Insights into Imaging
@insightsimaging.bsky.social
#InsightsIntoImaging is a gold open access journal owned by the European Society of Radiology and edited by Editor-in-Chief, Paola Clauser.
🟣 https://www.i3-journal.org/
Social Media Editor: @cannellaroberto.bsky.social
🟣 https://www.i3-journal.org/
Social Media Editor: @cannellaroberto.bsky.social
Predicting chemotherapy response in pediatric rhabdomyosarcoma
A T1CE MRI–based intratumoral + peritumoral radiomics model robustly predicts neoadjuvant chemo response in children (AUC up to 0.917).
A T1CE MRI–based intratumoral + peritumoral radiomics model robustly predicts neoadjuvant chemo response in children (AUC up to 0.917).
January 23, 2026 at 1:02 PM
Predicting chemotherapy response in pediatric rhabdomyosarcoma
A T1CE MRI–based intratumoral + peritumoral radiomics model robustly predicts neoadjuvant chemo response in children (AUC up to 0.917).
A T1CE MRI–based intratumoral + peritumoral radiomics model robustly predicts neoadjuvant chemo response in children (AUC up to 0.917).
📊 Predicting mesorectal lymph node metastasis in rectal cancer
A super-resolution multiparametric MRI radiomics model using tumor + peritumoral (3 mm) features achieved:
A super-resolution multiparametric MRI radiomics model using tumor + peritumoral (3 mm) features achieved:
January 23, 2026 at 10:01 AM
📊 Predicting mesorectal lymph node metastasis in rectal cancer
A super-resolution multiparametric MRI radiomics model using tumor + peritumoral (3 mm) features achieved:
A super-resolution multiparametric MRI radiomics model using tumor + peritumoral (3 mm) features achieved:
🧠This pictorial review highlights how cerebrospinal fluid (CSF) dynamics can create artifacts that:
🔹 Obscure or mimic pathology
🔹 Degrade image quality
🔹 Occasionally aid diagnosis
💡 Mitigation is straightforward: adjust phase encoding, use alternate sequences, or tweak MRI parameters.
🔹 Obscure or mimic pathology
🔹 Degrade image quality
🔹 Occasionally aid diagnosis
💡 Mitigation is straightforward: adjust phase encoding, use alternate sequences, or tweak MRI parameters.
January 22, 2026 at 10:01 AM
🧠This pictorial review highlights how cerebrospinal fluid (CSF) dynamics can create artifacts that:
🔹 Obscure or mimic pathology
🔹 Degrade image quality
🔹 Occasionally aid diagnosis
💡 Mitigation is straightforward: adjust phase encoding, use alternate sequences, or tweak MRI parameters.
🔹 Obscure or mimic pathology
🔹 Degrade image quality
🔹 Occasionally aid diagnosis
💡 Mitigation is straightforward: adjust phase encoding, use alternate sequences, or tweak MRI parameters.
#InsightsintoImaging Opinion Paper!
🧪🤖 AI is powerful - but science must guide it
Hospitals need auditable pipelines, preregistered evaluations, and external validation to ensure AI is robust, interpretable, and clinically meaningful.
🧪🤖 AI is powerful - but science must guide it
Hospitals need auditable pipelines, preregistered evaluations, and external validation to ensure AI is robust, interpretable, and clinically meaningful.
Keeping AI in medicine and radiology within the framework of scientific method: measuring to close the epistemic gap - Insights into Imaging
Insights into Imaging -
link.springer.com
January 21, 2026 at 1:01 PM
#InsightsintoImaging Opinion Paper!
🧪🤖 AI is powerful - but science must guide it
Hospitals need auditable pipelines, preregistered evaluations, and external validation to ensure AI is robust, interpretable, and clinically meaningful.
🧪🤖 AI is powerful - but science must guide it
Hospitals need auditable pipelines, preregistered evaluations, and external validation to ensure AI is robust, interpretable, and clinically meaningful.
