Breast Cancer Radiogenomics: Association of Enhancement Pattern at DCE MRI with Deregulation of mTOR Pathway
2020; Radiological Society of North America; Volume: 296; Issue: 2 Linguagem: Inglês
10.1148/radiol.2020201607
ISSN1527-1315
Autores Tópico(s)AI in cancer detection
ResumoHomeRadiologyVol. 296, No. 2 PreviousNext Reviews and CommentaryFree AccessEditorialBreast Cancer Radiogenomics: Association of Enhancement Pattern at DCE MRI with Deregulation of mTOR PathwayNariya Cho Nariya Cho Author AffiliationsFrom the Department of Radiology, Seoul National University Hospital, Seoul, Republic of Korea; Department of Radiology, Seoul National University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul 03080, Republic of Korea; and Institute of Radiation Medicine, Seoul National University Medical Research Center, Seoul, Republic of Korea.Address correspondence to the author (e-mail: [email protected]).Nariya Cho Published Online:May 26 2020https://doi.org/10.1148/radiol.2020201607MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In See also the article by Bismeijer et al in this issue.Dr Cho is a professor in the breast imaging section in the department of radiology at Seoul National University Hospital and Seoul National University College of Medicine. Her research interests focus on optimizing screening breast MRI and breast US. She has served as a member of Korean Society of Breast Imaging.Download as PowerPointOpen in Image Viewer Radiogenomics aims to identify imaging markers that predict risk, prognosis, and therapy response by incorporating imaging features and genetic data. By considering individual variability, this approach leads to better precision medicine for disease treatment and prevention (1). Helped by the advent of dynamic contrast material–enhanced, or DCE, breast MRI with high spatiotemporal resolution, next-generation high-spatial-resolution RNA sequencing, and advanced radiomics techniques, breast cancer radiogenomics is now the most active area of research in breast imaging.Radiomics is able to extract and select the quantitative features of medical images and draw hypotheses, which leads to better clinical decisions (2). The premise of radiomics is that genetic alterations in tumor biology regulate radiologic phenotypes. Genomics is a field focusing on the structure, function, evolution, mapping, and editing of entire genomes and their combined influences (1). In the last decade, the main genomics outcomes investigated in breast cancer radiogenomics are molecular subtypes and multigene assays. As the formal classification of molecular subtypes based on genomic analysis is costly and time consuming, the St Gallen International Expert Consensus panel provided surrogate molecular subtypes based on estrogen receptor, progesterone receptor, human epidermal growth factor receptor 2 (HER2), Ki-67, and multigene assay recurrence scores (3). This simplified surrogate classification (luminal A, luminal B, HER2, and triple-negative cancer) used for genetic outcomes in breast cancer radiogenomics is easy to obtain from routine immunohistochemistry results—resulting in its wide use. But there are large discordant rates among the formal and surrogate classifications, ranging from 41% to 100% (3). Multigene expression–based assays such as Oncotype DX, MammaPrint, PAM50/Prosigna, and EndoPredict are also ideal for genetic outcomes because they are commercially available and validated to provide prognostic or predictive information for clinical outcomes (3).Apart from these widely available tools for genetic outcomes, individual institutional gene expression profiling is also applied for breast cancer radiogenomics. Back in 2012, Yamamoto and colleagues (4) first explored the association among MRI features and gene expression profiles from 10 patients and found 21 MRI features correlated with 71% of filtered gene elements. They divided the 10 tumors into heterogeneous and homogeneous enhancement groups by using hierarchical clustering and found an