Insights into Imaging最新文献

{"title":"Diagnostic accuracy of a modified non-contrast O-RADS MRI scoring system: a retrospective study.","authors":"Yan Ren, Can Cui, Zhijing Zhang, Wei Liao, Chen Zhang, Yue Cheng, Wen Shen","doi":"10.1186/s13244-026-02425-y","DOIUrl":"10.1186/s13244-026-02425-y","url":null,"abstract":"<p><strong>Objective: </strong>To develop a modified non-contrast O-RADS MRI scoring system and validate its diagnostic performance.</p><p><strong>Materials and methods: </strong>The modified non-contrast O-RADS MRI scoring system was developed by incorporating the significant signs from the standard O-RADS MRI system. This study retrospectively enrolled the 176 adnexal masses in cohort A. MRI imaging features were compared between benign and malignant tumors within this cohort. The diagnostic performance of the two scoring models was assessed by estimating the area under the curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI). In addition, 352 adnexal lesions from cohorts B and C were included to further evaluate the diagnostic ability of the modified scoring system.</p><p><strong>Results: </strong>In cohort A, the papillary projections exhibiting the sea anemone sign were highly indicative of borderline lesions (p = 0.001). Diffusion restriction in solid components significantly distinguished benign from malignant masses (p = 0.001). The apparent diffusion coefficient (ADC) value of cystic components yielded a cutoff of 1.74 (p < 0.05). The DeLong test revealed no statistically significant difference in AUCs between the two systems in cohort A (p > 0.05). The NRI and IDI for the modified non-contrast O-RADS MRI scoring system were -4.13% and -0.04, respectively (all p > 0.05). In cohorts B and C, the AUCs were 0.933 and 0.910.</p><p><strong>Conclusion: </strong>The diagnostic efficacy of the modified non-contrast O-RADS MRI scoring system was comparable to that of the standard O-RADS MRI risk stratification system, facilitating the distinction between benign and malignant adnexal neoplasms.</p><p><strong>Key points: </strong>Question How can a modified non-contrast O-RADS MRI scoring system be developed, and can this scoring system accurately classify adnexal masses? Findings The modified non-contrast O-RADS MRI system demonstrates excellent diagnostic performance and reliable risk stratification of adnexal masses by evaluating key imaging features. Critical relevance statement The modified non-contrast O-RADS MRI scoring system contributes to the risk stratification of adnexal masses, particularly for patients who cannot undergo contrast agent administration. This system provides a reliable tool to optimize treatment strategies and improve patient outcomes.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13675434/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149206885","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The underestimated risk of additional ipsilateral BI-RADS 3 lesions identified by preoperative ultrasound in breast cancer patients.","authors":"Yue Hu, Hongli Wang, Mengzi Wu, Wei Zheng, Jingsi Mei, Xiaofang Jiang, Queran Lin, Ran Gu, Fengtao Liu, Jinna Lin, Chang Gong, Jun Shen, Qiang Liu","doi":"10.1186/s13244-026-02366-6","DOIUrl":"10.1186/s13244-026-02366-6","url":null,"abstract":"<p><strong>Objective: </strong>To determine the prevalence of malignancy of additional ultrasound BI-RADS category 3 lesions in women with newly diagnosed breast cancer.</p><p><strong>Materials and methods: </strong>Between July 2014 and June 2022, women with newly diagnosed breast cancer and additional ultrasound BI-RADS 3 lesions were retrospectively included. Malignancy was confirmed by pathology, and malignancy rates were assessed at both the patient and lesion levels. Independent risk factors for malignant BI-RADS 3 lesions were identified using multivariable logistic regression.</p><p><strong>Results: </strong>A total of 1458 women (median age, 45 years; IQR: 40-51) with 1995 ultrasound BI-RADS 3 lesions were included. At the patient level, malignancy was identified in 3.5% of patients (51/1458), with higher rates for ipsilateral than contralateral BI-RADS 3 lesions (9.8% [39/397] vs 1.1% [12/1061]; p < 0.001). At the lesion level, the overall malignancy rate was 2.7% (54/1995), similarly elevated for ipsilateral compared with contralateral lesions (7.5% [41/544] vs 0.9% [13/1451]; p < 0.001). Independent predictors for malignancy of ipsilateral BI-RADS 3 lesions included same-quadrant location as the primary lesion (OR, 4.3; 95% CI: 1.9, 9.7; p < 0.001), ultrasonographic architectural distortion of the primary lesion (OR, 8.5; 95% CI: 1.6, 46.1; p = 0.01), and mammographic suspicious calcification of the primary lesion (OR, 2.9; 95% CI: 1.4, 6.0; p = 0.004).