-pooling reached the AUC of 0.609. Our results implied that WSIs could be utilized to classify CMSs, and manual pixel-level annotation is not a necessity for computational pathology imaging analysis.

Consensus molecular subtypes and orphan non-coding RNAs in colorectal cancer.

e15651 Background: Colorectal cancer (CRC) is a highly heterogeneous disease with variable response to different treatment strategies. The consensus molecular subtype (CMS) classification system has been proposed as a gene expression-based framework aimed at capturing the genetic and molecular heterogeneity in CRCs and potentially stratifying patients into clinically relevant subgroups for treatment selection. We previously demonstrated that orphan non-coding RNAs (oncRNAs), a novel class of cancer enriched small non-coding RNAs (smRNA), exhibit cancer-specific expression. We also showed that oncRNAs can be used in a liquid biopsy assay for accurate early-stage detection of CRCs. We hypothesize that oncRNAs may be informative of CRC subtype-specific transcriptional signatures as defined by the CMS groups. Methods: To assess the effectiveness of oncRNAs for categorization of CMS group, we examined smRNA-seq data from The Cancer Genome Atlas (TCGA) for colon and rectal adenocarcinoma tumors ( n= 504), consisting of 77 CMS1, 216 CMS2, 71 CMS3, and 140 CMS4 tumors. CRC-specific oncRNA expression profiles were generated for all tumors in the TCGA cohort. We then trained a machine learning model on oncRNA expression profiles to predict CMS groups using a 5-fold cross validation setup. For our validation cohort, we applied our model to an independent CRC cohort ( n= 348) consisting of 56 CMS1, 101 CMS2, 59 CMS3, and 132 CMS4 CRC samples. Final CMS predictions for the validation cohort were made by averaging across the five training models. Results: Within TCGA testing folds, the model micro-averaged ROC-AUC for CMS classifications were 0.879 (95% CI: 0.86–0.90). In the validation cohort, micro-averaged AUC remained high at 0.915 (0.90–0.93). Sensitivities and AUCs for each subtype are provided in Table 1. The highest model-predicted likelihood for each CMS subtype had an accuracy of 73.6% (68.6%–78.1%), and the two most probable predictions had an accuracy of 91.1% (87.6%–93.9%). Conclusions: This study demonstrates the utility of oncRNA expression profiling to predict CMS for colorectal cancer. The generalizability of our machine learning model suggests that oncRNAs are informative of the biological differences across distinct CMS groups and may be used as a proxy for gene expression signatures to stratify colorectal cancers. Because smRNAs are more likely to be secreted into blood, we posit that an oncRNA-based liquid biopsy may open new opportunities to monitor patients and track changes in their tumor biology in real-time throughout their course of disease and treatment. [Table: see text]

Expression molecular subtype as a predictor of outcomes of chemoradiotherapy for esophageal cancer.

4030 Background: The standard treatment for advanced esophageal cancer is preoperative chemoradiotherapy (CRT) followed by surgery or definitive CRT (NCCN Guideline 2022 ver5). CRT has the advantage of preserving organs and functions and the prognosis for patients achieved complete response by CRT is good. However, it is currently impossible to predict the outcomes of CRT before treatment. We have performed RNA expression analysis of pre-treatment biopsy specimens of esophageal cancer and identified several intrinsic molecular subtypes, as well as E-type (high expression of squamous markers) with good prognosis for CRT and M2-type (high expression of stromal markers) with poor prognosis for CRT (US patent No 10,969,390 B2). Herein, we performed an integrated analysis to confirm the relationship between intrinsic expression molecular subtype classification and prognosis in the pooled population from two prospective cohort studies (UMIN000043637). Methods: Two cohort studies were as follows; 1) single institute prospective cohort study conducted at National Cancer Center Hospital East (n=157, UMIN-C000000459), 2) multicenter prospective cohort study conducted at 10 institutions (n=226, UMIN000016141). A total 383 cases met inclusion criteria and the Affymetrix U331Plus 2 panel (47,104 transcripts, 38,572 genes) was used to identify E-type and M2-type. The clinical course and outcome were blinded for subtype classification. Treatment was performed as usual at each institution, and treatment details and outcomes were examined by subtype. Results: Among 383 patients, 181 and 164 patients were treated by CRT and surgery, respectively. Ninety-nine percent (n=342) were squamous cell carcinoma. Stage I/II/III/IV were 13/61/86/21 and 17/63/81/3 in CRT group and surgery group, respectively. Among them, 76 (22%), 100 (29%), 169 (49%) patients were classified as E, M2 and others, respectively. Among the patients treated by CRT, CR rate of E-type and M2-type were 62.8% (95% CI, 47.8-75.7) and 33.3% (95% CI, 21.9-47.1) (p=0.006). The 5-year progression free survival (PFS) and overall survival (OS) rates of CRT group was 50.5% (95% CI, 34.7-64.4) and 59.8% (95% CI, 42.6-73.3) for E-type vs. 23.0% (95% CI, 12.0-36.2) and 29.0% (95%CI, 16.8-42.4) for M2-type, with statistically significant difference (p=0.011, p=0.003). Cox regression analysis showed that crude HR for PFS and OS of E vs M2 were 0.50 (95% CI, 0.296-0.859, p=0.012) and 0.41 (95% CI, 0.226-0750, p=0.004), respectively. In contrast, E and M2-type did not show the survival difference in surgery group. Conclusions: For the first time in the world, we confirmed that expression molecular subtype of esophageal cancer has the potential to predict the outcomes of CRT. The results may provide precision medicine in the selection of treatment for esophageal cancer.

