Deep learning kidney segmentation with very limited training data using a cascaded convolution neural network.

BACKGROUND: Deep learning segmentation requires large datasets with ground truth. Image annotation is time consuming and leads to shortages of ground truth data for clinical imaging. This study is to investigate the feasibility of kidney segmentation using deep learning convolution neural network (CNN) models trained with MR images from only a few subjects. METHODS: A total of 60 subjects from two cohorts were included in this study. The first cohort of 20 subjects from publicly available data was used for training and testing. The second cohort of 40 subjects with renal masses from our institution was used for testing only. A few-shot deep learning approach using 3D augmentation was investigated. T1-weighted images in the first cohort were used for training and testing. Cascaded CNN networks were trained using images from one, three, and six subjects, respectively. Images for the remaining subjects were used for testing. Images in the second cohort were utilized for testing only. Dice and Jaccard coefficients were generated to evaluate the performance of CNN models. Statistical analyses for segmentation metrics among different approaches were performed. RESULTS: Our approach achieved mean Dice coefficients of 0.85 using a single training subject and 0.91 with six training subjects. Compared to a single Unet, the cascaded network significantly improved the results using a single training subject (Dice, 0.759 vs. 0.835; p<0.001) and three subjects (0.864 vs. 0.893; p = 0.015) in the first cohort, and the results for the second cohort (0.821 vs. 0.873; p = 0.008). CONCLUSION: Our few-shot kidney segmentation approach using 3D augmentation achieved a good performance even using a single Unet. Furthermore, the cascaded network significantly improved the performance of segmentation and was superior to a single Unet in certain cases. Our approach provides a promising solution to segmentation in medical imaging when the number of ground truth masks is limited.

Long-Term Percutaneous Nephrostomy Management of Malignant Urinary Obstruction: Estimation of Optimal Exchange Frequency and Estimation of the Financial Impact of Patient Compliance.

PURPOSE: To estimate the least costly routine exchange frequency for percutaneous nephrostomies (PCNs) placed for malignant urinary obstruction, as measured by annual hospital charges, and to estimate the financial impact of patient compliance. MATERIALS AND METHODS: Patients with PCNs placed for malignant urinary obstruction were studied from 2011 to 2013. Exchanges were classified as routine or due to 1 of 3 complication types: mechanical (tube dislodgment), obstruction, or infection. Representative cases were identified, and median representative charges were used as inputs for the model. Accelerated failure time and Markov chain Monte Carlo models were used to estimate distribution of exchange types and annual hospital charges under different routine exchange frequency and compliance scenarios. RESULTS: Long-term PCN management was required in 57 patients, with 87 total exchange encounters. Median representative hospital charges for pyelonephritis and obstruction were 11.8 and 9.3 times greater, respectively, than a routine exchange. The projected proportion of routine exchanges increased and the projected proportion of infection-related exchanges decreased when moving from a 90-day exchange with 50% compliance to a 60-day exchange with 75% compliance, and this was associated with a projected reduction in annual charges. Projected cost reductions resulting from increased compliance were generally greater than reductions resulting from changes in exchange frequency. CONCLUSIONS: This simulation model suggests that the optimal routine exchange interval for PCN exchange in patients with malignant urinary obstruction is approximately 60 days and that the degree of reduction in charges likely depends more on patient compliance than exact exchange interval.