{"paper_id":"77a26371-f3ec-4b7f-b094-390c63b2582e","body_text":"Value of dynamic enhanced magnetic resonance image-based model in predicting low expression of HER-2 in breast cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Value of dynamic enhanced magnetic resonance image-based model in predicting low expression of HER-2 in breast cancer Lu Zheng, Chenyu Sun, Muzi Meng, Eric Chen, Tong Tang, Xiao Chen, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3151750/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective This study aimed to evaluate the feasibility of evaluating early low expression of HER-2 in patients with breast cancer by applying Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) based imaging features, which could potentially optimize treatment for patients. Method Clinical and pathology data of 294 female patients with invasive ductal carcinoma confirmed by puncture or surgical pathology were collected. Regions of interest (ROI) were mapped. Features were then extracted from the original Magnetic Resonance Imaging (MRI) image data. Relevant features were screened out by Mann-Whitney U test. Cross-validated LASSO regression was used for feature selection. Inner and outer 10-fold cross-validation (CV) models were used. The inner 10-fold CV was used to select the best model during the Linear SVC modeling in training set, and an outer 10-fold CV was used to validate the efficiency in validation set. Model performance was evaluated by using receiver operator curve (ROC) analysis. The average accuracy, sensitivity, and specificity were calculated. Results After model selection using the inner 10-fold CV in Linear SVC modeling and validation using the outer CV, the average accuracy, sensitivity, and specificity of the validation set were 79.6%, 73.7%, and 85.6%, respectively. The average area under curve (AUC) of ROC analysis was 0.87. The diagnostic efficiency of the replacement dataset after 1000 permutation tests was compared with the original dataset, and the average accuracy, sensitivity, and specificity were all less than 0.05. The differences were all statistically significant. The model established after cross-validation could classify patients as HER2 low expression or HER2 positive. The classification efficiency of the model was higher than the chance level. Conclusion DCE-MRI imaging model can help predict the low expression of HER2 receptor in breast cancer with a high predictive efficiency, which can provide a new method for clinical diagnosis of non-invasive HER2 status. Breast cancer low expression of human epidermal growth factor receptor 2 Radiomics DCE-MRI Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Since the mid20th century, the incidence of breast cancer in women has steadily increased by approximately 0.5% per year. According to Global Cancer Statistics 2020 [ 1 ], breast cancer is the \"number one killer\" of women worldwide. As of 2020, breast cancer is estimated to account for 2.26 million new instances of cancer worldwide and 10% of all malignancies in women [ 2 ]. Breast cancer is classified into four molecular subtypes based on the expression of human epithelial growth factor receptor 2 (HER2), hormone receptor (HR), and Ki-67. These subtypes are triple negative breast cancer (TNBC), HER2-positive breast cancer, Luminal A, and Luminal B. Although HER2-positive breast cancer has a poor prognosis and a high incidence of recurrence, the prognosis has improved since the advent of incorporating targeted medications such as trastuzumab, pertuzumab, and lapatinib into clinical practice. As HER2 detection technology improves and new targeted medications are deployed, the advantages of innovative antibody-drug conjugates (ADCs) like T-DXd continue to expand. Recent clinical trial results indicate that T-DXd induces a gradual treatment response in both the HER2-positive population and HER2-low-positive breast cancer [ 3 ]. Therefore, a precise method of evaluating HER2 status is crucial for breast cancer treatment. The immunohistochemical measurement of HER2 protein expression levels in tumor tissues is used to assess HER2-status. HER2-positive is characterized as IHC3 + or 2 + and positive on fluorescence in situ hybridization (FISH). Low HER2 expression is characterized by a negative FISH test and IHC1 + or 2+. Needle core biopsy specimens may not accurately assess the overall status of the tumor due to the heterogeneity of the tumor and the limited number of tissue samples. According to recent research [ 4 ], the accuracy rate for detecting receptor status in puncture biopsy specimens ranges from 64.2–98.8%. There is limited consistency between results from various laboratories since the diagnostic level of the pathologist dictates how accurate HER2 status testing is [ 5 ]. As a result, individuals with early-stage breast cancer urgently need a method to determine their HER-2 status that is both affordable and non-invasive. Gillies et al. [ 6 ] initially conceived of the idea of Imageomics in 2010, which was later expanded upon by Lambin et al[ 7 ]. Imageomics facilitates accurate tumor diagnosis by extracting high-throughput quantitative characteristics from pictures in order to create high-dimensional datasets, which are then used to mine tumor-related parameters such as molecular type, treatment effectiveness, and clinical outcomes. Breast cancer can only be visualized with conventional imaging, which does not reveal information on the genetic and molecular basis of the disease. Additionally, there are differences in how each radiologist subjectively interprets pictures. Imagingomics can extract high-throughput image characteristics from tumor images that are not visible to the human eye in order to more precisely define the heterogeneity of tumors. This transforms visual characteristics into numerical data. The main applications of imaging omics in breast cancer include diagnosis, assessment of treatment effectiveness, distinction of molecular subtype, and prognosis assessment. When combined with genetic, immunohistochemical, and pathology data, imagingomics has the potential to enhance clinical decision-making and improve patient outcomes. Using improved Magnetic Resonance Imaging (MRI) imaging to determine HER2 status is not widely reported at the present time, particularly in China. The purpose of this study is to determine whether improved MRI imaging could reliably predict HER2 low expression. A reliable HER2 status prediction model could be created to optimize treatment approach. Methods Data Collection From January 2019 to October 2022, clinical and pathology information on patients with invasive breast cancer that was verified by the pathology findings on Biopsy or surgical pathology confirmed puncture and an MRI scan before the puncture was gathered in our institution. Inclusion criteria included: (1) Female patients over the age of 18; (2) Patients who agreed to undergo breast MRI examination; (3) No other treatment was administered prior to the examination; (4) Adipose-suppressed T2WI, dynamic contrast-enhanced resonance imaging (DCE-MRI), and apparent diffusion coefficient, ADC, images were obtained by breast MRI prior to puncture or surgery; (5) All patients underwent air-core needle puncture or operation to identify non-special types of invasive ductal carcinoma; and (6) Complete clinical and pathological information was available. Exclusion criteria included: (1) Male patients; (2) patients with inflammatory, bilateral, lactation-related, and pregnancy-related breast cancer; (3) Any treatment prior to MRI scan, including needle biopsy, surgery, radiotherapy, chemotherapy, and endocrine therapy; (4) Poor image quality (such as artifacts); (5) patients whose tumor boundaries were not obvious; (6) Patients with claustrophobia. For patients with multiple tumors in the same breast, the largest tumor was selected for analysis. The study was approved by the Institutional Research Ethics Committee of The Second Affiliated Hospital of Anhui Medical University (Hefei, China), and written informed consent was obtained from all patients. This study was carried out in accordance with the Declaration of Helsinki. Collection of breast MRI image data Breast MRI images of all patients were obtained by a Siemens 3.0T MRI scanner (Verio, Siemens Healthcare, Erlangen, Germany) in a prone position using a specialized 8-channel breast coil. All patients underwent the same breast MRI protocol, which consisted of the following sequences: T1 and T2 weighted axial images, diffusion weighted images, and dynamic contrast-enhanced images with 6 phases (including a pre-injection non-contrast phase). The imaging parameters of the dynamic enhancement sequence