Artificial intelligence, machine learning, computer-aided diagnosis, and radiomics: advances in imaging towards to precision medicine
Open accessAbstract
Abstract The discipline of radiology and diagnostic imaging has evolved greatly in recent years. We have observed an exponential increase in the number of exams performed, subspecialization of medical fields, and increases in accuracy of the various imaging methods, making it a challenge for the radiologist to “know everything about all exams and regions”. In addition, imaging exams are no longer only qualitative and diagnostic, providing now quantitative information on disease severity, as well as identifying biomarkers of prognosis and treatment response. In view of this, computer-aided diagnosis systems have been developed with the objective of complementing diagnostic imaging and helping the therapeutic decision-making process. With the advent of artificial intelligence, “big data”, and machine learning, we are moving toward the rapid expansion of the use of these tools in daily life of physicians, making each patient unique, as well as leading radiology toward the concept of multidisciplinary approach and precision medicine. In this article, we will present the main aspects of the computational tools currently available for analysis of images and the principles of such analysis, together with the main terms and concepts involved, as well as examining the impact that the development of artificial intelligence has had on radiology and diagnostic imaging.
AI snapshot
Written from the paper’s full text.
TL;DR
This narrative review describes the computational tools used for medical image analysis—computer-aided diagnosis (CAD), content-based image retrieval (CBIR), machine learning/deep learning, and radiomics/radiogenomics—and argues they are moving radiology toward precision medicine. It presents key terms and principles rather than original data.
Key findings
No primary outcomes are reported; the paper is a conceptual review. Points it makes:
- CAD systems were first developed in the late 1980s from digitized radiography films, initially as a "second reading" for population-based cancer screening (nodules and microcalcifications on mammograms), later extended to pulmonary nodules on radiography/CT and to Alzheimer's disease in nuclear medicine.
- At the 2018 RSNA Annual Meeting there were 237 registered events related to artificial intelligence, machine learning, or deep learning, illustrating rapid growth of the field.
- Deep learning minimizes the need for preprocessing/segmentation but has disadvantages: it requires very large image sets (hundreds to thousands), greater dependence on exam quality and clinical data, and a "processing black box" logic that is hard to interpret.
- Radiomics extends CAD by associating quantitative image features with patient data and outcomes; radiogenomics correlates radiological with genomic patterns and can quantify tumor heterogeneity.
Methods
Narrative review of the literature on computer-aided image analysis, artificial intelligence, machine learning, CAD, CBIR, and radiomics/radiogenomics. No systematic search strategy, dataset, or sample is described. The paper explains technical principles (pixel/voxel matrices, segmentation, feature extraction into gray-level, texture, and shape groups, feature selection, supervised vs. unsupervised training, ANN/CNN architecture, cross-validation, and ROC/AUC performance assessment) and provides a glossary of key terms (Table 1).
Limitations
Not stated in the paper; the article does not include a limitations section. Evident from the text: it is a non-systematic narrative review with no original data, no defined search methodology, and no quantitative synthesis; its claims about future impact are explicitly framed as the author's belief ("We believe that..."), and it notes unresolved ethical and legal issues related to medical exam liability.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
0 of 46 cited sources checked · 0 supported · 0 flagged
- Findingcitation
There were 237 registered events related to AI, ML, or deep learning at the 2018 Annual Meeting of the Radiological Society of North America.
“That year, there were 237 registered events related to artificial intelligence, machine learning, or deep learning (https://rsna2018.rsna.org/program/index.cfm).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites https://rsna2018.rsna.org/program/index.cfm
Not checked in the free run.
- Premisecitation
Analysis of exams that produce hundreds of images, such as CT and MRI, poses challenges even for experienced specialists.
“Radiological imaging can be extremely complex, and it is recognized that the analysis of exams that produce hundreds of images, such as computed tomography (CT) and magnetic resonance imaging (MRI), poses challenges, even for experienced specialists(1-3).”
Quote in the paperSubstantial claim0 of 3 sources checked- cites [1] Azevedo-Marques PM, 2017
Not checked in the free run.
