Early detection of the autosomal dominant polycystic kidney disease (ADPKD) is crucial as it is one of the most common causes of end-stage renal disease (ESRD) and kidney failure. The total kidney volume (TKV) can be used as a biomarker to quantify disease progression. The TKV calculation requires accurate delineation of kidney volumes, which is usually performed manually by an expert physician. However, this is time-consuming and automated segmentation is warranted. Furthermore, the scarcity of large annotated datasets hinders the development of deep learning solutions. In this work, we address this problem by implementing three attention mechanisms into the U-Net to improve TKV estimation. Additionally, we implement a cosine loss function that works well on image classification tasks with small datasets. Lastly, we apply a technique called sharpness aware minimization (SAM) that helps improve the generalizability of networks. Our results show significant improvements (p-value < 0.05) over the reference kidney segmentation U-Net. We show that the attention mechanisms and/or the cosine loss with SAM can achieve a dice score (DSC) of 0.918, a mean symmetric surface distance (MSSD) of 1.20 mm with the mean TKV difference of −1.72%, and R${}^2$ of 0.96 while using only 100 MRI datasets for training and testing. Furthermore, we tested four ensembles and obtained improvements over the best individual network, achieving a DSC and MSSD of 0.922 and 1.09 mm, respectively.

Objectives: The goal of this study is to demonstrate the performance of radiomics and CNN-based classifiers in determining the primary origin of gastrointestinal liver metastases for visually indistinguishable lesions. Methods: In this retrospective, IRB-approved study, 31 pancreatic cancer patients with 861 lesions (median age [IQR]: 65.39 [56.87, 75.08], 48.4% male) and 47 colorectal cancer patients with 435 lesions (median age [IQR]: 65.79 [56.99, 74.62], 63.8% male) were enrolled. A pretrained nnU-Net performed automated segmentation of 1296 liver lesions. Radiomics features for each lesion were extracted using pyradiomics. The performance of several radiomics-based machine-learning classifiers was investigated for the lesions and compared to an image-based deep-learning approach using a DenseNet-121. The performance was evaluated by AUC/ROC analysis. Results: The radiomics-based K-nearest neighbor classifier showed the best performance on an independent test set with AUC values of 0.87 and an accuracy of 0.67. In comparison, the image-based DenseNet-121-classifier reached an AUC of 0.80 and an accuracy of 0.83. Conclusions: CT-based radiomics and deep learning can distinguish the etiology of liver metastases from gastrointestinal primary tumors. Compared to deep learning, radiomics based models showed a varying generalizability in distinguishing liver metastases from colorectal cancer and pancreatic adenocarcinoma.

(1) Background: Tumoral heterogeneity (TH) is a major challenge in the treatment of metastatic colorectal cancer (mCRC) and is associated with inferior response. Therefore, the identification of TH would be beneficial for treatment planning. TH can be assessed by identifying genetic alterations. In this work, a radiomics-based approach for assessment of TH in colorectal liver metastases (CRLM) in CT scans is demonstrated. (2) Methods: In this retrospective study, CRLM of mCRC were segmented and radiomics features extracted using pyradiomics. Unsupervised k-means clustering was applied to features and lesions. Feature redundancy was evaluated by principal component analysis and reduced by Pearson correlation coefficient cutoff. Feature selection was conducted by LASSO regression and visual analysis of the clusters by radiologists. (3) Results: A total of 47 patients’ (36% female, median age 64) CTs with 261 lesions were included. Five clusters were identified, and the categories small disseminated (n = 31), heterogeneous (n = 105), homogeneous (n = 64), mixed (n = 59), and very large type (n = 2) were assigned based on visual characteristics. Further statistical analysis showed correlation (p < 0.01) of clusters with sex, primary location, T- and N-status, and mutational status. Feature reduction and selection resulted in the identification of four features as a final set for cluster definition. (4) Conclusions: Radiomics features can characterize TH in liver metastases of mCRC in CT scans, and may be suitable for a better pretherapeutic classification of liver lesion phenotypes.

Background: Digital breast tomosynthesis (DBT) and abbreviated breast MRI (AB-MRI) offer superior diagnostic performance compared to conventional mammography in screening women with intermediate risk of breast cancer due to dense breast tissue. The aim of this model-based economic evaluation was to analyze whether AB-MRI is cost-effective in this cohort compared to DBT. Methods: Decision analysis and Markov simulations were used to model the cumulative costs and quality-adjusted life-years (QALYs) over a time horizon of 30 years. Model input parameters were adopted from recent literature. Deterministic and probabilistic sensitivity analyses were applied to test the stability of the model. Results: In the base-case scenario, the costs of an AB-MRI examination were defined to equal the costs of a full protocol acquisition. Two-yearly screening of women with dense breasts resulted in cumulative discounted costs of $8798 and $9505 for DBT and AB-MRI, and cumulative discounted effects of 19.23 and 19.27 QALYs, respectively, with an incremental cost-effectiveness ratio of $20,807 per QALY gained in the base-case scenario. By reducing the cost of an AB-MRI examination below a threshold of $241 in sensitivity analyses, AB-MRI would become cost-saving compared to DBT. Conclusion: In comparison to DBT, AB-MRI can be considered cost-effective up to a price per examination of $593 in screening patients at intermediate risk of breast cancer.

