Intravascular Ultrasound-Guided Transvenous Biopsy of Retroperitoneal Lymph Nodes: Diagnostic Accuracy and Safety Compared with CT-Guided Percutaneous Biopsy

PurposeTo compare the diagnostic accuracy and adverse event rates of intravascular ultrasound (US)-guided transvenous biopsy (TVB) versus those of computed tomography (CT)-guided percutaneous needle biopsy (PNB) for retroperitoneal (RP) lymph nodes.Materials and MethodsIn this single-institution, retrospective study, 32 intravascular US-guided TVB procedures and a sample of 34 CT-guided PNB procedures for RP lymph nodes where targets were deemed amenable to intravascular US-guided TVB were analyzed. Procedural metrics, including diagnostic accuracy, defined as diagnostic of malignancy or a clinically verifiable benign result, and adverse event rates were compared.ResultsThe targets of intravascular US-guided TVB were primarily aortocaval (47%, 15/32) or precaval (34%, 11/32), whereas those of CT-guided PNB were primarily right pericaval (44%, 15/34) or retrocaval (44%, 15/34) (P < .001). The targets of intravascular US-guided TVB averaged 2.4 cm in the long axis (range, 1.3–3.7 cm) compared with 2.9 cm (range, 1.4–5.7 cm) for those of CT-guided PNB (P = .02). There was no difference in the average number of needle passes (3.8 for intravascular US-guided TVB vs 3.9 for CT-guided PNB; P = .68). The diagnostic accuracy was 94% (30/32) and the adverse event rate was 3.1% (1/32) for intravascular US-guided TVB, similar to those of CT-guided PNB (accuracy, 91% [31/34]; adverse event rate, 2.9% [1/34]).ConclusionsIntravascular US-guided TVB had a diagnostic accuracy and adverse event rate similar to CT-guided PNB for RP lymph nodes, indicating that intravascular US-guided TVB may be as safe and effective as conventional biopsy approaches for appropriately selected targets.     

Image-Guided Percutaneous Thermal Ablation of Oligometastatic Ovarian and Non-Ovarian Gynecologic Tumors

PurposeTo assess the safety, feasibility, and efficacy of percutaneous thermal ablation (TA) in the treatment of metastatic gynecologic (GYN) tumors.Materials and MethodsA study cohort of 42 consecutive women (mean age, 59. years; range, 25–78 years) with metastatic GYN tumors (119 metastatic tumors) treated with radiofrequency (n = 47 tumors), microwave (n = 47 tumors), or cryogenic (n = 30 tumors) ablation from over 2,800 ablations performed from January 2001 to January 2019 was identified. The primary GYN neoplasms consisted of ovarian (27 patients; 77 tumors; mean tumor diameter [MTD], 2.50 cm), uterine (7 patients; 26 tumors; MTD, 1.89 cm), endometrial (5 patients; 10 tumors; MTD, 2.8 cm), vaginal (2 patients; 5 tumors; MTD, 2.40 cm), and cervical (1 patient; 1 tumor; MTD, 1.90 cm) cancers. In order of descending frequency, metastatic tumors treated by TA were located in the liver or liver capsule (74%), lungs (13%), and peritoneal implants (9%). Single tumors were also treated in the kidneys, rectus muscle, perirectal soft tissue (2.5%), and retroperitoneal lymph nodes (1.6%). All efficacy parameters of TA and definitions of major and minor adverse events are categorized by the latest Society of Interventional Radiology reporting standards.ResultsThe median follow-up of treated patients was 10 months. After the initial ablation, 95.6% of the patients achieved a complete tumor response confirmed by contrast-enhanced magnetic resonance imaging or computed tomography. On surveillance imaging, 8.5% of the ablated tumors developed local progression over a median follow-up period of 4.1 months. Five of 8 tumors with local recurrence underwent repeated treatment over a mean follow-up period of 18 months, and 4 of 5 tumors achieved complete eradication after 1 additional treatment session that resulted in a secondary efficacy of 80%. The overall technique efficacy of TA was 96.2% over a median follow-up period of 10 months.ConclusionsTA was safe and effective for the local control of metastatic GYN tumors in the lungs, abdomen, and pelvis, with an overall survival rate of 37.5 months and a local progression-free survival rate of 16.5 months, with only 4.8% of treated patients experiencing a major adverse event.     

