In vivo assessment of tumor heterogeneity in WHO 2016 glioma grades using diffusion kurtosis imaging: Diagnostic performance and improvement of feasibility in routine clinical practice.

PURPOSE: To assess the diagnostic performance of normalized and non-normalized diffusion kurtosis imaging (DKI) metrics extracted from different tumor volume data for grading glioma according to the integrated approach of the revised 2016 WHO classification. MATERIALS AND METHODS: Sixty patients with histopathologically confirmed glioma, who provided written informed consent, were retrospectively assessed between 01/2013 and 08/2016 from a prospective trial approved by the local institutional review board. Mean kurtosis (MK) and mean diffusivity (MD) metrics from DKI were assessed by two blinded physicians from four different volumes of interest (VOI): whole solid tumor including (VOItu-ed) and excluding perifocal edema (VOItu), infiltrative zone (VOIed), and single slice of solid tumor core (VOIslice). Intra-class correlation coefficient (ICC) was calculated to assess inter-rater agreement. One-way ANOVA was used to compare MK between 2016 CNS WHO tumor grades. Friedman's test compared MK and MD of each VOI. Spearman's correlation coefficient was used to correlate MK with 2016 CNS WHO tumor grades. ROC analysis was performed on MK for significant results. RESULTS: The MK assessment showed excellent inter-rater agreement for each VOI (ICC, 0.906-0.955). MK was significantly lower in IDHmutant astrocytoma (0.40±0.07), than in 1p/19q-confirmed oligodendroglioma (0.54±0.10, P=0.001) or IDHwild-type glioblastoma (0.68±0.13, P<0.001). MK and 2016 WHO tumor grades were strongly and positively correlated (VOItu-ed, r=0.684; VOItu, r=0.734; VOIed, r=0.625; VOIslice, r=0.698; P<0.001). CONCLUSIONS: Non-normalized MK values obtained from VOItu and VOIslice showed the best reproducibility and highest diagnostic performance for stratifying glioma according to the integrated approach of the recent 2016 WHO classification.

Repeated surgeries in invasive lobular breast cancer with preoperative MRI: Role of additional carcinoma in situ and background parenchymal enhancement.

OBJECTIVES: Analysing the influence of additional carcinoma in situ (CIS) and background parenchymal enhancement (BPE) in preoperative MRI on repeated surgeries in patients with invasive lobular carcinoma (ILC) of the breast. METHODS: Retrospective analysis of 106 patients (mean age 58.6±9.9years) with 108 ILC. Preoperative tumour size as assessed by MRI, mammography and sonography was recorded and compared to histopathology. In contrast-enhanced MRI, the degree of BPE was categorised by two readers. The influence of additionally detected CIS and BPE on the rate of repeated surgeries was analysed. RESULTS: Additional CIS was present in 45.4% of the cases (49/108). The degree of BPE was minimal or mild in 80% of the cases and moderate or marked in 20% of the cases. In 17 cases (15.7%) at least one repeated surgery was performed. In n=15 of these cases, repeated surgery was performed after BCT (n=9 re-excisions, n=6 conversions to mastectomy), in n=2 cases after initial mastectomy. The initial surgical procedure (p=0.008) and additional CIS (p=0.046) significantly influenced the rate of repeated surgeries, while tumour size, patient age and BPE did not (p=ns). CONCLUSIONS: Additional CIS was associated with a higher rate of repeated surgeries, whereas BPE had no influence.

Low-Dose Volume-Perfusion CT of the Brain: Effects of Radiation Dose Reduction on Performance of Perfusion CT Algorithms.