⏱️ Ultrafast breast MRI just got smarter!
January 21, 2026 at 10:01 AM
⏱️ Ultrafast breast MRI just got smarter!
Predicting hidden lymph node metastasis in gastric cancer
A new preoperative model combining CT-based extracellular volume (ECV) fraction and clinicopathological features showed:
A new preoperative model combining CT-based extracellular volume (ECV) fraction and clinicopathological features showed:
January 20, 2026 at 10:01 AM
Predicting hidden lymph node metastasis in gastric cancer
A new preoperative model combining CT-based extracellular volume (ECV) fraction and clinicopathological features showed:
A new preoperative model combining CT-based extracellular volume (ECV) fraction and clinicopathological features showed:
A new MRI-based DLR nomogram accurately predicts vessels encapsulating tumor clusters (VETC) and recurrence-free survival (RFS) in hepatocellular carcinoma.
January 19, 2026 at 1:02 PM
A new MRI-based DLR nomogram accurately predicts vessels encapsulating tumor clusters (VETC) and recurrence-free survival (RFS) in hepatocellular carcinoma.
🧠Deep learning for Alzheimer’s MRI? Beware the “Clever Hans” effect.
This study shows 3D CNNs classify AD using skull-stripping artifacts rather than true brain tissue texture - highlighting shortcut learning.
This study shows 3D CNNs classify AD using skull-stripping artifacts rather than true brain tissue texture - highlighting shortcut learning.
January 19, 2026 at 10:01 AM
🧠Deep learning for Alzheimer’s MRI? Beware the “Clever Hans” effect.
This study shows 3D CNNs classify AD using skull-stripping artifacts rather than true brain tissue texture - highlighting shortcut learning.
This study shows 3D CNNs classify AD using skull-stripping artifacts rather than true brain tissue texture - highlighting shortcut learning.
🦵💡 Hip abductor health in GTPS
Volumetric MRI shows higher fat infiltration in gluteus minimus & medius of hips with greater trochanteric pain syndrome (GTPS) vs asymptomatic hips.
Volumetric MRI shows higher fat infiltration in gluteus minimus & medius of hips with greater trochanteric pain syndrome (GTPS) vs asymptomatic hips.
January 16, 2026 at 1:01 PM
🦵💡 Hip abductor health in GTPS
Volumetric MRI shows higher fat infiltration in gluteus minimus & medius of hips with greater trochanteric pain syndrome (GTPS) vs asymptomatic hips.
Volumetric MRI shows higher fat infiltration in gluteus minimus & medius of hips with greater trochanteric pain syndrome (GTPS) vs asymptomatic hips.
🩻 MRI in penile cancer: why it matters
January 16, 2026 at 10:01 AM
🩻 MRI in penile cancer: why it matters
🩻🎯 How do you biopsy a lesion you can’t see on mammography or ultrasound?
This prospective study shows CEM-guided biopsy can accurately target enhancing-only breast lesions, with a 97% success rate and 57% malignancy yield - all with low radiation dose and short procedure time.
This prospective study shows CEM-guided biopsy can accurately target enhancing-only breast lesions, with a 97% success rate and 57% malignancy yield - all with low radiation dose and short procedure time.
January 15, 2026 at 10:00 AM
🩻🎯 How do you biopsy a lesion you can’t see on mammography or ultrasound?
This prospective study shows CEM-guided biopsy can accurately target enhancing-only breast lesions, with a 97% success rate and 57% malignancy yield - all with low radiation dose and short procedure time.
This prospective study shows CEM-guided biopsy can accurately target enhancing-only breast lesions, with a 97% success rate and 57% malignancy yield - all with low radiation dose and short procedure time.
📊 Ultrafast DCE-MRI in breast imaging
Early-phase semi-quantitative analysis showed noninferior performance to full pharmacokinetic modeling for differentiating benign from malignant lesions (AUC 0.93 vs 0.92).
Both approaches had limited value for molecular subtyping and pCR prediction.
