Review
Biochemistry & Molecular Biology
Davide Bellini, Marika Milan, Antonella Bordin, Roberto Rizzi, Marco Rengo, Simone Vicini, Alessandro Onori, Iacopo Carbone, Elena De Falco
Summary: Radiomics is a novel advanced approach to imaging, extracting quantitative and reproducible data from radiological images using sophisticated mathematical analysis. Radiogenomics, defined as the integration of radiology and genomics, explores the relationship between specific features extracted from radiological images and genetic or molecular traits of a particular disease. Despite advancements, standardized protocols in clinical practice are still lacking.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Review
Biochemistry & Molecular Biology
Matteo Ferro, Ottavio de Cobelli, Mihai Dorin Vartolomei, Giuseppe Lucarelli, Felice Crocetto, Biagio Barone, Alessandro Sciarra, Francesco Del Giudice, Matteo Muto, Martina Maggi, Giuseppe Carrieri, Gian Maria Busetto, Ugo Falagario, Daniela Terracciano, Luigi Cormio, Gennaro Musi, Octavian Sabin Tataru
Summary: Radiomics and genomics play crucial roles in prostate cancer research, enhancing clinical value through mathematical analysis and machine learning. Validation of recent findings in large, randomized cohorts can establish the role of radiogenomics in the future.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2021)
Review
Medicine, General & Internal
Dan Zheng, Xiujing He, Jing Jing
Summary: The heavy burden and mortality of breast cancer highlight the importance of early diagnosis and treatment. Imaging detection is a key tool in clinical practice for breast cancer screening, diagnosis, and treatment evaluation. The use of AI-assisted imaging diagnosis can improve efficiency and accuracy in recognizing, segmenting, and diagnosing tumor lesions.
JOURNAL OF CLINICAL MEDICINE
(2023)
Article
Biochemistry & Molecular Biology
Qian Liu, Pingzhao Hu
Summary: In precise medicine, computational frameworks for identifying prognostic biomarkers that capture the multi-genomic and phenotypic heterogeneity of breast cancer (BC) are of great value. However, previous radiogenomic studies have suffered from data incompleteness, feature subjectivity, and low interpretability. This study proposed a novel framework for identifying radiogenomic prognostic biomarkers for BC, addressing the limitations of previous studies. The framework includes an explainable DL model for image feature extraction, a Bayesian tensor factorization for multi-genomic feature extraction, a strategy to leverage unpaired data, and mediation analysis for further interpretation. The biomarkers identified by this framework outperformed traditional baseline radiogenomic biomarkers and had guaranteed interpretability through built-in and follow-up analyses.
COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL
(2022)
Article
Oncology
Huifang Chen, Xiaosong Lan, Tao Yu, Lan Li, Sun Tang, Shuling Liu, Fujie Jiang, Lu Wang, Yao Huang, Ying Cao, Wei Wang, Xiaoxia Wang, Jiuquan Zhang
Summary: This study developed a radiogenomics model to predict axillary lymph node metastasis in breast cancer by integrating transcriptomic data and MRI data. The results showed that the radiogenomics model outperformed the genomics and radiomics models in predicting metastasis in breast cancer.
