Radiomics signature for automatic hydronephrosis detection in unenhanced Low-Dose CT.

PURPOSE: To investigate the diagnostic performance of an automatic pipeline for detection of hydronephrosis on kidney's parenchyma on unenhanced low-dose CT of the abdomen. METHODS: This retrospective study included 95 patients with confirmed unilateral hydronephrosis in an unenhanced low-dose CT of the abdomen. Data were split into training (n = 67) and test (n = 28) cohorts. Both kidneys for each case were included in further analyses, whereas the kidney without hydronephrosis was used as control. Using the training cohort, we developed a pipeline consisting of a deep-learning model for automatic segmentation (a Convolutional Neural Network based on nnU-Net architecture) of the kidney's parenchyma and a radiomics classifier to detect hydronephrosis. The models were assessed using standard classification metrics, such as area under the ROC curve (AUC), sensitivity and specificity, as well as semantic segmentation metrics, including Dice coefficient and Jaccard index. RESULTS: Using manual segmentation of the kidney's parenchyma, hydronephrosis can be detected with an AUC of 0.84, a sensitivity of 75% and a specificity of 82%, a PPV of 81% and a NPV of 77%. Automatic kidney segmentation achieved a mean Dice score of 0.87 and 0.91 for the right and left kidney, respectively. Additionally, automatic segmentation achieved an AUC of 0.83, a sensitivity of 86%, specificity of 64%, PPV of 71%, and NPV of 82%. CONCLUSION: Our proposed radiomics signature using automatic kidney's parenchyma segmentation allows for accurate hydronephrosis detection on unenhanced low-dose CT scans of the abdomen independently of widened renal pelvis. This method could be used in clinical routine to highlight hydronephrosis to radiologists as well as clinicians, especially in patients with concurrent parapelvic cysts and might reduce time and costs associated with diagnosing hydronephrosis.

Non-operative management of blunt splenic trauma: The role of splenic artery embolization depending on the severity of parenchymal injury.

PURPOSE: To address the disagreement about the need for splenic artery embolization (SAE) in medium grade blunt splenic trauma this retrospective study evaluates the clinical outcome of non-operative management (NOM) and the possible impact of a more liberal indication for primary SAE. METHOD: From 01/2010 to 12/2019 186 patients presented with splenic injury on computed tomography (CT) after blunt abdominal trauma. The extent of splenic injuries according to Marmery, vascular pathologies, active bleeding as well as clinical and laboratory parameters were recorded and analyzed with regard to the success rates of NOM and SAE. Procedural complications and clinical outcome were noted. The number needed to treat (NNT) was determined for a possible extension of the indication for SAE to grade 3 injuries. RESULTS: Of 186 patients 126 were managed non-operatively, 47 underwent primary SAE and twelve splenectomy. NOM was successful in 119/126 (94 %) patients. Conversion rate was significantly higher in patients with active bleeding or vascular pathology. Patients with failed NOM had a significantly greater decrease in haemoglobin and haematocrit levels. Primary SAE was successful in 45/47 (96 %) cases. Major complications occurred in four cases (9%), all managed without sequela. The NNT in grade 3 splenic injuries equals 13. CONCLUSIONS: NOM of low to medium-grade blunt splenic trauma has a low failure rate. Presence of active haemorrhage is the most important predictor for failure of NOM. SAE should be reserved for high-grade injuries and visible vascular pathology or active bleeding to avoid a disproportionate increase in the NNT.

Differentiation of heroin and cocaine using dual-energy CT-an experimental study.

OBJECTIVE: To evaluate if heroin and cocaine can be distinguished using dual-energy CT. MATERIALS AND METHODS: Twenty samples of heroin and cocaine at different concentrations and standardized compression (SC) were scanned in dual-energy mode on a newest generation Dual Energy 64-row MDCT scanner. CT number, spectral graphs, and dual-energy index (DEI) were evaluated. Results were prospectively tested on six original samples from a body packer. Wilcoxon's test was used for statistical evaluation. RESULTS: Values are given as median and range. Under SC, the CT number of cocaine samples (-29.87 Hounsfield unit (HU) [-125.85; 16.16 HU]) was higher than the CT number of heroin samples (-184.37 HU [-199.81; -159.25 HU]; p < 0.01). Slope of spectral curves for cocaine was -2.36 HU/keV [-7.15; -0.67 HU/keV], and for heroin, 1.75 HU/keV [1.28; 2.5 HU/keV] (p < 0.01). DEI was 0.0352 [0.0081; 0.0528] for cocaine and significantly higher than for heroin samples (-0.0127 [-0.0097; -0.0159]; p < 0.001). While CT number was inconclusive, all six original packs were correctly classified after evaluation of the spectral curve and DEI. In contrast to the CT number, slope of the spectral curve and DEI were independent of concentration and compression. CONCLUSION: The slope of the spectral curve and the DEI from dual-energy CT data can be used to distinguish heroin and cocaine in vitro; these results are independent of compression and concentration in the measured range.