A novel federated deep learning scheme for glioma and its subtype classification.

Deep learning (DL) has shown promising results in molecular-based classification of glioma subtypes from MR images. DL requires a large number of training data for achieving good generalization performance. Since brain tumor datasets are usually small in size, combination of such datasets from different hospitals are needed. Data privacy issue from hospitals often poses a constraint on such a practice. Federated learning (FL) has gained much attention lately as it trains a central DL model without requiring data sharing from different hospitals. We propose a novel 3D FL scheme for glioma and its molecular subtype classification. In the scheme, a slice-based DL classifier, EtFedDyn, is exploited which is an extension of FedDyn, with the key differences on using focal loss cost function to tackle severe class imbalances in the datasets, and on multi-stream network to exploit MRIs in different modalities. By combining EtFedDyn with domain mapping as the pre-processing and 3D scan-based post-processing, the proposed scheme makes 3D brain scan-based classification on datasets from different dataset owners. To examine whether the FL scheme could replace the central learning (CL) one, we then compare the classification performance between the proposed FL and the corresponding CL schemes. Furthermore, detailed empirical-based analysis were also conducted to exam the effect of using domain mapping, 3D scan-based post-processing, different cost functions and different FL schemes. Experiments were done on two case studies: classification of glioma subtypes (IDH mutation and wild-type on TCGA and US datasets in case A) and glioma grades (high/low grade glioma HGG and LGG on MICCAI dataset in case B). The proposed FL scheme has obtained good performance on the test sets (85.46%, 75.56%) for IDH subtypes and (89.28%, 90.72%) for glioma LGG/HGG all averaged on five runs. Comparing with the corresponding CL scheme, the drop in test accuracy from the proposed FL scheme is small (-1.17%, -0.83%), indicating its good potential to replace the CL scheme. Furthermore, the empirically tests have shown that an increased classification test accuracy by applying: domain mapping (0.4%, 1.85%) in case A; focal loss function (1.66%, 3.25%) in case A and (1.19%, 1.85%) in case B; 3D post-processing (2.11%, 2.23%) in case A and (1.81%, 2.39%) in case B and EtFedDyn over FedAvg classifier (1.05%, 1.55%) in case A and (1.23%, 1.81%) in case B with fast convergence, which all contributed to the improvement of overall performance in the proposed FL scheme. The proposed FL scheme is shown to be effective in predicting glioma and its subtypes by using MR images from test sets, with great potential of replacing the conventional CL approaches for training deep networks. This could help hospitals to maintain their data privacy, while using a federated trained classifier with nearly similar performance as that from a centrally trained one. Further detailed experiments have shown that different parts in the proposed 3D FL scheme, such as domain mapping (make datasets more uniform) and post-processing (scan-based classification), are essential.

Principles of Molecular Utility for CMS Classification in Colorectal Cancer Management

Colorectal cancer (CRC) is the second cause of cancer-related deaths in both sexes globally and presents different clinical outcomes that are described by a range of genomic and epigenomic alterations. Despite the advancements in CRC screening plans and treatment strategies, the prognosis of CRC is dismal. In the last two decades, molecular biomarkers predictive of prognosis have been identified in CRC, although biomarkers predictive of treatment response are only available for specific biological drugs used in stage IV CRC. Translational clinical trials mainly based on “omic” strategies allowed a better understanding of the biological heterogeneity of CRCs. These studies were able to classify CRCs into subtypes mainly related to prognosis, recurrence risk, and, to some extent, also to treatment response. Accordingly, the comprehensive molecular characterizations of CRCs, including The Cancer Genome Atlas (TCGA) and consensus molecular subtype (CMS) classifications, were presented to improve the comprehension of the genomic and epigenomic landscapes of CRCs for a better patient management. The CMS classification obtained by the CRC subtyping consortium categorizes CRC into four consensus molecular subtypes (CMS1–4) characterized by different prognoses. In this review, we discussed the CMS classification in different settings with a focus on its relationships with precursor lesions, tumor immunophenotype, and gut microbiota, as well as on its role in predicting prognosis and/or response to pharmacological treatments, as a crucial step towards precision medicine.

