Purpose To develop a deep learning-based deformable registration method for dynamic contrast-enhanced (DCE) breast MRI that preserves tumor regions while maintaining global anatomic alignment during neoadjuvant chemotherapy (NAC) response assessment. Materials and Methods This retrospective study included internal and external cohorts of patients with breast cancer who were undergoing NAC. The internal cohort comprised patients who underwent DCE MRI from 2017 to 2020, and the external cohort was derived from the I-SPY2 trial. A conditional pyramid registration network integrating unsupervised keypoint detection with a volume-preserving mechanism was developed. Registration performance was evaluated using the Dice similarity coefficient (DSC), average landmark error, and tumor volume difference. A local-global biomarker derived from registered images was evaluated for predicting pathologic complete response (pCR) using the area under the receiver operating characteristic curve (AUC) and accuracy. Paired t tests were used for statistical comparisons. Results In 314 patients (all female; age, 50.6 years ± 12.0 [SD]) with 1630 scans in the internal cohort, the proposed method achieved a DSC of 0.95 ± 0.02, an average landmark error of 5.35 mm ± 3.46, and a tumor volume difference of 11.0% ± 10.7. In 100 patients (all female; age, 48.5 years ± 12.3) with 372 scans in the external cohort, the method achieved a DSC of 0.91 ± 0.09 and a tumor volume difference of 15.5% ± 13.8. Improvements in landmark distance and tumor preservation were statistically significant (P < .05) compared with most methods. For pCR prediction, incorporation of the proposed biomarker achieved an AUC of 0.81 ± 0.04 and an accuracy of 72.1% ± 5.0. Conclusion The proposed framework improved anatomic alignment while preserving tumor volume in longitudinal DCE breast MRI during NAC response assessment and enabled a registration-based biomarker for predicting treatment response. Keywords: MR Imaging, Image Postprocessing, Breast, Neural Networks, Radiomics, Prognosis, DCE Breast MRI, Deep Learning, Deformable Registration, Neoadjuvant Chemotherapy, Unsupervised Keypoint Detection Supplemental material is available for this article. © RSNA, 2026 See also the commentary by Zhang in this issue.

OBJECTIVES:To assess the impact of radiologist experience on the technical success, safety and effectiveness of CT-guided thermal ablation (TA) procedures for abdominal tumors, including hepatocellular carcinomas (HCCs), colorectal cancer liver metastases (CRLMs), and renal cell carcinomas (RCCs), when assisted by an electromagnetic navigation system (EMNS). MATERIAL AND METHODS:We retrospectively collected data for patients who had undergone CT-guided TA between 2020 and 2022, recording the characteristics of the lesions. Lesions were considered high-risk if they were located in the subphrenic or subcapsular areas of the liver or less than 1 cm from the bowel, bile duct, portal vein, vena cava or gallbladder and those located in the kidney in the anterior leaflet or close to the urinary tract. The radiologists who performed the procedures were classified according to whether they had more or less experience (more or less than five years of experience in percutaneous TA). Technical success was assessed immediately after treatment. Procedure data, response and complication rates were recorded. RESULTS:A total of 139 tumors were treated in 105 ablation sessions in 93 patients, consisting of 69 men and 24 women. Sixty-two percent of the tumors were HCCs, 27% were CRLMs, and 12% were RCCs. The median tumor size was 16 mm. A total of 68% of the tumors were located in high-risk areas. The technical success rate was 96.4%, with minor complications occurring in 20% of the procedures and major complications in 3.8%. The median follow-up was 12 months. The complete response rates were 96.6%, 93.1%, and 86.2% at 3, 6, and 12 months, respectively. There were no significant differences in response at one month (p = 0.706) or one year (p = 0.402), complications (p = 0.583), procedure time (p = 0.729), or the number of follow-up CT scans (p = 0.208) between more and less experienced radiologists. CONCLUSION:An EMNS enhances accuracy and standardisation in interventional procedures, enabling precise ablations, regardless of the radiologist's experience or location involved.