- cites [2] Akgül CB, 2011, Content-based image retrieval in radiology: current status and future directions, J Digit Imaging source
Not checked in the free run.
- cites [3] Doi K, 2007, Computer-aided diagnosis in medical imaging: historical review, current status and future potential, Comput Med Imaging Graph source
Not checked in the free run.
- Premisecitation
Many studies have shown that imaging features are significantly associated with patterns of gene expression and genetic mutations.
“Many studies have shown that imaging features are significantly associated with patterns of gene expression and genetic mutations, demonstrating that radiogenomic analysis can identify different biological mechanisms by means of mathematical and computational devices, enabling the decoding of disease phenotypes by noninvasive methods(48,49).”
Quote in the paperSubstantial claim0 of 2 sources checked- cites [48] Gevaert O, 2017, Predictive radiogenomics modeling of EGFR mutation status in lung cancer, Sci Rep source
Not checked in the free run.
- cites [49] Aerts HJWL, 2014, Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach, Nat Commun source
Not checked in the free run.
- Premisereasoning
Deep learning minimizes the need for preprocessing or segmentation.
“Therefore, in deep learning technique, the need for preprocessing or segmentation is minimized.”
Quote in the paperNot a substantial claim - Premisecitation
Traditional machine learning methods have limitations in data processing, mainly related to need for segmentation and development of feature extractors.
“Traditional machine learning methods have limitations in data processing, mainly related to the need for segmentation and development of feature extractors to represent images and serve as input for the classifiers(34).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [34] LeCun Y, 2015, Deep learning, Nature source
Not checked in the free run.
- Premisecitation
Machine learning methods have been applied to classifying images acquired with various imaging modalities, using a variety of features, for various diseases, and with tools such as CAD and radiomics.
“Machine learning methods have been applied to classifying images acquired with various imaging modalities, using a variety of features, for various diseases, and with tools such as CAD and radiomics(25,29,30).”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [25] Giger ML, 2018, Machine learning in medical imaging, J Am Coll Radiol source
Not checked in the free run.
- cites [29] Erickson BJ, 2017, Machine learning for medical imaging, Radiographics source
Not checked in the free run.
- cites [30] Choy G, 2018, Current applications and future impact of machine learning in radiology, Radiology source
Not checked in the free run.
- Premisecitation
Radiomics has the potential to provide a robust approach to imaging analysis in a noninvasive, rapid, timely, and affordable manner.
“In view of recent advances in targeted treatment and immunotherapy, particularly in the treatment of malignancies, the need for a robust approach to imaging analysis has become clear, and radiomics has the potential to provide this in a noninvasive, rapid, timely, and affordable manner(12).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [12] Giardino A, 2017, Role of imaging in the era of precision medicine, Acad Radiol source
Not checked in the free run.
- Premisecitation
Texture features have become particularly important because they can reflect details of a lesion identified in an image.
“Texture features have become particularly important because they can reflect the details of a lesion identified in an image(1,2).”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [1] Azevedo-Marques PM, 2017
Not checked in the free run.
- cites [2] Akgül CB, 2011, Content-based image retrieval in radiology: current status and future directions, J Digit Imaging source
Not checked in the free run.
- Premisereasoning
Lesions in lung radiological images often have adjacent opacities or structures such as vessels and mediastinum, which can lead to poor segmentation and poorer characterization by shape features.
“Lesions seen in radiological images of the lung, for example, often have adjacent opacities or structures such as vessels and the mediastinum (Figure 5), which can lead to poor segmentation and, consequently, poorer characterization by shape features.”
Quote in the paperNot a substantial claim - Premisecitation
CAD tools were initially developed to aid interpretation of radiological findings and early identification of diseases, especially breast and lung cancer.
“Tools of CAD were initially developed to aid in the interpretation of radiological findings and early identification of diseases, especially breast and lung cancer(7,10).”