Attention-based GCN integrates multi-omics data for breast cancer subtype classification and patient-specific gene marker identification.

Breast cancer is a heterogeneous disease and can be divided into several subtypes with unique prognostic and molecular characteristics. The classification of breast cancer subtypes plays an important role in the precision treatment and prognosis of breast cancer. Benefitting from the relation-aware ability of a graph convolution network (GCN), we present a multi-omics integrative method, the attention-based GCN (AGCN), for breast cancer molecular subtype classification using messenger RNA expression, copy number variation and deoxyribonucleic acid methylation multi-omics data. In the extensive comparative studies, our AGCN models outperform state-of-the-art methods under different experimental conditions and both attention mechanisms and the graph convolution subnetwork play an important role in accurate cancer subtype classification. The layer-wise relevance propagation (LRP) algorithm is used for the interpretation of model decision, which can identify patient-specific important biomarkers that are reported to be related to the occurrence and development of breast cancer. Our results highlighted the effectiveness of the GCN and attention mechanisms in multi-omics integrative analysis and the implement of the LRP algorithm can provide biologically reasonable insights into model decision.

Abstract LB298: Molecular phenotype classification of metastatic prostate cancer by cell-free DNA methylation analysis

Abstract Introduction: Metastatic castration-resistant prostate cancer (mCRPC) is a heterogeneous disease which can be classified into clinically relevant subtypes based on the expression of transcription factors (TF), such as the androgen receptor (AR) and neuroendocrine markers. Neuroendocrine prostate cancer (NEPC), characterized by gain of stem-like and neuroendocrine features and lack of AR expression is a clinically aggressive variant. Due to the absence of adequate biomarkers, NEPC is usually detected at a very advanced stage. There is mounting evidence that molecular subtype changes seen in NEPC are enforced by widespread epigenetic alterations, in particular DNA methylation changes. In this study, we aim to devise a novel DNA methylation-based assay for molecular subtyping and disease monitoring from cell-free DNA (cfDNA). Methods: We analyzed genome wide methylation patterns in 60 prostate cancer patient-derived xenograft (PDX) and 133 mCRPC tumors using array- and sequencing-based assays. We integrated DNA methylation with TF cistrome data to determine the landscape of methylation alterations at key lineage TF binding sites (TFBS). A linear regression model was trained on low-pass Enzymatic Methyl-Seq (EM-seq) cfDNA data derived from PDXs to identify molecular subtype specific DNA methylation changes at these TFBS. The model performance was optimized with in silico admixture experiments. This model was then used to discern tumor molecular phenotypes from cfDNA in three independent cohorts of mCRPC patients using low-pass whole genome bisulfite sequencing and EM-seq. Results: We observed a strong association between TFBS methylation and TF expression. For lineage specific TFs such as AR and ASCL1, we identified core sets of TFBSs whose differential methylation allowed for accurate assay-independent molecular subtype classification in tumor tissues. Applying an optimized quantitative model to mCRPC patients who underwent comprehensive tissue sampling by rapid autopsy we observed perfect subtype prediction from both tissue samples and cfDNA (AUC=1). A similar analytical performance was observed in additional clinical mCRPC cohorts with cfDNA. Conclusions: We show that methylation patterns at TFBSs can determine TF activity and can be used to classify molecular subtypes from both tumor tissue and cfDNA. For prostate cancer, we demonstrate that this approach can accurately detect NEPC by cost-effective low-pass EM-seq. More broadly, this study provides a novel analysis framework for robustly assessing molecular tumor phenotypes in cfDNA with applications in solid and liquid tumor diagnostics. Citation Format: Mohamed Adil, Brian Hanratty, Pallabi Mustafi, Ilsa Coleman, Radhika Patel, Anna-Lisa Doebley, Robert Patton, Eden Cruikshank, Patricia Galipeau, Ruth Dumpit, Martine Roudier, Jin-Yih Low, Navonil De Sarkar, Robert Montgomery, Eva Corey, Colm Morrissey, Peter Nelson, Gavin Ha, Michael Haffner. Molecular phenotype classification of metastatic prostate cancer by cell-free DNA methylation analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB298.