CT-Guided Percutaneous Cryoablation in Patients with Lung Nodules Mainly Composed of Ground-Glass Opacities

PurposeTo assess the safety and efficacy of cryoablation in patients with lung nodules mainly composed of ground-glass opacities (GGOs).Materials and MethodsIn this retrospective study, 50 patients (mean age, 65.0 years ± 12.3; 28 women) with a diagnosis of lung GGO nodules who underwent cryoablation were included (from June 2016 to June 2021). The local recurrence rate, incidence of regional metastases to lymph nodes, incidence of distant metastases, adverse events, and lung function condition were analyzed.ResultsFollow-up computed tomography (CT) was performed at a mean of 33 months (range, 3–60 months) after the cryoablation procedure. Outcomes were only evaluated in 30 patients. A total of 20 patients were excluded. Of these 20 patients, 10 patients had no cancer detected by histopathological analysis and a diagnosis was made using CT or positron emission tomography (PET)-CT, The other 10 patients had nodules with a diameter of <10 mm and a consolidation-to-tumor ratio of >0.25; thus, histopathological analysis was not performed because of the small nodule size, and patients were diagnosed using CT or PET-CT. The local recurrence rate was 0% (0 of 30). Evidence of regional metastases of lymph nodes was not found in any patients (0%; 0 of 30), and the incidence of distant metastases was 0% (0 of 30). No major adverse events were noted. Lung function recovered to normal within 1 month after cryoablation in all patients.ConclusionsCryoablation may serve as a safe and feasible option for the treatment of lung nodules mainly composed of GGOs.     

Anatomical Variations of the Left Adrenal Vein Encountered During Venous Sampling

PurposeTo identify anatomical variations in the left adrenal vein (LAV) and to evaluate the role of preprocedural contrast-enhanced computed tomography (CT) planning.MethodsThe length of the left adrenal central vein (LACV), the vessel that receives blood from all tributaries of the left adrenal gland, was measured using venograms of patients who had undergone adrenal venous sampling (AVS) for the diagnosis of primary aldosteronism between October 2017 and December 2019. The anatomical variants of the LAV were described and classified. Contrast-enhanced CT was used to evaluate the detection rate of the following: (a) confluence of the left inferior phrenic vein and the LAV and (b) the last tributary flowing into the LAV.ResultsIn total, 311 patients (143 men, 168 women; mean age: 49.3 years ± 11.0) were enrolled. Of them, 9 (2.9%) patients had anatomical variants lacking a LACV. In patients with a LACV (n = 302), the venographic LACV length was 9.0 mm ± 3.9 (<1 mm in 9 patients). The detection rate of the confluence of the left inferior phrenic vein and LAV, as determined using contrast-enhanced CT, was high (96.2%), whereas that of the last tributary flowing into the LAV was low (0.8%). In 4 of 18 patients with short or absent LACV, the variant was visualized using contrast-enhanced CT.ConclusionsIn some patients, the LACV is absent or short, which is an anatomical variation. Understanding venographic anatomical variations can help avoid misleading results resulting from a suboptimal sampling site in AVS. For some subtypes, contrast-enhanced CT may also help in planning the AVS procedure.     

Deep learning with convolutional neural network for estimation of the characterisation of coronary plaques: Validation using IB-IVUS

IntroductionDeep learning approaches have shown high diagnostic performance in image classifications, such as differentiation of malignant tumors and calcified coronary plaque. However, it is unknown whether deep learning is useful for characterizing coronary plaques without the presence of calcification using coronary computed tomography angiography (CCTA). The purpose of this study was to compare the diagnostic performance of deep learning with a convolutional neural network (CNN) with that of radiologists in the estimation of coronary plaques.MethodsWe retrospectively enrolled 178 patients (191 coronary plaques) who had undergone CCTA and integrated backscatter intravascular ultrasonography (IB-IVUS) studies. IB-IVUS diagnosed 81 fibrous and 110 fatty or fibro-fatty plaques. We manually captured vascular short-axis images of the coronary plaques as Portable Network Graphics (PNG) images (150 × 150 pixels). The display window level and width were 100 and 700 Hounsfield units (HU), respectively. The deep-learning system (CNN; GoogleNet Inception v3) was trained on 153 plaques; its performance was tested on 38 plaques. The area under the curve (AUC) obtained by receiver operating characteristic analysis of the deep learning system and by two board-certified radiologists was compared.ResultsWith the CNN, the AUC and the 95% confidence interval were 0.83 and 0.69–0.96, respectively; for radiologist 1 they were 0.61 and 0.42–0.80; for radiologist 2 they were 0.68 and 0.51–0.86, respectively. The AUC for CNN was significantly higher than for radiologists 1 (p = 0.04); for radiologist 2 it was not significantly different (p = 0.22).ConclusionDL-CNN performed comparably to radiologists for discrimination between fatty and fibro-fatty plaque on CCTA images.Implications for practiceThe diagnostic performance of the CNN and of two radiologists in the assessment of 191 ROIs on CT images of coronary plaques whose type corresponded with their IB-IVUS characterization was comparable.     

