Evaluation of the Applicability of an Artificial Intelligence System for Mammography Analysis Trained on Overseas Data for Japanese Domestic Data

Evaluation of the Applicability of an Artificial Intelligence System for Mammography Analysis Trained on Overseas Data for Japanese Domestic Data | 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 Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Evaluation of the Applicability of an Artificial Intelligence System for Mammography Analysis Trained on Overseas Data for Japanese Domestic Data Maya Makita, Kouzou Murakami, Wakana Murakami, Hiroko Takamatsu, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4855505/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 This study evaluated the performance of the Artificial Intelligence (AI)-based Computer-Aided Diagnosis system (AI-CAD), Lunit INSIGHT MMG, in detecting breast cancer from digital mammography images of Japanese women. We collected digital mammography images from two groups at Showa University Hospital. One group consisted of surgical and biopsy specimens of breast lesions between January and December 2019, and the other was digital mammography images taken at Showa University Hospital during the same period. The AI-CAD system was developed based on a convolutional neural network trained on over 200,000 cases, overseas of Japan. We analyzed the breast cancer detection capabilities and compared the results with the interpretations of the radiologists and breast surgeons. We used the area under the receiver operating characteristic (AUROC) curve to evaluate the data. We evaluated the performance of the Lunit INSIGHT MMG using a dataset of 676 breasts from 338 patients. Although no significant overall difference was observed, the radiologists reported increased sensitivity, specificity, and AUROC values, on average. The AI-CAD system trained on overseas data showed comparable effectiveness with Japanese data. Biological sciences/Cancer/Cancer imaging Health sciences/Health care/Diagnosis Health sciences/Medical research Artificial Intelligence Computer-Aided Diagnosis system Japan mammography breast cancer Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Breast cancer accounts for 15.5% of cancer-related deaths in women worldwide. However, early detection and treatment have significantly improved prognosis [ 1 ]. In recent years, artificial intelligence (AI) products have been developed to assist the interpreting of digital mammography images, with the expectation of enhancing diagnostic performance [ 2 – 4 ]. However, many of these AI systems have been developed using data collected outside Japan. While commercial software such as HOLOGIC ( https://www.hologic.com/hologic-products/breast-health-solutions/image-analytics ), iCAD ( https://www.icadmed.com/breast-health/ai-breast-cancer-detection/ ), and CureMetrix ( https://curemetrix.com/ ) have been approved overseas, there are limited reports on their usefulness, specifically in the Japanese population. The 2022 Japanese breast cancer treatment guidelines state that the use of AI-based Computer-Aided Diagnosis (AI-CAD) for image interpretation does not conclusively improve accuracy in Japan [ 5 ]. The Lunit INSIGHT for Mammography (v1.1.7.2; Lunit, South Korea) is an AI system designed exclusively for image interpretation in digital mammography using large-scale data. Previous studies have shown that it outperforms radiologists in detecting breast cancer from mammography images, and significantly improves diagnostic accuracy when combined with AI assistance [ 4 , 6 – 10 ]. Although the Lunit INSIGHT has been trained on data from various countries, its performance has not yet been explicitly validated in Japanese female populations. Hence, this study aimed to assess the breast cancer detection performance of the Lunit INSIGHT for mammography images of Japanese women and to investigate whether its use contributes to improved diagnostic accuracy for physicians. Results AI-CAD Diagnostic Accuracy Patient characteristics are presented in Table 1. The study included 338 patients with 676 breasts as the target data for validation. For each case, the distribution of background breast tissue was as follows: in malignant cases, 2 cases were fatty, 55 cases had scattered fibroglandular densities, 54 cases had heterogeneously dense tissue, and 8 cases had extremely dense tissue; in benign cases, 1 case was fatty, 29 cases had scattered fibroglandular densities, 62 cases had heterogeneously dense tissue, and 3 cases had extremely dense tissue; in normal cases, 1 case was fatty, 29 cases had scattered fibroglandular densities, 94 cases had heterogeneously dense tissue, and 0 cases had extremely dense tissue. The following results were based on per-breast counts. Among the malignant cases, 102 were invasive carcinomas, 22 were ductal carcinoma in situ (DCIS), 1 was papillary carcinoma (encapsulated), 1 was a special type, and 1 was challenging to determine as invasive or noninvasive. When summarizing the features across all breasts, in terms of masses, 50 were cancerous and 24 were non-cancerous; in terms of asymmetry, 34 were cancer and 13 were not cancer; in terms of architectural distortion, 62 were cancerous and 3 were non-cancerous; and in terms of calcifications requiring assessment of malignancy, 55 were cancerous and 26 were non-cancerous. In the diagnostic interpretations by the readers, the number of malignant cases classified as negative and positive were 10 (8%) and 117 (92%) cases, respectively, whereas for non-malignant cases, the number of negative and positive were 497 (91%) and 52 (9%) cases, respectively. In the AI-CAD system, the number of malignant cases classified as negative and positive were 27 (21%) and 100 (79%) cases, respectively, whereas for non-malignant cases, the number of negative and positive were 493 (90%) and 56 (10%) cases, respectively. The diagnostic performance of the AI-CAD for validation data is presented in Table 2. Their sensitivity, specificity, and AUROC of the AI-CAD system were 0.787, 0.894, and 0.897, respectively. After excluding 10 cases of malignancy judged as negative by GS, the sensitivity, specificity, and AUROC for the remaining data were 0.838, 0.894, and 0.930, respectively. Reading Experiment The characteristics of the patient groups used for reading are presented in Table 3. For individual cases, the background breast density for malignant cases comprised of 0 cases of fatty tissue, 7 cases of scattered fibroglandular density, 12 cases of heterogeneous high density, and 2 cases of extremely high density. For benign cases, there were 0 cases of fatty tissue, 4 cases of scattered fibroglandular densities, 5 cases of heterogeneous high density, and 0 cases of extremely high density. For normal cases, there were no cases of fatty tissue, three cases of scattered fibroglandular densities, seven cases of heterogeneous high density, and none of extremely high density. When summarizing the features of the findings across all breasts, there were 7 cases of masses with cancer, 3 cases of masses with no cancer, and 7 findings of asymmetry, 11 findings of architectural distortion, and 10 findings of calcifications requiring assessment, all with cancers. Among the malignant cases, there were 15 cases of invasive ductal carcinoma, 7 cases of DCIS, and 1 case of special type carcinoma. We also compared the AUROC values based on category classification and LOM scores before and after using AI-CAD (Table 4). The AUROC values based on the category assessment before and after using the AI-CAD system were 0.750 and 0.756, respectively, whereas those based on the LOM assessment were 0.750 and 0.761, respectively. Although no significant differences were observed, the AUROC values generally increased after using the AI-CAD system compared to before its use. Representative examples wherein the judgment of the readers were changed are shown in Figures 1 and 2. Discussion Based on applying the AI-CAD system to the verification data, it is anticipated that, despite the use of AI-CAD, detecting tricky cases visually will still be challenging. However, when such cases were excluded, the performance of the AI-CAD system was comparable to that of the radiologists in terms of detection capability. Specifically, when cases in which radiologists visually judged lesions as challenging to identify were excluded, the diagnostic performance of the AI-CAD system showed a sensitivity of 84% and a specificity of 90%. Although the version was different, this performance was similar to that of a previous study that used the Lunit INSIGHT MMG, reporting a sensitivity of 82.1% and a specificity of 90.3% [ 12 ]. Therefore, overseas-approved products can be utilized in Japan, but further research is essential. In the reading experiment, we investigated whether the use of AI-CAD improves the detection performance of radiologists. Despite no significant difference in the change in detection performance due to AI-assistance, the radiologists detected them with greater sensitivity, specificity, and AUROC values on average, indicating improved detection performance. Notably, cases with calcifications in the lesions were more manageable to judge by the radiologists. However, in cases where the background breast tissue had a high density and calcifications were easily overlooked, the AI-CAD was efficient in assisting the