FRONTIERS IN ONCOLOGY
(2022)
Article
Biochemistry & Molecular Biology
Sarah Fischer, Nicolas Spath, Mohamed Hamed
Summary: The heterogeneity of lung tumor nodules can be studied through the analysis of phenotypic characteristics in radiological images combined with transcriptome expression levels. By constructing a radiogenomic association map, potential links between gene and miRNA expression and image phenotypes were discovered. The study also found a correlation between gene regulatory networks and the formation of texture in lung tumors. Radiogenomic approaches have the potential to identify image biomarkers for genetic variation and expand our understanding of tumor heterogeneity.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Article
Biology
Jianing Xi, Donghui Sun, Cai Chang, Shichong Zhou, Qinghua Huang
Summary: Radiogenomics analysis can infer the genomic features of tumors from their radiogenomic associations through low-cost and non-invasive screening ultrasonic images, providing connections between genomics and radiomics. Existing studies mainly focus on the relationship between ultrasonic features and popular cancer genes, but overlook the many-to-many relationships and sample associations with tumor heterogeneity. To address these challenges, we propose an omics-to-omics joint knowledge association subtensor model that discovers cross-modal modules and identifies sample subgroups. Experimental results demonstrate the jointness of discovered modules, their association with tumorigenesis contribution, and their relation to cancer-related functions. In conclusion, our proposed model can effectively facilitate radiogenomic knowledge associations and promote the construction of explainable AI cancer diagnosis.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Review
Oncology
Bogdan Badic, Florent Tixier, Catherine Cheze Le Rest, Mathieu Hatt, Dimitris Visvikis
Summary: Colorectal carcinoma is characterized by intratumoral heterogeneity that can be assessed by radiogenomics, which combines high-throughput quantitative data from medical imaging with molecular analysis. This approach is expected to advance personalized medicine but is still in its early stages with many challenges to overcome. The integration of image-derived features-radiomics and genomic profiles-genomics is a rapidly evolving field known as radiogenomics, with various studies dedicated to colorectal cancer showing potential for enhancing clinical decision-making.
Article
Oncology
Qian Liu, Pingzhao Hu
Summary: This study aimed to predict clinical characteristics of breast cancer using deep radiomic features extracted from MRI images through deep learning technology. The results showed that deep radiomic features performed better in predicting clinical characteristics and had significant associations with genomic factors.
BIOMARKER RESEARCH
(2023)
Review
Biochemistry & Molecular Biology
Matteo Ferro, Gennaro Musi, Michele Marchioni, Martina Maggi, Alessandro Veccia, Francesco Del Giudice, Biagio Barone, Felice Crocetto, Francesco Lasorsa, Alessandro Antonelli, Luigi Schips, Riccardo Autorino, Gian Maria Busetto, Daniela Terracciano, Giuseppe Lucarelli, Octavian Sabin Tataru
Summary: Renal cancer management poses challenges throughout the entire process from diagnosis to treatment and follow-up. The differentiation between benign and malignant tissues in cases of small renal masses and cystic lesions can be problematic, even with imaging or renal biopsy. Recent advances in artificial intelligence, imaging techniques, and genomics provide potential for helping clinicians with risk stratification, treatment selection, follow-up strategies, and prognosis. However, further prospective studies with larger patient cohorts are needed to validate previous results and implement these techniques into clinical practice.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Review
Medicine, Research & Experimental
Zhen Liu, Kefeng Wu, Binhua Wu, Xiaoning Tang, Huiqing Yuan, Hao Pang, Yongmei Huang, Xiao Zhu, Hui Luo, Yi Qi
Summary: Imaging genomics plays a crucial role in understanding tumor characteristics and researching tumor treatment, particularly in aspects such as tumor screening, diagnosis, and prognosis evaluation.
BIOMEDICINE & PHARMACOTHERAPY
(2021)
Article
Oncology
Jingxian Duan, Yuanshen Zhao, Qiuchang Sun, Dong Liang, Zaiyi Liu, Xin Chen, Zhi-Cheng Li
Summary: This study developed a deep learning signature (DLS) from pretreatment MRI to predict responses to neoadjuvant chemotherapy in breast cancer patients and explored the biological pathways of DLS using paired MRI and proteomic sequencing data. The DLS showed high accuracy in predicting pathologic complete response (pCR) and revealed associations with biological functions facilitating pCR, potentially guiding personalized medication.
Article
Computer Science, Information Systems
Dong Sui, Maozu Guo, Xiaoxuan Ma, Julian Baptiste, Lei Zhang
Summary: This study presents a deep learning-based radiogenomic framework that can provide more relevant features and vivid results to intuitively demonstrate the connections among medical data.