Early-phase semi-quantitative analysis showed noninferior performance to full pharmacokinetic modeling for differentiating benign from malignant lesions (AUC 0.93 vs 0.92).
Both approaches had limited value for molecular subtyping and pCR prediction.
January 14, 2026 at 1:01 PM
📊 Ultrafast DCE-MRI in breast imaging
Early-phase semi-quantitative analysis showed noninferior performance to full pharmacokinetic modeling for differentiating benign from malignant lesions (AUC 0.93 vs 0.92).
Both approaches had limited value for molecular subtyping and pCR prediction.
Early-phase semi-quantitative analysis showed noninferior performance to full pharmacokinetic modeling for differentiating benign from malignant lesions (AUC 0.93 vs 0.92).
Both approaches had limited value for molecular subtyping and pCR prediction.
🧬📊 What if a CT scan could read a tumor’s DNA?
Radiomics and radiogenomics are unlocking hidden data in medical images to reveal renal cell carcinoma biology, heterogeneity, and prognosis—non-invasively. A new review shows how this tech is powering smarter, more personalized cancer care.
Radiomics and radiogenomics are unlocking hidden data in medical images to reveal renal cell carcinoma biology, heterogeneity, and prognosis—non-invasively. A new review shows how this tech is powering smarter, more personalized cancer care.
January 14, 2026 at 10:00 AM
🧬📊 What if a CT scan could read a tumor’s DNA?
Radiomics and radiogenomics are unlocking hidden data in medical images to reveal renal cell carcinoma biology, heterogeneity, and prognosis—non-invasively. A new review shows how this tech is powering smarter, more personalized cancer care.
Radiomics and radiogenomics are unlocking hidden data in medical images to reveal renal cell carcinoma biology, heterogeneity, and prognosis—non-invasively. A new review shows how this tech is powering smarter, more personalized cancer care.
🩺 New #InsightsintoImaging study in Biliary Atresia (BA).
A large two-center study of 1,352 infants achieved outstanding diagnostic accuracy for BA:
📈 AUC up to 0.972 across training, validation & test cohorts
📏 Optimal MR-TCT cutoff: 3.75 mm (AUC 0.828)
A large two-center study of 1,352 infants achieved outstanding diagnostic accuracy for BA:
📈 AUC up to 0.972 across training, validation & test cohorts
📏 Optimal MR-TCT cutoff: 3.75 mm (AUC 0.828)
January 13, 2026 at 10:00 AM
🩺 New #InsightsintoImaging study in Biliary Atresia (BA).
A large two-center study of 1,352 infants achieved outstanding diagnostic accuracy for BA:
📈 AUC up to 0.972 across training, validation & test cohorts
📏 Optimal MR-TCT cutoff: 3.75 mm (AUC 0.828)
A large two-center study of 1,352 infants achieved outstanding diagnostic accuracy for BA:
📈 AUC up to 0.972 across training, validation & test cohorts
📏 Optimal MR-TCT cutoff: 3.75 mm (AUC 0.828)
Breast screening saves lives - but over-detection and false positives remain major challenges across mammography, DBT, US, and MRI, especially for BI-RADS 3 and 4A lesions.
January 7, 2026 at 10:01 AM
Breast screening saves lives - but over-detection and false positives remain major challenges across mammography, DBT, US, and MRI, especially for BI-RADS 3 and 4A lesions.
📢 AI in Radiology: Guidance for Safe Use
The ESR eHealth & Informatics Subcommittee provides practical recommendations for post-market surveillance of AI as medical devices (AIaMDs).
The ESR eHealth & Informatics Subcommittee provides practical recommendations for post-market surveillance of AI as medical devices (AIaMDs).
January 6, 2026 at 10:02 AM
📢 AI in Radiology: Guidance for Safe Use
The ESR eHealth & Informatics Subcommittee provides practical recommendations for post-market surveillance of AI as medical devices (AIaMDs).