</p><p><strong>Conclusion: </strong>Additional ipsilateral ultrasound BI-RADS 3 lesions in women with newly diagnosed breast cancer demonstrated malignancy rates exceeding the 2% BI-RADS threshold, particularly when located in the same quadrant as the primary cancer.</p><p><strong>Key points: </strong>Question The malignancy risk of additional BI-RADS 3 lesions identified by preoperative ultrasound in breast cancer patients remains unclear. Findings Ipsilateral lesions, especially in the same quadrant, show malignancy rates exceeding 2%, while malignancy risk remains low for contralateral BI-RADS 3 lesions. Critical relevance statement Ipsilateral BI-RADS 3 lesions in the same quadrant as the primary cancer, or those with primary tumor architectural distortion or suspicious calcifications, warrant biopsy. This ensures precise surgical planning and avoids missing additional malignancy.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13675164/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149206974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large language model assistance for report-based carotid plaque AHA classification: a multi-reader study.","authors":"Mengfei Wang, Yixiao Tang, Yulu Yang, Yifan Tu, Zhongzheng Cao, Runjianya Ling, Kaiyue Sun, Jian Ma, Zhicheng Zhang, Ruipeng Zhang, Xiaoer Wei, Dan Wang, Yuehua Li","doi":"10.1186/s13244-026-02421-2","DOIUrl":"10.1186/s13244-026-02421-2","url":null,"abstract":"<p><strong>Objectives: </strong>To evaluate whether large language model assistance improves radiologists' agreement with a text-based reference standard in inferring the modified American Heart Association carotid plaque classification from free-text MRI reports, and whether a lightweight model can perform well under criteria-guided prompting.</p><p><strong>Materials and methods: </strong>This retrospective study included 433 patients from three institutions. In a report-based reader study, four radiologists classified MRI reports with and without large language model assistance. In standalone evaluation, a rule-based baseline and 10 models were compared under naive and criteria-guided prompting. The reference standard was expert consensus based on the report text.</p><p><strong>Results: </strong>Among 433 patients (mean age, 65.9 years ± 13.4; 248 males), AI assistance improved six-class agreement with the text-based reference standard. At the vessel level, junior radiologists improved from 61.75% to 83.12% (Qwen3-8B) and 87.75% (DeepSeek-R1); senior radiologists from 77.00% to 87.50% and 92.25%, respectively (all Holm-corrected p < 0.001). Patient-level agreement increased from 41.0% to 78.8% (junior) and from 59.8% to 87.8% (senior) with DeepSeek-R1. Agreement increased (quadratic-weighted κ: 0.231 without AI vs 0.775 with DeepSeek-R1). In standalone evaluation, criteria-guided prompting improved Qwen3-8B agreement from 38.91% to 81.52% (95% CI: 78.87-84.06%), remaining 5.89 percentage points below DeepSeek-R1 (87.41%; 95% CI: 85.10-89.61%).</p><p><strong>Conclusion: </strong>In this report-based classification setting, large language model assistance improved radiologists' agreement with the text-based reference standard, with larger gains in junior radiologists. When prompts explicitly encoded AHA criteria, an 8B local candidate model performed well, supporting its potential use as a standardization aid for text-based workflows.</p><p><strong>Key points: </strong>Question Report-based modified AHA classification of carotid plaques is experience-dependent and difficult to standardize, as mapping multi-sequence imaging findings to classification conclusions imposes substantial cognitive load. Findings Large language model assistance improved agreement with a text-based reference standard and raised inter-reader consistency (weighted κ 0.231-0.775), narrowing the junior-senior gap. Critical relevance statement Embedding explicit AHA criteria in the prompt let an 8B model approach 671B-model agreement (38.91-81.52%), supporting standardized text-based workflows without fine-tuning.