Clinicopathological features and prognoses of very young patients (≤35 years) with breast cancer: a retrospective population-based study in China.

Clinicopathological features and prognoses of very young patients (≤35 years) with breast cancer: a retrospective population-based study in China.

Cross-attention multi-branch CNN using DCE-MRI to classify breast cancer molecular subtypes.

The aim of this study is to improve the accuracy of classifying luminal or non-luminal subtypes of breast cancer by using computer algorithms based on DCE-MRI, and to validate the diagnostic efficacy of the model by considering the patient's age of menarche and nodule size. DCE-MRI images of patients with non-specific invasive breast cancer admitted to the Second Affiliated Hospital of Dalian Medical University were collected. There were 160 cases in total, with 84 cases of luminal type (luminal A and luminal B and 76 cases of non-luminal type (HER 2 overexpressing and triple negative). Patients were grouped according to thresholds of nodule sizes of 20 mm and age at menarche of 14 years. A cross-attention multi-branch net CAMBNET) was proposed based on the dataset to predict the molecular subtypes of breast cancer. Diagnostic performance was assessed by accuracy, sensitivity, specificity, F1 and area under the ROC curve (AUC). And the model is visualized with Grad-CAM. Several classical deep learning models were included for diagnostic performance comparison. Using 5-fold cross-validation on the test dataset, all the results of CAMBNET are significantly higher than the compared deep learning models. The average prediction recall, accuracy, precision, and AUC for luminal and non-luminal types of the dataset were 89.11%, 88.44%, 88.52%, and 96.10%, respectively. For patients with tumor size <20 mm, the CAMBNET had AUC of 83.45% and ACC of 90.29% for detecting triple-negative breast cancer. When classifying luminal from non-luminal subtypes for patients with age at menarche years, our CAMBNET model achieved an ACC of 92.37%, precision of 92.42%, recall of 93.33%, F1of 92.33%, and AUC of 99.95%. The CAMBNET can be applied in molecular subtype classification of breasts. For patients with menarche at 14 years old, our model can yield more accurate results when classifying luminal and non-luminal subtypes. For patients with tumor sizes ≤20 mm, our model can yield more accurate result in detecting triple-negative breast cancer to improve patient prognosis and survival.

Phenotypic subtypes predict outcomes in colorectal cancer

BackgroundColorectal cancer (CRC) is the second leading cause of cancer-related deaths globally. The Colorectal Cancer Subtyping Consortium used the transcriptome-based method to classify CRC according to four molecular subtypes, each showing different genomic alterations and prognoses: CMS1 (microsatellite instable [MSI] immune), CMS2 (canonical), CMS3 (metabolic), and CMS4 (mesenchymal). To expedite the clinical implementation of such methods, easier and preferably tumor phenotype–based methods are needed. In this study, we describe a method to divide patients into four phenotypic subgroups using immunohistochemistry. Moreover, we analyze disease-specific survival (DSS) among different phenotypic subtypes and the associations between the phenotypic subtypes and clinicopathological variables.MethodsWe categorized 480 surgically treated CRC patients into four phenotypic subtypes (immune, canonical, metabolic, and mesenchymal) using the immunohistochemically determined CD3–CD8 tumor–stroma index, proliferation index, and tumor–stroma percentage. We analyzed survival rates for the phenotypic subtypes in different clinical patient subgroups using the Kaplan–Meier method and Cox regression analysis. Associations between phenotypic subtypes and clinicopathological variables were examined using the chi-square test.ResultsPatients with immune subtype tumors exhibited the best 5-year DSS, while mesenchymal subtype tumors accompanied the worst prognosis. The prognostic value of the canonical subtype showed wide variation among different clinical subgroups. Immune subtype tumors were associated with being female, stage I disease, and a right-side colon location. Metabolic tumors, however, were associated with pT3 and pT4 tumors, and being male. Finally, a mesenchymal subtype associated with stage IV disease, a mucinous histology, and a rectal tumor location.ConclusionsPhenotypic subtype predicts patient outcome in CRC. Associations and prognostic values for subtypes resemble the transcriptome-based consensus molecular subtypes (CMS) classification. In our study, the immune subtype stood out with its exceptionally good prognosis. Moreover, the canonical subtype showed wide variability among clinical subgroups. Further studies are needed to investigate the concordance between transcriptome-based classification systems and the phenotypic subtypes.