Natural History of Incidental Enhancing Nodules on Cone-Beam Computed Tomography during Transarterial Therapy of Hepatocellular Carcinoma

PurposeTo evaluate the natural history of incidental enhancing nodules (IENs) on contrast-enhanced cone-beam computed tomography (CT) during transarterial treatment of hepatocellular carcinoma (HCC).Material and MethodsA single-center retrospective analysis of 100 patients with HCC who underwent contrast-enhanced cone-beam CT prior to transarterial treatment from August 2015 to June 2019 was performed. Inclusion criteria were patients with segmental distribution sublobar HCC, contrast-enhanced cone-beam CT of the target lesion and nontarget liver parenchyma, and follow-up cross-sectional imaging. Patients with IENs ≥3 mm that did not meet imaging criteria for HCC were analyzed. Exclusion criteria included biphenotypic tumors and IEN present inside the treated area of the liver.ResultsFifty-six patients demonstrated 154 IENs on contrast-enhanced cone-beam CT, of which 13 IENs (8.5%) progressed to HCC. The mean primary tumor size was 29 mm (range: 10.2–189 mm). Ten patients had ≥4 IENs, and 46 patients had 1–3 IENs. The mean IEN size was 6.8 mm (range: 3.0–16.3 mm). The median follow-up interval after contrast-enhanced cone-beam CT was 282 days (interquartile range: 143–522). Increased alpha-fetoprotein before treatment (≥15.5 ng/mL, P = .035), having ≥4 IENs (P = .020), and hepatitis C virus (P = .015) were significantly correlated with IEN progression to HCC. No statistically significant differences were identified in baseline neutrophil-to-lymphocyte ratio, targeted HCC characteristics (size, macrovascular invasion, infiltrative pattern, enhancement pattern, and satellite lesions), and IEN size between those with IEN progression to HCC and those without.ConclusionsMost IENs of ≥3 mm on contrast-enhanced cone-beam CT in patients with segmental distribution sublobar HCC do not progress to HCC. Patients with segmental distribution sublobar HCC with ≥4 IENs, alpha-fetoprotein elevation (≥15.5 ng/mL), or hepatitis C virus have an increased risk of IEN progression to HCC.     

Prediction of Response of Hepatocellular Carcinoma to Radioembolization: Machine Learning Using Preprocedural Clinical Factors and MR Imaging Radiomics

PurposeTo create and evaluate the ability of machine learning–based models with clinicoradiomic features to predict radiologic response after transarterial radioembolization (TARE).Materials and Methods82 treatment-naïve patients (65 responders and 17 nonresponders; median age: 65 years; interquartile range: 11) who underwent selective TARE were included. Treatment responses were evaluated using the European Association for the Study of the Liver criteria at 3-month follow-up. Laboratory, clinical, and procedural information were collected. Radiomic features were extracted from pretreatment contrast-enhanced T1-weighted magnetic resonance images obtained within 3 months before TARE. Feature selection consisted of intraclass correlation, followed by Pearson correlation analysis and finally, sequential feature selection algorithm. Support vector machine, logistic regression, random forest, and LightGBM models were created with both clinicoradiomic features and clinical features alone. Performance metrics were calculated with a nested 5-fold cross-validation technique. The performances of the models were compared by Wilcoxon signed-rank and Friedman tests.ResultsIn total, 1,128 features were extracted. The feature selection process resulted in 12 features (8 radiomic and 4 clinical features) being included in the final analysis. The area under the receiver operating characteristic curve values from the support vector machine, logistic regression, random forest, and LightGBM models were 0.94, 0.94, 0.88, and 0.92 with clinicoradiomic features and 0.82, 0.83, 0.82, and 0.83 with clinical features alone, respectively. All models exhibited significantly higher performances when radiomic features were included (P = .028, .028, .043, and .028, respectively).ConclusionsBased on clinical and imaging-based information before treatment, machine learning–based clinicoradiomic models demonstrated potential to predict response to TARE.     