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13675435/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149206864","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Patient acceptance and experience with contrast-enhanced mammography in women with a personal history of breast cancer: findings from the COMBO trial.","authors":"Giulia Vatteroni, Riccardo Levi, Marco Filippini, Martina Costa, Rubina Manuela Trimboli, Daniela Bernardi","doi":"10.1186/s13244-026-02434-x","DOIUrl":"10.1186/s13244-026-02434-x","url":null,"abstract":"<p><strong>Objectives: </strong>To evaluate patient acceptance, tolerability and experience with contrast-enhanced mammography (CEM) in women with a personal history of breast cancer undergoing routine post-surgical surveillance.</p><p><strong>Materials and methods: </strong>This prospective cohort study, conducted within the COMBO Trial, included consecutive women with a personal history of breast cancer undergoing CEM instead of digital mammography (DM) for routine surveillance from January 2023 to February 2025. Immediately after the examination, participants completed an anonymous 16-question questionnaire assessing discomfort, pain, exam duration, pre-examination reassurance, information received, anxiety, contrast injection, satisfaction and comparative experience with prior DM. Responses used Likert-type scales. Descriptive and non-parametric analyses were performed (p < 0.05 statistically significant).</p><p><strong>Results: </strong>Of 969 eligible women, 936 (mean age 58.2 years, SD ± 10.2 years) completed the questionnaire. CEM was perceived as comparable to prior DM in terms of discomfort (631/936; 67.4%), pain (725/936; 77.4%), positioning (717/936; 76.6%) and duration (601/936; 64.3%). Overall satisfaction was high (900/936; 96.2%), and 932/936 (99.6%) of participants would recommend CEM. Although 433/936 (46.3%) of women reported moderate-to-high pre-examination anxiety, higher anxiety was significantly associated with greater perceived discomfort, pain, uncomfortable positioning, longer examination duration, lower reassurance, lower comfort during contrast medium administration, and lower overall satisfaction (all p < 0.001).</p><p><strong>Conclusions: </strong>In this large cohort of breast cancer survivors, CEM demonstrated high acceptance, satisfaction and tolerability in post-surgical follow-up. Although pre-examination anxiety negatively influenced examination experience, it did not reduce patients' willingness to undergo CEM, supporting its integration into personalised post-treatment breast cancer surveillance.</p><p><strong>Key points: </strong>Question Evidence on patient acceptance and tolerability of CEM for routine post-surgical surveillance in women with previous breast cancer is limited. Findings In 936 breast cancer survivors, CEM showed high satisfaction and tolerability, while pre-examination anxiety was associated with a less favourable experience. Critical Relevance Statement High patient acceptance supports CEM as a feasible surveillance option after breast cancer treatment, while the association between anxiety and poorer experience highlights the radiologist's role in pre-examination communication and reassurance.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13675458/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149206945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Environmental impact of a radiology department: greenhouse gas emissions and primary energy accounting and strategies for sustainable practice.","authors":"Stefania Rizzo, Christoph Rüesch, Marco Budin, Rolf Wyttenbach, Ermidio Rezzonico, Filippo Del Grande","doi":"10.1186/s13244-026-02436-9","DOIUrl":"10.1186/s13244-026-02436-9","url":null,"abstract":"<p><strong>Objectives: </strong>To assess the environmental impact of a radiology department's operations (primary energy consumption and greenhouse gas emissions (GHG); to identify the main levers for reduction.</p><p><strong>Materials and methods: </strong>Data collection included measured and calculated energy consumption of imaging machines (2 MRI, 2 CT, 1 interventional radiology (IR) suite, 2 US, 3 radiography machines), workstations, lighting, ventilation; cooling and heating for imaging systems and rooms; materials/pharmaceuticals; external services (e.g. laundry and information technology (IT)). The electrical energy consumption of building systems, lighting, and IT equipment was primarily determined through measurements; other energy consumption figures were estimated using technical data and meteorological data. For the calculation of GHG emissions and primary energy consumption, a specialised tool using specific conversion factors from established environmental databases was developed.