Multi-modal image analysis using deep learning (DL) lays the foundation for neoadjuvant treatment (NAT) response monitoring. However, existing methods prioritize extracting multi-modal features to enhance predictive performance, with limited consideration on real-world clinical applicability, particularly in longitudinal NAT scenarios with multi-modal data. Here, we propose the Multi-modal Response Prediction (MRP) system, designed to mimic real-world physician assessments of NAT responses in breast cancer. To enhance feasibility, MRP integrates cross-modal knowledge mining and temporal information embedding strategy to handle missing modalities and remain less affected by different NAT settings. We validated MRP through multi-center studies and multinational reader studies. MRP exhibited comparable robustness to breast radiologists, outperforming humans in predicting pathological complete response in the Pre-NAT phase (Delta AUROC 14% and 10% on in-house and external datasets, respectively). Furthermore, we assessed MRP's clinical utility impact on treatment decision-making. MRP may have profound implications for enrolment into NAT trials and determining surgery extensiveness. Deep learning for medical image analysis is a promising new avenue to predict treatment response, however the clinical application of these methods has been so far limited. Here, the authors propose a model to predict chemotherapy response in breast cancer in real world clinical settings.

Clinicians compare breast DCE-MRI after neoadjuvant chemotherapy (NAC) with pre-treatment scans to evaluate the response to NAC. Clinical evidence supports that accurate longitudinal deformable registration without deforming treated tumor regions is key to quantifying tumor changes. We propose a conditional pyramid registration network based on unsupervised keypoint detection and selective volume-preserving to quantify changes over time. In this approach, we extract the structural and the abnormal keypoints from DCE-MRI, apply the structural keypoints for the registration algorithm to restrict large deformation, and employ volume-preserving loss based on abnormal keypoints to keep the volume of the tumor unchanged after registration. We use a clinical dataset with 1630 MRI scans from 314 patients treated with NAC. The results demonstrate that our method registers with better performance and better volume preservation of the tumors. Furthermore, a local-global-combining biomarker based on the proposed method achieves high accuracy in pathological complete response (pCR) prediction, indicating that predictive information exists outside tumor regions. The biomarkers could potentially be used to avoid unnecessary surgeries for certain patients. It may be valuable for clinicians and/or computer systems to conduct follow-up tumor segmentation and response prediction on images registered by our method. Our code is available on .

Thoracic surgical procedures are increasing in recent years, and there are different types of lung resections. Postsurgical complications vary depending on the type of resection and the time elapsed, with imaging techniques being key in the postoperative follow-up. Multidisciplinary management of these patients throughout the perioperative period is essential to ensure an optimal surgical outcome. This pictorial review will review the different thoracic surgical techniques, normal postoperative findings and postsurgical complications.

Thoracic surgical procedures are increasing in recent years, and there are different types of lung resections. Postsurgical complications vary depending on the type of resection and the time elapsed, with imaging techniques being key in the postoperative follow-up. Multidisciplinary management of these patients throughout the perioperative period is essential to ensure an optimal surgical outcome. This pictorial review will review the different thoracic surgical techniques, normal postoperative findings and postsurgical complications. (c) 2023 SERAM. Published by Elsevier Espa & ntilde;a, S.L.U. All rights reserved.

Machine learning models are increasingly used in the medical domain to study the association between risk factors and diseases to support practitioners in understanding health outcomes. In this paper, we showcase the use of machine-learned staged tree models for investigating complex asymmetric dependence structures in health data. Staged trees are a specific class of generative, probabilistic graphical models that formally model asymmetric conditional independence and non-regular sample spaces. An investigation of the risk factors in invasive fungal infections demonstrates the insights staged trees provide to support medical decision-making.