Abstract HER2-08: HER2-08 Molecular characterization of HER-2 low patients identifies basal-enriched subset with poor clinical outcomes in real-world data

Abstract Background: Recent groundbreaking work has shown that patients with lower levels of HER2-expression (HER2-low) may benefit from treatment with trastuzumab deruxtecan—an HER2 antibody-drug conjugate FDA-approved for treatment for HER2-positive (HER2+) patients and thus can represent a new molecular subtype. In fact, this HER2-low patient population is enriched with luminal disease but is clinically heterogeneous and outcomes have therefore not been extensively characterized due to the lack of annotated multimodal real-world data (RWD). We used Tempus RWD to identify unique HER2-low subtypes using RNA sequencing and compare outcomes across subtypes. Methods: We retrospectively analyzed 1,545 breast cancer samples from the Tempus database tested with the Tempus xT assay that includes whole-exome capture RNA-seq. Only tumors with known HER2 status determined via immunohistochemistry [IHC] and/or FISH were included. A HER2 RNA gene signature was developed by comparing HER2- (n=464, IHC 0+) and HER2+ (n=161, IHC 3+ or IHC 2+ and FISH+) patients—controlling for HR status—to identify genes associated with HER2 over-expression. This HER2 signature was used to stratify independent HER2-low samples (n=920, determined by IHC 1+ or IHC2+ and FISH-) via hierarchical clustering. Treatment use in this cohort was not assessed. Clusters were subsequently assessed according to clinical, demographic, and molecular factors including PAM50 molecular subtype classification of the RNA signatures. Real-world progression-free survival (rwPFS) was evaluated based on progression and death captured through manual expert abstraction for a subset of stage 4 patients (n=336) and estimated via Kaplan-Meier analysis. Results: HER2-low patients were clustered according to our HER2 expression signature identifying three distinct molecular clusters (Table 1). Stage and demographic distributions (race, ethnicity, age) were similar across clusters. Of note, cluster 3 (n=186, 20% of the HER2-low population) was significantly enriched in hormone receptor negative (HR–) patients (p&amp;lt; 1e-5) and had lower ERBB2 RNA expression (p&amp;lt; 1e-5). Interestingly, molecular characterization using PAM50 demonstrated that cluster 3 was predominantly a basal-like subtype, whereas luminal-like samples dominated cluster 1 and 2, and cluster 2 had the largest composition of HER2-like samples (Table 1). Strikingly, cluster 3 stage 4 patients (n=57) had a median rwPFS that was significantly shorter (&amp;gt;12 months) than cluster 1 and 2 stage 4 patients (n=279, HR&amp;gt;1.66, p&amp;lt; 2e-2, Table 1). Conclusions: Tempus multimodal RWD reveals that HER2-low breast cancers are comprised of distinct molecular subtypes. In a preliminary analysis, a cluster of HER2-low, predominantly basal-like patients demonstrated dramatically worse outcomes than other clusters. These data further emphasize the importance of using RNA expression to fully characterize clinically relevant subpopulations. Further prospective studies are urgently needed to assess treatment response in this heterogenous emerging HER2-low distinct population. Citation Format: Talal Ahmed, Daniel Stover, Adam Hockenberry, Matthew Mackay, Halla Nimeiri, James Chen, Rotem Ben-Shachar, Massimo Cristofanilli. HER2-08 Molecular characterization of HER-2 low patients identifies basal-enriched subset with poor clinical outcomes in real-world data [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr HER2-08.

Abstract P3-05-10: Disease features, genomic profiles and outcomes of younger vs. older Chinese hormone receptor positive (HR+), HER2 negative (HER2-) metastatic breast cancer (MBC) patients