The direct costs of coronary CT angiography relative to contrast-enhanced thoracic CT: Time-driven activity-based costing

BackgroundCoronary CT angiography (CCTA) and contrast-enhanced thoracic CT (CECT) are distinctly different diagnostic procedures that involve intravenous contrast-enhanced CT of the chest. The technical component of these procedures is reimbursed at the same rate by the Centers for Medicare and Medicaid Services (CMS). This study tests the hypothesis that the direct costs of performing these exams are significantly different.MethodsDirect costs for both procedures were measured using a time-driven activity-based costing (TDABC) model. The exams were segmented into four phases: preparation, scanning, post-scan monitoring, and image processing. Room occupancy and direct labor times were collected for scans of 54 patients (28 CCTA and 26 CECT studies), in seven medical facilities within the USA and used to impute labor and equipment cost. Contrast material costs were measured directly. Cost differences between the exams were analyzed for significance and variability.ResultsMean CCTA duration was 3.2 times longer than CECT (121 and 37 ​min, respectively. p ​< ​0.01). Mean CCTA direct costs were 3.4 times those of CECT ($189.52 and $55.28, respectively, p ​< ​0.01). Both labor and capital equipment costs for CCTA were significantly more expensive (6.5 and 1.8-fold greater, respectively, p ​< ​0.001). Segmented by procedural phase, CCTA was both longer and more expensive for each (p ​< ​0.01). Mean direct costs for CCTA exceeded the standard CMS technical reimbursement of $182.25 without accounting for indirect or overhead costs.ConclusionThe direct cost of performing CCTA is significantly higher than CECT, and thus reimbursement schedules that treat these procedures similarly undervalue the resources required to perform CCTA and possibly decrease access to the procedure.     

Evaluation of an Integrated Spectroscopy and Classification Platform for Point-of-Care Core Needle Biopsy Assessment: Performance Characteristics from Ex Vivo Renal Mass Biopsies

PurposeTo evaluate a transmission optical spectroscopy instrument for rapid ex vivo assessment of core needle cancer biopsies (CNBs) at the point of care.Materials and MethodsCNBs from surgically resected renal tumors and nontumor regions were scanned on their sampling trays with a custom spectroscopy instrument. After extracting principal spectral components, machine learning was used to train logistic regression, support vector machines, and random decision forest (RF) classifiers on 80% of randomized and stratified data. The algorithms were evaluated on the remaining 20% of the data set held out during training. Binary classification (tumor/nontumor) was performed based on a decision threshold. Multinomial classification was also performed to differentiate between the subtypes of renal cell carcinoma (RCC) and account for potential confounding effects from fat, blood, and necrotic tissue. Classifiers were compared based on sensitivity, specificity, and positive predictive value (PPV) relative to a histopathologic standard.ResultsA total of 545 CNBs from 102 patients were analyzed, yielding 5,583 spectra after outlier exclusion. At the individual spectra level, the best performing algorithm was RF with sensitivities of 96% and 92% and specificities of 90% and 89%, for the binary and multiclass analyses, respectively. At the full CNB level, RF algorithm also showed the highest sensitivity and specificity (93% and 91%, respectively). For RCC subtypes, the highest sensitivity and PPV were attained for clear cell (93.5%) and chromophobe (98.2%) subtypes, respectively.ConclusionsEx vivo spectroscopy imaging paired with machine learning can accurately characterize renal mass CNB at the time of tissue acquisition.     

Comparing Radiomics features of tumour and healthy liver tissue in a limited CT dataset: A machine learning study