A 52-year-old patient admitted to our hospital 1 month earlier with a severe SARS-CoV-2 infection (requiring admission to the intensive care unit and mechanical intubation for 2 weeks, where he was placed in the prone position for several cycles) complained of chest discomfort. Physical examination showed an indurated anterior chest wall scar (Figure 1). Anterior chest wall ultrasound (Figure 2) and thoracic computed tomography (Figure 3) were performed.Figure 2Ultrasound image (obtained with a high-frequency linear probe) of the indurated area of the anterior chest wall shows thickening of the skin (arrows) and heterogeneous increase in the echogenicity of the subcutaneous fat tissue (asterisk).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3Axial computed tomography image of the thorax shows subcutaneous fat stranding of the anterior chest wall (long arrows), a skin defect in the left inframammary region (asterisk), and lack of involvement of the rib cartilage and muscular planes (short arrows).View Large Image Figure ViewerDownload Hi-res image Download (PPT) Anterior chest wall fat necrosis in a patient with COVID-19 following prolonged prone position. Prone position is frequently used in patients with respiratory failure, but this position can lead to complications, such as pressure ulcers.1Moore Z. Patton D. Avsar P. et al.Prevention of pressure ulcers among individuals cared for in the prone position: lessons for the COVID-19 emergency.J Wound Care. 2020; 29: 312-320Crossref PubMed Scopus (57) Google Scholar In general, there is a lack of knowledge among health care personnel for the prevention of pressure ulcers in patients with COVID-19 in the prone position, aggravated by the difficulty in mobilizing intubated patients in this position. Anterior chest wall pressure ulcers are rare, and imaging techniques can help to rule out the involvement of deeper structures.2Ibarra G. Rivera A. Fernandez-Ibarburu B. et al.Prone position pressure sores in the COVID-19 pandemic: the Madrid experience.J Plast Reconstr Aesthet Surg. 2021; 74: 2141-2148Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar In the literature, to our knowledge, we have not found a radiological description of fat necrosis as a complication of prolonged prone position in a patient with COVID-19. Clinicians should be aware of the risk of pressure ulcers in the anterior chest wall in patients with COVID-19 placed in the prone position. A biopsy of the indurated area confirmed fat necrosis. The patient was treated with local dressings and antiseptics and responded favorably to the treatment.

The authors declare that there is no conflict of interest regarding the publication of this article. The authors confirm that the data supporting the findings of this study are available within the article (and/or) its supplementary materials.