Quote in the paperNot a substantial claim0 of 2 sources checked - Premisecitation
Radiogenomics has been used in studies that analyze tumor heterogeneity, which is related to disease recurrence and treatment resistance.
“Radiogenomics has also been used in studies that analyze tumor heterogeneity (i.e., the presence of multiple tissue and genetic subregions within the same tumor), which is related to disease recurrence and treatment resistance.”
Quote in the paperNot a substantial claim
- May combine two claims. 12
- Classified as probably not a substantial claim. 13
May combine two claims.
PremisecitationIn oncology, imaging information can indicate histological type of tumor, staging, presence of mutations, chance of treatment response, risk of recurrence, and expected survival.
“Those changes have been especially prominent in oncology, showing that the information provided by imaging studies can go far beyond determining whether a lesion is benign or malignant, now being able to indicate the histological type of the tumor, staging, presence of mutations, chance of treatment response, risk of recurrence, and expected survival(7-9).”
Quote in the paperSubstantial claim0 of 3 sources checked- cites [7] Gillies RJ, 2016, Radiomics: images are more than pictures, they are data, Radiology source
Not checked in the free run.
- cites [8] Koenigkam Santos M, 2014, Morphological computed tomography features of surgically resectable pulmonary squamous cell carcinomas: impact on prognosis and comparison with adenocarcinomas, Eur J Radiol source
Not checked in the free run.
- cites [9] Cooper WA, 2011, What's new in non-small cell lung cancer for pathologists: the importance of accurate subtyping, EGFR mutations and ALK rearrangements, Pathology source
Not checked in the free run.
May combine two claims.
PremisecitationImaging exams now provide quantitative information on disease severity and identify biomarkers of prognosis and treatment response.
“In addition, imaging exams are no longer only qualitative and diagnostic, having begun to provide quantitative information on disease severity, as well as identifying biomarkers of prognosis and treatment response(4-6).”
Quote in the paperSubstantial claim0 of 3 sources checked- cites [4] Koenigkam-Santos M, 2013, Endobronchial valves in severe emphysematous patients: CT evaluation of lung fissures completeness, treatment radiological response and quantitative emphysema analysis, Radiol Bras source
Not checked in the free run.
- cites [5] Koenigkam-Santos M, 2016, Quantitative computed tomography analysis of the airways in patients with cystic fibrosis using automated software: correlation with spirometry in the evaluation of severity, Radiol Bras source
Not checked in the free run.
- cites [6] Wada DT, 2017, Use of computed tomography and automated software for quantitative analysis of the vasculature of patients with pulmonary hypertension, Radiol Bras source
Not checked in the free run.
May combine two claims.
PremisecitationCBIR can be a useful tool in daily clinical practice because it can aid radiologists in diagnostic interpretation or increase confidence level, using a decision model based on similar exams.
“It can be a very useful tool in daily clinical practice, because it can aid radiologists in diagnostic interpretation of exams, or increase their confidence level, because it uses a decision model based on similar exams(22).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [22] Ponciano-Silva M, 2013
Not checked in the free run.
May combine two claims.
PremisereasoningAI/ML tools have been applied to automated detection of pulmonary nodules, pneumonia, pneumothorax, and pleural effusion on conventional chest radiography; detection and quantification of emphysema, estimation of lung nodule malignancy risk, chance of local invasion by lung cancer, and cardiovascular risk on chest CT; and automated analysis of cardiac function on cardiac MRI.
“Examples include the following: automated detection of pulmonary nodules, pneumonia, pneumothorax, and pleural effusion on conventional chest radiography; detection and quantification of emphysema, estimation of lung nodule malignancy risk, chance of local invasion by lung cancer, and cardiovascular risk on chest CT; and automated analysis of cardiac function on cardiac MRI.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationThe volume of health data has been growing rapidly, characterizing a big data era of health, and electronic data are available in large quantities in information systems of large hospitals and other health care centers.
