Radiologia Medica最新文献

Purpose: Build machine learning (ML) models able to predict pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) in breast cancer (BC) patients based on conventional and radiomic signatures extracted from baseline [¹⁸F]FDG PET/CT.

Material and methods: Primary tumor and the most significant lymph node metastasis were manually segmented in baseline [¹⁸F]FDG PET/CT of 52 newly diagnosed BC patients. Clinical parameters, NAC and conventional semiquantitative PET parameters were collected. The standard of reference considered was surgical pCR after NAC (ypT0;ypN0). Eight-hundred-fifty-four radiomic features (RFts) were extracted from both PET and CT datasets, according to IBSI; robust RFTs were selected. The cohort was split in training (70%) and validation (30%) sets. Four ML Models (Clinical Model, CT Model, PET Model_T and PET Model_T + N) each one with 3 learners (Random Forest (RF), Neural Network and Stochastic Gradient Descendent) were trained and tested using RFts and clinical signatures. PET Models were built considering robust RFTs extracted from either primary tumor alone (PET Model_T) or also including the reference lymph node (PET Model_T + N).

Results: 72 pathological uptakes (52 primary BC and 20 lymph node metastasis) at [¹⁸F]FDG PET/CT were segmented. pCR occurred in 44.2% cases. Twelve, 46 and 141 robust RFts were selected from CT Model, PET Model_T and PET Model_T + N, respectively. PET Models showed better performance than CT and Clinical Models. The best performances were obtained by the RF algorithm of the PET Model_T + N (AUC = 0.83;CA = 0.74;TP = 78%;TN = 72%).

Conclusion: ML models trained on PET/CT radiomic features extracted from primary BC and lymph node metastasis could concur in the prediction of pCR after NAC and improve BC management.

Purpose: Bodyweight loss is commonly found in Nasopharyngeal Carcinoma patients during Concurrent Chemo-radiotherapy (CCRT) and has implications for treatment decisions. However, the prognostic value of this weight loss remains uncertain. We addressed it by proposing a novel index Weight Censorial Score (WCS) that characterizes the patient-specific CCRT response on actual to estimated weight loss.

Methods: A retrospective study included 315 patients from two independent hospitals. An Estimated WCS (eWCS) was obtained through linear regression of image and dosimetry features. The eWCS was converted to an estimated net weight loss (nWL), with its accuracy evaluated. The Determined WCS (dWCS) was calculated by centering and scaling the post-RT actual nWL with patient's pre-RT body information. The ratio of dWCS to eWCS (WCS ratio) reflected the actual to estimated weight loss of a patient. The prognostic ability of WCS ratio dichotomized at 1 was evaluated.

Results: The mean absolute error of estimated to actual nWL was 1.84 kg. Patients who had their actual WL larger than estimated WL were found to have significantly worse OS (p = 0.005, HR = 3.35[1.45-7.73]), PFS (p = 0.038, HR = 1.86[1.03-3.35]), and DMFS (p = 0.050, HR = 2.20[1.00-4.85]), respectively, in multivariable cox analysis. They were also found not to benefit from adjuvant chemotherapy (p = 0.572), whereas the adjuvant chemotherapy provided significant PFS benefit in patients with actual WL smaller than estimated WL (p = 0.036, HR = 0.53[0.29-0.96]).

Conclusion: The nWL of patient during CCRT can be reasonably estimated by dosimetry factors at pre-RT stage. The prognostic value of the actual to expected weight loss holds promise for highlighting vulnerable patients after CCRT.

Purpose: To develop an artificial intelligence (AI) algorithm for automated measurements of spinopelvic parameters on lateral radiographs and compare its performance to multiple experienced radiologists and surgeons.

Methods: On lateral full-spine radiographs of 295 consecutive patients, a two-staged region-based convolutional neural network (R-CNN) was trained to detect anatomical landmarks and calculate thoracic kyphosis (TK), lumbar lordosis (LL), sacral slope (SS), and sagittal vertical axis (SVA). Performance was evaluated on 65 radiographs not used for training, which were measured independently by 6 readers (3 radiologists, 3 surgeons), and the median per measurement was set as the reference standard. Intraclass correlation coefficient (ICC), mean absolute error (MAE), and standard deviation (SD) were used for statistical analysis; while, ANOVA was used to search for significant differences between the AI and human readers.

Results: Automatic measurements (AI) showed excellent correlation with the reference standard, with all ICCs within the range of the readers (TK: 0.92 [AI] vs. 0.85-0.96 [readers]; LL: 0.95 vs. 0.87-0.98; SS: 0.93 vs. 0.89-0.98; SVA: 1.00 vs. 0.99-1.00; all p < 0.001). Analysis of the MAE (± SD) revealed comparable results to the six readers (TK: 3.71° (± 4.24) [AI] v.s 1.86-5.88° (± 3.48-6.17) [readers]; LL: 4.53° ± 4.68 vs. 2.21-5.34° (± 2.60-7.38); SS: 4.56° (± 6.10) vs. 2.20-4.76° (± 3.15-7.37); SVA: 2.44 mm (± 3.93) vs. 1.22-2.79 mm (± 2.42-7.11)); while, ANOVA confirmed no significant difference between the errors of the AI and any human reader (all p > 0.05). Human reading time was on average 139 s per case (range: 86-231 s).