The ESR eHealth & Informatics Subcommittee provides practical recommendations for post-market surveillance of AI as medical devices (AIaMDs).
📢 RADUCATION: Digital Learning for Radiology Residents
📚 A peer-developed platform offering curriculum-based content and original board exam questions.
📚 A peer-developed platform offering curriculum-based content and original board exam questions.
January 5, 2026 at 12:02 PM
📢 RADUCATION: Digital Learning for Radiology Residents
📚 A peer-developed platform offering curriculum-based content and original board exam questions.
📚 A peer-developed platform offering curriculum-based content and original board exam questions.
📢 New study in #InsightsintoImaging
A study of 193 patients shows that deep learning reconstruction (DLR) for diffusion-weighted imaging (DWI) can halve acquisition time while improving image quality and lesion differentiation.
A study of 193 patients shows that deep learning reconstruction (DLR) for diffusion-weighted imaging (DWI) can halve acquisition time while improving image quality and lesion differentiation.
January 5, 2026 at 9:02 AM
📢 New study in #InsightsintoImaging
A study of 193 patients shows that deep learning reconstruction (DLR) for diffusion-weighted imaging (DWI) can halve acquisition time while improving image quality and lesion differentiation.
A study of 193 patients shows that deep learning reconstruction (DLR) for diffusion-weighted imaging (DWI) can halve acquisition time while improving image quality and lesion differentiation.
📢 New study in #InsightsintoImaging
🩻 Predicting PRCC Recurrence with Preoperative CT
A multicenter study (n=266) developed a CT-based radiological score using tumor margin irregularity and regional lymph node size to predict papillary renal cell carcinoma recurrence after surgery.
✨ Key points:
🩻 Predicting PRCC Recurrence with Preoperative CT
A multicenter study (n=266) developed a CT-based radiological score using tumor margin irregularity and regional lymph node size to predict papillary renal cell carcinoma recurrence after surgery.
✨ Key points:
January 2, 2026 at 1:01 PM
📢 New study in #InsightsintoImaging
🩻 Predicting PRCC Recurrence with Preoperative CT
A multicenter study (n=266) developed a CT-based radiological score using tumor margin irregularity and regional lymph node size to predict papillary renal cell carcinoma recurrence after surgery.
✨ Key points:
🩻 Predicting PRCC Recurrence with Preoperative CT
A multicenter study (n=266) developed a CT-based radiological score using tumor margin irregularity and regional lymph node size to predict papillary renal cell carcinoma recurrence after surgery.
✨ Key points:
A multicenter study (1,182 patients) shows a MIL model combining radiomics & deep learning can accurately predict invasiveness of ground-glass nodules in lung adenocarcinoma.
✨ MIL-DL-Rad model: robust AUC up to 0.936, supporting precise preoperative planning.
✨ MIL-DL-Rad model: robust AUC up to 0.936, supporting precise preoperative planning.
January 2, 2026 at 10:01 AM
A multicenter study (1,182 patients) shows a MIL model combining radiomics & deep learning can accurately predict invasiveness of ground-glass nodules in lung adenocarcinoma.
✨ MIL-DL-Rad model: robust AUC up to 0.936, supporting precise preoperative planning.
✨ MIL-DL-Rad model: robust AUC up to 0.936, supporting precise preoperative planning.
🫀 Lipomatous Metaplasia: A Strong Predictor After Myocardial Infarction
A multicenter study of 1,702 post-MI patients shows that lipomatous metaplasia (LM) detected on CCTA is a significant predictor of major adverse cardiovascular events (MACE)—even stronger than CMR parameters and infarct size.
A multicenter study of 1,702 post-MI patients shows that lipomatous metaplasia (LM) detected on CCTA is a significant predictor of major adverse cardiovascular events (MACE)—even stronger than CMR parameters and infarct size.