were: TR (repetition time) 4.34 ms, TE (time to echo) 1.52 ms, averages 1, concatenations 1, FoV (field of view) read 360 mm, FoV phase 93.8%, and slice thickness 1.0 mm. Determination of HER2 receptor expression status This study adhered to the clinical practice recommendations for HER2 detection in breast cancer issued by the American Society of Clinical Oncology (ASCO) and the Association of American Pathologists (CAP). HER2 receptor status was assessed by immunohistochemistry (IHC) or in situ hybridization (FISH) in all patients in our study. IHC results of 0/1 + are regarded as negative for HER2. HER2 status is deemed positive when the IHC result is 3+. Further FISH testing was done to determine the HER2 receptor status if the IHC result was 2+. Delineation of the receiver of interest (ROI) Breast MRI images of each patient were evaluated on a PACS workstation. Phase 2 of contrast-enhanced MRI (first phase after contrast agent injection) was selected as the contour image. At this stage, the MRI scans of every patient were exported and saved as DICOM files for upload to ITK-SNAP ( http://www.itksnap.org/ ) [ 8 ]. The receiver of interest (ROI) was manually drawn on the MRI images by two skilled radiologists (with five- and ten-years’ experience of breast cancer diagnosis) on each slice and exported as a ROI mask image. All patients' DICOM images as well as the images used to draw ROI masks were saved in batches. The characteristics of the ROI mask were extracted using pyradiomics. The extracted image features include original features, features transformed by wavelet, and Log transformation features (including shape, Gray-level run-length matrix (GLRLM), Gray-level cooccurrence matrix (GLCM), Neighborhood gray-tone difference matrix (NGTDM), Gray-level difference matrix (GLDM), and Gray-level size zone matrix (GLSZM)). The intraobserver and interobserver agreement Intraclass correlation coefficient (ICC) was used to test the reliability and reproducibility of intra-observer and inter-observer data features [ 9 ]. Twenty patients’ MRI data were selected randomly for ICC validation analysis. Two experienced diagnostic radiologists (Physician A and Physician B) independently sketched the ROI. The extracted features were used to measure inter-observer ICC. Two weeks later, the ROI was sketched again by physician A without referring to the previous sketch results. The resulting features were used to measure interobserver ICC. Features satisfying interobserver and interobserver ICC ≥ 0.75 were preserved. Statistical analysis The statistical analysis was completed with the aid of Python 3.9.0 and R 3.5.1 (The R Project for Statistical Computing (r-project.org)) software. Data reproducibility within and between observers was assessed using ICC. The normal distribution-conforming measurement results were reported as mean standard deviation (x ± s). A pipeline from Sandra Vieira [ 10 ] was used during the model building. The process of analytical and model construction was as follows: (1) The Mann-Whitney U test was used to check for correlation characteristics, and features with p value < 0.05 were obtained. (2) The dimensions of the omics characteristics were successively screened and reduced using the least absolute shrinkage and selection operator (LASSO) approach according to the retention characteristic of the ideal cut-off point. Then the data was randomly separated into training sets and test sets with a ratio of 7:3. Data was normalized in each separate set. An inner 10-fold cross-validation (CV) and outer 10-fold CV were used for model construction. (3) The training sets were modeled using Linear SVC (Support Vector Machines (SVM)). The optimum linear kernel \"C\" value was chosen using grid search and the 10-fold CV approach to establish the best SVM model. An outer 10-fold CV was used to validate the efficacy of the SVM model in the test set and diagnostic value was shown with Receiver Operator Characteristic (ROC) curve. In parallel CV repetitions of 10, ten optimum models were produced. The average iteration's performance measurements served as a representation of overall performance. (4) Permutation test was used to determine if the model performance assessment index had statistical significance. After 1000 replacements, the evaluation index of the data set's performance was determined. The P-value was calculated by dividing the number of permutations by the number of times its performance was better than that of the SVM model. P < 0.05 was considered statistically significant.Details are shown in Fig. 1 . Results General Information A total of 294 cases of non-specific invasive breast cancer were collected in this study. There were 40 cases (13.6%) with zero expression of HER. There were 141 patients with low HER2 expression (48.0%), including 89 patients with HER21 + type (30.0%), 52 patients with HER22 + and FISH- (18.0%), and 113 patients with HER2 positive HER23 +/ HER22 + and FISH- (38.4%). Interobserver and Intraobserver Agreement The ICC of the interobserver of the two independent readers ranged from − 0.25 to 0.99, and the ICC of the intraobserver of the same reader ranged from − 0.04 to 0.99. A total of 858 features were retained, and all showed high interobserver and intraobserver ICCs with ICCs ≥ 0.75. Thus, features extracted by reader A were used for further analysis. Construction of model 135 features were retained after relevant features were eliminated using the Mann-Whitney U test. 24 eigenvalues were kept after the LASSOCV regression. During the LinearSVC model construction, ten best models were selected and validated in test set. The average accuracy, sensitivity, and specificity of the test set were 79.6%, 73.7%, and 85.6%, respectively, following 10-fold CV Linear SVC modeling (Table 1 ). The ROC analysis showed a mean area under the curve (AUC) of 0.87 of the models. Figure 2 depicts LASSO regression of feature selection. Figure 3 depicts the 10-fold CV ROC curve for the specific value index of the test set. Figure 4 depicts the corresponding features' regression coefficients. Table 1 diagnostic efficiencies of the LinearSVC model in 10-fold CV specificity sensitivity accuracy ppv npv auc(95%CI) fold_1 0.94 0.83 0.90 0.91 0.89 0.90(0.78-1.0) fold_2 1.00 0.75 0.90 1.00 0.86 0.89(0.74-1.00) fold_3 0.94 0.83 0.90 0.91 0.89 0.92(0.82-1.00) fold_4 0.89 0.67 0.80 0.80 0.80 0.75(0.54–0.95) fold_5 0.78 0.82 0.79 0.69 0.88 0.85(0.70–0.99) fold_6 0.94 0.73 0.86 0.89 0.85 0.88(0.76-1.00) fold_7 0.84 0.70 0.79 0.70 0.84 0.82(0.66–0.97) fold_8 0.83 1.00 0.90 0.79 1.00 0.97(0.93-1.00) fold_9 0.83 1.00 0.90 0.79 1.00 0.97(0.92-1.00) fold_10 0.78 0.91 0.83 0.71 0.93 0.86(0.73–0.99) Diagnostic performance of the permutation data set After 1000 permutation tests, the replacement data set's average accuracy, sensitivity, and specificity of the diagnostic effectiveness were less than 0.05 in comparison to the original data set. HER2 low expression and HER2 positive in breast cancer patients may be classified using the model developed following cross-validation. The prediction model's classification effectiveness was greater than the opportunity level. Discussion As of now, 45–55% of patients with breast cancer have low articulation of HER2. The treatment status of HER2-low-positive breast cancer patients has changed as a result of the development of new targeted drugs [ 11 ]. In HER2-low-positive breast cancer, anti-antibody drug conjugates (ADCs) like Trastuzumab deruxtecan (T-DXd, DS-8201) have shown promising therapeutic results. The identification of HER2-low-positive breast cancer has become a hot topic. Imaging was used to assist in the determination of molecular typing because biopsy is invasive and may not represent the characteristics of the entire tumor. Customary imaging tests depend on the demonstrative level of the specialist, which is subjective and not reproducible. Imaging omics has recently been used by some researchers to combine various algorithms, examination techniques, and clinical data. To better identify breast cancer's molecular typing, multimodal omics were developed. Ultrasound, MRI, and mammography (MG) are the primary screening methods for breast cancer this time. According to previous research [ 12 ], mammography has a low accuracy in predicting HER2 receptor expression status (95% CI of AUC is approximately 0.55 to 0.62, and sensitivity is approximately 0.11 to 0.50). The most common cause of poor accuracy in mammography is compression of the breast tissue [ 13 ]. Because breast tissue and lesions overlap, mammography cannot effectively detect lesions. In particular, tiny lesions are frequently missed in the background of dense mammary glands. In the