IntroductionLiver cancer lesions on Computed Tomography (CT) withholds a great amount of data, which is not visible to the radiologists and radiographer. Radiomics features can be extracted from the lesions and used to train Machine Learning (ML) algorithms to predict between tumour and liver tissue. The purpose of this study was to investigate and classify Radiomics features extracted from liver tumours and normal liver tissue in a limited CT dataset.MethodsThe Liver Tumour Segmentation Benchmark (LiTS) dataset consisting of 131 CT scans of the liver with segmentations of tumour tissue and healthy liver was used to extract Radiomic features. Extracted Radiomic features included size, shape, and location extracted with morphological and statistical techniques according to the International Symposium on Biomedical Imaging manual. Relevant features was selected with chi2 correlation and principal component analysis (PCA) with tumour and healthy liver tissue as outcome according to a consensus between three experienced radiologists. Logistic regression, random forest and support vector machine was used to train and validate the dataset with a 10-fold cross-validation method and the Grid Search as hyper-parameter tuning. Performance was evaluated with sensitivity, specificity and accuracy.ResultsThe performance of the ML algorithms achieved sensitivities, specificities and accuracy ranging from 96.30% (95% CI: 81.03%–99.91%) to 100.00% (95% CI: 86.77%–100.00%), 91.30% (95% CI: 71.96%–98.93%) to 100.00% (95% CI: 83.89%–100.00%)and 94.00% (95% CI: 83.45%–98.75%) to 100.00% (95% CI: 92.45%–100.00%), respectively.ConclusionML algorithms classifies Radiomics features extracted from healthy liver and tumour tissue with perfect accuracy. The Radiomics signature allows for a prognostic biomarker for hepatic tumour screening on liver CT.Implications for practiceDifferentiation between tumour and liver tissue with Radiomics ML algorithms have the potential to increase the diagnostic accuracy, assist in the decision-making of supplementary multiphasic enhanced medical imaging, as well as for developing novel prognostic biomarkers for liver cancer patients.     

Influence of coronary stenosis location on diagnostic performance of machine learning-based fractional flow reserve from CT angiography

BackgroundCompared with invasive fractional flow reserve (FFR), coronary CT angiography (cCTA) is limited in detecting hemodynamically relevant lesions. cCTA-based FFR (CT-FFR) is an approach to overcome this insufficiency by use of computational fluid dynamics. Applying recent innovations in computer science, a machine learning (ML) method for CT-FFR derivation was introduced and showed improved diagnostic performance compared to cCTA alone. We sought to investigate the influence of stenosis location in the coronary artery system on the performance of ML-CT-FFR in a large, multicenter cohort.MethodsThree hundred and thirty patients (75.2% male, median age 63 years) with 502 coronary artery stenoses were included in this substudy of the MACHINE (Machine Learning Based CT Angiography Derived FFR: A Multi-Center Registry) registry. Correlation of ML-CT-FFR with the invasive reference standard FFR was assessed and pooled diagnostic performance of ML-CT-FFR and cCTA was determined separately for the following stenosis locations: RCA, LAD, LCX, proximal, middle, and distal vessel segments.ResultsML-CT-FFR correlated well with invasive FFR across the different stenosis locations. Per-lesion analysis revealed improved diagnostic accuracy of ML-CT-FFR compared with conventional cCTA for stenoses in the RCA (71.8% [95% confidence interval, 63.0%–79.5%] vs. 54.8% [45.7%–63.8%]), LAD (79.3 [73.9–84.0] vs. 59.6 [53.5–65.6]), LCX (84.1 [76.0–90.3] vs. 63.7 [54.1–72.6]), proximal (81.5 [74.6–87.1] vs. 63.8 [55.9–71.2]), middle (81.2 [75.7–85.9] vs. 59.4 [53.0–65.6]) and distal stenosis location (67.4 [57.0–76.6] vs. 51.6 [41.1–62.0]).ConclusionIn a multicenter cohort with high disease prevalence, ML-CT-FFR offered improved diagnostic performance over cCTA for detecting hemodynamically relevant stenoses regardless of their location.     

Applying patient characteristics,stent-graft selection,and pre-operative computed tomographic angiography data to a machine learning algorithm: Is endoleak prediction possible?

IntroductionThis study aims to predict endoleak after endovascular aneurysm repair (EVAR) using machine learning (ML) integration of patient characteristics, stent-graft configuration, and a selection of vessel lengths, diameters and angles measured using pre-operative computed tomography angiography (CTA).MethodsWe evaluated 1-year follow-up CT scans (arterial and delayed phases) in patients who underwent EVAR for the presence or absence of an endoleak. We also obtained data on the patient characteristics, stent-graft selection, and preoperative CT vessel morphology (diameter, length, and angle). The extreme gradient boosting (XGBoost) for the ML system was trained on 30 patients with endoleaks and 81 patients without. We evaluated 5217 items in 111 patients with abdominal aortic aneurysms, including the patient characteristics, stent-graft configuration and vascular morphology acquired using pre-EVAR abdominal CTA. We calculated the area under the curve (AUC) of our receiver operating characteristic analysis using the ML method.ResultsThe AUC, accuracy, 95% confidence interval (CI), sensitivity, and specificity were 0.88, 0.88, 0.79–0.97, 0.85, and 0.91 for ML applying XGBoost, respectively.ConclusionsThe diagnostic performance of the ML method was useful when factors such as the patient characteristics, stent-graft configuration and vessel length, diameter and angle of the vessels were considered from pre-EVAR CTA.Implications for practiceBased on our findings, we suggest that this is a potential application of ML for the interpretation of abdominal CTA scans in patients with abdominal aortic aneurysms scheduled for EVAR.     