radiologists. However, there are points of reflection in the case shown in Fig. 2. While radiologists could identify findings in both the craniocaudal and mediolateral oblique (MLO) views, their confidence in malignancy appeared low. In contrast, the AI only indicated findings in the MLO view and provided a low abnormality score of 12. Two radiologists changed the category from 1 to 3, and three changed it from 3 to 1. This may be due to their interpretation of the image as part of the normal structures, possibly due to low abnormality score and unidirectional indications. However, considering the actual findings, the AI-CAD system should not have downgraded this case to Category 1, as this would have been considered a positive judgment. Cases in which radiologists are uncertain based on individual findings may lead to differing judgments based on their understanding and trust in the AI-CAD system. To achieve results consistent with the AI-CAD, emphasizing the interpretation of abnormal scores and providing training for such judgments may be necessary. This case-control study may not be the most appropriate evaluation method for the AI-CAD system, which is primarily intended for screening purposes. In this study, a significant proportion of malignant and benign cases was evaluated based on pathological results, and there were many cases in which assessment based solely on images was challenging. In the reading experiment conducted in this study, no specific practice time was provided for using the AI-CAD system in conjunction with reading. There was variability in the participants’ experiences with AI-CAD, and observing consistent trends in diagnostic accuracy improvement was challenging. Additionally, the limited number of cases in the reading experiments may have contributed to this variation. Feedback from the participants indicated that the AI-CAD system increased confidence in diagnosing malignancy in cases suspected to be malignant or decreased confidence in diagnosing non-malignant lesions. We did not precisely measure the reading time or impose time constraints on the change in confidence levels before and after using the AI-CAD system. Therefore, timesaving effects were not evaluated in this study. The application of the AI-CAD system trained on overseas data to Japanese data demonstrated comparable effectiveness. Further research, including larger sample sizes, long-term studies, and comparisons of reading times in real-world reading environments, may bring us closer to the clinical implementation of AI-CAD for mammography screening in Japan. Materials and Methods This retrospective study was approved by the Ethics Review Committee of Showa University (Approval Number: 3426). The research used digital mammography images obtained from patients who visited the Showa University Hospital Breast Surgery Department. Images were collected retrospectively. We used the mammography category classification and background breast evaluation criteria for image diagnosis, based on the Breast Imaging Reporting and Data System (BI-RADS) categories of the American College of Radiology [ 13 ]. All research procedures were performed in accordance with the Declaration of Helsinki and local regulations. Computer-Aided Diagnosis Software The AI-CAD system used in this study is the Lunit INSIGHT MMG. It detects regions suggestive of breast cancer on mammograms and marks areas indicative of malignant lesions. The system displays an abnormality score (range, 0–100) for qualitative assessment, which assists radiologists in the diagnosis. The Lunit INSIGHT MMG was developed based on a convolutional neural network, utilizing training data from over 200,000 cases analyzed in South Korea, United States, and United Kingdom. Patients and Dataset We collected patient data from two groups at the Showa University Hospital. Group A contains the surgical and biopsy specimens of breast lesions obtained between January and December 2019 (2,813 cases). Group B contains digital mammography images obtained at the Showa University Hospital between the same time period (4,390 cases). The collected data were categorized into three groups: malignant, benign, and normal. We selected patients from Group A who underwent pathological examination for breast lesions and reached a final diagnosis of either malignant or benign. For Group B, we classified patients as normal if both sides were diagnosed as BI-RADS Category 1 by the interpreting radiologist and were re-evaluated as Category 1 again between January and December 2021. We excluded patients who met the following criteria: 1) male patients; 2) patients with imaging of only one breast; and 3) patients with a history of breast surgery, including breast augmentation. In addition, for both the malignant and benign groups, we excluded cases based on the following criteria: 1) pathological examination was performed on non-breast tissue, 2) pathological diagnosis was made after neoadjuvant chemotherapy, and 3) the pathological diagnosis was difficult to determine as benign or malignant. The number of cases collected according to criteria was 821, 380, and 179 in the malignant, benign, and normal groups, respectively. Among these, 150 cases of image data were extracted from each group in ascending order of examination date, with further exclusion of cases due to specific reasons, such as the inability to analyze images from computed radiography or a time gap of more than 6 months between pathological images. The final dataset included 119, 95, and 124 cases in the malignant, benign, and normal groups, respectively (Figs. 3 and 4). It is important to note that patients in the malignant and benign groups used digital mammography images from the most recent examination date before surgery or biopsy, including those taken at other facilities (Table 5 ). Validation of the AI-CAD System Two radiologists with 11 and 2 years of experience validated the collected data. First, they referred to clinical information and breast magnetic resonance imaging findings that could be confirmed at our hospital to create the gold standard data (GS) for the mammography evaluation of each case. Next, they compared the GS with the results of the AI-CAD system to verify whether the regions showing significant abnormality scores in the AI-CAD matched the lesions indicated by the GS. The background breast density for each case was assessed by three radiologists with 11, 6, and 6 years of experience. Reading Experiment From the verification data, 40 cases were randomly selected (21 malignant, 9 benign, and 10 normal). Seven radiologists with mammography reading experience (3 radiologists specializing in diagnostic radiology with 21, 16, and 8 years of experience; 2 radiologists specializing in radiology with 6 years of experience each; and 2 radiology residents with 4 and 3 years of experience) and six breast surgeons (with 23, 18, 11, 3, 3, and 2 years of experience) participated. Each participant read the same 40 cases independently. Without referring to the AI-CAD evaluations, the participants assessed the findings, calculated using the Likelihood of Malignancy (LOM) scores (11-point scoring system), and categorized both breasts based solely on their interpretation. Participants then reviewed the AI-CAD evaluations and repeated the same assessment. Readers repeated steps 1 and 2 for each case without time constraints. Statistical Analysis For mammography evaluations by radiologists, the category scores of 1 and 2 were classified as “negative” and 3, 4, and 5 were “positive.” Based on prior research, the abnormality score cutoff of the AI-CAD system was set at 10, which classifies scores < 10 as “negative” and scores ≥ 10 as “positive.” We calculated the sensitivity and specificity using the true positive, which was defined as malignant according to the final pathological diagnosis for each breast. Diagnostic performance was analyzed using a receiver operating characteristic (ROC) curve. The evaluation of the reading experiments assessed whether there was a change in the area under the ROC (AUROC) curve for each reader before and after reading using the Wilcoxon signed-rank test. Statistical analysis was performed using the EZR software version 1.62, and an extension of the R software and R Commander [ 14 ]. Statistical significance was set at p < 0.05. Declarations Additional information Competing Interests: This research was funded by FUJIFILM. Ethical Approval: This retrospective study was approved by the Ethics Review Committee of Showa University (Approval Number: 3426). Informed Consent: Informed consent was obtained from all individual participants included in the study. Author Contribution conceptualization；M.M. and K.M.data curation; M.M. , K.M., H.T., T.K., and A.S. Hiroko Takamatsu, Takahiro Kanai, Atsuhito Sekimotoformal analysis; M.Minvestigation; all authorsproject administration; M.M.supervision; K.M.Writing – review & editing, all authors. Acknowledgement We wish to thank Mr. Ryuji Hisanaga and Ms. Chiori Murata (FUJIFILM Corporation) for their useful help and contributions in this research. We also would like to thank Editage (www.editage.jp) for English language editing. Data Availability The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request. References Sung, H. et al . Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71, 209–249 (2021). Sechopoulos, I., Teuwen, J. & Mann, R. Artificial Intelligence for breast cancer detection in mammography and digital breast tomosynthesis: State of the art. Semin. Cancer Biol. 72, 214–225 (2021). Sechopoulos, I. & Mann, R. M. Stand-alone artificial intelligence - The future of breast cancer screening? Breast 49, 254–260 (2020). Kim, H. E. et al . Changes in cancer detection and false-positive recall in mammography using artificial intelligence: A retrospective, multireader study. Lancet Digit. Health 2, e138–e148 (2020). Japan breast cancer society. Is It Useful to Use AI Software for Image Interpretation in Mammography Breast Cancer Screening (FRQ2)? Breast Cancer Treatment Guidelines , (2022) Edition. https://jbcs.xsrv.jp/guideline/2022/k_index/frq2/ . Park, G. E., Kang, B. J., Kim, S. H. & Lee, J. Retrospective review of missed cancer detection and its mammography findings with artificial-intelligence-based, computer-aided diagnosis. Diagnostics (Basel) 12 (2022). Dembrower, K. et al . Effect of artificial intelligence-based triaging of breast cancer screening mammograms on cancer detection and radiologist workload: A retrospective simulation study. Lancet Digit. Health 2, e468–e474 (2020). Salim, M. et al . External evaluation of 3 commercial artificial intelligence algorithms for independent assessment of screening mammograms. JAMA Oncol. 6, 1581–1588 (2020). Lee, J. H. et al . Improving the performance of radiologists using artificial intelligence-based detection support software for mammography: A multi-reader study. Korean J. Radiol. 23, 505–516 (2022). Freeman, K. et al . Use of artificial intelligence for image analysis in breast cancer screening programmes: Systematic review of test accuracy. BMJ 374, n1872 (2021). UICC. TNM Classification of Malignant Tumours . 8th ed. Wiley-Blackwell (U.I.C.C. International Union Against Cancer, 2017). Yoon, J. H. et al. Artificial intelligence-based computer-assisted detection/diagnosis (AI-CAD) for screening mammography: Outcomes of AI-CAD in the mammographic interpretation workflow. Eur. J. Radiol. Open 11, 100509 (2023). American College of Radiology. ACR BI-RADS Atlas . 5th ed. (American College of Radiol., Reston, VA, 2013). Kanda, Y. Investigation of the freely available easy-to-use software “EZR” for medical statistics. Bone Marrow Transplant. 48, 452–458 (2013). Tables Table 1. Patient demographics (verification data) Overall Malignant Benign Normal N =338 n =119 n =95 n =124 Age group, n (%) ≤50 years 159 (47) 58 (49) 71 (75) 30 (24) 51–65 years 112 (33) 34 (29) 19 (20) 59 (48) >65 years 67 (20) 27 (23) 5 (5) 35 (28) Density grade, n (%) Fatty 4 (1) 2 (2) 1 (1) 1 (1) Scattered fibrolandular dense 113 (33) 55 (46) 29 (31) 29 (23) Heterogeneous dense 210 (62) 54 (45) 62 (65) 94 (76) Extremely dense 11 (3) 8 (7) 3 (3) 0 Characteristics of the imaging features of the breasts Overall Cancer Not cancer N =676 n =127 n =549 Interpretation of the radiologists, n (%) Category 1, 2 507 (75) 10 (8) 497 (91) Category 3, 4, 5 169 (25) 117 (92) 52 (9) AI-CAD abnormality score, n (%) <10% 520 (77) 27 (21) 493 (90) ≥10％ 156 (23) 100 (79) 56 (10) Histology, n (%) Invasive cancers 102 (80) DCIS 22 (17) Papillary carcinoma (encapsulated) 1 (1) Special type 1 (1) Others a 1 (1) Mammography findings across all breasts Cancer Not cancer Mass, n 50 24 Asymmetry, n 34 13 Distortion, n 62 3 Calcifications, n 55 26 a Challenging to determine as invasive or noninvasive AI-CAD, Artificial Intelligence-based Computer-Aided Diagnosis system; DCIS, ductal carcinoma in situ. Table 2. Diagnostic performance of the AI-CAD for validation data A B Sensitivity 0.787 0.838 Specificity 0.894 0.894 AUROC 0.897 0.930 (A) represents the performance of AI-CAD for all data. (B) represents the performance of AI-CAD for data excluding gold standard data-negative malignant cases. AUROC, area under the receiver operating characteristic curve; AI-CAD, Artificial Intelligence-based Computer-Aided Diagnosis system. Table 3. Patient demographics (reading experiment data) Overall Malignant Benign Normal N =40 n =21 n =9 n =10 Age group, n ≤50 years 20 12 6 2 51–65 years 16 7 2 7 >65 years 4 2 1 1 Density grade, n Fatty 0 0 0 0 Scattered fibrolandular dense 14 7 4 3 Heterogeneous dense 24 12 5 7 Extremely dense 2 2 0 0 Characteristics of the imaging features of the breasts Cancer Histology, n Invasive cancers 15 DCIS 7 Special type 1 UICC classifications, n a T is(DCIS) 7 1mi 2 1a 2 1b 2 1c 2 2 6 3 0 4a 1 Unknown 1 N 0 19 1 3 Unknown 1 Mammography findings across all breasts Cancer Not cancer Mass, n 7 3 Asymmetry, n 7 0 Distortion, n 11 0 Calcifications, n 10 0 a The UICC TNM Classification of Malignant Tumors[11]. UICC, Union for International Cancer Control; DCIS, ductal carcinoma in situ; T, primary tumor site and size; N, regional lymph node involvement; M, metastasis. Table 4. Performance readers with and without AI. A AUROC B AUROC No. without AI with AI No. without AI with AI 1 0.765 0.754 1 0.768 0.763 2 0.732 0.732 2 0.748 0.754 3 0.759 0.743 3 0.78 0.768 4 0.749 0.73 4 0.735 0.739 5 0.797 0.802 5 0.803 0.813 6 0.773 0.764 6 0.778 0.773 7 0.743 0.74 7 0.733 0.727 8 0.768 0.784 8 0.752 0.775 9 0.721 0.778 9 0.711 0.767 10 0.735 0.708 10 0.736 0.735 11 0.754 0.795 11 0.754 0.796 12 0.708 0.729 12 0.712 0.731 13 0.742 0.768 13 0.741 0.753 p-value p-value Average 0.750 0.756 0.53 Average 0.750 0.761 0.108 (A) represents patients evaluated using the category scores. (B) represents the patients evaluated using the Likelihood of Malignancy score. AI, artificial intelligence; AUROC, area under the receiver operating characteristic curve. Table 5. MMG facilities Corporate name Product name number Canon medical systems MAMMOREX Pe・ru・ru DIGITAL 7 Carestream Health unknown 1 FUJIFILM unknown 20 GE HealthCare Senographe Essential 216 GE HealthCare (Japan) Senographe Pristina 2 Senographe DS 15 Hologic Lorad Selenia 3 Hologic Japan Selenia Dimentions 22 KONICA MINOLTA unknown 2 Siemens Mammomat Inspiration 5 Unknown 45 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 Advisory Board 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-4855505","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":348819334,"identity":"33116aeb-dc0a-40d0-a2ef-d638a6a38301","order_by":0,"name":"Maya Makita","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIiWNgGAWjYNACHgYGNmbm4x8+ANls7ERqkeBjb0tjnAHSwkykPRJyPGfMmHlATEJadNvPGH5gkDlcxyaRYPbY5tc2eT5mBsYPH3NwazE7k2MswcBzWAKoJd04t++2YRszA7PkzG14tBzIMYBpOSCd23ObEaiFjZkXn5bzb4x/QLQkNkhb9ty2J6zlRo4ZxBaew2zSDD9uJxKh5VmZRQJPumQbexuzYW/D7eQ2ZsZm/H45n7z5xscea375Zv6PD378uW07v7354IePeLQwMHAYMCT2QNmMbWCyAZ96IGB/wMDwA8b5Q0DxKBgFo2AUjEgAAI7qTBVnX8NsAAAAAElFTkSuQmCC","orcid":"","institution":"Showa University","correspondingAuthor":true,"prefix":"","firstName":"Maya","middleName":"","lastName":"Makita","suffix":""},{"id":348819335,"identity":"75f4e7ce-3ab2-4e43-9961-deaf372cfaf5","order_by":1,"name":"Kouzou Murakami","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Kouzou","middleName":"","lastName":"Murakami","suffix":""},{"id":348819339,"identity":"800c0e1c-ac81-41ed-8952-fe21f013d004","order_by":2,"name":"Wakana Murakami","email":"","orcid":"","institution":"Nanahoshi clinic","correspondingAuthor":false,"prefix":"","firstName":"Wakana","middleName":"","lastName":"Murakami","suffix":""},{"id":348819340,"identity":"c1060aa1-e5ef-4623-90c5-7743836a6b45","order_by":3,"name":"Hiroko Takamatsu","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Hiroko","middleName":"","lastName":"Takamatsu","suffix":""},{"id":348819341,"identity":"393575a4-85e0-4976-9ae4-b673250d03af","order_by":4,"name":"Takahiro Kanai","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Takahiro","middleName":"","lastName":"Kanai","suffix":""},{"id":348819342,"identity":"5cbe1bf9-b342-4410-9382-49f65732bcc9","order_by":5,"name":"Atsuhito Sekimoto","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Atsuhito","middleName":"","lastName":"Sekimoto","suffix":""},{"id":348819346,"identity":"917a50d1-f5d8-4ce1-9359-6c3fe01517e4","order_by":6,"name":"Yoshinori Ito","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Yoshinori","middleName":"","lastName":"Ito","suffix":""},{"id":348819347,"identity":"2ae8ce2e-4763-4bcd-8296-b7c93c97e2e8","order_by":7,"name":"Yoshimitsu Ohgiya","email":"","orcid":"","institution":"Showa University","correspondingAuthor":false,"prefix":"","firstName":"Yoshimitsu","middleName":"","lastName":"Ohgiya","suffix":""}],"badges":[],"createdAt":"2024-08-04 05:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4855505/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4855505/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":66119029,"identity":"369bc25c-5aa8-4a9b-a8f7-443baea02ea0","added_by":"auto","created_at":"2024-10-08 01:11:44","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":33717,"visible":true,"origin":"","legend":"\u003cp\u003eA 30-year-old female with right invasive ductal carcinoma. \u0026nbsp;\u0026nbsp;The breast tissue exhibited high density, particularly in the craniocaudal \u0026nbsp;\u0026nbsp;view. Identifying the lesion containing calcifications was challenging or had \u0026nbsp;\u0026nbsp;low confidence. Artificial intelligence assistance improved the Likelihood of \u0026nbsp;\u0026nbsp;Malignancy score in 8 of 13 readings. The magnification and heat color map show \u0026nbsp;suspected malifnancy.