Review
Radiology, Nuclear Medicine & Medical Imaging
Ah Young Park, Bo Kyoung Seo, Mi-Ryung Han
Summary: Microvascular ultrasound (US) techniques offer high sensitivity and spatial resolution for detailed visualization of low-flow vessels and are used in breast lesion evaluation. Microvascular US imaging with contrast agents can amplify flow signals, facilitating hemodynamic evaluation of breast lesions.
KOREAN JOURNAL OF RADIOLOGY
(2021)
Article
Medicine, Research & Experimental
Ming Fan, Kailang Wang, You Zhang, Yuanyuan Ge, Zhong Lu, Lihua Li
Summary: This study aimed to identify radiogenomic signatures of cellular tumor-stroma heterogeneity (TSH) to improve breast cancer management and prognosis analysis. Cell subpopulations were estimated using gene expression data, and the relative difference in cell subpopulations between the tumor and stroma was used as a biomarker to categorize patients into good- and poor-survival groups. A radiogenomic signature-based model utilizing dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) was developed to target TSH, and its clinical significance in relation to survival outcomes was independently validated.
JOURNAL OF TRANSLATIONAL MEDICINE
(2023)
Article
Biophysics
Albert C. Yeh, Paul O'Donnell, Gary Schoch, Paul J. Martin, Chris McFarland, Jeannine S. McCune, Jason P. Cooper, Kris Doney, Mary E. D. Flowers, Mohamed L. Sorror, Frederick R. Appelbaum, Barry E. Storer, Ted Gooley, H. Joachim Deeg
Summary: In this study, long-term outcomes in MDS and AML patients receiving different conditioning regimens were evaluated, with BuFluTHY conditioning associated with lower rates of chronic GVHD and long-term mixed T-cell chimerism, but potentially higher risks of relapse and non-relapse mortality.
BONE MARROW TRANSPLANTATION
(2022)
Review
Radiology, Nuclear Medicine & Medical Imaging
Jordan D. Fuhrman, Naveena Gorre, Qiyuan Hu, Hui Li, Issam El Naqa, Maryellen L. Giger
Summary: The development of medical imaging AI for evaluating COVID-19 patients shows potential in enhancing clinical decision making, with developers utilizing explainability techniques to increase user trust and clinical translation potential.
Article
Oncology
Gabriela E. S. Felix, Rodrigo Santa Cruz Guindalini, Yonglan Zheng, Tom Walsh, Elisabeth Sveen, Taisa Manuela Machado Lopes, Juliana Cortes, Jing Zhang, Polyanna Carozo, Irlania Santos, Thais Ferreira Bonfim, Bernardo Garicochea, Maria Betania Pereira Toralles, Roberto Meyer, Eduardo Martins Netto, Kiyoko Abe-Sandes, Mary-Claire King, Ivana Lucia de Oliveira Nascimento, Olufunmilayo Olopade
Summary: The study reveals the prevalence and heterogeneous spectrum of pathogenic variants among self-reported African-descended women in Northeast Brazil, with high-risk genes including BRCA1, BRCA2, PALB2, and TP53.
BREAST CANCER RESEARCH AND TREATMENT
(2022)
Article
Multidisciplinary Sciences
Rodrigo Santa Cruz Guindalini, Danilo Vilela Viana, Joao Paulo Fumio Whitaker Kitajima, Vinicius Marques Rocha, Rossana Veronica Mendoza Lopez, Yonglan Zheng, Erika Freitas, Fabiola Paoli Mendes Monteiro, Andre Valim, David Schlesinger, Fernando Kok, Olufunmilayo Olopade, Maria Aparecida Azevedo Koike Folgueira
Summary: Genetic diversity of germline variants in breast cancer predisposition genes was explored in a study of 1663 Brazilian patients. Pathogenic/likely pathogenic variants were found in 20.1% of participants, with TP53, BRCA1, and BRCA2 being the most commonly mutated genes. The TP53 R337H variant was strongly associated with breast cancer risk, while MUTYH variants showed no association. About 46.1% of patients had variants of uncertain significance (VUS).