</p><p><strong>Results: </strong>Our hospital switched from coal-based electricity to renewable electrical energy (hydropower), leading to a reduction in primary energy consumption (-64%) and GHG emissions (-92%) of the radiology department. Following this change, without a reduction in direct energy consumption, the MRI systems showed a marked reduction in primary energy consumption, although still having the highest primary energy consumption per exam (49.2 kWh compared to 13.9 kWh (CT1) and 48.9 kWh (IR)). After the transition to renewable energy, the major source of GHG emissions shifted from electrical energy toward materials/pharmaceuticals.</p><p><strong>Conclusion: </strong>Our findings support the transition to renewable energy sources for electricity, as a key strategy for reducing GHG emissions and primary energy demand of radiology departments. Following this shift, selecting materials and pharmaceuticals with lower emissions and primary energy factors can further mitigate the environmental impact.</p><p><strong>Key points: </strong>Question Radiology departments are significant sources of energy consumption and GHG emissions. Findings Adoption of renewable energy reduces primary energy consumption and GHG emissions. Critical relevance statement GHG emissions in a radiology department can be significantly reduced by transitioning to renewable energy sources. After this shift, emissions will mainly depend on the materials and pharmaceuticals used, such as disposable supplies and contrast agents.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13672723/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149193254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The added value of post-mortem MRI in forensic imaging: an observational study in the Netherlands.","authors":"Cédrique A S Gregoire, Maartje M A Eggen, Geneviève A J C Crombag, Alida A Postma, J Martijn Nobel","doi":"10.1186/s13244-026-02418-x","DOIUrl":"10.1186/s13244-026-02418-x","url":null,"abstract":"<p><strong>Objectives: </strong>In the Netherlands, post-mortem radiology aids forensic investigation, using a full-body post-mortem computed tomography (PMCT) scan and a post-mortem magnetic resonance imaging (PMMRI) scan in specific cases. Application of PMMRI varies between countries and institutions in the absence of international guidelines. This study analyses current use and added value of PMMRI in a single forensic imaging institution in the Netherlands.</p><p><strong>Materials and methods: </strong>All post-mortem cases scanned between July 2024 and July 2025 were included, except non-accidental injury protocol cases. The (contra-)indications for additional PMMRI scans in addition to PMCT were recorded. On the same day as image evaluation, added value was scored by the forensic radiologist and consultant using the following options: additional finding, exclusion of a finding, confirmation of PMCT and no added value.</p><p><strong>Results: </strong>PMMRI was conducted in 47.4% of all cases, indicated by suspected strangulation, penetrating injuries, blunt force trauma and lesions seen on PMCT. In suspected strangulation, PMMRI neck showed additional findings or exclusion of findings in 31/46 cases, while PMMRI brain had the lowest added value in 13/25 cases. PMMRI thorax and abdomen showed additional findings in 14/20 and 4/9 cases of penetrating trauma, respectively. Additional findings were found in 10/15 PMMRI brain scans and 6/8 PMMRI thorax scans, based on PMCT.</p><p><strong>Conclusion: </strong>Application of PMMRI in addition to PMCT is currently used in a targeted manner. In cases of strangulation and penetrating injuries, additional or excluded findings were found. Therefore, adding PMMRI in these specific cases can be of added value.