PURPOSE: We aimed to compare different computed tomography (CT) perfusion post-processing algorithms regarding image quality of perfusion maps from low-dose volume perfusion CT (VPCT) and their diagnostic performance regarding the detection of ischemic brain lesions. METHODS AND MATERIALS: We included VPCT data of 21 patients with acute stroke (onset < 6h), which were acquired at 80 kV and 180 mAs. Low-dose VPCT datasets with 72 mAs (40 % of original dose) were generated using realistic low-dose simulation. Perfusion maps (cerebral blood volume (CBV); cerebral blood flow (CBF) from original and low-dose datasets were generated using two different commercially available post-processing methods: deconvolution-based method (DC) and maximum slope algorithm (MS). The resulting DC and MS perfusion maps were compared regarding perfusion values, signal-to-noise ratio (SNR) as well as image quality and diagnostic accuracy as rated by two blinded neuroradiologists. RESULTS: Quantitative perfusion parameters highly correlated for both algorithms and both dose levels (r ≥ 0.613, p < 0.001). Regarding SNR levels and image quality of the CBV maps, no significant differences between DC and MS were found (p ≥ 0.683). Low-dose MS CBF maps yielded significantly higher SNR levels (p < 0.001) and quality scores (p = 0.014) than those of DC. Low-dose CBF and CBV maps from both DC and MS yielded high sensitivity and specificity for the detection of ischemic lesions (sensitivity ≥ 0.82, specificity ≥ 0.90). CONCLUSION: Our results indicate that both methods produce diagnostically sufficient perfusion maps from simulated low-dose VPCT. However, MS produced CBF maps with significantly higher image quality and SNR than DC, indicating that MS might be more suitable for low-dose VPCT imaging.

Frequency Selective Non-Linear Blending to Improve Image Quality in Liver CT.

Purpose: To evaluate the effects of a new frequency selective non-linear blending (NLB) algorithm on the contrast resolution of liver CT with low intravascular concentration of iodine contrast. Materials and Methods: Our local ethics committee approved this retrospective study. The informed consent requirement was waived. CT exams of 25 patients (60 % female, mean age: 65 ±â€Š16 years of age) with late phase CT scans of the liver were included as a model for poor intrahepatic vascular contrast enhancement. Optimal post-processing settings to enhance the contrast of hepatic vessels were determined. Outcome variables included signal-to-noise (SNR) and contrast-to-noise ratios (CNR) of hepatic vessels and SNR of liver parenchyma of standard and post-processed images. Image quality was quantified by two independent readers using Likert scales. Results: The post-processing settings for the visualization of hepatic vasculature were optimal at a center of 115HU, delta of 25HU, and slope of 5. Image noise was statistically indifferent between standard and post-processed images. The CNR between the hepatic vasculature (HV) and liver parenchyma could be significantly increased for liver veins (CNRStandard 1.62 ±â€Š1.10, CNRNLB 3.6 ±â€Š2.94, p = 0.0002) and portal veins (CNRStandard 1.31 ±â€Š0.85, CNRNLB 2.42 ±â€Š3.03, p = 0.046). The SNR of liver parenchyma was significantly higher on post-processed images (SNRNLB 11.26 ±â€Š3.16, SNRStandard 8.85 ± 2.27, p = 0.008). The overall image quality and depiction of HV were significantly higher on post-processed images (NLBDHV: 4 [3 - 4.75], StandardDHV: 2 [1.3 - 2.5], p = < 0.0001; NLBIQ: 4 [4 - 4], StandardIQ: 2 [2 - 3], p = < 0.0001). Conclusion: The use of a frequency selective non-linear blending algorithm increases the contrast resolution of liver CT and can improve the visibility of the hepatic vasculature in the setting of a low contrast ratio between vessels and the parenchyma. Key Points: • Using the new frequency selective non-linear blending algorithm is feasible in contrast-enhanced liver CT.• Optimal post-processing settings make it possible to significantly increase the contrast resolution of liver CT without affecting image noise.• Especially in low contrast CT images, the novel algorithm is capable of significantly increasing image quality. Citation Format: • Bongers MN, Bier G, Kloth C et al. Frequency Selective Non-Linear Blending to Improve Image Quality in Liver CT. Fortschr Röntgenstr 2016; 188: 1163 - 1168.