Background: Colorectal cancer is among the most prevalent cancer entities worldwide, with every second patient developing liver metastases during their illness. For local treatment of liver metastases, a surgical approach as well as ablative treatment options, such as microwave ablation (MWA) and radiofrequency ablation (RFA), are available. The aim of this study is to evaluate the cost-effectiveness of RFA, MWA and surgery in the treatment of liver metastases of oligometastatic colorectal cancer (omCRC) that are amenable for all investigated treatment modalities. Methods: A decision analysis based on a Markov model assessed lifetime costs and quality-adjusted life years (QALY) related to the treatment strategies RFA, MWA and surgical resection. Input parameters were based on the best available and most recent evidence. Probabilistic sensitivity analyses (PSA) were performed with Monte Carlo simulations to evaluate model robustness. The percentage of cost-effective iterations was determined for different willingness-to-pay (WTP) thresholds. Results: The base-case analysis showed that surgery led to higher long-term costs compared to RFA and MWA (USD 41,848 vs. USD 36,937 vs. USD 35,234), while providing better long-term outcomes than RFA, yet slightly lower than MWA (6.80 vs. 6.30 vs. 6.95 QALYs for surgery, RFA and MWA, respectively). In PSA, MWA was the most cost-effective strategy for all WTP thresholds below USD 80,000 per QALY. Conclusions: In omCRC patients with liver metastases, MWA and surgery are estimated to provide comparable efficacy. MWA was identified as the most cost-effective strategy in intermediate resource settings and should be considered as an alternative to surgery in high resource settings.

Background: To compare the diagnostic characteristics between arterial phase imaging versus portal venous phase imaging, applying polychromatic T3D images and low keV virtual monochromatic images using a 1st generation photon-counting CT detector, of CT in patients with hepatocellular carcinoma (HCC). Methods: Consecutive patients with HCC, with a clinical indication for CT imaging, were prospectively enrolled. Virtual monoenergetic images (VMI) were reconstructed at 40 to 70 keV for the PCD-CT. Two independent, blinded radiologists counted all hepatic lesions and quantified their size. The lesion-to-background ratio was quantified for both phases. SNR and CNR were determined for T3D and low VMI images; non-parametric statistics were used. Results: Among 49 oncologic patients (mean age 66.9 ± 11.2 years, eight females), HCC was detected in both arterial and portal venous scans. The signal-to-noise ratio, the CNR liver-to-muscle, the CNR tumor-to-liver, and CNR tumor-to-muscle were 6.58 ± 2.86, 1.40 ± 0.42, 1.13 ± 0.49, and 1.53 ± 0.76 in the arterial phase and 5.93 ± 2.97, 1.73 ± 0.38, 0.79 ± 0.30, and 1.36 ± 0.60 in the portal venous phase with PCD-CT, respectively. There was no significant difference in SNR between the arterial and portal venous phases, including between “T3D” and low keV images (p > 0.05). CNR_{tumor-to-liver} differed significantly between arterial and portal venous contrast phases (p < 0.005) for both “T3D” and all reconstructed keV levels. CNR_{liver-to-muscle} and CNR_{tumor-to-muscle} did not differ in either the arterial or portal venous contrast phases. CNR_{tumor-to-liver} increased in the arterial contrast phase with lower keV in addition to SD. In the portal venous contrast phase, CNR_{tumor-to-liver} decreased with lower keV; whereas, CNR_{tumor-to-muscle} increased with lower keV in both arterial and portal venous contrast phases. CTDI and DLP mean values for the arterial upper abdomen phase were 9.03 ± 3.59 and 275 ± 133, respectively. CTDI and DLP mean values for the abdominal portal venous phase were 8.75 ± 2.99 and 448 ± 157 with PCD-CT, respectively. No statistically significant differences were found concerning the inter-reader agreement for any of the (calculated) keV levels in either the arterial or portal-venous contrast phases. Conclusions: The arterial contrast phase imaging provides higher lesion-to-background ratios of HCC lesions using a PCD-CT; especially, at 40 keV. However, the difference was not subjectively perceived as significant.