European Journal of Radiology最新文献

Objective: We aimed to construct a risk prediction model for PSL after posterior lumbar fusion using machine learning and radiomic methods.

Summary of background information: Posterior lumbar fusion surgery is a routine procedure for the treatment of lumbar degenerative disease (LDD). Pedicle screw loosening (PSL) is a common complication after posterior lumbar fusion that can lead to postoperative pain and infection in the implant area.

Methods: A total of 304 patients who underwent posterior lumbar fusion for LDD at Sun Yat-sen Memorial Hospital were reviewed in this study. 80 patients at the Seventh Affiliated Hospital of SYSU and Guangzhou First People's Hospital made up the external validation datasets. Preoperative demographic and surgical information was collected. One-year-postoperative radiological follow-up was conducted, and patients were divided into PSL and non-PSL groups. Clinical and surgical information was subjected to Student's t test for feature selection. Lumbar CT images were analyzed using radiomic methods, and PSL prediction models were constructed by machine learning methods. The best model was selected and externally validated on the data from the two other hospitals.

Results: All 304 patients were included. Age (p＜0.001), Preoperative lumbar lordosis (LL) (p = 0.006), mean CT Hounsfield units (p = 0.007), and number of fixed segments (p < 0.001) differed between the two groups. Logistic regression revealed that the number of fixed segments was an independent risk factor (OR = 2.147, p < 0.001). Radiomic features were selected after feature extraction and selection. After model training and testing, the clinical + radiomic model showed acceptable predictive performance (AUC 0.894). Its AUCs in the two external validation datasets were 0.821 and 0.892 respectively.

Conclusion: Combining clinical and radiomic features can better predict the risk of PSL after posterior lumbar fusion surgery. This investigation revalidated the risk factors for PSL after posterior lumbar fusion as well.

Objectives: To evaluate the diagnostic performance of iodine density map (IDM) using dual-energy CT (DECT) in differentiating colorectal cancer from benign colorectal wall thickening.

Material and methods: This IRB-approved dual-center retrospective exploratory study included 71 consecutive patients with colorectal wall thickening due to tumoral or non-tumoral origin, confirmed by colonoscopy and assessed with DECT. Thirty-eight had pathology-proven colorectal adenocarcinoma, and 33 had non-neoplastic thickening (inflammation, physiologic collapse, or post-radiotherapy change). Iodine density values were measured from regions of interest on three consecutive CT slices on portal venous phase, normalized to aortic iodine concentration. Diagnostic performance was determined by ROC analysis, and inter-reader agreement was evaluated with intraclass correlation coefficients (ICC) and limits of agreement (LOA).

Results: Mean IDM value was significantly higher in tumors than in benign thickening (2.31 ± 0.42 vs 1.43 ± 0.24 mg/mL; p < 0.001). Mean normalized iodine density map (NIDM) value was likewise elevated in tumors (0.46 ± 0.10 vs 0.31 ± 0.07; p < 0.001). ROC analysis demonstrated excellent performance for both IDM (AUC 0.98; optimal cutoff 1.72 mg/mL; sensitivity 92.1%, specificity 90.9%, NPV 90.9%) and NIDM (AUC 0.88; cutoff 0.35; sensitivity 92.1%, specificity 69.7%, NPV 88.7%). There was no significant difference between the inflammatory and collapsed-wall subgroups. Inter-reader agreement was excellent (ICC: 0.93 for IDM; 0.92 for NIDM).

Conclusion: IDM and NIDM on DECT provide robust, reproducible markers that differentiate colorectal carcinoma from benign wall thickening with high diagnostic accuracy. These quantitative parameters may improve diagnostic confidence and reduce unnecessary colonoscopies, supporting their integration into colorectal cancer evaluation.

Objective: To explore the clinical value of parameters of dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging (DWI) in the diagnosis of pancreatic cancer and its invasiveness prediction.

Materials and methods: A total of 122 participants, comprising 58 patients with pathologically confirmed pancreatic cancer (median age 60.50 years [interquartile range, 54.00-65.25 years]; 38 males), 34 pancreatitis patients (median age 55.50 years[interquartile range, 43.00-67.25 years]; 23 males), and 30 healthy volunteers (median age 52.00 years[interquartile range, 36.00-62.25 years]; 16 males), underwent DWI and DCE-MRI scans. The values of K^trans (volume transfer constant), K_ep (rate constant), V_e (extravascular extracellular volume fraction), V_p (plasma volume fraction), TTP (Time to Peak), MAX Conc (Maximum Concentration), AUC (Area Under the Curve), and MAXSlope (Maximum Slope) from DCE sequence as well as the values of ADC and eADC from DWI sequence were collected. We also collected data on tumor invasion and metastasis within the pancreatic cancer cohort and divided it into ten subgroups. We analyzed the differences in parameters between pancreatic cancer, pancreatitis, and normal groups, and differences in parameters within pathologic subgroups of the pancreatic cancer group to distinguish among the groups. The area under the receiver operating characteristic curve (AUC) was used to assess diagnostic performance.