Safety and Efficacy of Radiofrequency Ablation for Nonfunctioning Benign Thyroid Nodules in Children and Adolescents in 14 Patients over a 10-Year Period

PurposeTo evaluate the efficacy and safety of radiofrequency (RF) ablation for nonfunctioning benign thyroid nodules in children and adolescents.Materials and MethodsFourteen pediatric patients (10 female, 4 male; mean age 15.7 ± 2.3 years, range 12–19 years) with nonfunctioning benign thyroid nodules (mean longest diameter 3.7 ± 1.1 cm, range 2.0–5.6 cm) treated with the use of RF ablation from 2005 to 2015 were evaluated. The inclusion criteria for RF ablation therapy were (i) age <20 years, (ii) benign cytological confirmation by ≥2 separate fine-needle aspiration or core needle biopsies, (iii) pressure symptoms or cosmetic problems caused by thyroid nodules, (iv) absence of any suspicious feature as determined with the use of ultrasound (US), (v) normal serum levels of thyroid hormone and thyrotropin, and (vi) follow-up of >6 months. RF ablation was performed with the use of an RF generator and an 18-gauge internally cooled electrode. RF ablation was performed under local anesthesia without conscious sedation or general anesthesia. Changes in nodules on follow-up US, changes in symptomatic and cosmetic scores, and complications arising during or after RF ablation were evaluated.ResultsMean follow-up period was 36.9 ± 21.7 months (range 6–69 months). At last follow-up visits, mean longest nodule diameter and volume had decreased significantly (3.7 ± 1.1 cm vs 1.4 ± 0.9 cm and 14.6 ± 13.3 mL vs 1.7 ± 4.4 mL; P < 0.001). Both cosmetic and compressive symptoms significantly improved (3.8 ± 0.6 vs 1.4 ± 0.6 and 3.4 ± 1.0 vs 0.1 ± 0.4; P < 0.001). The mean number of ablation sessions was 2.1 ± 1.2 (range 1–5 sessions) and no major complication was encountered during or after RF ablation.ConclusionsRF ablation might be a safe and effective treatment modality for nonfunctioning benign thyroid nodules in children and adolescents.     

Improved Utilization Following Conversion of a Fluoroscopy Suite to Hybrid CT/Angiography System

PurposeTo assess changes in operational utilization following conversion of a single IR suite to a hybrid CT/angiography (Angio-CT) system at an academic tertiary care center.Materials and MethodsThe total number of interventional procedures and diagnostic CT examinations performed in 29 rooms (20 diagnostic radiology, 7 IR, 2 shared between divisions) was calculated in the 24 months before conversion of an IR suite to Angio-CT and 12 months after conversion. The total number of IR procedures (global IR/month) and diagnostic CT scans per month (global CT/month) in both before and after conversion periods was calculated and defined as baseline institutional growth. This was compared against the change in the number of IR procedures performed in the before and after periods in the converted room (Angio-CT/month) as well as the number of diagnostic CT scans performed in the shared rooms (shared CT/month).ResultsThe percent change in global CT and global IR from the before to the after periods was 39.2% and 3.1%, respectively. Shared CT per month and Angio-CT per month increased by 46.7% and 12.0% across the same time periods, respectively. The ratio of the percent increase in Angio-CT per month to percent increase in global IR per month was 3.87. The ratio of the percent increase in shared CT per month to percent increase in global CT per month was 1.19.ConclusionsOperational utilization improved in both diagnostic radiology and IR sections following conversion of a conventional fluoroscopic IR suite to an Angio-CT room.     