SCIENTIFIC REPORTS
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Xavier M. Keutgen, Hui Li, Kelvin Memeh, Julian Conn Busch, Jelani Williams, Li Lan, David Sarne, Brendan Finnerty, Peter Angelos, Thomas J. Fahey, Maryellen L. Giger
Summary: This study utilized machine learning to analyze indeterminate thyroid nodules on ultrasound, aiming to improve cancer diagnosis. The results showed that texture features extracted from grayscale ultrasound images can effectively classify indeterminate thyroid nodules.
JOURNAL OF MEDICAL IMAGING
(2022)
Article
Biochemistry & Molecular Biology
Yoo Jane Han, Jing Zhang, Ashley Hardeman, Margaret Liu, Olga Karginova, Roger Romero, Galina F. Khramtsova, Yonglan Zheng, Dezheng Huo, Olufunmilayo Olopade
Summary: Women of African ancestry have the highest mortality from triple-negative breast cancer (TNBC). The SNP rs13074711 regulates TNFSF10 expression in TNBC cells, and its genotype is associated with TNFSF10 expression in breast tumors. TNFSF10 expression is consistently lower in African Americans compared with European Americans. Loss of TNFSF10 decreases apoptosis of TNBC cells in response to type I interferons and poly(I:C). TNFSF10 plays an important role in the regulation of antiviral immune responses in TNBC.
HUMAN MOLECULAR GENETICS
(2023)
Article
Multidisciplinary Sciences
Ashley A. Hardeman, Yoo Jane Han, Tatyana A. Grushko, Jeffrey Mueller, Maria J. Gomez, Yonglan Zheng, Olufunmilayo I. Olopade
Summary: This study found that MELK gene expression is highly upregulated in aggressive forms of breast cancer, particularly in basal-like breast cancer (BLBC). The increase in MELK expression appears to be driven by gene copy number gains rather than epigenetic modifications. The findings suggest that MELK overexpression and copy number gains could be potential diagnostic and prognostic markers for identifying patients with more aggressive breast cancer.
Article
Radiology, Nuclear Medicine & Medical Imaging
Hui Li, Heather M. Whitney, Yu Ji, Alexandra Edwards, John Papaioannou, Peifang Liu, Maryellen L. Giger
Summary: This study aimed to demonstrate the impact of continuous learning on the performance of artificial intelligence in distinguishing malignant from benign lesions in breast dynamic contrast-enhanced magnetic resonance imaging. The results showed a significant positive trend in classification performance with continuous learning, and the diagnostic performance improved over time as the number of training cases increased.
JOURNAL OF MEDICAL IMAGING
(2022)
Article
Immunology
Simone A. Minnie, Olivia G. Waltner, Kathleen S. Ensbey, Nicole S. Nemychenkov, Christine R. Schmidt, Shruti S. Bhise, Samuel R. W. Legg, Gabriela Campoy, Luke D. Samson, Rachel D. Kuns, Ting Zhou, John D. Huck, Slavica Vuckovic, Danniel Zamora, Albert Yeh, Andrew Spencer, Motoko Koyama, Kate A. Markey, Steven W. Lane, Michael Boeckh, Aaron M. Ring, Scott N. Furlan, Geoffrey R. Hill
Summary: This study elucidates the mechanisms of resistance of tumors to T cell-mediated antitumor effects after alloBMT and proposes an immunotherapy approach targeting stem-like memory T cells to enhance antitumor immunity.
SCIENCE IMMUNOLOGY
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Yu Ji, Heather M. Whitney, Hui Li, Peifang Liu, Maryellen L. Giger, Xuening Zhang
Summary: This study aims to classify breast cancer tumors into molecular subtypes using radiomic features extracted from dynamic contrast-enhanced MRI scans and investigate the impact of disagreement between immunohistochemical and St Gallen standards on artificial intelligence classification. The results show that the performance of using radiomic features from dynamic contrast-enhanced MRI for classification is better.