Conclusion: Our AI algorithm provides spinopelvic measurements accurate within the variability of experienced readers, but with the potential to save time and increase reproducibility.

Purpose: This study aimed to assess the posterior cruciate ligament (PCL) angle in anterior cruciate ligament (ACL) deficient knees and correlate it with anatomical and demographic factors such as tibial slope, anterior tibial translation, age, gender, and time of injury.

Material and methods: Patients were eligible for inclusion if they were clinically diagnosed with an ACL tear confirmed by MRI. For each patient, the following parameters were evaluated: PCL angle (PCLA), medial tibial slope (MTS), lateral tibial slope (LTS), medial anterior tibial translation (MATT), and lateral anterior tibial translation (LATT).

Results: A total of 193 patients were included in the study, comprising 91 (47.2%) females and 102 (52.8%) males, with a mean age of 30.27 ± 12.54 years. The mean time from injury to MRI was 14.18 ± 55.77 days. In the overall population, the mean PCL angle was 128.72 ± 10.33°, the mean medial tibial slope was 3.57 ± 2.33°, and the mean lateral tibial slope was 6.07 ± 3.52°. The mean medial and lateral anterior tibial translations were 4.76 ± 2.02 mm and 7.01 ± 2.48 mm, respectively. In 190 cases (98.4%), the PCL angle was ≥ 105°. The PCL angle negatively correlated with medial and lateral anterior tibial translation (p < 0.05). Females exhibited a higher PCL angle compared to males (p = 0.019).

Conclusion: In the context of ACL lesions, the PCL angle has a normal value in acute injuries (> 105°) and decreases over time. The PCL angle is negatively correlated with anterior tibial translation, and females have a higher PCL angle compared to males.

Level of evidence iv: Retrospective Cohort.

Purpose: To compare the performance of ultrafast MRI with standard MRI in classifying histological factors and subtypes of invasive breast cancer among radiologists with varying experience.

Methods: From October 2021 to November 2022, this prospective study enrolled 225 participants with 233 breast cancers before treatment (NCT06104189 at clinicaltrials.gov). Tumor segmentation on MRI was performed independently by two readers (R1, dedicated breast radiologist; R2, radiology resident). We extracted 1618 radiomic features and four kinetic features from ultrafast and standard images, respectively. Logistic regression algorithms were adopted for prediction modeling, following feature selection by the least absolute shrinkage and selection operator. The performance of predicting histological factors and subtypes was evaluated using the area under the receiver-operating characteristic curve (AUC). Performance differences between MRI methods and radiologists were assessed using the DeLong test.

Results: Ultrafast MRI outperformed standard MRI in predicting HER2 status (AUCs [95% CI] of ultrafast MRI vs standard MRI; 0.87 [0.83-0.91] vs 0.77 [0.64-0.90] for R1 and 0.88 [0.83-0.91] vs 0.77 [0.69-0.84] for R2) (all P < 0.05). Both ultrafast MRI and standard MRI showed comparable performance in predicting hormone receptors. Ultrafast MRI exhibited superior performance to standard MRI in classifying subtypes. The classification of the luminal subtype for both readers, the HER2-overexpressed subtype for R2, and the triple-negative subtype for R1 was significantly better with ultrafast MRI (P < 0.05).

Conclusion: Ultrafast MRI-based radiomics holds promise as a noninvasive imaging biomarker for classifying hormone receptors, HER2 status, and molecular subtypes compared to standard MRI, regardless of radiologist experience.

Purpose: Thread release of the carpal tunnel is the most recent of several minimally invasive ultrasound-guided carpal tunnel release techniques. The purpose of this article is to provide a step-by-step guide for minimally invasive, ultrasound-guided thread release of the carpal tunnel focused on transecting the transverse carpal ligament with minimal damage to the palmar aponeurosis on anatomical specimens.

Methods: Fifteen ultrasound-guided carpal tunnel thread releases were performed on the wrists of soft-embalmed anatomical specimens, which were dissected immediately after the intervention. The procedures were performed by two musculoskeletal radiologists with 25 and 8 years of experience, respectively, in interventional radiology. Ultrasound visibility, completeness of transection, and damage to surrounding structures were evaluated on a score from 1 to 3.

Results: We achieved a complete transection of the transverse carpal ligament in 11 of 15 interventions (73%) and an incomplete transection in the remaining four (27%). No neural or vascular structures were harmed. In two cases (13%), there was irrelevant damage to flexor tendons. The ultrasound visibility was rough in five specimens (33.3%), moderate in five (33.3%), and optimal in five (33.3%). Essential structures were delineated in all cases.

Conclusion: Thread release of the carpal tunnel leads to only minimal damage to skin, as well as structures within the carpal tunnel and the palmar aponeurosis, promising a low amount of postinterventional complications.

Relevance statement: Our study showed that minimally invasive ultrasound-guided thread release of the carpal tunnel is a feasible approach in the anatomical model. The results may provide a basis for further research and refinement of this technique.