diagnosis of breast cancer, MRI has medium specificity and high sensitivity due to its high spatial and tissue resolution. From a single sequence to a multi-sequence, multi-modal sequence, MRI sequences are included. The construction of the model shifts from a single model to a joint model. From the tumor itself to the surrounding tissue and even the lymph nodes, the features' region of interest (ROI) was taken. The study's goals ranged from predicting cancer to determining the types of molecules. Numerous omics studies have emerged in recent years as a result of image omics and other omics. To further enhance the prediction effect of molecular classification and optimize the performance of various classifiers, Li et al. proposed a multi-model-based recursive feature elimination strategy [ 14 ]. The Luminal A, Luminal B, HER-2, and Basal-like subtypes had an identification accuracy of 0.91, 0.89, 0.83, and 0.87, respectively. Dynamic enhanced magnetic resonance imaging (DCE-MRI) is an upgraded technique for detecting breast cancer. DCE-MRI showed the highest detection accuracy when using multi-phase time series imaging to reflect the blood perfusion of breast cancer lesions. DCE-MRI can characterize soft tissue with high resolution to give a more precise portrayal of the morphological and structural aspects of breast cancer lesions. Moreover, patient exposure to radiation during the examination is greatly reduced by the DCE-X-ray component without ionizing radiation. DCE-MRI is essential in the field of breast cancer imaging [ 15 ]. In this study, the application of DCE-MRI to assess HER2 status was further explored. After 10-fold cross-validation Linear SVC modeling, the average accuracy of the validation set was 79.6%, the average sensitivity was 73.7%, the average specificity was 85.6%, and the average AUC of ROC analysis was 0.87. After 1000 substitution tests, the replacement data set's average accuracy, sensitivity, and specificity were less than 0.05 compared to the original data set. This shows that the cross-validated model can categorize patients with breast cancer as having high or low expression of the HER2 gene, and that its classification accuracy is more than the opportunity level. The results of this study present a novel technique for non-invasive diagnosis of breast cancer with low HER2 expression. There are still some limitations to this study. First of all, this research, which was a single-center evaluation, only included 294 cases. Only a small number of patients lacked HER2 expression, and breast cancer patients without or with little HER2 expression were not categorized. In the future, the sample size could be increased to confirm the model's capacity for forecasting. Secondly, only patients with invasive ductal carcinoma of the breast were included in this research. The model requires more patients with other types of breast cancer in order to obtain the highest level of predicting power. Thirdly, this study employed manual layer-by-layer mapping of lesions, which is not only arduous and time-consuming, but also changes depending on the reader. Reproductions of various radiological aspects are made. Fourthly, this study’s focus was primarily on predicting HER2 status. Further studies can analyze the prediction of the Ki-67 proliferation index [ 16 ], an optimization model of clinical coupled genetic data, and imaging characteristics around tumors [ 17 ] in order to further explore the importance of imaging omics in predicting the prognosis of breast cancer patients. Finally, because this research was retrospective in nature, more prospective investigations will be required to confirm these prognostic models. In conclusion, our research suggested that HER2 low expression status in breast cancer patients can be predicted using radiological features extracted from pre-treatment MRI images. To better understand the relationship between imaging omics features and HER-2 status, more studies involving multiple centers, larger sample sizes, and various imaging modalities are required. Declarations Acknowledgements None. Author Disclosure Statement Authors declare no conflicts of interest. The authors declare that they have no competing interests. Data Availability Statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Corresponding author should be contacted if someone wants to request the data from this study. Funding Sources This research is supported by National Natural Science Foundation of China (No. 82200225), Research Project of Young Scholars of Breast Cancer in China (No.320.6750.2021-10-25), Research Fund of Anhui Medical University (No. 2018xkj038) and Key Project of Higher Education Humanities and Social Sciences of Anhui Province (No.SK2021A0167). Patient consent for publication Not applicable. Ethics approval and consent to participate The study was approved by the Institutional Research Ethics Committee of The Second Affiliated Hospital of Anhui Medical University (Hefei, China), and written informed consent was obtained from all patients. This study was carried out in accordance with the Declaration of Helsinki. Author Contributions Statement LZ and RZ designed the experiment. CS and ZW collected data. MM and EC analyzed the data. LZ, TT and XC wrote the main manuscript text. All the authors reviewed the manuscript. References SungH FerlayJ. SiegelRL, etal.Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortalit world wide for 36 cancers in 185 countries [J]. CA Cancer JClin. 2021;71(3):209–49. Siegel RL, Miller KD, Wagle NS, Jemal A, Cancer statistics. 2023; CA Cancer J Clin. 2023;1–32.DOI: 10.3322/caac.21763 . Modi S, Park H, Murthy RK et al. Antitumor activity and safety of trastuzumab deruxtecan in patients with HER2-low-expressing advanced breast cancer: results from a phase â… b study [J].J Clin Oncol 2020,38(17):1887–96. DOI: 10.1200/JCO.19.02318 . Lorgis V, Algros MP, Villanueva C, et al. Discordance in early breast cancer for tumour grade, estrogen receptor, progesteron receptors and human epidermal receptor-2 status between core needle biopsy and surgical excisional primary tumour.[J]. Breast. 2011;20(3):284–7. Orlando L, Viale G, Bria E, et al. Discordance in pathology report after central pathology review: Implications for breast cancer adjuvant treatment[J]. Breast. 2016;30:151–5. Gillies RJ, Anderson AR, Gatenby RA, et al. The biology underlying molecular imaging in oncology: from genome to anatome and back again [. J] Clin Radiol. 2010;65:517–21. 10.1016/j.crad.2010.04.005 . Lambin P, Rios -Velazquez E, Leijenaar R et al. Radiomics:extracting more information from medical images using advanced feature analysis[J].EurJCancer,2012,48:441–6DOI: 10.1016/j.ejca.2011.11.036 . Yushkevich PA, Piven J, Hazlett HC, et al. User-guided3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage. 2006;31(3):1116–28. Xu X, Zhu H, Li R et al. Whole-liver histogram and texture analysis on T1 maps improves the risk stratification of advanced fibrosis inNAFLD[J].EurRadiol.2021.31(3):1748–59. VieraS,Garcia-Dias R. Pinaya WHL.A step-by-step tutorial on how to build a mechine learning model[J].Mach Learn,2020:343–70. Modi S, Park H, Murthy RK, et al. Antitumor activity and safety of trastuzumab deruxtecan in patients with HER2 -low -expressing advanced breast cancer: results from a phase Ib study [J]. J Clin Oncol. 2020;38(17):1887–96. Son J, Si EL, Kim EK, et al. Prediction of breast cancer molecular subtypes using radiomics signatures of synthetic mammography from digital breast tomosynthesis[J]. Sci Rep. 2020;10:21566. Dullum JR, Lewis EC, Mayer JA. Rates and Correlates of Discomfort Associated with Mammography1[J]. Radiology. 2000;214:547–52. Li W, Yu K, Feng CL, et al. Molecular subtypes recognition of breast cancer in dynamic contrast-enhanced breast magnetic resonance imaging phenotypes from radiomics data. Comput Math Methods Med. 2019;2019:6978650. Marino MA, Helbich T, Baltzer P, et al. Multiparametric MRI of the breast: A review [J]. J Magn Reson Imaging. 2018;47(2):301–15. ACHILONU O J, SINGH E, NIMAKO G et al. Rule-based information extraction from free-text pathology reports reveals trends in South African female breast cancer molecular subtypes and Ki-67 expression [J]. Biomed Res Int, 2022, 2022: 6157861. Lin J et al. Chao You,Yi Xiao,. Radiogenomic analysis reveals tumor heterogeneity of triple-negative breast cancer [J]. Cell Reports Medicine 3, July 19, 2022, 3(19): 100694. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-3151750\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":true,\"archivedVersions\":[],\"articleType\":\"Research Article\",\"associatedPublications\":[],\"authors\":[{\"id\":218915176,\"identity\":\"b7f920dc-1734-4d5d-84ca-6244647f1992\",\"order_by\":0,\"name\":\"Lu Zheng\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Lu\",\"middleName\":\"\",\"lastName\":\"Zheng\",\"suffix\":\"\"},{\"id\":218915178,\"identity\":\"3369b42d-3809-4078-b81c-592d52047795\",\"order_by\":1,\"name\":\"Chenyu Sun\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Chenyu\",\"middleName\":\"\",\"lastName\":\"Sun\",\"suffix\":\"\"},{\"id\":218915180,\"identity\":\"937d63e2-9380-4310-b8d9-eeceafaab36b\",\"order_by\":2,\"name\":\"Muzi Meng\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"American University of the Caribbean School of Medicine\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Muzi\",\"middleName\":\"\",\"lastName\":\"Meng\",\"suffix\":\"\"},{\"id\":218915182,\"identity\":\"0c6ee30d-4066-47db-9d05-d026977b434a\",\"order_by\":3,\"name\":\"Eric Chen\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"University of Washington School of Medicine\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Eric\",\"middleName\":\"\",\"lastName\":\"Chen\",\"suffix\":\"\"},{\"id\":218915184,\"identity\":\"8d7484d8-fb69-450b-827a-4d5bfa0f85c1\",\"order_by\":4,\"name\":\"Tong Tang\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Tong\",\"middleName\":\"\",\"lastName\":\"Tang\",\"suffix\":\"\"},{\"id\":218915186,\"identity\":\"1b4158d8-d417-4866-845e-aa8ac94b1799\",\"order_by\":5,\"name\":\"Xiao Chen\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Xiao\",\"middleName\":\"\",\"lastName\":\"Chen\",\"suffix\":\"\"},{\"id\":218915187,\"identity\":\"af9e9fcc-ad16-475e-8ef4-f9b475186175\",\"order_by\":6,\"name\":\"Zhitao Wang\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":false,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Zhitao\",\"middleName\":\"\",\"lastName\":\"Wang\",\"suffix\":\"\"},{\"id\":218915188,\"identity\":\"983d9cd3-b110-4be8-91fa-9735cd858f80\",\"order_by\":7,\"name\":\"Ru Zhao\",\"email\":\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3ElEQVRIiWNgGAWjYJCCAx8q2HjY2JsPABnE6WA8OOMMnxw/z7FEIIM4LcyHedvkjCVn5Bgf5m0hQj3/tDMGh3nOmCVuOHPmwwHeBgZ5frED+LVI3M4xODinIi1xw/HeDQckdzAYzpydQMAaoJYDb84cA9pydsMBwzMMCQa3CWiRB2nhbfufuOFGzoMDiW1EaDEAOYy3jQ3kfYYDB4nRYng7rQAYtmygQDY42HBGgrBf5G4nb/4AjcrHn/9U2MjzSxPQwsDAYYDMkyCkHATYHxCjahSMglEwCkYyAAAxb1UpWIWyBQAAAABJRU5ErkJggg==\",\"orcid\":\"\",\"institution\":\"The Second Hospital of Anhui Medical University\",\"correspondingAuthor\":true,\"submittingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Ru\",\"middleName\":\"\",\"lastName\":\"Zhao\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2023-07-08 11:59:21\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-3151750/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-3151750/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":40260139,\"identity\":\"e4872912-f3e5-46a3-b118-ca7baca9ebad\",\"added_by\":\"auto\",\"created_at\":\"2023-07-19 14:37:01\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":218194,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eFlowchart of Statistical analysis\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Figure1.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-3151750/v1/4417eb82cd932fe25e311e37.png\"},{\"id\":40260136,\"identity\":\"3022751e-8504-4812-922e-d6134a7d86db\",\"added_by\":\"auto\",\"created_at\":\"2023-07-19 14:37:00\",\"extension\":\"png\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":104010,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eLASSO regression of feature selection\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Figure2.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-3151750/v1/8d3cfca1aed92fef61a790cd.png\"},{\"id\":40260137,\"identity\":\"f6e531f5-8b7d-4fee-80e4-1be9424e03d3\",\"added_by\":\"auto\",\"created_at\":\"2023-07-19 14:37:00\",\"extension\":\"png\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":130594,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eROC curves of each verification set in 10-fold cross-validation\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Figure3.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-3151750/v1/311bd56ddf626fd846b0eb77.png\"},{\"id\":40260138,\"identity\":\"b9b37362-36c8-44fc-9d67-60bc2888e7b2\",\"added_by\":\"auto\",\"created_at\":\"2023-07-19 14:37:00\",\"extension\":\"png\",\"order_by\":4,\"title\":\"Figure 4\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":169465,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003eThe weight coefficient of the LinearSVC model\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Figure4.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-3151750/v1/89a5d02d50fd9127b36482d6.png\"},{\"id\":65692076,\"identity\":\"5a96d689-d0e7-4f9a-9565-bdfb9125f6cf\",\"added_by\":\"auto\",\"created_at\":\"2024-10-01 10:24:04\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":1044468,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-3151750/v1/55d49c05-55f0-44cc-a3c9-cd9fcba18b53.pdf\"}],\"financialInterests\":\"No competing interests reported.\",\"formattedTitle\":\"Value of dynamic enhanced magnetic resonance image-based model in predicting low expression of HER-2 in breast cancer\",\"fulltext\":[{\"header\":\"Introduction\",\"content\":\"\\u003cp\\u003eSince the mid20th century, the incidence of breast cancer in women has steadily increased by approximately 0.5% per year. According to Global Cancer Statistics 2020 [\\u003cspan citationid=\\\"CR1\\\" class=\\\"CitationRef\\\"\\u003e1\\u003c/span\\u003e], breast cancer is the \\\"number one killer\\\" of women worldwide. As of 2020, breast cancer is estimated to account for 2.26\\u0026nbsp;million new instances of cancer worldwide and 10% of all malignancies in women [\\u003cspan citationid=\\\"CR2\\\" class=\\\"CitationRef\\\"\\u003e2\\u003c/span\\u003e]. Breast cancer is classified into four molecular subtypes based on the expression of human epithelial growth factor receptor 2 (HER2), hormone receptor (HR), and Ki-67. These subtypes are triple negative breast cancer (TNBC), HER2-positive breast cancer, Luminal A, and Luminal B. Although HER2-positive breast cancer has a poor prognosis and a high incidence of recurrence, the prognosis has improved since the advent of incorporating targeted medications such as trastuzumab, pertuzumab, and lapatinib into clinical practice. As HER2 detection technology improves and new targeted medications are deployed, the advantages of innovative antibody-drug conjugates (ADCs) like T-DXd continue to expand. Recent clinical trial results indicate that T-DXd induces a gradual treatment response in both the HER2-positive population and HER2-low-positive breast cancer [\\u003cspan citationid=\\\"CR3\\\" class=\\\"CitationRef\\\"\\u003e3\\u003c/span\\u003e].\\u003c/p\\u003e \\u003cp\\u003eTherefore, a precise method of evaluating HER2 status is crucial for breast cancer treatment. The immunohistochemical measurement of HER2 protein expression levels in tumor tissues is used to assess HER2-status. HER2-positive is characterized as IHC3\\u0026thinsp;+\\u0026thinsp;or 2\\u0026thinsp;+\\u0026thinsp;and positive on fluorescence in situ hybridization (FISH). Low HER2 expression is characterized by a negative FISH test and IHC1\\u0026thinsp;+\\u0026thinsp;or 2+. Needle core biopsy specimens may not accurately assess the overall status of the tumor due to the heterogeneity of the tumor and the limited number of tissue samples. According to recent research [\\u003cspan citationid=\\\"CR4\\\" class=\\\"CitationRef\\\"\\u003e4\\u003c/span\\u003e], the accuracy rate for detecting receptor status in puncture biopsy specimens ranges from 64.2\\u0026ndash;98.8%. There is limited consistency between results from various laboratories since the diagnostic level of the pathologist dictates how accurate HER2 status testing is [\\u003cspan citationid=\\\"CR5\\\" class=\\\"CitationRef\\\"\\u003e5\\u003c/span\\u003e]. As a result, individuals with early-stage breast cancer urgently need a method to determine their HER-2 status that is both affordable and non-invasive.\\u003c/p\\u003e \\u003cp\\u003eGillies et al. [\\u003cspan citationid=\\\"CR6\\\" class=\\\"CitationRef\\\"\\u003e6\\u003c/span\\u003e] initially conceived of the idea of Imageomics in 2010, which was later expanded upon by Lambin et al[\\u003cspan citationid=\\\"CR7\\\" class=\\\"CitationRef\\\"\\u003e7\\u003c/span\\u003e]. Imageomics facilitates accurate tumor diagnosis by extracting high-throughput quantitative characteristics from pictures in order to create high-dimensional datasets, which are then used to mine tumor-related parameters such as molecular type, treatment effectiveness, and clinical outcomes.