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4855505/v1/10286274fcacc1ddd7f1250d.jpg"},{"id":66117628,"identity":"eb56414c-a393-43f5-aa1f-9c15e0ae270f","added_by":"auto","created_at":"2024-10-08 00:55:45","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":34962,"visible":true,"origin":"","legend":"\u003cp\u003eA 69-year-old female with invasive ductal carcinoma in the \u0026nbsp;\u0026nbsp;left lower inner quadrant of the breast. Artificial intelligence observations \u0026nbsp;\u0026nbsp;identified findings suggestive of malignancy in two individuals, while the \u0026nbsp;\u0026nbsp;categorization was downgraded from Category 3 to 1 in three individuals. Arrow point the suspected \u0026nbsp;\u0026nbsp;malignancy.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4855505/v1/15075fa1367ffc84e2d08c37.jpg"},{"id":66117625,"identity":"c6647590-2d09-43c0-b050-1f42bb02c7aa","added_by":"auto","created_at":"2024-10-08 00:55:44","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":67179,"visible":true,"origin":"","legend":"\u003cp\u003ePatients dataset (malignant and benign)\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4855505/v1/91c8a217e50f74260d92fe94.jpg"},{"id":66117626,"identity":"64d0d910-729d-4a9e-8677-92a3d7c22b42","added_by":"auto","created_at":"2024-10-08 00:55:45","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":48469,"visible":true,"origin":"","legend":"\u003cp\u003ePatients dataset (normal)\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4855505/v1/77271a2a324462617fe4a3de.jpg"},{"id":73890685,"identity":"634796e7-a633-450b-9416-59623d0a4a06","added_by":"auto","created_at":"2025-01-15 15:32:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1012201,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4855505/v1/8756a104-0302-4269-bf12-f2a60f526bf2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Evaluation of the Applicability of an Artificial Intelligence System for Mammography Analysis Trained on Overseas Data for Japanese Domestic Data","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBreast cancer accounts for 15.5% of cancer-related deaths in women worldwide. However, early detection and treatment have significantly improved prognosis [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. In recent years, artificial intelligence (AI) products have been developed to assist the interpreting of digital mammography images, with the expectation of enhancing diagnostic performance [\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. However, many of these AI systems have been developed using data collected outside Japan.\u003c/p\u003e \u003cp\u003eWhile commercial software such as HOLOGIC (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.hologic.com/hologic-products/breast-health-solutions/image-analytics\u003c/span\u003e\u003cspan address=\"https://www.hologic.com/hologic-products/breast-health-solutions/image-analytics\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), iCAD (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.icadmed.com/breast-health/ai-breast-cancer-detection/\u003c/span\u003e\u003cspan address=\"https://www.icadmed.com/breast-health/ai-breast-cancer-detection/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and CureMetrix (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://curemetrix.com/\u003c/span\u003e\u003cspan address=\"https://curemetrix.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) have been approved overseas, there are limited reports on their usefulness, specifically in the Japanese population.\u003c/p\u003e \u003cp\u003eThe 2022 Japanese breast cancer treatment guidelines state that the use of AI-based Computer-Aided Diagnosis (AI-CAD) for image interpretation does not conclusively improve accuracy in Japan [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. The Lunit INSIGHT for Mammography (v1.1.7.2; Lunit, South Korea) is an AI system designed exclusively for image interpretation in digital mammography using large-scale data. Previous studies have shown that it outperforms radiologists in detecting breast cancer from mammography images, and significantly improves diagnostic accuracy when combined with AI assistance [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan additionalcitationids=\"CR7 CR8 CR9\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Although the Lunit INSIGHT has been trained on data from various countries, its performance has not yet been explicitly validated in Japanese female populations. Hence, this study aimed to assess the breast cancer detection performance of the Lunit INSIGHT for mammography images of Japanese women and to investigate whether its use contributes to improved diagnostic accuracy for physicians.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eAI-CAD Diagnostic Accuracy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePatient characteristics are presented in Table 1. The study included 338 patients with 676 breasts as the target data for validation. For each case, the distribution of background breast tissue was as follows: in malignant cases, 2 cases were fatty, 55 cases had scattered fibroglandular densities, 54 cases had heterogeneously dense tissue, and 8 cases had extremely dense tissue; in benign cases, 1 case was fatty, 29 cases had scattered fibroglandular densities, 62 cases had heterogeneously dense tissue, and 3 cases had extremely dense tissue; in normal cases, 1 case was fatty, 29 cases had scattered fibroglandular densities, 94 cases had heterogeneously dense tissue, and 0 cases had extremely dense tissue.\u003c/p\u003e\n\u003cp\u003eThe following results were based on per-breast counts. Among the malignant cases, 102 were invasive carcinomas, 22 were ductal carcinoma in situ (DCIS), 1 was papillary carcinoma (encapsulated), 1 was a special type, and 1 was challenging to determine as invasive or noninvasive. When summarizing the features across all breasts, in terms of masses, 50 were cancerous and 24 were non-cancerous; in terms of asymmetry, 34 were cancer and 13 were not cancer; in terms of architectural distortion, 62 were cancerous and 3 were non-cancerous; and in terms of calcifications requiring assessment of malignancy, 55 were cancerous and 26 were non-cancerous.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; In the diagnostic interpretations by the readers, the number of malignant cases classified as negative and positive were 10 (8%) and 117 (92%) cases, respectively, whereas for non-malignant cases, the number of negative and positive were 497 (91%) and 52 (9%) cases, respectively. In the AI-CAD system, the number of malignant cases classified as negative and positive were 27 (21%) and 100 (79%) cases, respectively, whereas for non-malignant cases, the number of negative and positive were 493 (90%) and 56 (10%) cases, respectively. The diagnostic performance of the AI-CAD for validation data is presented in Table 2. Their sensitivity, specificity, and AUROC of the AI-CAD system were 0.787, 0.894, and 0.897, respectively. After excluding 10 cases of malignancy judged as negative by GS, the sensitivity, specificity, and AUROC for the remaining data were 0.838, 0.894, and 0.930, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReading Experiment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe characteristics of the patient groups used for reading are presented in Table 3. For individual cases, the background breast density for malignant cases comprised of 0 cases of fatty tissue, 7 cases of scattered fibroglandular density, 12 cases of heterogeneous high density, and 2 cases of extremely high density. For benign cases, there were 0 cases of fatty tissue, 4 cases of scattered fibroglandular densities, 5 cases of heterogeneous high density, and 0 cases of extremely high density. For normal cases, there were no cases of fatty tissue, three cases of scattered fibroglandular densities, seven cases of heterogeneous high density, and none of extremely high density. When summarizing the features of the findings across all breasts, there were 7 cases of masses with cancer, 3 cases of masses with no cancer, and 7 findings of asymmetry, 11 findings of architectural distortion, and 10 findings of calcifications requiring assessment, all with cancers. Among the malignant cases, there were 15 cases of invasive ductal carcinoma, 7 cases of DCIS, and 1 case of special type carcinoma. We also compared the AUROC values based on category classification and LOM scores before and after using AI-CAD (Table 4). The AUROC values based on the category assessment before and after using the AI-CAD system were 0.750 and 0.756, respectively, whereas those based on the LOM assessment were 0.750 and 0.761, respectively. Although no significant differences were observed, the AUROC values generally increased after using the AI-CAD system compared to before its use. Representative examples wherein the judgment of the readers were changed are shown in Figures 1 and 2.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eBased on applying the AI-CAD system to the verification data, it is anticipated that, despite the use of AI-CAD, detecting tricky cases visually will still be challenging. However, when such cases were excluded, the performance of the AI-CAD system was comparable to that of the radiologists in terms of detection capability. Specifically, when cases in which radiologists visually judged lesions as challenging to identify were excluded, the diagnostic performance of the AI-CAD system showed a sensitivity of 84% and a specificity of 90%. Although the version was different, this performance was similar to that of a previous study that used the Lunit INSIGHT MMG, reporting a sensitivity of 82.1% and a specificity of 90.3% [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Therefore, overseas-approved products can be utilized in Japan, but further research is essential.