“The volume of health data has been growing at a rapid pace in recent years, characterizing what some authors call the “big data era” of health, and those electronic data are available in large quantities in information systems of large hospitals and other health care centers(45).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [45] Filonenko E, 2018, Big data: the next era of informatics and data science in medical imaging - a literature review, J Clin Exp Radiol
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationMachine learning allows identification of patterns seen in previous cases and experiments, as occurs with human intelligence.
“Machine learning allows the identification of patterns seen in previous cases and experiments, as occurs with human intelligence(29).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [29] Erickson BJ, 2017, Machine learning for medical imaging, Radiographics source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationSome authors have started to develop and suggest the use of a radiomic signature, in which most significant features of different categories are chosen, analyzed, and tested with accurate protocols and reproducible algorithms.
“Given the large number of features and numerous processing possibilities, some authors have started to develop and suggest the use of a radiomic signature, in which the most significant features of different categories are chosen, analyzed, and tested with accurate protocols and reproducible algorithms(46).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [46] Coroller TP, 2015, CT-based radiomic signature predicts distant metastasis in lung adenocarcinoma, Radiother Oncol source
Not checked in the free run.
May combine two claims.
InterpretationcitationRadiogenomic analysis can identify different biological mechanisms by means of mathematical and computational devices, enabling decoding of disease phenotypes by noninvasive methods.
“Many studies have shown that imaging features are significantly associated with patterns of gene expression and genetic mutations, demonstrating that radiogenomic analysis can identify different biological mechanisms by means of mathematical and computational devices, enabling the decoding of disease phenotypes by noninvasive methods(48,49).”
Quote in the paperSubstantial claim0 of 2 sources checked- cites [48] Gevaert O, 2017, Predictive radiogenomics modeling of EGFR mutation status in lung cancer, Sci Rep source
Not checked in the free run.
- cites [49] Aerts HJWL, 2014, Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach, Nat Commun source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.
InterpretationcitationWith AI and big data, we are moving toward reducing limitations of CAD, homogenizing and expanding use of CAD tools, making each patient unique, and leading radiology toward multidisciplinary approach and precision medicine.
“With the advent of artificial intelligence and “big data”, we are moving toward reducing those limitations, homogenizing and expanding the use of CAD tools in daily routine of physicians, making each patient unique, and leading radiology toward the concepts of a multidisciplinary approach and precision medicine(10-14).”
Quote in the paperNot a substantial claim0 of 5 sources checked- cites [10] Aerts HJWL, 2016, The potential of radiomic-based phenotyping in precision medicine: a review, JAMA Oncol source
Not checked in the free run.
- cites [11] Constanzo J, 2017, Radiomics in precision medicine for lung cancer, Transl Lung Cancer Res source
Not checked in the free run.
- cites [12] Giardino A, 2017, Role of imaging in the era of precision medicine, Acad Radiol source
Not checked in the free run.
- cites [13] Aerts HJWL, 2016, Defining a radiomic response phenotype: a pilot study using targeted therapy in NSCLC, Sci Rep source
Not checked in the free run.
- cites [14] Keek SA, 2018, A review on radiomics and the future of theranostics for patient selection in precision medicine, Br J Radiol source
Not checked in the free run.
- Rests on an assumption: There are still limitations to the use of CAD tools in routine clinical practice.
May combine two claims.
PremisecitationMany studies using radiomics in different types of imaging studies to evaluate different diseases exist; most are related to oncology, such as lung and kidney neoplasms on CT and PET/CT, and prostate cancer, breast cancer, glioma, and hepatocellular carcinoma on MRI.
“In the medical literature, there are already many studies using radiomics in different types of imaging studies to evaluate different diseases. Most of those studies are related to oncology, such as the study of lung and kidney neoplasms on CT and positron-emission tomography/CT scans, as well as prostate cancer, breast cancer, glioma, and hepatocellular carcinoma on MRI(7,14,36).”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [7] Gillies RJ, 2016, Radiomics: images are more than pictures, they are data, Radiology source
Not checked in the free run.
- cites [14] Keek SA, 2018, A review on radiomics and the future of theranostics for patient selection in precision medicine, Br J Radiol source
Not checked in the free run.