Results: The most prominent parameter to differentiate pancreatic cancer from pancreatitis was TTP, with AUC of 0.897, sensitivity of 86.2% and specificity of 91.2%. K^trans values were higher in the group with organ invasion than in the group without organ invasion (p = 0.007), higher in the group with nerve invasion than in the group without nerve invasion (p = 0.021), and lower in the group with implantation metastasis than in the group without implantation metastasis (p = 0.009) for pancreatic cancer.

Conclusion: DCE-MRI and DWI parameters were of clinical value in the diagnosis of pancreatic cancer. DCE parameters, especially K_trans, can be used to predict the occurrence of invasion or metastasis in pancreatic cancer.

Background: The central vein sign (CVS) is a neuroimaging biomarker in multiple sclerosis (MS) with high diagnostic specificity. CVS is best detected with high-quality susceptibility-sensitive MRI sequences. For concurrent detection of lesions and veins, FLAIR* was developed as a post-processing method to provide contrast for T2 hyperintense lesions (FLAIR) and paramagnetic hypointense veins (T2*-weighted). Occasionally, CVS-like features have been noted on FLAIR, but the reliability of this finding is unknown.

Objective: To compare the central FLAIR hypointensity to FLAIR* CVS.

Methods: Scans from the CentrAl Vein Sign in MS (CAVS-MS) pilot study were included for the analysis. A blinded rater assessed all lesions for CVS on 3-tesla post-contrast FLAIR*. A second blinded rater assessed the same lesions for central hypointensity on FLAIR images alone. Counts were compared between methods. The same approach was applied for a subset with available non-contrast FLAIR* lesion ratings.

Results: With post-contrast FLAIR* CVS as the standard (n= 92; 1737 lesions), central FLAIR hypointensity demonstrated concordance of 64%, with sensitivity of 34% (95% CI, 30-37%) and specificity of 83% (95% CI, 81-85%). With non-contrast FLAIR* CVS as the standard (n= 38; 768 lesions), FLAIR demonstrated sensitivity of 40% (95% CI, 33-47%) and specificity of 85% (95% CI, 82-88%). Select 6 (≥6 central hypointense lesions) FLAIR was 59% accurate for a diagnosis of MS, with a lower specificity (63% vs. 90%, p= 0.008) in comparison to post-contrast FLAIR*.

Conclusions: Assessment of CVS on FLAIR alone is unreliable and requires susceptibility-sensitive sequences to be clinically useful.

Rationale and objectives: Adenomyosis significantly impairs quality of life, yet optimal uterine-sparing management remains uncertain. While both uterine artery embolization (UAE) and dienogest show promise, no direct comparison exists. This study compared their 6-month efficacy in treating symptomatic adenomyosis.

Patients and methods: This prospective, single-center, randomized controlled trial enrolled 60 patients with MRI-confirmed adenomyosis desiring uterine preservation between January and July 2025, randomly assigned (1:1) to UAE (n = 30) or 2 mg/day dienogest (n = 30). Co-primary outcomes were complete resolution of chronic pelvic pain and heavy menstrual bleeding at 6 months. Secondary outcomes included visual analog scale (VAS) dysmenorrhea scores, Uterine Fibroid Symptom and Quality of Life (UFS-QoL) symptom severity and health-related quality of life (HRQOL) scores, hemoglobin levels, junctional zone thickness, and uterine volume measured at baseline, 3 months, and 6 months.

Results: UAE achieved higher complete resolution rates for chronic pelvic pain (93.3% vs. 66.7%; p = 0.010; relative risk [RR] = 1.40) and heavy menstrual bleeding (90.0% vs. 56.7%; p = 0.004; RR = 1.59), with composite success in 86.7% versus 50.0% (RR = 1.73; number needed to treat = 3). At 6 months, UAE was associated with significantly better outcomes for VAS dysmenorrhea (1.27 ± 1.23 vs. 2.10 ± 1.25; p = 0.016), symptom severity scores (31.48 ± 9.78 vs. 45.36 ± 9.55; p < 0.001), HRQOL scores (66.25 ± 7.35 vs. 50.82 ± 10.11; p < 0.001), hemoglobin levels (10.55 ± 1.49 vs. 10.19 ± 1.55 g/dL; p = 0.010), junctional zone reduction (33.5% vs. 14.0%; p = 0.005), and uterine volume reduction (17.6% vs. 7.3%; p = 0.006).

Conclusion: At 6 months, UAE achieved better symptom relief and anatomical outcomes than dienogest, supporting UAE as an effective uterine-sparing therapy for symptomatic adenomyosis.

Objectives: To compare image quality between low-dose thin-slice deep-learning reconstruction (DLR) and standard-dose thick-slice hybrid iterative reconstruction (HIR) in pediatric abdominal CT.