Abstract Background: Breast cancer in younger women has been characterized as a more aggressive disease with relatively poor outcomes compared to breast cancer in older women. This pattern has been attributed to a variety of clinicopathologic characteristics including the enrichment of aggressive HER2+ and TNBC subtypes in younger patients. However, similar to older breast cancer patients, the majority of young breast cancer patients are HR+/HER2-. Here we have evaluated the clinicopathologic characteristics and genomic profiles of real-world HR+/HER2- MBC patients to examine the determinants of outcome for younger vs. older patients in this subtype. Patients and Methods: This study included 65 Chinese patients presenting with HR+/HER2- MBC at the Peking University Cancer Hospital. Blood samples were collected upon metastatic disease diagnosis before treatment and profiled with the targeted 152-gene PredicineCARETM sequencing panel. Fourteen patients were &amp;lt; 40 years, 19 were 40-50 years, and 32 were &amp;gt; 50 years at the time of primary cancer diagnosis. Kaplan-Meier survival analysis was performed to analyze outcomes including overall survival (OS), disease free survival (DFS) and progression free survival (PFS) in association with clinicopathologic and genomic variables. OS was also evaluated in association with age in separate multivariate models using Cox proportional hazards regression. Results: Significant variation across age groups was observed for several clinicopathological features, including molecular subtype classification (p = 0.045), de novo Stage IV disease (p = 0.011), type of adjuvant endocrine therapy (p = 5.56E-06), resistance to adjuvant endocrine therapy (p = 0.015) and receipt of adjuvant chemotherapy (p = 0.042). Luminal B status was most frequent in patients &amp;lt; 40 years (92.31%), whereas de novo Stage IV disease was more prevalent in patients &amp;gt; 50 years (37.5%). Aromatase inhibitors (AIs) were administered as adjuvant endocrine therapy (ET) to 73.68% of women &amp;gt; 50 years vs. 16.67% and 0% of women aged 40-50 years and &amp;lt; 40 years, respectively. Resistance to adjuvant ET was observed in 23.08% of patients &amp;lt; 40 years but was not observed in the other age groups. Adjuvant chemotherapy was received by a higher proportion of patients in the 40-50 year age group (94.44%) compared to patients &amp;lt; 40 years (69.23%) and patients &amp;gt; 50 years (60%). No differences in somatic gene alteration frequencies were observed across age groups, menopausal status, germline mutation status or tumor grade. However, a higher frequency of FGFR1 alterations was observed in patients classified as Luminal B vs. A (p = 0.048). Comparison of profiles across women who received adjuvant ET revealed a higher prevalence of FGFR1 (p = 0.012), ATM (p = 0.047) and CCND2 (p = 0.04) alterations in patients treated with AIs vs. SERMS. In addition, a higher frequency of APC alterations was observed in patients with high (≥ 20%) vs. low (&amp;lt; 20%) Ki67 index (p = 0.035). At the univariate level OS was significantly associated with age (p = 0.039). OS was shortest in patients &amp;gt; 50 years, intermediate in patients &amp;lt; 40 years, and longest in patients 40-50 years. Across all patients, shorter OS was also associated with de novo Stage IV disease (p = 0.0001), Luminal B subtype (p = 0.008), adjuvant ET with AIs vs SERMS (p = 0.028) and the presence of an FGFR1 (p = 0.028) or CCND2 (p = 0.033) alteration. In multivariate analyses, OS remained significantly longer for patients aged 40-50 years after adjustment for de novo Stage IV disease, Luminal B subtype and FGFR1 status (all p &amp;lt; 0.05). Conclusions: In this group of real-world HR+/HER2- MBC breast cancer patients, younger age was not associated with worse outcomes. While current guidelines recommend treatment decisions based on tumor biology rather than age, young HR+ breast cancer patients are more likely to receive chemotherapy. Our findings support the development of biomarker-driven treatment strategies for these patients. Citation Format: Jinhao Wang, Yaxin Liu, Bin Shao, Hang Dong, Tiantian Zheng, Pan Du, Shidong Jia, Bonnie King, Jing Wang, Xiaoran Liu, Huiping Li. Disease features, genomic profiles and outcomes of younger vs. older Chinese hormone receptor positive (HR+), HER2 negative (HER2-) metastatic breast cancer (MBC) patients [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P3-05-10.

Improving Noninvasive Classification of Molecular Subtypes of Adult Gliomas With Diffusion-Weighted MR Imaging: An Externally Validated Machine Learning Algorithm.

Genetic testing for molecular markers of gliomas sometimes is unavailable because of time-consuming and expensive, even limited tumor specimens or nonsurgery cases. To train a three-class radiomic model classifying three molecular subtypes including isocitrate dehydrogenase (IDH) mutations and 1p/19q-noncodeleted (IDHmut-noncodel), IDH wild-type (IDHwt), IDH-mutant and 1p/19q-codeleted (IDHmut-codel) of adult gliomas and investigate whether radiomic features from diffusion-weighted imaging (DWI) could bring additive value. Retrospective. A total of 755 patients including 111 IDHmut-noncodel, 571 IDHwt, and 73 IDHmut-codel cases were divided into training (n=480) and internal validation set (n=275); 139 patients including 21 IDHmut-noncodel, 104 IDHwt, and 14 IDHmut-codel cases were utilized as external validation set. A 1.5T or 3.0 T/multiparametric MRI, including T1-weighted (T1), T1-weighted gadolinium contrast-enhanced (T1c), T2-weighted (T2), fluid attenuated inversion recovery (FLAIR), and DWI. The performance of multiparametric radiomic model (random-forest model) using 22 selected features from T1, T2, FLAIR, T1c images and apparent diffusion coefficient (ADC) maps, and conventional radiomic model using 20 selected features from T1, T2, FLAIR, and T1c images was assessed in internal and external validation sets by comparing probability values and actual incidence. Mann-Whitney U test, Chi-Squared test, Wilcoxon test, receiver operating curve (ROC), and area under the curve (AUC); DeLong analysis. P < 0.05 was statistically significant. The multiparametric radiomic model achieved AUC values for IDHmut-noncodel, IDHwt, and IDHmut-codel of 0.8181, 0.8524, and 0.8502 in internal validation set and 0.7571, 0.7779, and 0.7491 in external validation set, respectively. Multiparametric radiomic model showed significantly better diagnostic performance after DeLong analysis, especially in classifying IDHwt and IDHmut-noncodel subtypes. Radiomic features from DWI could bring additive value and improve the performance of conventional MRI-based radiomic model for classifying the molecular subtypes especially IDHmut-noncodel and IDHwt of adult gliomas. Stage 2.