</p><p><strong>Key points: </strong>Question What is the added value of post-mortem MRI in forensic cases, in the absence of international guidelines and standardised indications? Findings Additional findings were observed in post-mortem MRI examinations indicated for soft tissue injuries caused by penetrating force and in cases of suspected strangulation. Critical Relevance Statement The criminal investigation can be optimised by using post-mortem MRI in addition to post-mortem CT, either by showing additional findings on post-mortem MRI not seen on CT, or by excluding findings with higher certainty.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13672765/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149193252","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Burnout and psychosocial working conditions in a multidisciplinary hospital radiology team: a cross-sectional study using the job demands-resources framework.","authors":"Thomas Dertnig, Patrick Kutschar, Hans-Peter Wiesinger","doi":"10.1186/s13244-026-02433-y","DOIUrl":"10.1186/s13244-026-02433-y","url":null,"abstract":"<p><strong>Objectives: </strong>Burnout is a major concern in hospital radiology, but data on psychosocial work conditions remain limited. Using the job demands-resources (JD-R) framework, this study examined burnout and its associations with psychosocial work factors across professional groups in a German university hospital radiology department.</p><p><strong>Materials and methods: </strong>An exploratory cross-sectional survey was conducted among radiologists and non-physician staff. Burnout was measured using the Copenhagen Burnout Inventory (CBI); psychosocial work conditions were assessed with COPSOQ III. Descriptive analyses, correlations, and hierarchical linear regressions were performed to evaluate associations between job demands, job resources, and burnout dimensions (personal, work-related, and patient-related burnout).</p><p><strong>Results: </strong>Sixty-three participants (≤ 28 to ≥ 61 years; 37 women) were included. Personal and work-related burnout exceeded patient-related burnout. Non-physician staff reported higher personal burnout than radiologists. Job demands strongly correlated with personal (r = 0.686) and work-related burnout (r = 0.748), whereas associations with resources were weaker. In regression analyses, job demands were strongly associated with personal (R² = 0.62) and work-related burnout (R² = 0.60), without significant incremental resource effects. For patient-related burnout, job resources demonstrated an independent inverse association (total R² = 0.30). Higher burnout was associated with lower job satisfaction and greater sickness absence.</p><p><strong>Conclusion: </strong>In this exploratory single-centre study, burnout in hospital radiology was primarily associated with psychosocial job demands, particularly work-life conflict, workload, and role conflicts, whereas job resources showed weaker associations except for patient-related burnout. These findings suggest a predominantly demand-driven pattern within the JD-R framework and indicate that psychosocial working conditions may represent targets for intervention.</p><p><strong>Key points: </strong>Question Which psychosocial job demands and resources show the strongest associations with burnout among radiology staff in a university hospital department? Findings Work-life conflict, quantitative demands, and role conflicts showed the strongest associations with personal and work-related burnout among radiology staff. Critical relevance statement Burnout in hospital radiology was predominantly associated with organisational stressors, highlighting the relevance of psychosocial working conditions for organisational monitoring and intervention planning.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13672748/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149193250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Estimation of glomerular filtration rate by standard renal CT imaging.","authors":"Kristoffer Huitfeldt Sola, Helena Genberg, Joshua Webster-Ford, Anders Ullén, Torkel B Brismar","doi":"10.1186/s13244-026-02431-0","DOIUrl":"10.1186/s13244-026-02431-0","url":null,"abstract":"<p><strong>Objectives: </strong>To develop and evaluate a method to predict glomerular filtration rate (GFR) from clinical renal CT.