December 30, 2025 at 10:01 AM
🫀 Lipomatous Metaplasia: A Strong Predictor After Myocardial Infarction
A multicenter study of 1,702 post-MI patients shows that lipomatous metaplasia (LM) detected on CCTA is a significant predictor of major adverse cardiovascular events (MACE)—even stronger than CMR parameters and infarct size.
A multicenter study of 1,702 post-MI patients shows that lipomatous metaplasia (LM) detected on CCTA is a significant predictor of major adverse cardiovascular events (MACE)—even stronger than CMR parameters and infarct size.
🩺 Mapping Obstructions in Azoospermia with Ultrasound
A retrospective study (n=166) shows that combined transscrotal–transrectal ultrasonography can accurately predict obstruction sites in patients with obstructive azoospermia.
A retrospective study (n=166) shows that combined transscrotal–transrectal ultrasonography can accurately predict obstruction sites in patients with obstructive azoospermia.
December 29, 2025 at 2:01 PM
🩺 Mapping Obstructions in Azoospermia with Ultrasound
A retrospective study (n=166) shows that combined transscrotal–transrectal ultrasonography can accurately predict obstruction sites in patients with obstructive azoospermia.
A retrospective study (n=166) shows that combined transscrotal–transrectal ultrasonography can accurately predict obstruction sites in patients with obstructive azoospermia.
🏥 Preoperative AI for Papillary Thyroid Carcinoma
A multicenter study (n=1,215) developed a habitat imaging + multimodal AI model using CT, ultrasound, clinical, and immune data to predict low-risk PTC patients before surgery.
A multicenter study (n=1,215) developed a habitat imaging + multimodal AI model using CT, ultrasound, clinical, and immune data to predict low-risk PTC patients before surgery.
December 29, 2025 at 10:01 AM
🏥 Preoperative AI for Papillary Thyroid Carcinoma
A multicenter study (n=1,215) developed a habitat imaging + multimodal AI model using CT, ultrasound, clinical, and immune data to predict low-risk PTC patients before surgery.
A multicenter study (n=1,215) developed a habitat imaging + multimodal AI model using CT, ultrasound, clinical, and immune data to predict low-risk PTC patients before surgery.
🎁 Radvent Day 24! 🎁
Neuropathic skeletal injuries in kids aren’t just rare curiosities: they show how sensory loss can lead to major joint damage, repeated fractures, and deformities that are easily missed without imaging.
Neuropathic skeletal injuries in kids aren’t just rare curiosities: they show how sensory loss can lead to major joint damage, repeated fractures, and deformities that are easily missed without imaging.
December 24, 2025 at 9:01 AM
🎁 Radvent Day 24! 🎁
Neuropathic skeletal injuries in kids aren’t just rare curiosities: they show how sensory loss can lead to major joint damage, repeated fractures, and deformities that are easily missed without imaging.
Neuropathic skeletal injuries in kids aren’t just rare curiosities: they show how sensory loss can lead to major joint damage, repeated fractures, and deformities that are easily missed without imaging.
🎄✨ Radvent Day 23 — The “Contrast Conundrum” for Kidney Patients ✨🎄
Behind today’s advent door: intravenous iodinated contrast media (ICM) in patients with kidney disease - a classic radiology balancing act! Guidelines exist, but they vary more than holiday cookie recipes. 🍪
Behind today’s advent door: intravenous iodinated contrast media (ICM) in patients with kidney disease - a classic radiology balancing act! Guidelines exist, but they vary more than holiday cookie recipes. 🍪
December 23, 2025 at 9:01 AM
🎄✨ Radvent Day 23 — The “Contrast Conundrum” for Kidney Patients ✨🎄
Behind today’s advent door: intravenous iodinated contrast media (ICM) in patients with kidney disease - a classic radiology balancing act! Guidelines exist, but they vary more than holiday cookie recipes. 🍪
Behind today’s advent door: intravenous iodinated contrast media (ICM) in patients with kidney disease - a classic radiology balancing act! Guidelines exist, but they vary more than holiday cookie recipes. 🍪