\\u003c/p\\u003e \\u003cp\\u003eBreast cancer can only be visualized with conventional imaging, which does not reveal information on the genetic and molecular basis of the disease. Additionally, there are differences in how each radiologist subjectively interprets pictures. Imagingomics can extract high-throughput image characteristics from tumor images that are not visible to the human eye in order to more precisely define the heterogeneity of tumors. This transforms visual characteristics into numerical data.\\u003c/p\\u003e \\u003cp\\u003eThe main applications of imaging omics in breast cancer include diagnosis, assessment of treatment effectiveness, distinction of molecular subtype, and prognosis assessment. When combined with genetic, immunohistochemical, and pathology data, imagingomics has the potential to enhance clinical decision-making and improve patient outcomes. Using improved Magnetic Resonance Imaging (MRI) imaging to determine HER2 status is not widely reported at the present time, particularly in China. The purpose of this study is to determine whether improved MRI imaging could reliably predict HER2 low expression. A reliable HER2 status prediction model could be created to optimize treatment approach.\\u003c/p\\u003e\"},{\"header\":\"Methods\",\"content\":\"\\u003cdiv id=\\\"Sec3\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eData Collection\\u003c/h2\\u003e \\u003cp\\u003eFrom January 2019 to October 2022, clinical and pathology information on patients with invasive breast cancer that was verified by the pathology findings on Biopsy or surgical pathology confirmed puncture and an MRI scan before the puncture was gathered in our institution. Inclusion criteria included: (1) Female patients over the age of 18; (2) Patients who agreed to undergo breast MRI examination; (3) No other treatment was administered prior to the examination; (4) Adipose-suppressed T2WI, dynamic contrast-enhanced resonance imaging (DCE-MRI), and apparent diffusion coefficient, ADC, images were obtained by breast MRI prior to puncture or surgery; (5) All patients underwent air-core needle puncture or operation to identify non-special types of invasive ductal carcinoma; and (6) Complete clinical and pathological information was available. Exclusion criteria included: (1) Male patients; (2) patients with inflammatory, bilateral, lactation-related, and pregnancy-related breast cancer; (3) Any treatment prior to MRI scan, including needle biopsy, surgery, radiotherapy, chemotherapy, and endocrine therapy; (4) Poor image quality (such as artifacts); (5) patients whose tumor boundaries were not obvious; (6) Patients with claustrophobia. For patients with multiple tumors in the same breast, the largest tumor was selected for analysis. The study was approved by the Institutional Research Ethics Committee of The Second Affiliated Hospital of Anhui Medical University (Hefei, China), and written informed consent was obtained from all patients. This study was carried out in accordance with the Declaration of Helsinki.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec4\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eCollection of breast MRI image data\\u003c/h2\\u003e \\u003cp\\u003eBreast MRI images of all patients were obtained by a Siemens 3.0T MRI scanner (Verio, Siemens Healthcare, Erlangen, Germany) in a prone position using a specialized 8-channel breast coil. All patients underwent the same breast MRI protocol, which consisted of the following sequences: T1 and T2 weighted axial images, diffusion weighted images, and dynamic contrast-enhanced images with 6 phases (including a pre-injection non-contrast phase). The imaging parameters of the dynamic enhancement sequence were: TR (repetition time) 4.34 ms, TE (time to echo) 1.52 ms, averages 1, concatenations 1, FoV (field of view) read 360 mm, FoV phase 93.8%, and slice thickness 1.0 mm.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec5\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eDetermination of HER2 receptor expression status\\u003c/h2\\u003e \\u003cp\\u003eThis study adhered to the clinical practice recommendations for HER2 detection in breast cancer issued by the American Society of Clinical Oncology (ASCO) and the Association of American Pathologists (CAP). HER2 receptor status was assessed by immunohistochemistry (IHC) or in situ hybridization (FISH) in all patients in our study. IHC results of 0/1\\u0026thinsp;+\\u0026thinsp;are regarded as negative for HER2. HER2 status is deemed positive when the IHC result is 3+. Further FISH testing was done to determine the HER2 receptor status if the IHC result was 2+.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec6\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eDelineation of the receiver of interest (ROI)\\u003c/h2\\u003e \\u003cp\\u003eBreast MRI images of each patient were evaluated on a PACS workstation. Phase 2 of contrast-enhanced MRI (first phase after contrast agent injection) was selected as the contour image. At this stage, the MRI scans of every patient were exported and saved as DICOM files for upload to ITK-SNAP (\\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003ehttp://www.itksnap.org/\\u003c/span\\u003e\\u003cspan address=\\\"http://www.itksnap.org/\\\" targettype=\\\"URL\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e) [\\u003cspan citationid=\\\"CR8\\\" class=\\\"CitationRef\\\"\\u003e8\\u003c/span\\u003e]. The receiver of interest (ROI) was manually drawn on the MRI images by two skilled radiologists (with five- and ten-years\\u0026rsquo; experience of breast cancer diagnosis) on each slice and exported as a ROI mask image. All patients' DICOM images as well as the images used to draw ROI masks were saved in batches. The characteristics of the ROI mask were extracted using pyradiomics. The extracted image features include original features, features transformed by wavelet, and Log transformation features (including shape, Gray-level run-length matrix (GLRLM), Gray-level cooccurrence matrix (GLCM), Neighborhood gray-tone difference matrix (NGTDM), Gray-level difference matrix (GLDM), and Gray-level size zone matrix (GLSZM)).\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec7\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eThe intraobserver and interobserver agreement\\u003c/h2\\u003e \\u003cp\\u003eIntraclass correlation coefficient (ICC) was used to test the reliability and reproducibility of intra-observer and inter-observer data features [\\u003cspan citationid=\\\"CR9\\\" class=\\\"CitationRef\\\"\\u003e9\\u003c/span\\u003e]. Twenty patients\\u0026rsquo; MRI data were selected randomly for ICC validation analysis. Two experienced diagnostic radiologists (Physician A and Physician B) independently sketched the ROI. The extracted features were used to measure inter-observer ICC. Two weeks later, the ROI was sketched again by physician A without referring to the previous sketch results. The resulting features were used to measure interobserver ICC. Features satisfying interobserver and interobserver ICC\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.75 were preserved.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec8\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eStatistical analysis\\u003c/h2\\u003e \\u003cp\\u003eThe statistical analysis was completed with the aid of Python 3.9.0 and R 3.5.1 (The R Project for Statistical Computing (r-project.org)) software. Data reproducibility within and between observers was assessed using ICC. The normal distribution-conforming measurement results were reported as mean standard deviation (x\\u0026thinsp;\\u0026plusmn;\\u0026thinsp;s). A pipeline from Sandra Vieira [\\u003cspan citationid=\\\"CR10\\\" class=\\\"CitationRef\\\"\\u003e10\\u003c/span\\u003e] was used during the model building. The process of analytical and model construction was as follows: (1) The Mann-Whitney U test was used to check for correlation characteristics, and features with p value\\u0026thinsp;\\u0026lt;\\u0026thinsp;0.05 were obtained. (2) The dimensions of the omics characteristics were successively screened and reduced using the least absolute shrinkage and selection operator (LASSO) approach according to the retention characteristic of the ideal cut-off point. Then the data was randomly separated into training sets and test sets with a ratio of 7:3. Data was normalized in each separate set. An inner 10-fold cross-validation (CV) and outer 10-fold CV were used for model construction. (3) The training sets were modeled using Linear SVC (Support Vector Machines (SVM)). The optimum linear kernel \\\"C\\\" value was chosen using grid search and the 10-fold CV approach to establish the best SVM model. An outer 10-fold CV was used to validate the efficacy of the SVM model in the test set and diagnostic value was shown with Receiver Operator Characteristic (ROC) curve. In parallel CV repetitions of 10, ten optimum models were produced. The average iteration's performance measurements served as a representation of overall performance. (4) Permutation test was used to determine if the model performance assessment index had statistical significance. After 1000 replacements, the evaluation index of the data set's performance was determined. The P-value was calculated by dividing the number of permutations by the number of times its performance was better than that of the SVM model. P\\u0026thinsp;\\u0026lt;\\u0026thinsp;0.05 was considered statistically significant.Details are shown in Fig.