- cites [36] Lambin P, 2017, Radiomics: the bridge between medical imaging and personalized medicine, Nat Rev Clin Oncol source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe aim of CAD is to improve accuracy and consistency of diagnostic imaging by using suggestion of a response (second opinion) provided primarily by image processing, computer vision, and machine learning techniques.
“The aim of CAD is to improve the accuracy and consistency of diagnostic imaging by using suggestion of a response (“second opinion”) provided primarily by image processing, computer vision, and machine learning techniques(1,3,38).”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites [1] Azevedo-Marques PM, 2017
Not checked in the free run.
- cites [3] Doi K, 2007, Computer-aided diagnosis in medical imaging: historical review, current status and future potential, Comput Med Imaging Graph source
Not checked in the free run.
- cites [38] Ferreira Jr JR, 2017, Selecting relevant 3D image features of margin sharpness and texture for lung nodule retrieval, Int J Comput Assist Radiol Surg source
Not checked in the free run.
Classified as probably not a substantial claim. May combine two claims.
PremisecitationRadiomics associates quantitative characteristics of images with patient data and clinical outcomes, allowing diagnosis and providing information regarding prognosis and treatment response.
“Radiomics has been described as an extension of CAD that associates the quantitative characteristics (features) of images with patient data and clinical outcomes, not only allowing the diagnosis to be made but also providing information regarding the prognosis and treatment response(7,25).”
Quote in the paperSubstantial claim0 of 2 sources checkedProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationDeep learning has been gaining attention in the scientific community most recently.
“One of the areas of artificial intelligence that has been gaining attention in the scientific community most recently is deep learning (16).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [16] Bartholmai BJ, 2015, Pulmonary nodule characterization, including computer analysis and quantitative features, J Thorac Imaging source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationCBIR has been described as one of the most promising computational tools because it has potential for clinical application, teaching, and research.
“Because it has potential for clinical application, as well as for application in teaching and research, CBIR has been described as one of the most promising computational tool(42,43).”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [42] Ferreira Junior JR, 2015
Not checked in the free run.
- cites [43] Bugatti PH, 2014, PRoSPer: perceptual similarity queries in medical CBIR systems through user profiles, Comput Biol Med source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Premisenone givenTools employing AI, ML, deep learning have been applied primarily in CAD, CBIR, and radiomics/radiogenomics.
“In the field of radiology and diagnostic imaging, such tools have been applied primarily in CAD, content-based image retrieval (CBIR), and radiomics/radiogenomics.”
Quote in the paperNot a substantial claimMay combine two claims.
Premiseown dataWith development of AI and new ML tools, auxiliary diagnostic systems have expanded greatly and have been used in many different tasks, with all medical imaging modalities.
“With the development of artificial intelligence and new machine learning tools, auxiliary diagnostic systems have expanded greatly and have been used in many different tasks, with all medical imaging modalities.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationreasoningPossibilities for developing computerized auxiliary diagnostic tools have become almost unlimited.
“Therefore, possibilities for developing computerized auxiliary diagnostic tools have become almost unlimited(39,40).”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites [39] Nam JG, 2019, Development and validation of deep learning-based automatic detection algorithm for malignant pulmonary nodules on chest radiographs, Radiology source
Not checked in the free run.
- cites [40] Tao Q, 2019, Deep learning-based method for fully automatic quantification of left ventricle function from cine MR images: a multivendor, multicenter study, Radiology source
Not checked in the free run.
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisereasoningCharacterizing object shapes quantitatively is complicated because it depends on efficiency of segmentation algorithms.
“Characterizing object shapes quantitatively is a complicated task because it depends on efficiency of the segmentation algorithms.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Premisenone givenDigitization of radiological environments occurred first; then evolution of computer vision and AI led to development of auxiliary diagnostic systems; more recently, maturation of computational models has provided support to clinical decision-making and prognostic prediction.