</p><p><strong>Materials and methods: </strong>Medical records of adult living kidney donors operated between 2013 and 2022 and patients considered for neoadjuvant cisplatin prior to cystectomy between 2019 and 2024 were retrospectively reviewed. Inclusion criteria were (1) renal CT obtained at 100 or 120 kVp before and 5-15 min after contrast administration, and (2) iohexol clearance test within 2 months of imaging. The iohexol clearance test was performed on the same day as the cystatin C and creatinine measurements. Estimated GFR values included cystatin C-based eGFR (eGFR_{cystatin C}), creatinine-based eGFR using the Cockcroft-Gault formula (eGFR_{Cockcroft-Gault}), and eGFR using the Revised Lund-Malmö formula (eGFR_LMR18), which is based on both creatinine and cystatin C. Manual and automatic segmentations of the renal pelvis and bladder were performed in both the excretory and unenhanced phases.</p><p><strong>Results: </strong>In total, 139 patients were included in the analysis (mean age 52 ± 13 years; 60% women). Multivariate linear regression analysis combining the segmentation-derived variable with body surface area and age gave an adjusted R-square of 0.77 to mGFR_iohexol (p < 0.001). GFR_CT had the strongest relationship to mGFR_iohexol with a mean absolute error of 12.8 mL/min, compared with 15.6 mL/min for eGFR_{Cockcroft-Gault}, 14.2 mL/min for eGFR_{cystatin C}, and 13.0 mL/min for eGFR_LMR18. Manual and automatic segmentations had similar performance. Patients evaluated for cisplatin treatment had greater variation in CT-estimated split renal function than donors.</p><p><strong>Conclusion: </strong>Renal function can be estimated with high precision from clinical renal CT.</p><p><strong>Key points: </strong>Question Accurate GFR assessment is critical, but current methods are either labor-intensive or imprecise. Can renal CT imaging provide a reliable alternative? Findings CT-derived eGFR showed the strongest correlation with iohexol clearance (R² = 0.77), with lower error compared to creatinine- and cystatin C-based estimates. Critical relevance statement Routine CT imaging can provide precise estimates of kidney function, supporting donor evaluations, chemotherapy dosing, and critical care decisions when accurate renal function is essential.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13669286/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149181666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimizing AI-assisted decision-making in mammography screening: focusing on uncertainties.","authors":"Nasrin Ahmadinejad, Nahid Sadighi, Rassa Ghavami Modegh, Mehran Arab Ahmadi, Mohammadhosein Akhlaghpasand, Alireza Dizaji, Amin Rezaee, Maryam Rahmani, Arvin Arian, Fahimeh Zeinalkhani, Salome Maghsudlu, Saeedeh Shokri, Niloofar Pashaeifakhri, Hossein Ghanaati, Ali Sefidmouy, Mahdi Ghaznavi, Hamed Dashti, Hamid R Rabiee, Masoumeh Gity","doi":"10.1186/s13244-026-02411-4","DOIUrl":"10.1186/s13244-026-02411-4","url":null,"abstract":"<p><strong>Objective: </strong>To evaluate an AI-based decision-support system for mammography interpretation, assessing its impact on diagnostic performance.</p><p><strong>Materials and methods: </strong>This retrospective study analyzed 130,838 mammograms from 35,709 patients with confirmed follow-up outcomes across multiple centers, using the Digital Database for Screening Mammography (DDSM) for external validation. Representing one of the largest datasets from the Middle East, it evaluates model generalizability in underrepresented populations. An AI super-model incorporated separate UNet-based networks for lesion segmentation, a high-resolution model (4096) for calcifications, a multi-resolution ensemble (4096, 2048, 1024) for macro-scale lesions, and a risk-assessment model for malignancy prediction. Radiologists interpreted images under five scenarios: unaided, AI-assisted, AI as the final decision-maker in uncertain cases, union of radiologist and AI decisions in uncertain cases, and standalone AI.</p><p><strong>Results: </strong>AI assistance significantly enhanced performance across all scenarios. Based on aggregated reader performance, AUC improved from 83.04% (95% CI: 79.8-86.3) unaided to 88.40% (95% CI: 85.6-91.2) in the union scenario (p < 0.001). Sensitivity increased from 54.8% (95% CI: 48.6-61.0) to 83.9% (95% CI: 79.3-88.4) (p < 0.001), and specificity rose from 83.6% (95% CI: 82.0-85.3) to 91.1% (95% CI: 89.8-92.4) (p < 0.001). Multi-resolution segmentation effectively detected calcifications and challenging lesions. DDSM validation confirmed model robustness. AI showed particular promise in aiding less-experienced radiologists in underserved areas to improve early cancer detection.