Article
Oncology
Hui Li, Kayla Robinson, Li Lan, Natalie Baughan, Chun-Wai Chan, Matthew Embury, Gary J. Whitman, Randa El-Zein, Isabelle Bedrosian, Maryellen L. Giger
Summary: The use of a long short-term memory (LSTM) network for analyzing annual screening mammograms may accurately identify women at risk of future breast cancer. By extracting radiomic and deep-learning-based features and inputting them into LSTM recurrent networks, classifiers can determine whether a future lesion would be malignant or benign. Incorporating temporal information into radiomic analyses improves classification performance, and evaluation of both affected and contralateral breasts can inform on the future risk of breast cancer.
Article
Oncology
Frederick M. Howard, James Dolezal, Sara Kochanny, Galina Khramtsova, Jasmine Vickery, Andrew Srisuwananukorn, Anna Woodard, Nan Chen, Rita Nanda, Charles M. Perou, Olufunmilayo I. Olopade, Dezheng Huo, Alexander T. Pearson
Summary: Gene expression-based recurrence assays are recommended for guiding chemotherapy in hormone receptor-positive, HER2-negative breast cancer, but their high cost and limited availability pose challenges. This study presents a deep learning model that utilizes digital histology and clinical risk factors to predict recurrence assay results and the risk of recurrence, surpassing the performance of established clinical nomograms. The model can identify patients with excellent prognoses who may not require further genomic testing.
Article
Medicine, General & Internal
Fangyuan Zhao, Minoru Miyashita, Masaya Hattori, Toshio Yoshimatsu, Frederick Howard, Kristiyana Kaneva, Ryan Jones, Joshua S. K. Bell, Gini F. Fleming, Nora Jaskowiak, Rita Nanda, Yonglan Zheng, Dezheng Huo, Olufunmilayo I. Olopade
Summary: This study investigated racial disparities in achieving pathologic complete response (pCR) after neoadjuvant chemotherapy in breast cancer patients and identified factors contributing to these disparities. The study found that Black patients had lower odds of achieving pCR compared to White patients in the hormone receptor-negative/ERBB2+ subtype, and they were more likely to have MAPK pathway alterations, a potential mechanism of anti-ERBB2 therapy resistance. Additionally, significant differences in tumor mutational burden and somatic alterations in several genes were observed between primary and residual tumors.
Article
Oncology
Minoru Miyashita, Joshua S. K. Bell, Stephane Wenric, Ezgi Karaesmen, Brooke Rhead, Matthew Kase, Kristiyana Kaneva, Francisco M. M. de la Vega, Yonglan Zheng, Toshio F. F. Yoshimatsu, Galina Khramtsova, Fang Liu, Fangyuan Zhao, Frederick M. M. Howard, Rita Nanda, Nike Beaubier, Kevin P. P. White, Dezheng Huo, Olufunmilayo I. I. Olopade
Summary: This study compared the molecular features of HR+/HER2- breast cancer and TNBC subtypes between patients of African and European ancestries, and found significant differences in genetic mutations, gene expression, and transcriptional signatures. These findings contribute to promoting equity in personalized cancer treatment by providing more accurate treatment strategies for diverse populations.
BREAST CANCER RESEARCH
(2023)
Article
Hematology
Nahid Rashid, Elizabeth F. Krakow, Albert C. Yeh, Masumi Ueda Oshima, Lynn Onstad, Laura Connelly-Smith, Phuong Vo, Marco Mielcarek, Stephanie J. Lee
Summary: Grade III-IV acute graft-versus-host disease (aGVHD) is associated with high rates of late medical comorbidities, worse physical and mental functioning, lower survival rates, and higher nonrelapse mortality in patients who have undergone allogeneic hematopoietic cell transplantation (HCT) and survived for at least 1 year. Continued monitoring and supportive care are important for the prevention of late effects and improvement of survival in patients recovering from severe aGVHD.
TRANSPLANTATION AND CELLULAR THERAPY
(2022)