\u003c/p\u003e \u003cp\u003eIn the reading experiment, we investigated whether the use of AI-CAD improves the detection performance of radiologists. Despite no significant difference in the change in detection performance due to AI-assistance, the radiologists detected them with greater sensitivity, specificity, and AUROC values on average, indicating improved detection performance. Notably, cases with calcifications in the lesions were more manageable to judge by the radiologists. However, in cases where the background breast tissue had a high density and calcifications were easily overlooked, the AI-CAD was efficient in assisting the radiologists.\u003c/p\u003e \u003cp\u003eHowever, there are points of reflection in the case shown in Fig.\u0026nbsp;2. While radiologists could identify findings in both the craniocaudal and mediolateral oblique (MLO) views, their confidence in malignancy appeared low. In contrast, the AI only indicated findings in the MLO view and provided a low abnormality score of 12. Two radiologists changed the category from 1 to 3, and three changed it from 3 to 1. This may be due to their interpretation of the image as part of the normal structures, possibly due to low abnormality score and unidirectional indications. However, considering the actual findings, the AI-CAD system should not have downgraded this case to Category 1, as this would have been considered a positive judgment. Cases in which radiologists are uncertain based on individual findings may lead to differing judgments based on their understanding and trust in the AI-CAD system. To achieve results consistent with the AI-CAD, emphasizing the interpretation of abnormal scores and providing training for such judgments may be necessary.\u003c/p\u003e \u003cp\u003eThis case-control study may not be the most appropriate evaluation method for the AI-CAD system, which is primarily intended for screening purposes. In this study, a significant proportion of malignant and benign cases was evaluated based on pathological results, and there were many cases in which assessment based solely on images was challenging. In the reading experiment conducted in this study, no specific practice time was provided for using the AI-CAD system in conjunction with reading. There was variability in the participants\u0026rsquo; experiences with AI-CAD, and observing consistent trends in diagnostic accuracy improvement was challenging. Additionally, the limited number of cases in the reading experiments may have contributed to this variation. Feedback from the participants indicated that the AI-CAD system increased confidence in diagnosing malignancy in cases suspected to be malignant or decreased confidence in diagnosing non-malignant lesions. We did not precisely measure the reading time or impose time constraints on the change in confidence levels before and after using the AI-CAD system. Therefore, timesaving effects were not evaluated in this study.\u003c/p\u003e \u003cp\u003eThe application of the AI-CAD system trained on overseas data to Japanese data demonstrated comparable effectiveness. Further research, including larger sample sizes, long-term studies, and comparisons of reading times in real-world reading environments, may bring us closer to the clinical implementation of AI-CAD for mammography screening in Japan.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e This retrospective study was approved by the Ethics Review Committee of Showa University (Approval Number: 3426). The research used digital mammography images obtained from patients who visited the Showa University Hospital Breast Surgery Department. Images were collected retrospectively. We used the mammography category classification and background breast evaluation criteria for image diagnosis, based on the Breast Imaging Reporting and Data System (BI-RADS) categories of the American College of Radiology [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. All research procedures were performed in accordance with the Declaration of Helsinki and local regulations.\u003c/p\u003e\n\u003ch3\u003eComputer-Aided Diagnosis Software\u003c/h3\u003e\n\u003cp\u003eThe AI-CAD system used in this study is the Lunit INSIGHT MMG. It detects regions suggestive of breast cancer on mammograms and marks areas indicative of malignant lesions. The system displays an abnormality score (range, 0\u0026ndash;100) for qualitative assessment, which assists radiologists in the diagnosis. The Lunit INSIGHT MMG was developed based on a convolutional neural network, utilizing training data from over 200,000 cases analyzed in South Korea, United States, and United Kingdom.\u003c/p\u003e\n\u003ch3\u003ePatients and Dataset\u003c/h3\u003e\n\u003cp\u003eWe collected patient data from two groups at the Showa University Hospital. Group A contains the surgical and biopsy specimens of breast lesions obtained between January and December 2019 (2,813 cases). Group B contains digital mammography images obtained at the Showa University Hospital between the same time period (4,390 cases).\u003c/p\u003e \u003cp\u003eThe collected data were categorized into three groups: malignant, benign, and normal. We selected patients from Group A who underwent pathological examination for breast lesions and reached a final diagnosis of either malignant or benign. For Group B, we classified patients as normal if both sides were diagnosed as BI-RADS Category 1 by the interpreting radiologist and were re-evaluated as Category 1 again between January and December 2021.\u003c/p\u003e \u003cp\u003eWe excluded patients who met the following criteria: 1) male patients; 2) patients with imaging of only one breast; and 3) patients with a history of breast surgery, including breast augmentation. In addition, for both the malignant and benign groups, we excluded cases based on the following criteria: 1) pathological examination was performed on non-breast tissue, 2) pathological diagnosis was made after neoadjuvant chemotherapy, and 3) the pathological diagnosis was difficult to determine as benign or malignant. The number of cases collected according to criteria was 821, 380, and 179 in the malignant, benign, and normal groups, respectively. Among these, 150 cases of image data were extracted from each group in ascending order of examination date, with further exclusion of cases due to specific reasons, such as the inability to analyze images from computed radiography or a time gap of more than 6 months between pathological images. The final dataset included 119, 95, and 124 cases in the malignant, benign, and normal groups, respectively (Figs.\u0026nbsp;3 and 4).\u003c/p\u003e \u003cp\u003eIt is important to note that patients in the malignant and benign groups used digital mammography images from the most recent examination date before surgery or biopsy, including those taken at other facilities (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eValidation of the AI-CAD System\u003c/h2\u003e \u003cp\u003eTwo radiologists with 11 and 2 years of experience validated the collected data. First, they referred to clinical information and breast magnetic resonance imaging findings that could be confirmed at our hospital to create the gold standard data (GS) for the mammography evaluation of each case. Next, they compared the GS with the results of the AI-CAD system to verify whether the regions showing significant abnormality scores in the AI-CAD matched the lesions indicated by the GS. The background breast density for each case was assessed by three radiologists with 11, 6, and 6 years of experience.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eReading Experiment\u003c/h2\u003e \u003cp\u003eFrom the verification data, 40 cases were randomly selected (21 malignant, 9 benign, and 10 normal). Seven radiologists with mammography reading experience (3 radiologists specializing in diagnostic radiology with 21, 16, and 8 years of experience; 2 radiologists specializing in radiology with 6 years of experience each; and 2 radiology residents with 4 and 3 years of experience) and six breast surgeons (with 23, 18, 11, 3, 3, and 2 years of experience) participated. Each participant read the same 40 cases independently. Without referring to the AI-CAD evaluations, the participants assessed the findings, calculated using the Likelihood of Malignancy (LOM) scores (11-point scoring system), and categorized both breasts based solely on their interpretation. Participants then reviewed the AI-CAD evaluations and repeated the same assessment. Readers repeated steps 1 and 2 for each case without time constraints.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eFor mammography evaluations by radiologists, the category scores of 1 and 2 were classified as \u0026ldquo;negative\u0026rdquo; and 3, 4, and 5 were \u0026ldquo;positive.\u0026rdquo; Based on prior research, the abnormality score cutoff of the AI-CAD system was set at 10, which classifies scores\u0026thinsp;\u0026lt;\u0026thinsp;10 as \u0026ldquo;negative\u0026rdquo; and scores\u0026thinsp;\u0026ge;\u0026thinsp;10 as \u0026ldquo;positive.