Ablation of Cervical Lymph Nodes in Patients with Thyroid Cancer: A Comparison between Cryoablation and Percutaneous Ethanol Injection

The purpose of this study was to retrospectively evaluate the safety and efficacy of cryoablation and compare the outcomes with those of percutaneous ethanol injection (PEI) for the treatment of metastatic cervical lymph nodes (CLNs) in patients with thyroid cancer. The study included 24 patients with 47 CLNs treated with PEI and 7 patients with 11 CLNs treated with cryoablation. Three of 7 (42.9%) patients did not respond to PEI and progressed to cryoablation. There were more local recurrences in CLNs treated with PEI (7/47, 14.9%) compared with cryoablation (0/11, 0%), but this did not reach significance (P = .33). There was no difference in mild/moderate (3/24, 12.5% vs 2/7, 28.6%; P = .31) or severe (1/24, 4.2% vs 0/7, 0%; P = 1) adverse events in the PEI and cryoablation cohorts. The number of treatments required for CLNs treated with PEI (2 ± 1.1) was significantly greater than those for CLNs treated with cryoablation (1 ± 0) (P = .002). These limited data suggest that the treatment of metastatic CLNs with cryoablation or PEI may both be safe and effective; however, further data are needed to confirm superiority of cryoablation.     

Robotically-Assisted Sonic Therapy for Renal Ablation in a Live Porcine Model: Initial Preclinical Results

PurposeTo demonstrate the feasibility of Robotically Assisted Sonic Therapy (RAST)—a noninvasive and nonthermal focused ultrasound therapy based on histotripsy—for renal ablation in a live porcine model.Materials and MethodsRAST ablations (n = 11) were performed in 7 female swine: 3 evaluated at 1 week (acute) and 4 evaluated at 4 weeks (chronic). Treatment groups were acute bilateral (3 swine, 6 ablations with immediate computed tomography [CT] and sacrifice); chronic single kidney (3 swine, 3 ablations; CT at day 0, week 1, and week 4 after treatment, followed by sacrifice); and chronic bilateral (1 swine, 2 ablations). Treatments were performed using a prototype system (VortxRx; HistoSonics, Inc) and targeted a 2.5-cm-diameter sphere in the lower pole of each kidney, intentionally including the central collecting system.ResultsMean treatment time was 26.4 minutes. Ablations had a mean diameter of 2.4 ± 0.3 cm, volume of 8.5 ± 2.4 cm³_, and sphericity index of 1.00. Median ablation volume decreased by 96.1% over 4 weeks. Histology demonstrated complete lysis with residual blood products inside the ablation zone. Temporary collecting system obstruction by thrombus was observed in 4/11 kidneys (2 acute and 2 chronic) and resolved by 1 week. There were no urinary leaks, main vessel thromboses, or adjacent organ injuries on imaging or necropsy.ConclusionsIn this normal porcine model, renal RAST demonstrated complete histologic destruction of the target renal tissue while sparing the urothelium.     

Machine Learning Offers Exciting Potential for Predicting Postprocedural Outcomes: A Framework for Developing Random Forest Models in IR

PurposeTo demonstrate that random forest models trained on a large national sample can accurately predict relevant outcomes and may ultimately contribute to future clinical decision support tools in IR.Materials and MethodsPatient data from years 2012–2014 of the National Inpatient Sample were used to develop random forest machine learning models to predict iatrogenic pneumothorax after computed tomography–guided transthoracic biopsy (TTB), in-hospital mortality after transjugular intrahepatic portosystemic shunt (TIPS), and length of stay > 3 days after uterine artery embolization (UAE). Model performance was evaluated with area under the receiver operating characteristic curve (AUROC) and maximum F1 score. The threshold for AUROC significance was set at 0.75.ResultsAUROC was 0.913 for the TTB model, 0.788 for the TIPS model, and 0.879 for the UAE model. Maximum F1 score was 0.532 for the TTB model, 0.357 for the TIPS model, and 0.700 for the UAE model. The TTB model had the highest AUROC, while the UAE model had the highest F1 score. All models met the criteria for AUROC significance.ConclusionsThis study demonstrates that machine learning models may suitably predict a variety of different clinically relevant outcomes, including procedure-specific complications, mortality, and length of stay. Performance of these models will improve as more high-quality IR data become available.     