</p><p><strong>Conclusions: </strong>AI in mammography interpretation improves diagnostic accuracy and reduces uncertainty, supporting early breast cancer detection and optimized screening workflows.</p><p><strong>Key points: </strong>Question Can an AI-based decision-support system improve the accuracy of mammography interpretation, reduce diagnostic uncertainty, and support less-experienced radiologists? Findings In this retrospective multicenter study, AI-assisted mammography improved sensitivity and specificity, increased cancer detection, and reduced recalls and false positives. Critical relevance statement AI integration in mammography enhances diagnostic accuracy, reduces unnecessary recalls, and supports more confident and efficient clinical decision-making, particularly by improving the performance of less-experienced radiologists in underserved and resource-limited settings where expert interpretation may be limited.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661836/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149181771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reproducibility and diagnostic performance of ADC for assessing response to neoadjuvant immunotherapy in muscle-invasive bladder cancer: a multi-ROI comparison analysis from the phase-Ib NABUCCO trial.","authors":"Joyce Greidanus, Carlotta Bernabei, Kalina Chupetvloska, Georgios Agrotis, Laura S Mertens, Bas W G van Rhijn, Henk G van der Poel, Michiel S van der Heijden, Regina G H Beets-Tan, Andrea Delli Pizzi, Thierry N Boellaard, Stefano Trebeschi","doi":"10.1186/s13244-026-02397-z","DOIUrl":"10.1186/s13244-026-02397-z","url":null,"abstract":"<p><strong>Objective: </strong>MRI is increasingly used to assess treatment response in muscle-invasive bladder cancer (MIBC), but only qualitatively. This study evaluates the reproducibility and predictive value of quantitative apparent diffusion coefficient (ADC), thought to be a non-invasive surrogate for tumor cell density, in patients with MIBC receiving neoadjuvant immunotherapy.</p><p><strong>Materials and methods: </strong>MRI data from 35 patients, enrolled in the Phase-Ib NABUCCO trial, were analyzed. ADC maps were derived from pre- and post-treatment MRI scans. Three independent radiologists assessed multiple regions of interest (ROIs) to evaluate inter-observer agreement and predictive performance for pathological complete response (pCR). ADC values were normalized using muscle tissue as a reference. Additionally, we analyzed the correlation between ADC and cell densities, derived from multiplex immunofluorescence (mIF) data.</p><p><strong>Results: </strong>ADC measurements from the whole-tumor ROI (n = 25) demonstrated the highest reproducibility (0.98, 95% CI = [0.95-0.99]), but lacked predictive value (AUC = 0.31, 95% CI = [0.12-0.53], p = 0.109). The most hypointense ROI (n = 22) showed the strongest predictive value for pCR (AUC = 0.80, 95% CI = [0.57-1], p = 0.025), though with limited reproducibility. The difference between baseline and follow-up ADC values in the hypointense region (n = 20) again showed the highest diagnostic accuracy (AUC = 0.80, 95% CI = [0.58-0.98], p = 0.015). Although muscle normalization is proposed to reduce scanner variability, it did not improve predictive performance in this cohort and occasionally reduced reproducibility. No significant correlation was found between whole-tumor ADC and cell density.</p><p><strong>Conclusion: </strong>Despite analyzing different measurement techniques for ADC values, none was simultaneously reproducible and predictive, nor provided insights into the biology of the tumor. While ADC shows promise in MIBC, there is an urgent need for more standardized, robust imaging biomarkers.</p><p><strong>Key points: </strong>Question Non-invasive biomarkers are urgently needed to guide bladder-preserving strategies in MIBC. Findings DWI-MRI shows promise, but standardized robust imaging biomarkers are essential for reliable response assessment. Critical relevance statement Standardized or AI-assisted ROI segmentation could improve ADC reproducibility and diagnostic accuracy in multicenter studies.</p>","PeriodicalId":13639,"journal":{"name":"Insights into Imaging","volume":"17 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661850/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149181720","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}