HomeRadioGraphicsVol. 43, No. 11 PreviousNext Musculoskeletal ImagingRadioGraphics FundamentalsRadiographic Evaluation of Bone TumorsAbel González-Huete , Alba Salgado-Parente, Carlos Suevos-Ballesteros, Elisa Antolinos-Macho, Sofía Ventura-Díaz, Antonio Michael-Fernández, Javier Blázquez-Sánchez, José Acosta-BatlleAbel González-Huete , Alba Salgado-Parente, Carlos Suevos-Ballesteros, Elisa Antolinos-Macho, Sofía Ventura-Díaz, Antonio Michael-Fernández, Javier Blázquez-Sánchez, José Acosta-BatlleAuthor AffiliationsFrom the Department of Radiology, Hospital Universitario Ramón y Cajal, Ctra de Colmenar Viejo km 9,100, Madrid 28034, Spain.Address correspondence to A.G.H. (email: [email protected], [email protected]).Abel González-Huete Alba Salgado-ParenteCarlos Suevos-BallesterosElisa Antolinos-MachoSofía Ventura-DíazAntonio Michael-FernándezJavier Blázquez-SánchezJosé Acosta-BatllePublished Online:Oct 12 2023https://doi.org/10.1148/rg.230048MoreSectionsFull textPDF ToolsAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In AbstractRadiography remains the primary imaging modality for diagnosis of bone tumors and tumorlike lesions, and a structured approach to evaluation is based on lesion location, density, matrix, margins, number, periosteal reaction, cortical involvement, and soft-tissue components in accordance with patient age.Suggested ReadingsAmerican College of Radiology Committee on Bone-RADS. Bone-RADS v2023 Assessment Categories and Scoring System. https://www.acr.org/-/media/ACR/Files/RADS/Bone-RADS/Bone-RADS-v2023-Assessment-Categories-Table-_Final.pdf. Accessed March 8, 2023. Google ScholarCostelloe CM, Madewell JE. Radiography in the initial diagnosis of primary bone tumors. AJR Am J Roentgenol 2013;200(1):3–7. Crossref, Medline, Google ScholarHelms CA. Fundamentals of skeletal radiology. 5th ed. Philadelphia, Pa: Elsevier–Health Sciences Division, 2021. Google ScholarLalam R, Bloem JL, Noebauer-Huhmann IM, et al. ESSR consensus document for detection, characterization, and referral pathway for tumors and tumorlike lesions of bone. Semin Musculoskelet Radiol 2017;21(5):630–647. Crossref, Medline, Google ScholarLodwick GS, Wilson AJ, Farrell C, Virtama P, Smeltzer FM, Dittrich F. Estimating rate of growth in bone lesions: observer performance and error. Radiology 1980;134(3):585–590. Link, Google ScholarMadewell JE, Ragsdale BD, Sweet DE. Radiologic and pathologic analysis of solitary bone lesions. Part I: internal margins. Radiol Clin North Am 1981;19(4):715–748. Crossref, Medline, Google ScholarMehta K, McBee MP, Mihal DC, England EB. Radiographic analysis of bone tumors: a systematic approach. Semin Roentgenol 2017;52(4):194–208. Crossref, Medline, Google ScholarMhuircheartaigh JN, Lin YC, Wu JS. Bone tumor mimickers: a pictorial essay. Indian J Radiol Imaging 2014;24(3):225–236. Crossref, Medline, Google ScholarMiller TT. Bone tumors and tumorlike conditions: analysis with conventional radiography. Radiology 2008;246(3):662–674. Link, Google ScholarRana RS, Wu JS, Eisenberg RL. Periosteal reaction. AJR Am J Roentgenol 2009;193(4):W259–W272. Crossref, Medline, Google ScholarArticle HistoryReceived: Mar 17 2023Revision requested: Apr 26 2023Revision received: May 9 2023Accepted: May 16 2023Published online: Oct 12 2023 FiguresReferencesRelatedDetailsRecommended Articles Pediatric Benign Bone Tumors: What Does the Radiologist Need to Know?: Pediatric ImagingRadioGraphics2017Volume: 37Issue: 3pp. 1001-1002Imaging Review of Normal and Abnormal Skeletal MaturationRadioGraphics2022Volume: 42Issue: 3pp. 861-879Common Skeletal Neoplasms and Nonneoplastic Lesions at 18F-FDG PET/CTRadioGraphics2021Volume: 42Issue: 1pp. 250-267Periosteal Pathologic Conditions: Imaging Findings and PathophysiologyRadioGraphics2022Volume: 43Issue: 2Multitask Deep Learning for Segmentation and Classification of Primary Bone Tumors on RadiographsRadiology2021Volume: 301Issue: 2pp. 398-406See More RSNA Education Exhibits The Role of Conventional Radiography in the Diagnosis of Bone TumorsDigital Posters2022Bone Tumors and Tumor-Like Lesions on CT: A Primer for the Emergency RadiologistDigital Posters2019Recognizing Pediatric Bone Tumors: Growing Your KnowledgeDigital Posters2022 RSNA Case Collection Telangiectatic Osteosarcoma RSNA Case Collection2021Metacarpal EnchondromaRSNA Case Collection2020Monostotic Fibrous DysplasiaRSNA Case Collection2021 Vol. 43, No. 11 Slide PresentationMetrics Altmetric Score PDF download

COVID-19 raises D-dimer (DD) levels even in the absence of pulmonary embolism (PE), resulting in an increase in computed tomography pulmonary angiogram (CTPA) requests. Our purpose is to determine whether there are differences between DD values in PE-positive and PE-negative COVID-19 patients and, if so, to establish a new cutoff value which accurately determines when a CTPA is needed. This study retrospectively analyzed all COVID-19 patients who underwent a CTPA due to suspected PE between March 1 and April 30, 2020, at Ramón y Cajal University Hospital, Madrid (Spain). DD level comparisons between PE-positive and PE-negative groups were made using Student’s t test. The optimal DD cutoff value to predict PE risk in COVID-19 patients was calculated in the ROC curve. Two hundred forty-two patients were included in the study. One hundred fifty-one (62%) were men and the median age was 68 years (IQR 55–78). An increase of DD (median 3260; IQR 1203–9625 ng/mL) was detected in 205/242 (96%) patients. 73/242 (30%) of the patients were diagnosed with PE on CTPA. The DD median value was significantly higher (p < .001) in the PE-positive group (7872, IQR 3150–22,494 ng/mL) compared with the PE-negative group (2009, IQR 5675–15,705 ng/mL). The optimal cutoff value for DD to predict PE was 2903 ng/mL (AUC was 0.76 [CI 95% 0.69–0.83], sensitivity 81%). The overall mortality rate was 16% (39/242). A higher threshold (2903 ng/mL) for D-dimer could predict the risk of PE in COVID-19 patients with a sensitivity of 81%.