\\u0026nbsp;\\u003cspan refid=\\\"Fig1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e.\\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"Results\",\"content\":\"\\u003cdiv id=\\\"Sec10\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eGeneral Information\\u003c/h2\\u003e \\u003cp\\u003eA total of 294 cases of non-specific invasive breast cancer were collected in this study. There were 40 cases (13.6%) with zero expression of HER. There were 141 patients with low HER2 expression (48.0%), including 89 patients with HER21\\u0026thinsp;+\\u0026thinsp;type (30.0%), 52 patients with HER22\\u0026thinsp;+\\u0026thinsp;and FISH- (18.0%), and 113 patients with HER2 positive HER23 +/ HER22\\u0026thinsp;+\\u0026thinsp;and FISH- (38.4%).\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec11\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eInterobserver and Intraobserver Agreement\\u003c/h2\\u003e \\u003cp\\u003eThe ICC of the interobserver of the two independent readers ranged from \\u0026minus;\\u0026thinsp;0.25 to 0.99, and the ICC of the intraobserver of the same reader ranged from \\u0026minus;\\u0026thinsp;0.04 to 0.99. A total of 858 features were retained, and all showed high interobserver and intraobserver ICCs with ICCs\\u0026thinsp;\\u0026ge;\\u0026thinsp;0.75. Thus, features extracted by reader A were used for further analysis.\\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec12\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eConstruction of model\\u003c/h2\\u003e \\u003cp\\u003e135 features were retained after relevant features were eliminated using the Mann-Whitney U test. 24 eigenvalues were kept after the LASSOCV regression. During the LinearSVC model construction, ten best models were selected and validated in test set. The average accuracy, sensitivity, and specificity of the test set were 79.6%, 73.7%, and 85.6%, respectively, following 10-fold CV Linear SVC modeling (Table\\u0026nbsp;\\u003cspan refid=\\\"Tab1\\\" class=\\\"InternalRef\\\"\\u003e1\\u003c/span\\u003e). The ROC analysis showed a mean area under the curve (AUC) of 0.87 of the models. Figure\\u0026nbsp;\\u003cspan refid=\\\"Fig2\\\" class=\\\"InternalRef\\\"\\u003e2\\u003c/span\\u003e depicts LASSO regression of feature selection. Figure\\u0026nbsp;\\u003cspan refid=\\\"Fig3\\\" class=\\\"InternalRef\\\"\\u003e3\\u003c/span\\u003e depicts the 10-fold CV ROC curve for the specific value index of the test set. Figure\\u0026nbsp;\\u003cspan refid=\\\"Fig4\\\" class=\\\"InternalRef\\\"\\u003e4\\u003c/span\\u003e depicts the corresponding features' regression coefficients.\\u003c/p\\u003e \\u003cp\\u003e \\u003cdiv class=\\\"gridtable\\\"\\u003e\\u003ctable float=\\\"Yes\\\" id=\\\"Tab1\\\" border=\\\"1\\\"\\u003e \\u003ccaption language=\\\"En\\\"\\u003e \\u003cdiv class=\\\"CaptionNumber\\\"\\u003eTable 1\\u003c/div\\u003e \\u003cdiv class=\\\"CaptionContent\\\"\\u003e \\u003cp\\u003ediagnostic efficiencies of the LinearSVC model in 10-fold CV\\u003c/p\\u003e \\u003c/div\\u003e \\u003c/caption\\u003e \\u003ccolgroup cols=\\\"7\\\"\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c1\\\" colnum=\\\"1\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c2\\\" colnum=\\\"2\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c3\\\" colnum=\\\"3\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c4\\\" colnum=\\\"4\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c5\\\" colnum=\\\"5\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"char\\\" char=\\\".\\\" class=\\\"colspec\\\" colname=\\\"c6\\\" colnum=\\\"6\\\"\\u003e\\u003c/div\\u003e \\u003cdiv align=\\\"left\\\" class=\\\"colspec\\\" colname=\\\"c7\\\" colnum=\\\"7\\\"\\u003e\\u003c/div\\u003e \\u003cthead\\u003e \\u003ctr\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c1\\\"\\u003e\\u0026nbsp;\\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003especificity\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003esensitivity\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003eaccuracy\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003eppv\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003enpv\\u003c/p\\u003e \\u003c/th\\u003e \\u003cth align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003eauc(95%CI)\\u003c/p\\u003e \\u003c/th\\u003e \\u003c/tr\\u003e \\u003c/thead\\u003e \\u003ctbody\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_1\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.94\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.91\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.90(0.78-1.0)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_2\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.75\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.86\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.89(0.74-1.00)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_3\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.94\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.91\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.92(0.82-1.00)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_4\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.67\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.80\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.80\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.80\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.75(0.54\\u0026ndash;0.95)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_5\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.82\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.69\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.88\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.85(0.70\\u0026ndash;0.99)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_6\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.94\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.73\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.86\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.89\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.85\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.88(0.76-1.00)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_7\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.84\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.70\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.70\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.84\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.82(0.66\\u0026ndash;0.97)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_8\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.97(0.93-1.00)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_9\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.90\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.79\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e1.00\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.97(0.92-1.00)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003ctr\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c1\\\"\\u003e \\u003cp\\u003efold_10\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c2\\\"\\u003e \\u003cp\\u003e0.78\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c3\\\"\\u003e \\u003cp\\u003e0.91\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c4\\\"\\u003e \\u003cp\\u003e0.83\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c5\\\"\\u003e \\u003cp\\u003e0.71\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"char\\\" char=\\\".\\\" colname=\\\"c6\\\"\\u003e \\u003cp\\u003e0.93\\u003c/p\\u003e \\u003c/td\\u003e \\u003ctd align=\\\"left\\\" colname=\\\"c7\\\"\\u003e \\u003cp\\u003e0.86(0.73\\u0026ndash;0.99)\\u003c/p\\u003e \\u003c/td\\u003e \\u003c/tr\\u003e \\u003c/tbody\\u003e \\u003c/colgroup\\u003e \\u003c/table\\u003e\\u003c/div\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003cp\\u003e \\u003c/p\\u003e \\u003c/div\\u003e \\u003cdiv id=\\\"Sec13\\\" class=\\\"Section2\\\"\\u003e \\u003ch2\\u003eDiagnostic performance of the permutation data set\\u003c/h2\\u003e \\u003cp\\u003eAfter 1000 permutation tests, the replacement data set's average accuracy, sensitivity, and specificity of the diagnostic effectiveness were less than 0.05 in comparison to the original data set. HER2 low expression and HER2 positive in breast cancer patients may be classified using the model developed following cross-validation. The prediction model's classification effectiveness was greater than the opportunity level.