Methods: 82 children (≤6 years) who underwent contrast-enhanced abdominal CT with standard-dose (STD, n = 41) or low-dose (LD, n = 41) protocol matched by age and weight were retrospectively identified. STD and LD images were reconstructed at 3.0-mm using HIR (STD-HIR/3.0 and LD-HIR/3.0, respectively). LD images were also reconstructed at 0.5-mm using HIR (LD-HIR/0.5) and DLR (LD-DLR/0.5). Size-specific dose estimate (SSDE) was compared between groups. Image noise and contrast-to-noise ratio (CNR) were quantified. Noise power spectrum (NPS) and edge-rise slope (ERS) were employed for noise texture and edge sharpness measures, respectively. For subjective evaluation, noise magnitude, noise texture, edge sharpness, delineation of small structures, and diagnostic confidence were rated on a 5-point scale (1 = undiagnostic, 5 = best).

Results: SSDE was on average 59.5% lower in LD than in STD group (1.7 ± 0.4 vs. 4.2 ± 0.8 mGy, p < 0.001). Image noise was lower in LD-DLR/0.5 compared to STD-HIR/3.0, LD-HIR/3.0, and LD-HIR/0.5 (8.7 ± 1.5, 9.6 ± 1.1, 11.7 ± 1.9, and 17.9 ± 2.3 HU, respectively, all p ≤ 0.017). LD-DLR/0.5 showed equivalent CNR (e.g., liver CNR: 7.1 ± 2.1 vs. 7.0 ± 2.8, p = 0.827) and higher ERS (64.3 ± 23.9 vs. 53.3 ± 10.6 HU/mm, p = 0.011) with similar average NPS frequency (0.281 ± 0.042 vs. 0.291 ± 0.027 mm^-1, p = 0.177) compared to STD-HIR/3.0. Subjective scores for all criteria were higher in LD-DLR/0.5 than in STD-HIR/3.0 (e.g., diagnostic confidence score: 4.5 ± 0.5 vs. 3.4 ± 0.4, p < 0.001).

Conclusion: Low-dose thin-slice DLR improved edge sharpness, small structure visualization, and diagnostic confidence in pediatric abdominal CT without increasing noise compared to standard-dose thick-slice HIR.

Objectives: To develop and validate CT-based radiomics models for the identification of chronic pancreatitis (CP) and selected CP-related complications, and to explore associations between radiomics-derived features and clinical measures.

Material & methods: This retrospective, multicenter study included a training cohort of 349 subjects (201 CP; 148 healthy controls) from Aalborg University Hospital. Test cohort comprised 109 subjects, including 41 pancreas-healthy controls from Aalborg and an external cohort of 68 CP patients from Bergen. Portal venous phase CT scans were automatically segmented, and radiomics features were extracted using PyRadiomics. AI models were trained to classify CP and identify CP-related complications, including exocrine pancreatic insufficiency (EPI), diabetes, and pain. Model performance was assessed using area under the receiver operating characteristic curve (AUC) with 95% confidence intervals (CI).

Results: The CP classification model demonstrated high discriminative performance with an AUC of 0.97 (95% CI: 0.94-0.99). The EPI model showed moderate discriminative performance (AUC 0.80, 95% CI: 0.66-0.88). In contrast, the diabetes and pain models demonstrated lower discriminative performance, with AUCs of 0.63 (95% CI: 0.47-0.77) and 0.59 (95% CI: 0.37-0.67), respectively. Radiomics-derived probability scores correlated significantly with fecal elastase levels (p < 0.001) and increased with greater functional disease severity (p = 0.004).

Conclusion: CT-based radiomics can accurately classify CP and reflect exocrine functional impairment. However, performance for diabetes and pain was limited, and clinical utility beyond established clinical assessments remains to be demonstrated.

Pelvic floor dysfunction encompasses a spectrum of disorders characterized by organ descent, muscular weakness, and impaired coordination across the anterior, middle, and posterior compartments. MRI defecography has become an established non-invasive technique for comprehensive assessment of both static anatomy and dynamic function of the pelvic floor. This review describes the MRI defecography technique, protocol components, and quantitative parameters that provide an objective evaluation of pelvic floor dysfunction. Standard MRI defecography protocol incorporates resting, contraction, straining, and evacuation phases, with single-shot fast spin-echo and real-time sequences enabling visualization of pelvic structures throughout motion. Quantitative parameters, including the anorectal angle, pubococcygeal line, H and M lines, minimal prolapse level, and levator plate angle, allow for objective evaluation of pelvic floor dysfunction. MRI defecography is particularly valuable in complex and postoperative scenarios, allowing detection of cystoceles, rectoceles, enteroceles, uterine or vaginal prolapse, intussusception, and post-surgical complications. In addition, MRI-defecography highlights the importance of stabilizing structures such as the urogenital diaphragm, endopelvic fascia, and levator ani complex. By integrating anatomic and functional findings, MRI defecography supports individualized therapeutic planning, guides surgical decision-making, and improves long-term outcomes. MRI defecography has thus emerged as a cornerstone in the multidisciplinary management of pelvic floor dysfunction.