association between heterogeneous enhancement and immune-related genes. In a later study, the same group used next-generation RNA sequencing and automated imaging segmentation software. In breast tumors from 70 patients, they found an association between enhancing rim fraction score and long noncoding RNA expression related to metastasis-free survival (5).Another group integrated molecular data from the Cancer Imaging Archive and MRI data from the Cancer Genome Atlas for 91 breast tumors. That group reported that tumor size, shape, margin, and kinetics at MRI were associated with the transcriptional activities of the various genetic pathways (6). Radiogenomics research based on complete RNA sequencing is scarce because it requires extensive genetic testing and there is no clinical implication, contrary to the molecular subtypes or multigene expression-based assays such as Oncotype DX, MammaPrint, PAM50/Prosigna, and EndoPredict. But radiogenomics research based on complete RNA sequencing might be paramount to broadening our insights into the linkage among genomic biology and imaging traits.In this issue of Radiology, Bismeijer and colleagues (7) report breast cancer radiogenomics research by linking MRI phenotypes and genomics using complete RNA sequencing from 295 patients, which is a relatively large sample size. They used tumor RNA collected from surgical specimens for sequencing. To minimize interobserver variability, they used 21 computer-generated MRI features. From those 21 features, they extracted seven MRI factors through factor analysis: tumor size, shape, initial enhancement, late enhancement, smoothness of enhancement, sharpness, and sharpness variation. This extraction removed multicollinearity among MRI features and achieved the fewest number of imaging variables to explain the largest number of genetic features. The authors then performed a pathway analysis of the genes to identify the associations among biologic processes and MRI factors. After regressing the MRI factors on the expression of all genes, the authors determined MRI factor-gene pairs and quantified the strength of the association and enrichment of the gene set for each given pair.As a result, they found that larger and more irregular tumors showed increased expression levels of cell cycle and DNA damage checkpoint genes (false discovery rate <0.25; normalized enrichment score [NES], 2.15). The false discovery rate is the expected proportion of false-positive findings in multiphenotype studies, often used in genomic analyses. The NES is the effect size of the gene set enrichment. The larger the NES is, the stronger the associations. Second, increased smoothness of enhancement, smaller tumor size, and more irregular tumor shape were associated with the expression of genes related to the extracellular matrix (false discovery rate <0.25; NES, 2.25). Last, low initial enhancement, increased smoothness of enhancement, and low sharpness of the tumor boundary were associated with the expression of proteins that are part of the ribosome, a target of anticancer drugs (false discovery rate <0.25; NES, 1.95).This study by Bismeijer et al elucidates plausible associations among gene expression profiling and MRI phenotypes. Their findings of an association between size phenotypes or proliferation and multigene assay results are consistent with the findings of a prior 2016 Radiology study (8). But unlike that study, Bismeijer and colleagues (7) performed a complete transcriptome analysis and used a subset of genes included in the multigene assays. They also showed that increased smoothness of enhancement, smaller tumor size, and more irregular tumor shape were associated with the expression of genes related to the extracellular matrix and collagen production. Their results are in line with a previous study (9) reporting an association between spiculated tumors with intense desmoplasia and persistent enhancement (compared with nonspiculated tumors) and another recent study (10) suggesting stromal types and