“First, there was the digitization of radiological environments. Then, evolution of computer vision techniques and artificial intelligence led to the development of auxiliary diagnostic systems. More recently, maturation of computational models has provided support to the clinical decision-making and prognostic prediction processes.”
Quote in the paperNot a substantial claimMay combine two claims.
PremisecitationRadiogenomics is able to quantify spatial complexity of the tumor and identify phenotypic/genotypic subregions.
“Radiogenomics is able to quantify the spatial complexity of the tumor and identify these phenotypic/genotypic subregions(50) , as shown in Figure 10.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [50] Lee G, 2017, Radiomics and its emerging role in lung cancer research, imaging biomarkers and clinical management: state of the art, Eur J Radiol source
Not checked in the free run.
Probably not a substantial claim: may not pass S2 (checkable), S4 (based).
PremisereasoningRadiology has undergone significant advances due to the technological revolution.
“Radiology has undergone significant advances due to the technological revolution that is taking place in the world.”
Quote in the paperNot a substantial claim
The paper relies on this without support.
There are still limitations to the use of CAD tools in routine clinical practice.
“Although such tools have enormous potential, there are still limitations to their use in routine clinical practice.”
The paper relies on this without support.
The use of CBIR is more effective than searching the Internet via browser or specialized radiology sites for similar texts or cases.
“The use of CBIR takes the place of such strategies and is more effective, rapidly providing physicians with cases similar to that represented by the image for which they seek definition.”
The paper relies on this without support.
Deep learning has disadvantages: need for a very large set of images; greater dependence on exam quality and clinical data; and difficulty in identifying the logic used.
“However, the method also has disadvantages, such as the need for a very large set of images (hundreds to thousands); greater dependence on exam quality and clinical data; and difficulty in identifying the logic used (“processing black box”).”
The paper relies on this without support.
Those challenges have increased in recent years, with the exponential increase in the number of exams performed, subspecialization of medical fields, and increased accuracy of imaging methods, making it difficult for radiologists to know everything about all exams and regions.
“Those challenges have increased in recent years, with the exponential increase in the number of exams performed, subspecialization of medical fields, and increased accuracy of imaging methods, making it difficult for radiologists to “know everything about all exams and regions”.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Content-based image retrieval helps by searching for images with contents similar to a reference image using information derived from the images.
“CBIR helps by searching for images with contents similar to a reference image using information derived from the images [22].”
From Artificial Intelligence Revolutionizing the Field of Medical Education (2023) · cites this paper as [22]
The citation lists haven’t been fetched yet
References 53 · Cited by 214
- Authors
- Santos, M. K., Ferreira Júnior, J. R., Wada, D. T., Tenório, A. P. M., Nogueira-Barbosa, M. H., & Marques, P. M. D. A.
- Affiliations
- Universidade de São Paulo (BR)
- Universidade de São Paulo, Brazil; Universidade de São Paulo, Brazil
- Universidade de São Paulo, Brazil
- Corresponding author
- Santos, Marcel Koenigkam
- ORCID
- Venue
- Radiologia Brasileira
- Publisher
- FapUNIFESP (SciELO)
- Published
- 2019-12
- Volume
- 52
- Issue
- 6
- Pages
- 387-396
- ISSN
- 0100-3984 (ISSN-L)
- 1678-7099
- PMC
- PMC7007049
- OpenAlex
- W2977158709
- Semantic Scholar
- 654c066fa63a129c8c0d07ed37e762ad44765db5
- Cited by
- 214
- References
- 53
- Citations by year
- 2026: 15 · 2025: 37 · 2024: 35 · 2023: 41 · 2022: 35 · 2021: 31 · 2020: 19 · 2019: 1
- FWCI
- 1.05
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- Locations
- Radiologia Brasileira (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) (publishedVersion, other-oa)
- Americanae (AECID Library) (submittedVersion, other-oa)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
- Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
- Keywords
- artificial intelligence, machine learning, computer-aided diagnosis, radiomics, precision medicine, diagnostic imaging, quantitative imaging, image analysis, big data, treatment decision-making, radiology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