\u0026rdquo; We calculated the sensitivity and specificity using the true positive, which was defined as malignant according to the final pathological diagnosis for each breast. Diagnostic performance was analyzed using a receiver operating characteristic (ROC) curve. The evaluation of the reading experiments assessed whether there was a change in the area under the ROC (AUROC) curve for each reader before and after reading using the Wilcoxon signed-rank test. Statistical analysis was performed using the EZR software version 1.62, and an extension of the R software and R Commander [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Statistical significance was set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003e \u003cb\u003eAdditional information\u003c/b\u003e \u003c/h2\u003e \u003cp\u003e \u003cstrong\u003eCompeting Interests:\u003c/strong\u003e \u003cp\u003eThis research was funded by FUJIFILM.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eEthical Approval:\u003c/strong\u003e \u003cp\u003eThis retrospective study was approved by the Ethics Review Committee of Showa University (Approval Number: 3426).\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eInformed Consent:\u003c/strong\u003e \u003cp\u003e Informed consent was obtained from all individual participants included in the study.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003econceptualization；M.M. and K.M.data curation; M.M. , K.M., H.T., T.K., and A.S. Hiroko Takamatsu, Takahiro Kanai, Atsuhito Sekimotoformal analysis; M.Minvestigation; all authorsproject administration; M.M.supervision; K.M.Writing \u0026ndash; review \u0026amp; editing, all authors.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe wish to thank Mr. Ryuji Hisanaga and Ms. Chiori Murata (FUJIFILM Corporation) for their useful help and contributions in this research. We also would like to thank Editage (www.editage.jp) for English language editing.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSung, H. \u003cem\u003eet al\u003c/em\u003e. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71, 209\u0026ndash;249 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSechopoulos, I., Teuwen, J. \u0026amp; Mann, R. Artificial Intelligence for breast cancer detection in mammography and digital breast tomosynthesis: State of the art. Semin. Cancer Biol. 72, 214\u0026ndash;225 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSechopoulos, I. \u0026amp; Mann, R. M. Stand-alone artificial intelligence - The future of breast cancer screening? Breast 49, 254\u0026ndash;260 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim, H. E. \u003cem\u003eet al\u003c/em\u003e. Changes in cancer detection and false-positive recall in mammography using artificial intelligence: A retrospective, multireader study. Lancet Digit. Health 2, e138\u0026ndash;e148 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJapan breast cancer society. \u003cem\u003eIs It Useful to Use AI Software for Image Interpretation in Mammography Breast Cancer Screening (FRQ2)? Breast Cancer Treatment Guidelines\u003c/em\u003e, (2022) Edition. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://jbcs.xsrv.jp/guideline/2022/k_index/frq2/\u003c/span\u003e\u003cspan address=\"https://jbcs.xsrv.jp/guideline/2022/k_index/frq2/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark, G. E., Kang, B. J., Kim, S. H. \u0026amp; Lee, J. Retrospective review of missed cancer detection and its mammography findings with artificial-intelligence-based, computer-aided diagnosis. Diagnostics (Basel) 12 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDembrower, K. \u003cem\u003eet al\u003c/em\u003e. Effect of artificial intelligence-based triaging of breast cancer screening mammograms on cancer detection and radiologist workload: A retrospective simulation study. Lancet Digit. Health 2, e468\u0026ndash;e474 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalim, M. \u003cem\u003eet al\u003c/em\u003e. External evaluation of 3 commercial artificial intelligence algorithms for independent assessment of screening mammograms. JAMA Oncol. 6, 1581\u0026ndash;1588 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, J. H. \u003cem\u003eet al\u003c/em\u003e. Improving the performance of radiologists using artificial intelligence-based detection support software for mammography: A multi-reader study. Korean J. Radiol. 23, 505\u0026ndash;516 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFreeman, K. \u003cem\u003eet al\u003c/em\u003e. Use of artificial intelligence for image analysis in breast cancer screening programmes: Systematic review of test accuracy. BMJ 374, n1872 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUICC. \u003cem\u003eTNM Classification of Malignant Tumours\u003c/em\u003e. 8th ed. Wiley-Blackwell (U.I.C.C. International Union Against Cancer, 2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoon, J. H. \u003cem\u003eet al.\u003c/em\u003e Artificial intelligence-based computer-assisted detection/diagnosis (AI-CAD) for screening mammography: Outcomes of AI-CAD in the mammographic interpretation workflow. Eur. J. Radiol. Open 11, 100509 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmerican College of Radiology. \u003cem\u003eACR BI-RADS Atlas\u003c/em\u003e. 5th ed. (American College of Radiol., Reston, VA, 2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanda, Y. Investigation of the freely available easy-to-use software \u0026ldquo;EZR\u0026rdquo; for medical statistics. Bone Marrow Transplant. 48, 452\u0026ndash;458 (2013).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1.\u003c/strong\u003e Patient demographics (verification data)\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003eMalignant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003eBenign\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003eNormal\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eN\u003c/em\u003e=338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=119\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=124\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eAge group, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026le;50 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e159 (47)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e58 (49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e71 (75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e30 (24)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;51\u0026ndash;65 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e112 (33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e34 (29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e19 (20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e59 (48)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026gt;65 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e67 (20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e27 (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e5 (5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e35 (28)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eDensity grade, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Fatty\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e4 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e2 (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e1 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e1 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Scattered fibrolandular dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e113 (33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e55 (46)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e29 (31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e29 (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Heterogeneous dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e210 (62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e54 (45)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e62 (65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e94 (76)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Extremely dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e11 (3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e8 (7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e3 (3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"100%\" colspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003eCharacteristics of the imaging features of the breasts\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003eCancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003eNot cancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eN\u003c/em\u003e=676\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=127\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=549\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eInterpretation of the radiologists, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Category 1, 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e507 (75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e10 (8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e497 (91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Category 3, 4, 5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e169 (25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e117 (92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e52 (9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eAI-CAD abnormality score, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026lt;10%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e520 (77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e27 (21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e493 (90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026ge;10％\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e156 (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e100 (79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e56 (10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eHistology, \u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Invasive cancers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e102 (80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003eDCIS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e22 (17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003ePapillary carcinoma (encapsulated)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e1 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"58.657243816254415%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eSpecial type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e1 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"58.657243816254415%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eOthers \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e1 (1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"100%\" colspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003eMammography findings across all breasts\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003eCancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003eNot cancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Mass, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Asymmetry, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Distortion, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"45.40636042402827%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Calcifications, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.250883392226148%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.07773851590106%\" valign=\"top\"\u003e\n \u003cp\u003e55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.190812720848056%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003csup\u003ea\u0026nbsp;\u003c/sup\u003eChallenging\u0026nbsp;to determine as invasive or noninvasive\u003c/p\u003e\n\u003cp\u003eAI-CAD, Artificial Intelligence-based Computer-Aided Diagnosis system; DCIS, ductal carcinoma in situ.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"400\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"100%\" colspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 2.\u0026nbsp;\u003c/strong\u003eDiagnostic performance of the AI-CAD for validation data\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"37.34335839598997%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"31.829573934837093%\"\u003e\n \u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.827067669172934%\"\u003e\n \u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"37.34335839598997%\"\u003e\n \u003cp\u003eSensitivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"31.829573934837093%\"\u003e\n \u003cp\u003e0.787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.827067669172934%\"\u003e\n \u003cp\u003e0.838\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"37.34335839598997%\"\u003e\n \u003cp\u003eSpecificity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"31.829573934837093%\"\u003e\n \u003cp\u003e0.894\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.827067669172934%\"\u003e\n \u003cp\u003e0.894\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"37.34335839598997%\"\u003e\n \u003cp\u003eAUROC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"31.829573934837093%\"\u003e\n \u003cp\u003e0.897\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.827067669172934%\"\u003e\n \u003cp\u003e0.930\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e(A) represents the performance of AI-CAD for all data. (B) represents the\u0026nbsp;performance of AI-CAD for data excluding gold standard data-negative malignant cases.\u003c/p\u003e\n\u003cp\u003eAUROC, area under the receiver operating characteristic curve; AI-CAD, Artificial Intelligence-based Computer-Aided Diagnosis system.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"100%\" colspan=\"6\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 3.\u003c/strong\u003e Patient demographics (reading experiment data)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003eMalignant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003eBenign\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003eNormal\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eN\u003c/em\u003e=40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e=10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eAge group, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026le;50 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;51\u0026ndash;65 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026gt;65 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eDensity grade, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Fatty\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Scattered fibrolandular dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Heterogeneous dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Extremely dense\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"74.38162544169612%\" colspan=\"4\" valign=\"top\"\u003e\n \u003cp\u003eCharacteristics of the imaging features of the breasts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003eCancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003eHistology, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eInvasive cancers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.59717314487632%\" colspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eDCIS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eSpecial type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eUICC classifications, \u003cem\u003en\u003c/em\u003e \u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003eT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003eis(DCIS)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e1mi\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e1b\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e1c\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e4a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.59717314487632%\" colspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eMammography findings across all breasts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.380281690140846%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.809859154929576%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.380281690140846%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.725352112676056%\" valign=\"top\"\u003e\n \u003cp\u003eCancer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"25.704225352112676%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eNot cancer\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Mass, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"30.864197530864196%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.403880070546737%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.75485008818342%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.522045855379188%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.051146384479718%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Asymmetry, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Distortion, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"44.169611307420496%\" colspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Calcifications, \u003cem\u003en\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.42756183745583%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.784452296819786%\" valign=\"top\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.54416961130742%\" valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"13.074204946996467%\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003csup\u003ea\u0026nbsp;\u003c/sup\u003eThe\u0026nbsp;UICC TNM Classification of Malignant Tumors[11].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUICC, Union for International Cancer Control; DCIS, ductal carcinoma in situ; T, primary tumor site and size; N, regional lymph node involvement; M, metastasis.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"100%\" colspan=\"8\" valign=\"top\" style=\"width: 61.1458%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 4.\u0026nbsp;\u003c/strong\u003ePerformance readers with and without AI.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.46031746031746%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.869488536155202%\" colspan=\"2\" style=\"width: 15.8053%;\"\u003e\n \u003cp\u003eAUROC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.347442680776014%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.46031746031746%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"21.869488536155202%\" colspan=\"2\" style=\"width: 15.8053%;\"\u003e\n \u003cp\u003eAUROC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.347442680776014%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eNo.