stromal programmed death-ligand 1 expression status of breast cancer determines clinical outcomes.The most promising result of this study is that initial enhancement, smoothness of enhancement, and sharpness of the tumor boundary were associated with the expression of ribosomal proteins required for ribogenesis regulated by the mammalian target of rapamycin, or mTOR, pathway. Meanwhile, the development of anticancer drugs to block the mTOR pathway is underway in the field of medical oncology. Given the results from this study, in the future, the MRI features of enhancement kinetics and tumor boundaries might identify patients who are candidates for drugs targeting the mTOR pathway.A limitation of this study was the application of hypothesis-driven radiogenomics, which uses a limited number of imaging phenotypes and specific genetic alterations, thereby reducing the chances of identifying unprecedented insights for complex tumor biology (although the approach is efficient, unlike hierarchical clustering). Also, as there is substantial intrainstitutional or interinstitutional heterogeneity from various hardware or scanning protocols, cross validation or external validation using a larger cohort is necessary. In future radiogenomics studies, image standardization or normalization is essential. Furthermore, contrary to genetic sequencing, as MRI phenotypes can be measured repeatedly and three-dimensionally, further studies to correlate the topographic sequencing of whole tumors with MRI phenotypes to reflect intratumoral heterogeneity might be valuable for evaluating the genetic evolution of tumors following chemotherapy.In conclusion, the application of radiogenomics in breast cancer is accelerating. Bismeijer and colleagues (7) found strong associations between the gene expression profiling of biologic processes (proliferation, extracellular matrix, and protein synthesis) and MRI factors (tumor size, smoothness of enhancement, tumor shape, and sharpness of tumor boundary). The association between enhancement characteristics or sharpness of the tumor boundary and ribosomal proteins suggests the possibility of noninvasive imaging markers for targeted therapy related to the mTOR pathway. Although further validation is necessary, this study adds to the evidence that MRI phenotypes can reflect underlying gene expression by using RNA sequencing.Disclosures of Conflicts of Interest: N.C. disclosed no relevant relationships.References1. Pinker K, Chin J, Melsaether AN, Morris EA, Moy L. Precision medicine and radiogenomics in breast cancer: new approaches toward diagnosis and treatment. Radiology 2018;287(3):732–747. Link, Google Scholar2. Gillies RJ, Kinahan PE, Hricak H. Radiomics: images are more than pictures, they are data. Radiology 2016;278(2):563–577. Link, Google Scholar3. Grimm LJ, Mazurowski MA. Breast cancer radiogenomics: current status and future directions. Acad Radiol 2020;27(1):39–46. Crossref, Medline, Google Scholar4. Yamamoto S, Maki DD, Korn RL, Kuo MD. Radiogenomic analysis of breast cancer using MRI: a preliminary study to define the landscape. AJR Am J Roentgenol 2012;199(3):654–663. Crossref, Medline, Google Scholar5. Yamamoto S, Han W, Kim Y, et al. Breast cancer: radiogenomic biomarker reveals associations among dynamic contrast-enhanced MR imaging, long noncoding RNA, and metastasis. Radiology 2015;275(2):384–392. Link, Google Scholar6. Zhu Y, Li H, Guo W, et al. Deciphering genomic underpinnings of quantitative MRI-based radiomic phenotypes of invasive breast carcinoma. Sci Rep 2015;5(1):17787. Crossref, Medline, Google Scholar7. Bismeijer T, van der Velden BHM, Canisius S, et al. Radiogenomic analysis of breast cancer by linking MRI phenotypes with tumor gene expression. Radiology 2020;296:277–287. Link, Google Scholar8. Li H, Zhu Y, Burnside ES, et al. MR imaging radiomics signatures for predicting the risk of breast cancer recurrence as given by research versions of MammaPrint, Oncotype DX, and PAM50 gene assays. Radiology 2016;281(2):382–391. Link, Google