Percutaneous CT-Guided Abdominal and Pelvic Biopsies: Comparison of an Electromagnetic Navigation System and CT Fluoroscopy

PurposeTo compare electromagnetic navigation (EMN) with computed tomography (CT) fluoroscopy for guiding percutaneous biopsies in the abdomen and pelvis.Materials and MethodsA retrospective matched-cohort design was used to compare biopsies in the abdomen and pelvis performed with EMN (consecutive cases, n = 50; CT-Navigation; Imactis, Saint-Martin-d’Hères, France) with those performed with CT fluoroscopy (n = 100). Cases were matched 1:2 (EMN:CT fluoroscopy) for target organ and lesion size (±10 mm).ResultsThe population was well-matched (age, 65 vs 65 years; target size, 2.0 vs 2.1 cm; skin-to-target distance, 11.4 vs 10.7 cm; P > .05, EMN vs CT fluoroscopy, respectively). Technical success (98% vs 100%), diagnostic yield (98% vs 95%), adverse events (2% vs 5%), and procedure time (33 minutes vs 31 minutes) were not statistically different (P > .05). Operator radiation dose was less with EMN than with CT fluoroscopy (0.04 vs 1.2 μGy; P < .001), but patient dose was greater (30.1 vs 9.6 mSv; P < .001) owing to more helical scans during EMN guidance (3.9 vs 2.1; P < .001). CT fluoroscopy was performed with a mean of 29.7 tap scans per case. In 3 (3%) cases, CT fluoroscopy was performed with gantry tilt, and the mean angle out of plane for EMN cases was 13.4°.ConclusionsPercutaneous biopsies guided by EMN and CT fluoroscopy were closely matched for technical success, diagnostic yield, procedure time, and adverse events in a matched cohort of patients. EMN cases were more likely to be performed outside of the gantry plane. Radiation dose to the operator was higher with CT fluoroscopy, and patient radiation dose was higher with EMN. Further study with a wider array of procedures and anatomic locations is warranted.     

Image-Guided Percutaneous Needle Biopsy for Benign and Malignant Bone Tumors: Systematic Review and Meta-Analysis

PurposeTo compare the diagnostic yield and accuracy of both image-guided core-needle biopsy (CNB) and fine-needle biopsy and evaluate the benefit of performing fine-needle biopsy in addition to CNB in patients with suspected benign and malignant bone tumors.Materials and MethodsA systematic search was performed on March 10, 2021, to determine whether fine-needle aspiration (FNA) plays any role when performed alone or in combination with CNB. The included studies were aggregated for the pooled estimates of diagnostic yield and histologic accuracy of image-guided percutaneous needle biopsy of bone tumors. Twenty-nine studies published between 1996 and 2021 were included.ResultsWhen all patients with bone tumors were included, the rates of diagnostic yield and accuracy of FNA and CNB were 88.5% and 82.5% and 91.4% and 92.7%, respectively; the rates of both the methods combined were 96.5% and 94.1%, respectively; and for the lytic subgroup, the rates of diagnostic yield and accuracy of CNB and both the methods combined were 94.3% and 100% and 98.9% and 90.4%, respectively. A P value of <.05 was considered statistically significant.ConclusionThe present meta-analysis showed that core biopsy alone outperformed fine-needle biopsy alone in all categories of benign and malignant tumors. Additionally, the diagnostic yield was improved when FNA was used in addition to CNB for lytic bone lesions.     

Computed tomography reference dimensions for identification of stented surgical mitral bioprostheses valve size

BackgroundSelection of the transcatheter heart valve size for a mitral valve-in-valve procedure is based on the type and manufacturer’s labelled size. However, accurate information of surgical heart valve (SHV) size may not be available in the patient’s medical record. The purpose of this study is to establish reference data for computed tomography (CT) dimensions of commonly used mitral SHV in order to determine the manufacturer’s labelled size from a cardiac CT data set.MethodsCT datasets of 105 patients with surgical mitral bioprosthesis and available manufacturer labeled datasets were included in the analysis. CT derived valve dimensions were assessed by two observers using multiplanar reformats aligned with the basal sewing ring. A circular region of interest was used in a standardized fashion to minimize influence of image acquisition and reconstruction parameters. Interobserver variability was assessed by Bland-Altman analysis.ResultsThe CT-derived dimensions were stratified by valve size and type, and SHV properties were demonstrated for 5 common valve types. Variability of measurements was small and inter-observer limits of agreement were narrow. Stratified by SHV type, no overlap was noted for CT-derived dimensions among different SHV sizes . A reference table of CT characteristics of surgical mitral bioprosthesis types was created.ConclusionThe study provides reference CT data for determining the manufacturers’ labeled SHV size across a range of commonly used mitral SHVs. The findings will be important to help identify types of surgical mitral bioprosthesis utilizing CT characteristics for patients without SHV size documentation.