Transcriptomic Analyses of Pretreatment Tumor Biopsy Samples, Response to Neoadjuvant Chemoradiotherapy, and Survival in Patients With Advanced Rectal Cancer

Neoadjuvant chemoradiotherapy (CRT) is the standard of care for advanced rectal cancer. Yet, estimating response to CRT remains an unmet clinical challenge. To investigate and better understand the transcriptomic factors associated with response to neoadjuvant CRT and survival in patients with advanced rectal cancer. A single-center, retrospective, case series was conducted at a comprehensive cancer center. Pretreatment biopsies from 298 patients with rectal cancer who were later treated with neoadjuvant CRT between April 1, 2004, and September 30, 2020, were analyzed by RNA sequencing. Data analysis was performed from July 1, 2021, to May 31, 2022. Chemoradiotherapy followed by total mesorectal excision or watch-and-wait management. Transcriptional subtyping was performed by consensus molecular subtype (CMS) classification. Immune cell infiltration was assessed using microenvironment cell populations-counter (MCP-counter) scores and single-sample gene set enrichment analysis (ssGSEA). Patients with surgical specimens of tumor regression grade 3 to 4 or whose care was managed by the watch-and-wait approach for more than 3 years were defined as good responders. Of the 298 patients in the study, 205 patients (68.8%) were men, and the median age was 61 (IQR, 52-67) years. Patients classified as CMS1 (6.4%) had a significantly higher rate of good response, albeit survival was comparable among the 4 subtypes. Good responders exhibited an enrichment in various immune-related pathways, as determined by ssGSEA. Microenvironment cell populations-counter scores for cytotoxic lymphocytes were significantly higher for good responders than nonresponders (median, 0.76 [IQR, 0.53-1.01] vs 0.58 [IQR, 0.43-0.83]; P < .001). Cytotoxic lymphocyte MCP-counter score was independently associated with response to CRT, as determined in the multivariable analysis (odds ratio, 3.81; 95% CI, 1.82-7.97; P < .001). Multivariable Cox proportional hazards regression analysis, including postoperative pathologic factors, revealed the cytotoxic lymphocyte MCP-counter score to be independently associated with recurrence-free survival (hazard ratio [HR], 0.38; 95% CI, 0.16-0.92; P = .03) and overall survival (HR, 0.16; 95% CI, 0.03-0.83; P = .03). In this case series of patients with rectal cancer treated with neoadjuvant CRT, the cytotoxic lymphocyte score in pretreatment biopsy samples, as computed by RNA sequencing, was associated with response to CRT and survival. This finding suggests that the cytotoxic lymphocyte score might serve as a biomarker in personalized multimodal rectal cancer treatment.

Classification of the mitochondrial ribosomal protein-associated molecular subtypes and identified a serological diagnostic biomarker in hepatocellular carcinoma.

The objective of this study was to sort out innovative molecular subtypes associated with mitochondrial ribosomal proteins (MRPs) to predict clinical therapy response and determine the presence of circulating markers in hepatocellular carcinoma (HCC) patients. Using an unsupervised clustering method, we categorized the relative molecular subtypes of MRPs in HCC patients. The prognosis, biological properties, immune checkpoint inhibitor and chemotherapy response of the patients were clarified. A signature and nomogram were developed to evaluate the prognosis. Enzyme-linked immunosorbent assay (ELISA) measured serum mitochondrial ribosomal protein L9 (MRPL9) levels in liver disease patients and normal individuals. Receiver operating characteristic (ROC) curves were conducted to calculate the diagnostic effect. The Cell Counting Kit 8 was carried out to examine cell proliferation, and flow cytometry was used to investigate the cell cycle. Transwell assay was applied to investigate the potential of cell migration and invasion. Western blot detected corresponding changes of biological markers. Participants were classified into two subtypes according to MRPs expression levels, which were characterized by different prognoses, biological features, and marked differences in response to chemotherapy and immune checkpoint inhibitors. Serum MRPL9 was significantly higher in HCC patients than in normal individuals and the benign liver disease group. ROC curve analysis showed that MRPL9 was superior to AFP and Ferritin in differentiating HCC from healthy and benign patients, or alone. Overexpressed MRPL9 could enhance aggressiveness and facilitate the G1/S progression in HCC cells. We constructed novel molecular subtypes based on MRPs expression in HCC patients, which provided valuable strategies for the prediction of prognosis and clinical personalized treatment. MRPL9 might act as a reliable circulating diagnostic biomarker and therapeutic target for HCC patients.