Scholar9. Gokalp G, Topal U, Yildirim N, Tolunay S. Malignant spiculated breast masses: dynamic contrast enhanced MR (DCE-MR) imaging enhancement characteristics and histopathological correlation. Eur J Radiol 2012;81(2):203–208. Crossref, Medline, Google Scholar10. Zhai Q, Fan J, Lin Q, et al. Tumor stromal type is associated with stromal PD-L1 expression and predicts outcomes in breast cancer. PLoS One 2019;14(10):e0223325. Crossref, Medline, Google ScholarArticle HistoryReceived: Apr 14 2020Revision requested: Apr 22 2020Revision received: Apr 23 2020Accepted: Apr 28 2020Published online: May 26 2020Published in print: Aug 2020 FiguresReferencesRelatedDetailsCited ByThe value, diagnostic efficacy and clinical significance of functional magnetic resonance imaging in evaluating the efficacy of neoadjuvant chemotherapy in patients with triple negative breast cancerXiaopingHe, ZongshengWang, YingZhou, YongliFeng2023 | Frontiers in Oncology, Vol. 13A Focus on the Synergy of Radiomics and RNA Sequencing in Breast CancerDavideBellini, MarikaMilan, AntonellaBordin, RobertoRizzi, MarcoRengo, SimoneVicini, AlessandroOnori, IacopoCarbone, ElenaDe Falco2023 | International Journal of Molecular Sciences, Vol. 24, No. 8Application of DCE-MRI radiomics signature analysis in differentiating molecular subtypes of luminal and non-luminal breast cancerTingHuang, BingFan, YingyingQiu, RuiZhang, XiaolianWang, ChaoxiongWang, HuashanLin, TingYan, WentaoDong2023 | Frontiers in Medicine, Vol. 10Contemporary Medical ImagingMártonKolossváry, PálMaurovich-Horvat2022Radiogenomics analysis reveals the associations of dynamic contrast-enhanced–MRI features with gene expression characteristics, PAM50 subtypes, and prognosis of breast cancerWenlongMing, YanhuiZhu, YunfeiBai, WanjunGu, FuyuLi, ZixiHu, TiansongXia, ZuoleiDai, XiafeiYu, HuameiLi, YuGu, ShaoxunYuan, RongxinZhang, HaitaoLi, WenyongZhu, JianingDing, XiaoSun, YunLiu, HongdeLiu, XiaoanLiu2022 | Frontiers in Oncology, Vol. 12Identifying Associations between DCE-MRI Radiomic Features and Expression Heterogeneity of Hallmark Pathways in Breast Cancer: A Multi-Center Radiogenomic StudyWenlongMing, YanhuiZhu, FuyuLi, YunfeiBai, WanjunGu, YunLiu, XiaoSun, XiaoanLiu, HongdeLiu2022 | Genes, Vol. 14, No. 1A Comprehensive Review on Radiomics and Deep Learning for Nasopharyngeal Carcinoma ImagingSongLi, Yu-QinDeng, Zhi-LingZhu, Hong-LiHua, Ze-ZhangTao2021 | Diagnostics, Vol. 11, No. 9Accompanying This ArticleRadiogenomic Analysis of Breast Cancer by Linking MRI Phenotypes with Tumor Gene ExpressionMay 26 2020RadiologyRecommended Articles Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and TreatmentRadiology2018Volume: 287Issue: 3pp. 732-747Impact of Molecular Subtype Definitions on AI Classification of Breast Cancer at MRIRadiology2023Volume: 307Issue: 1Heterogeneous Enhancement Patterns of Tumor-adjacent Parenchyma at MR Imaging Are Associated with Dysregulated Signaling Pathways and Poor Survival in Breast CancerRadiology2017Volume: 285Issue: 2pp. 401-413Radiogenomic Analysis of Breast Cancer by Linking MRI Phenotypes with Tumor Gene ExpressionRadiology2020Volume: 296Issue: 2pp. 277-287Intratumoral Spatial Heterogeneity at Perfusion MR Imaging Predicts Recurrence-free Survival in Locally Advanced Breast Cancer Treated with Neoadjuvant ChemotherapyRadiology2018Volume: 288Issue: 1pp. 26-35See More RSNA Education Exhibits Breast Cancer Imaging and Risk Profiles in Women with Moderate Risk Genetic Mutations: A Case-Based Review  Digital Posters2019The New Era for Breast Cancer Screening: Abbreviated Breast Magnetic Resonance ImagingDigital Posters2019Risk-y Business: ATMs to CHEKs: Understanding Risk of Breast Cancer in Patients with Genetic MutationsDigital Posters2019 RSNA Case Collection Malignancy on abbreviated screening breast MRIRSNA Case Collection2020Solitary breast plasmacytomaRSNA Case Collection2021Granular Cell Tumor of the BreastRSNA Case Collection2021 Vol. 296, No. 2 Metrics Altmetric Score PDF download
Referência(s)