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003ewithout AI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003ewith AI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eNo.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003ewithout AI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003ewith AI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.765\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.754\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.768\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.763\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.732\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.732\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.748\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.754\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.759\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.743\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.768\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.749\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.735\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.739\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.797\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.802\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.803\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.813\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.773\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.764\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.778\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.773\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.743\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.733\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.727\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.768\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.752\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.775\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.721\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.778\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.711\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.767\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.735\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.708\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.736\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.735\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.754\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.795\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.754\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.796\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.708\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.729\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.712\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.731\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.742\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.768\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.741\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.753\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" valign=\"top\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" valign=\"top\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" valign=\"top\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" valign=\"top\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" valign=\"top\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" valign=\"top\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" valign=\"top\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eAverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.750\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.756\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.398945518453427%\" style=\"width: 12.5928%;\"\u003e\n \u003cp\u003eAverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.478031634446397%\" style=\"width: 8.9949%;\"\u003e\n \u003cp\u003e0.750\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.490333919156415%\" style=\"width: 6.8104%;\"\u003e\n \u003cp\u003e0.761\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"9.314586994727593%\" style=\"width: 6.6819%;\"\u003e\n \u003cp\u003e0.108\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e(A) represents patients evaluated using the category scores. (B) represents the patients evaluated using the Likelihood of Malignancy score.\u003c/p\u003e\n\u003cp\u003eAI, artificial intelligence; AUROC, area under the receiver operating characteristic curve.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 5.\u003c/strong\u003e MMG facilities\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"525\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eCorporate name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\" valign=\"top\"\u003e\n \u003cp\u003eProduct name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003enumber\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eCanon medical systems\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eMAMMOREX Pe・ru・ru DIGITAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eCarestream Health\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003e\u003cem\u003eunknown\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eFUJIFILM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003e\u003cem\u003eunknown\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\"\u003e\n \u003cp\u003eGE\u0026nbsp;HealthCare\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eSenographe Essential\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e216\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\"\u003e\n \u003cp\u003eGE\u0026nbsp;HealthCare (Japan)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eSenographe Pristina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eSenographe DS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\"\u003e\n \u003cp\u003eHologic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eLorad Selenia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eHologic Japan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eSelenia Dimentions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\"\u003e\n \u003cp\u003eKONICA MINOLTA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003e\u003cem\u003eunknown\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003eSiemens\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\"\u003e\n \u003cp\u003eMammomat Inspiration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"35.42857142857143%\" valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eUnknown\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"52.19047619047619%\" valign=\"top\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"12.380952380952381%\" valign=\"top\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\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":"[email protected]","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":"Artificial Intelligence, Computer-Aided Diagnosis system, Japan, mammography, breast cancer","lastPublishedDoi":"10.21203/rs.3.rs-4855505/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4855505/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study evaluated the performance of the Artificial Intelligence (AI)-based Computer-Aided Diagnosis system (AI-CAD), Lunit INSIGHT MMG, in detecting breast cancer from digital mammography images of Japanese women. We collected digital mammography images from two groups at Showa University Hospital. One group consisted of surgical and biopsy specimens of breast lesions between January and December 2019, and the other was digital mammography images taken at Showa University Hospital during the same period. The AI-CAD system was developed based on a convolutional neural network trained on over 200,000 cases, overseas of Japan. We analyzed the breast cancer detection capabilities and compared the results with the interpretations of the radiologists and breast surgeons. We used the area under the receiver operating characteristic (AUROC) curve to evaluate the data. We evaluated the performance of the Lunit INSIGHT MMG using a dataset of 676 breasts from 338 patients. Although no significant overall difference was observed, the radiologists reported increased sensitivity, specificity, and AUROC values, on average. The AI-CAD system trained on overseas data showed comparable effectiveness with Japanese data.\u003c/p\u003e","manuscriptTitle":"Evaluation of the Applicability of an Artificial Intelligence System for Mammography Analysis Trained on Overseas Data for Japanese Domestic Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-08 00:55:39","doi":"10.21203/rs.3.rs-4855505/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","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":"66c944e8-021e-48a2-9d84-ee566146298a","owner":[],"postedDate":"October 8th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":37013075,"name":"Biological sciences/Cancer/Cancer imaging"},{"id":37013076,"name":"Health sciences/Health care/Diagnosis"},{"id":37013077,"name":"Health sciences/Medical research"}],"tags":[],"updatedAt":"2025-01-15T15:23:58+00:00","versionOfRecord":[],"versionCreatedAt":"2024-10-08 00:55:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4855505","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4855505","identity":"rs-4855505","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}