A developed ant colony algorithm for cancer molecular subtype classification to reveal the predictive biomarker in the renal cell carcinoma

Background: Recently, researchers have been attracted in identifying the crucial genes related to cancer, which plays important role in cancer diagnosis and treatment. However, in performing the cancer molecular subtype classification task from cancer gene expression data, it is challenging to obtain those significant genes due to the high dimensionality and high noise of data. Moreover, the existing methods always suffer from some issues such as premature convergence. Methods: To address those problems, we propose a new ant colony optimization (ACO) algorithm called DACO to classify the cancer gene expression datasets, identifying the essential genes of different diseases. In DACO, first, we propose the initial pheromone concentration based on the weight ranking vector to accelerate the convergence speed; then, a dynamic pheromone volatility factor is designed to prevent the algorithm from getting stuck in the local optimal solution; finally, the pheromone update rule in the Ant Colony System is employed to update the pheromone globally and locally. To demonstrate the performance of the proposed algorithm in classification, different existing approaches are compared with the proposed algorithm on eight high-dimensional cancer gene expression datasets. Results: The experiment results show that the proposed algorithm performs better than other effective methods in terms of classification accuracy and the number of feature sets. It can be used to address the classification problem effectively. Moreover, a renal cell carcinoma dataset is employed to reveal the biological significance of the proposed algorithm from a number of biological analyses. Conclusion: The results demonstrate that CAPS may play a crucial role in the occurrence and development of renal clear cell carcinoma.

Improved Delineation of Colorectal Cancer Molecular Subtypes and Functional Profiles with a 62-Gene Panel.

: Well-selected gene panel and representation can capture both the continuum of cancer cell states and tumor microenvironment status.

Value of genomics- and radiomics-based machine learning models in the identification of breast cancer molecular subtypes: a systematic review and meta-analysis.

In the era of precision therapy, early classification of breast cancer (BRCA) molecular subtypes has clinical significance for disease management and prognosis. We explored the accuracy of machine learning (ML) models for early classification of BRCA molecular subtypes through a systematic review of the literature currently available. We retrieved relevant studies published in PubMed, EMBASE, Cochrane, and Web of Science until 15 April 2022. A prediction model risk of bias assessment tool (PROBAST) was applied for the assessment of risk of bias of a genomics-based ML model, and the Radiomics Quality Score (RQS) was simultaneously used to evaluate the quality of this radiomics-based ML model. A random effects model was adopted to analyze the predictive accuracy of genomics-based ML and radiomics-based ML for Luminal A, Luminal B, Basal-like or triple-negative breast cancer (TNBC), and human epidermal growth factor receptor 2 (HER2). The PROSPERO of our study was prospectively registered (CRD42022333611). Of the 38 studies were selected for analysis, 14 ML models were based on gene-transcriptomic, with only 4 external validations; and 43 ML models were based on radiomics, with only 14 external validations. Meta-analysis results showed that c-statistic values of the ML based on radiomics for the identification of BRCA molecular subtypes Luminal A, Luminal B, Basal-like or TNBC, and HER2 were 0.76 [95% confidence interval (CI): 0.60-0.96], 0.78 (95% CI: 0.69-0.87), 0.89 (95% CI: 0.83-0.91), and 0.83 (95% CI: 0.81-0.86), respectively. The c-statistic values of ML based on the gene-transcriptomic analysis cohort for the identification of the previously described BRCA molecular subtypes were 0.96 (95% CI: 0.93-0.99), 0.96 (95% CI: 0.93-0.99), 0.98 (95% CI: 0.95-1.00), and 0.97 (95% CI: 0.96-0.98) respectively. Additionally, the sensitivity of the ML model based on radiomics for each molecular subtype ranged from 0.79 to 0.85, while the sensitivity of the ML model based on gene-transcriptomic was between 0.92 and 0.99. Both radiomics and gene transcriptomics produced ideal effects on BRCA molecular subtype prediction. Compared with radiomics, gene transcriptomics yielded better prediction results, but radiomics was simpler and more convenient from a clinical point of view.