\\u003c/p\\u003e \\u003c/div\\u003e\"},{\"header\":\"Discussion\",\"content\":\"\\u003cp\\u003eAs of now, 45\\u0026ndash;55% of patients with breast cancer have low articulation of HER2. The treatment status of HER2-low-positive breast cancer patients has changed as a result of the development of new targeted drugs [\\u003cspan citationid=\\\"CR11\\\" class=\\\"CitationRef\\\"\\u003e11\\u003c/span\\u003e]. In HER2-low-positive breast cancer, anti-antibody drug conjugates (ADCs) like Trastuzumab deruxtecan (T-DXd, DS-8201) have shown promising therapeutic results. The identification of HER2-low-positive breast cancer has become a hot topic. Imaging was used to assist in the determination of molecular typing because biopsy is invasive and may not represent the characteristics of the entire tumor. Customary imaging tests depend on the demonstrative level of the specialist, which is subjective and not reproducible. Imaging omics has recently been used by some researchers to combine various algorithms, examination techniques, and clinical data. To better identify breast cancer's molecular typing, multimodal omics were developed.\\u003c/p\\u003e \\u003cp\\u003eUltrasound, MRI, and mammography (MG) are the primary screening methods for breast cancer this time. According to previous research [\\u003cspan citationid=\\\"CR12\\\" class=\\\"CitationRef\\\"\\u003e12\\u003c/span\\u003e], mammography has a low accuracy in predicting HER2 receptor expression status (95% CI of AUC is approximately 0.55 to 0.62, and sensitivity is approximately 0.11 to 0.50). The most common cause of poor accuracy in mammography is compression of the breast tissue [\\u003cspan citationid=\\\"CR13\\\" class=\\\"CitationRef\\\"\\u003e13\\u003c/span\\u003e]. Because breast tissue and lesions overlap, mammography cannot effectively detect lesions. In particular, tiny lesions are frequently missed in the background of dense mammary glands. In the diagnosis of breast cancer, MRI has medium specificity and high sensitivity due to its high spatial and tissue resolution. From a single sequence to a multi-sequence, multi-modal sequence, MRI sequences are included. The construction of the model shifts from a single model to a joint model. From the tumor itself to the surrounding tissue and even the lymph nodes, the features' region of interest (ROI) was taken. The study's goals ranged from predicting cancer to determining the types of molecules. Numerous omics studies have emerged in recent years as a result of image omics and other omics. To further enhance the prediction effect of molecular classification and optimize the performance of various classifiers, Li et al. proposed a multi-model-based recursive feature elimination strategy [\\u003cspan citationid=\\\"CR14\\\" class=\\\"CitationRef\\\"\\u003e14\\u003c/span\\u003e]. The Luminal A, Luminal B, HER-2, and Basal-like subtypes had an identification accuracy of 0.91, 0.89, 0.83, and 0.87, respectively.\\u003c/p\\u003e \\u003cp\\u003eDynamic enhanced magnetic resonance imaging (DCE-MRI) is an upgraded technique for detecting breast cancer. DCE-MRI showed the highest detection accuracy when using multi-phase time series imaging to reflect the blood perfusion of breast cancer lesions. DCE-MRI can characterize soft tissue with high resolution to give a more precise portrayal of the morphological and structural aspects of breast cancer lesions. Moreover, patient exposure to radiation during the examination is greatly reduced by the DCE-X-ray component without ionizing radiation. DCE-MRI is essential in the field of breast cancer imaging [\\u003cspan citationid=\\\"CR15\\\" class=\\\"CitationRef\\\"\\u003e15\\u003c/span\\u003e]. In this study, the application of DCE-MRI to assess HER2 status was further explored. After 10-fold cross-validation Linear SVC modeling, the average accuracy of the validation set was 79.6%, the average sensitivity was 73.7%, the average specificity was 85.6%, and the average AUC of ROC analysis was 0.87. After 1000 substitution tests, the replacement data set's average accuracy, sensitivity, and specificity were less than 0.05 compared to the original data set. This shows that the cross-validated model can categorize patients with breast cancer as having high or low expression of the HER2 gene, and that its classification accuracy is more than the opportunity level. The results of this study present a novel technique for non-invasive diagnosis of breast cancer with low HER2 expression.\\u003c/p\\u003e \\u003cp\\u003eThere are still some limitations to this study. First of all, this research, which was a single-center evaluation, only included 294 cases. Only a small number of patients lacked HER2 expression, and breast cancer patients without or with little HER2 expression were not categorized. In the future, the sample size could be increased to confirm the model's capacity for forecasting. Secondly, only patients with invasive ductal carcinoma of the breast were included in this research. The model requires more patients with other types of breast cancer in order to obtain the highest level of predicting power. Thirdly, this study employed manual layer-by-layer mapping of lesions, which is not only arduous and time-consuming, but also changes depending on the reader. Reproductions of various radiological aspects are made. Fourthly, this study\\u0026rsquo;s focus was primarily on predicting HER2 status. Further studies can analyze the prediction of the Ki-67 proliferation index [\\u003cspan citationid=\\\"CR16\\\" class=\\\"CitationRef\\\"\\u003e16\\u003c/span\\u003e], an optimization model of clinical coupled genetic data, and imaging characteristics around tumors [\\u003cspan citationid=\\\"CR17\\\" class=\\\"CitationRef\\\"\\u003e17\\u003c/span\\u003e] in order to further explore the importance of imaging omics in predicting the prognosis of breast cancer patients. Finally, because this research was retrospective in nature, more prospective investigations will be required to confirm these prognostic models.\\u003c/p\\u003e \\u003cp\\u003eIn conclusion, our research suggested that HER2 low expression status in breast cancer patients can be predicted using radiological features extracted from pre-treatment MRI images. To better understand the relationship between imaging omics features and HER-2 status, more studies involving multiple centers, larger sample sizes, and various imaging modalities are required.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eAcknowledgements\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eNone.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthor Disclosure Statement \\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eAuthors declare no conflicts of interest. The authors declare that they have no competing interests.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eData Availability Statement\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Corresponding author should be contacted if someone wants to request the data from this study.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eFunding Sources\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThis research is supported by National Natural Science Foundation of China (No. 82200225), Research Project of Young Scholars of Breast Cancer in China (No.320.6750.2021-10-25), Research Fund of Anhui Medical University (No. 2018xkj038) and Key Project of Higher Education Humanities and Social Sciences of Anhui Province (No.SK2021A0167).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003ePatient consent for publication\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eNot applicable.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eEthics approval and consent to participate\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe study was approved by the Institutional Research Ethics Committee of The Second Affiliated Hospital of Anhui Medical University (Hefei, China), and written informed consent was obtained from all patients. This study was carried out in accordance with the Declaration of Helsinki.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthor Contributions Statement\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eLZ and RZ designed the experiment. CS and ZW collected data. MM and EC analyzed the data. LZ, TT and XC wrote the main manuscript text. All the authors reviewed the manuscript.