Background: Uterine fibroids are prevalent benign pelvic tumors, often causing debilitating symptoms that impair quality of life. Uterine fibroid embolization (UFE) is a consolidated minimally invasive treatment option. The purpose of this study is to report our experience with polyethylene glycol microspheres (HydroPearl) in UFE for symptomatic patients.

Methods: This single-center retrospective study evaluated 37 consecutive patients with symptomatic uterine fibroids referred to our institution since November 2016 to February 2020 for UFE with HydroPearl microspheres. All patients included completed a comprehensive pre-UFE clinical assessment and underwent a magnetic resonance imaging (MRI) pre- and post-procedure. Technical success, procedure-related complications, clinical outcomes, follow-up MRI findings, and patient satisfaction were evaluated.

Results: A technical success rate of 97% with complete bilateral uterine artery embolization was achieved. Significant improvements were observed in menorrhagia, bulk-type symptoms, abdominal pain, and urinary dysfunctions after UFE. Post-procedural MRI assessments demonstrated reductions in uterine and dominant fibroid volumes, indicating successful devascularization. No immediate procedural complications were reported. Symptoms interfering with everyday activities significantly improved after the procedure. Patient satisfaction was high, with 89% expressing satisfaction with the treatment and 84% indicating purpose to repeat the procedure if necessary.

Conclusions: Uterine artery embolization with HydroPearl is a safe and effective treatment option for symptomatic uterine fibroids. A significant improvement in menorrhagia and bulk-type symptoms after the procedure was observed correlated by a reduction in diameters and volumes of both the uterus and the main fibroid in post-procedural MRI.

Objectives: To evaluate the at-risk organs that require protection during percutaneous cryoablation (PCA) of renal tumours and the correlation with patient and target lesion characteristics, type of protective measure used and postoperative outcomes.

Materials and methods: Single-centre retrospective review of patients with renal tumours who underwent PCA between 2008 and 2020. Final analysis included 374 tumours. Patient, tumour, and procedure technical details were extracted and analysed. At-risk organs were classified according to tumour location relative to kidney side, pyelic axis, and polar lines.

Results: There were 171 (46.0%) tumours in the left kidney, and 194 (52.0%) in the right. Cryoprotection was required for 272 (272/374; 73.0%) tumours, with hydrodissection (216/374; 58.0%) being the most common technique. Protective measures were used for 82 (82/93; 88.0%) tumours in under/normal-weight patients and 143 (143/196; 73.0%) in overweight/obese ones (P = 0.004). In the left kidney, colon was the most common at-risk organ (63/171; 37.0%), followed by spleen (21/171; 12.3%), small bowel (21/171; 12.3%), ureter (19/171; 11.1%), abdominal wall (15/171; 8.8%), psoas muscle (10/171; 5.8%), and pancreas (9/171; 5.3%). In the right kidney, common at-risk organs were the colon (67/194; 35.0%), liver (50/194; 25.7%), ureter (15/194; 15.5%), diaphragm (16/194; 8.2%), abdominal wall (14/194; 7.2%), and duodenum (12/194; 6.1%). No cryoinjuries to at-risk organs occurred.

Conclusion: Hydrodissection is the most common cryoprotective measure used for renal tumour PCA. Under/normal-weight patients are more likely to require cryoprotection. The colon is the most common adjacent at-risk organ requiring protection for both right- and left-sided tumours.

Background: Accurate differentiation between benign and malignant pancreatic lesions is critical for effective patient management. This study aimed to develop and validate a novel deep learning network using baseline computed tomography (CT) images to predict the classification of pancreatic lesions.

Methods: This retrospective study included 864 patients (422 men, 442 women) with confirmed histopathological results across three medical centers, forming a training cohort, internal testing cohort, and external validation cohort. A novel hybrid model, Multi-Scale Large Kernel Attention with Mobile Vision Transformer (MVIT-MLKA), was developed, integrating CNN and Transformer architectures to classify pancreatic lesions. The model's performance was compared with traditional machine learning methods and advanced deep learning models. We also evaluated the diagnostic accuracy of radiologists with and without the assistance of the optimal model. Model performance was assessed through discrimination, calibration, and clinical applicability.

Results: The MVIT-MLKA model demonstrated superior performance in classifying pancreatic lesions, achieving an AUC of 0.974 (95% CI 0.967-0.980) in the training set, 0.935 (95% CI 0.915-0.954) in the internal testing set, and 0.924 (95% CI 0.902-0.945) in the external validation set, outperforming traditional models and other deep learning models (P < 0.05). Radiologists aided by the MVIT-MLKA model showed significant improvements in diagnostic accuracy and sensitivity compared to those without model assistance (P < 0.05). Grad-CAM visualization enhanced model interpretability by effectively highlighting key lesion areas.

Conclusion: The MVIT-MLKA model efficiently differentiates between benign and malignant pancreatic lesions, surpassing traditional methods and significantly improving radiologists' diagnostic performance. The integration of this advanced deep learning model into clinical practice has the potential to reduce diagnostic errors and optimize treatment strategies.