\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\u003cli\\u003e\\u003cspan\\u003eSungH FerlayJ. SiegelRL, etal.Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortalit world wide for 36 cancers in 185 countries [J]. CA Cancer JClin. 2021;71(3):209\\u0026ndash;49.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSiegel RL, Miller KD, Wagle NS, Jemal A, Cancer statistics. 2023; CA Cancer J Clin. 2023;1\\u0026ndash;32.DOI: \\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003e10.3322/caac.21763\\u003c/span\\u003e\\u003cspan address=\\\"10.3322/caac.21763\\\" targettype=\\\"DOI\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eModi S, Park H, Murthy RK et al. Antitumor activity and safety of trastuzumab deruxtecan in patients with HER2-low-expressing advanced breast cancer: results from a phase \\u0026acirc;\\u0026#133;\\u0026nbsp;b study [J].J Clin Oncol 2020,38(17):1887\\u0026ndash;96. DOI:\\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003e10.1200/JCO.19.02318\\u003c/span\\u003e\\u003cspan address=\\\"10.1200/JCO.19.02318\\\" targettype=\\\"DOI\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLorgis V, Algros MP, Villanueva C, et al. Discordance in early breast cancer for tumour grade, estrogen receptor, progesteron receptors and human epidermal receptor-2 status between core needle biopsy and surgical excisional primary tumour.[J]. Breast. 2011;20(3):284\\u0026ndash;7.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eOrlando L, Viale G, Bria E, et al. Discordance in pathology report after central pathology review: Implications for breast cancer adjuvant treatment[J]. Breast. 2016;30:151\\u0026ndash;5.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eGillies RJ, Anderson AR, Gatenby RA, et al. The biology underlying molecular imaging in oncology: from genome to anatome and back again [. J] Clin Radiol. 2010;65:517\\u0026ndash;21. \\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003e10.1016/j.crad.2010.04.005\\u003c/span\\u003e\\u003cspan address=\\\"10.1016/j.crad.2010.04.005\\\" targettype=\\\"DOI\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLambin P, Rios -Velazquez E, Leijenaar R et al. Radiomics:extracting more information from medical images using advanced feature analysis[J].EurJCancer,2012,48:441\\u0026ndash;6DOI: \\u003cspan class=\\\"ExternalRef\\\"\\u003e\\u003cspan class=\\\"RefSource\\\"\\u003e10.1016/j.ejca.2011.11.036\\u003c/span\\u003e\\u003cspan address=\\\"10.1016/j.ejca.2011.11.036\\\" targettype=\\\"DOI\\\" class=\\\"RefTarget\\\"\\u003e\\u003c/span\\u003e\\u003c/span\\u003e.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eYushkevich PA, Piven J, Hazlett HC, et al. User-guided3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage. 2006;31(3):1116\\u0026ndash;28.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eXu X, Zhu H, Li R et al. Whole-liver histogram and texture analysis on T1 maps improves the risk stratification of advanced fibrosis inNAFLD[J].EurRadiol.2021.31(3):1748\\u0026ndash;59.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eVieraS,Garcia-Dias R. Pinaya WHL.A step-by-step tutorial on how to build a mechine learning model[J].Mach Learn,2020:343\\u0026ndash;70.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eModi S, Park H, Murthy RK, et al. Antitumor activity and safety of trastuzumab deruxtecan in patients with HER2 -low -expressing advanced breast cancer: results from a phase Ib study [J]. J Clin Oncol. 2020;38(17):1887\\u0026ndash;96.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eSon J, Si EL, Kim EK, et al. Prediction of breast cancer molecular subtypes using radiomics signatures of synthetic mammography from digital breast tomosynthesis[J]. Sci Rep. 2020;10:21566.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eDullum JR, Lewis EC, Mayer JA. Rates and Correlates of Discomfort Associated with Mammography1[J]. Radiology. 2000;214:547\\u0026ndash;52.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLi W, Yu K, Feng CL, et al. Molecular subtypes recognition of breast cancer in dynamic contrast-enhanced breast magnetic resonance imaging phenotypes from radiomics data. Comput Math Methods Med. 2019;2019:6978650.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eMarino MA, Helbich T, Baltzer P, et al. Multiparametric MRI of the breast: A review [J]. J Magn Reson Imaging. 2018;47(2):301\\u0026ndash;15.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eACHILONU O J, SINGH E, NIMAKO G et al. Rule-based information extraction from free-text pathology reports reveals trends in South African female breast cancer molecular subtypes and Ki-67 expression [J]. Biomed Res Int, 2022, 2022: 6157861.\\u003c/span\\u003e\\u003c/li\\u003e \\u003cli\\u003e\\u003cspan\\u003eLin J et al. Chao You,Yi Xiao,. Radiogenomic analysis reveals tumor heterogeneity of triple-negative breast cancer [J]. Cell Reports Medicine 3, July 19, 2022, 3(19): 100694.\\u003c/span\\u003e\\u003c/li\\u003e\\u003c/ol\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":false,\"hideJournal\":true,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true},\"keywords\":\"Breast cancer, low expression of human epidermal growth factor receptor 2, Radiomics, DCE-MRI\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-3151750/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-3151750/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003ch2\\u003eObjective\\u003c/h2\\u003e \\u003cp\\u003eThis study aimed to evaluate the feasibility of evaluating early low expression of HER-2 in patients with breast cancer by applying Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) based imaging features, which could potentially optimize treatment for patients.\\u003c/p\\u003e\\u003ch2\\u003eMethod\\u003c/h2\\u003e \\u003cp\\u003eClinical and pathology data of 294 female patients with invasive ductal carcinoma confirmed by puncture or surgical pathology were collected. Regions of interest (ROI) were mapped. Features were then extracted from the original Magnetic Resonance Imaging (MRI) image data. Relevant features were screened out by Mann-Whitney U test. Cross-validated LASSO regression was used for feature selection. Inner and outer 10-fold cross-validation (CV) models were used. The inner 10-fold CV was used to select the best model during the Linear SVC modeling in training set, and an outer 10-fold CV was used to validate the efficiency in validation set. Model performance was evaluated by using receiver operator curve (ROC) analysis. The average accuracy, sensitivity, and specificity were calculated.\\u003c/p\\u003e\\u003ch2\\u003eResults\\u003c/h2\\u003e \\u003cp\\u003eAfter model selection using the inner 10-fold CV in Linear SVC modeling and validation using the outer CV, the average accuracy, sensitivity, and specificity of the validation set were 79.6%, 73.7%, and 85.6%, respectively. The average area under curve (AUC) of ROC analysis was 0.87. The diagnostic efficiency of the replacement dataset after 1000 permutation tests was compared with the original dataset, and the average accuracy, sensitivity, and specificity were all less than 0.05. The differences were all statistically significant. The model established after cross-validation could classify patients as HER2 low expression or HER2 positive. The classification efficiency of the model was higher than the chance level.\\u003c/p\\u003e\\u003ch2\\u003eConclusion\\u003c/h2\\u003e \\u003cp\\u003eDCE-MRI imaging model can help predict the low expression of HER2 receptor in breast cancer with a high predictive efficiency, which can provide a new method for clinical diagnosis of non-invasive HER2 status.\\u003c/p\\u003e\",\"manuscriptTitle\":\"Value of dynamic enhanced magnetic resonance image-based model in predicting low expression of HER-2 in breast cancer\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2023-07-19 14:36:56\",\"doi\":\"10.21203/rs.3.rs-3151750/v1\",\"editorialEvents\":[{\"type\":\"communityComments\",\"content\":0}],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"5f6faa30-2611-4e66-b2d7-3bf691bb5d35\",\"owner\":[],\"postedDate\":\"July 19th, 2023\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"posted\",\"subjectAreas\":[],\"tags\":[],\"updatedAt\":\"2024-10-01T10:23:45+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2023-07-19 14:36:56\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-3151750\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-3151750\",\"identity\":\"rs-3151750\",\"version\":[\"v1\"]},\"buildId\":\"369fNeqWncA4NS6XSWjrt\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}