Reduced-Dose Deep Learning Reconstruction for Abdominal CT of Liver Metastases

Background Patients with acute non-traumatic abdominal pain often undergo abdominal computed tomography (CT). However, abdominal CT is associated with high radiation exposure. Purpose To evaluate diagnostic performance of a reduced-dose 100 kVp CT protocol with advanced modeled iterative reconstruction as compared to a linearly blended 120 kVp protocol for assessment of acute, non-traumatic abdominal pain. Material and Methods Two radiologists assessed 100 kVp and linearly blended 120 kVp series of 112 consecutive patients with acute non-traumatic pain (onset < 48 h) regarding image quality, noise, and artifacts on a five-point Likert scale. Both radiologists assessed both series for abdominal pathologies and for diagnostic confidence. Both 100 kVp and linearly blended 120 kVp series were quantitatively evaluated regarding radiation dose and image noise. Comparative statistics and diagnostic accuracy was calculated using receiver operating curve (ROC) statistics, with final clinical diagnosis/clinical follow-up as reference standard. Results Image quality was high for both series without detectable significant differences ( P = 0.157). Image noise and artifacts were rated low for both series but significantly higher for 100 kVp ( P ≤ 0.021). Diagnostic accuracy was high for both series (120 kVp: area under the curve [AUC] = 0.950, sensitivity = 0.958, specificity = 0.941; 100 kVp: AUC ≥ 0.910, sensitivity ≥ 0.937, specificity = 0.882; P ≥ 0.516) with almost perfect inter-rater agreement (Kappa = 0.939). Diagnostic confidence was high for both dose levels without significant differences (100 kVp 5, range 4–5; 120 kVp 5, range 3–5; P = 0.134). The 100 kVp series yielded 26.1% lower radiation dose compared with the 120 kVp series (5.72 ± 2.23 mSv versus 7.75 ± 3.02 mSv, P < 0.001). Image noise was significantly higher in reduced-dose CT (13.3 ± 2.4 HU versus 10.6 ± 2.1 HU; P < 0.001). Conclusion Reduced-dose abdominal CT using 100 kVp yields excellent image quality and high diagnostic accuracy for the assessment of acute non-traumatic abdominal pain.

Download Full-text

Deep Learning Based Liver Cancer Segmentation from Computed Tomography Images

10.21203/rs.3.rs-65573/v1 ◽

2020 ◽

Author(s):

Yodit Abebe Ayalew ◽

Kinde Anlay Fante ◽

Mohammed Aliy

Keyword(s):

Computed Tomography ◽

Deep Learning ◽

Liver Cancer ◽

Ct Scan ◽

Tumor Segmentation ◽

Cancer Death ◽

Liver Segmentation ◽

Abdominal Ct ◽

Computed Tomography Images ◽

Abdominal Ct Scan

Abstract Background: Liver cancer is the sixth most common cancer worldwide. According to WHO data in 2017, the liver cancer death in Ethiopia reached 1040 (0.16%) from all cancer deaths. Hepatocellular carcinoma (HCC), primary liver cancer causes the death of around 700,000 people each year worldwide and this makes it the third leading cause of cancer death. HCC is occurred due to cirrhosis and hepatitis B or C viruses. Liver cancer mostly diagnosed with a computed tomography (CT) scan. But, the detection of the tumor from the CT scan image is difficult since tumors have similar intensity with nearby tissues and may have a different appearance depending on their type, state, and equipment setting. Nowadays deep learning methods have been used for the segmentation of liver and its tumor from the CT scan images and they are more efficient than those traditional methods. But, they are computationally expensive and need many labeled samples for training, which are difficult in the case of biomedical images. Results: A deep learning-based segmentation algorithm is employed for liver and tumor segmentation from abdominal CT scan images. Three separate UNet models, one for liver segmentation and the others two for tumor segmentation from the segmented liver and directly from the abdominal CT scan image were used. A dice score of 0.96 was obtained for liver segmentation. And a dice score of 0.74 and 0.63 was obtained for segmentation of tumor from the liver and from abdominal CT scan image respectively. Conclusion: The research improves the liver tumor segmentation that will help the physicians in the diagnosis and detection of liver tumors and in designing a treatment plan for the patient. And for the patient, it increases the patients’ chance of getting treatment and decrease the mortality rate due to liver cancer.

Download Full-text

Development of Deep Learning-Based Automatic Detection Algorithm for Adrenal Nodules on Contrast-Enhanced Abdominal CT Scans

Journal of the Endocrine Society ◽

10.1210/jendso/bvab048.166 ◽

2021 ◽

Vol 5 (Supplement_1) ◽

pp. A82-A83

Author(s):

Chang Ho Ahn ◽

Tae Woo Kim ◽

Kyungmin Jo ◽

Sung Hye Kong ◽

Jinhee Kim ◽

...

Keyword(s):

Deep Learning ◽

Adrenal Gland ◽

External Validation ◽

Automatic Detection ◽

Ct Scans ◽

Average Sensitivity ◽

Free Response ◽

Abdominal Ct ◽

Contrast Enhanced ◽

Adrenal Nodule

Abstract Objective: Adrenal nodules are often incidentally detected on abdominal computed tomography (CT) scans due to their asymptomatic nature. We aimed to develop an automatic detection program for adrenal nodules on abdominal CT scans using deep learning algorithms. Methods: We retrospectively analyzed abdominal CT scans performed at two university-affiliated hospitals (n = 483 and n = 514, respectively) from 2006 to 2019. This dataset was randomly divided into training set (181 CTs without adrenal nodule and 362 CTs with adrenal nodule) and test set (291 CTs without adrenal nodule and 163 CTs with adrenal nodule). All CT scans were contrast-enhanced and the phase with the highest contrast between adrenal gland and adjacent normal tissues was selected for multi-phase CT. The core algorithm of our deep learning algorithm for adrenal nodule (DLAAN) was MULAN (Multitask Universal Lesion Analysis Network) algorithm whose backbone was a convolutional neural network. DLAAN was composed of two stages. The first stage was to detect the CT slice where normal adrenal gland or adrenal nodule were located. The second stage was for fine localization of adrenal nodule on the corresponding CT slice. The performance of DLAAN was evaluated using the area under the receiver operating characteristic curve (AUROC) for patient-level classification and free-response ROC for nodule-level localization. The figure of merit for free-response ROC was calculated as an average sensitivity when 0.5, 1, 2, and 4 false positives per slice were allowed. Results: The AUROC of DLAAN was 0.927 (95% confidence interval: 0.900–0.955). With a threshold probability of 0.9, the sensitivity and specificity were 86.5% and 89.0%, respectively. When left and right adrenal nodules were analyzed separately, the AUROC was 0.910 for left adrenal nodule and 0.957 for right adrenal nodule, respectively. The accuracy of DLAAN according to the size of adrenal nodule was 0.890, 0.734, 0.981, 1.00 and 1.00 for no adrenal nodule, adrenal nodule sized 1–2 cm, 2–3 cm, 3–4 cm and > 4 cm, respectively. The performance of DLAAN for the localization of adrenal nodule which was estimated by average sensitivity was 0.812. The number of CTs with at least one false positive nodule was 93/454 (20.5%). Conclusion: Our proof of concept study of deep learning-based automatic detection of adrenal nodule on contrast-enhanced abdominal CT scans showed high accuracy for both the classification of patients with or without adrenal nodule and the localization of adrenal nodule, although the performance of the algorithm decreased for small sized adrenal nodules. External validation with different CT settings and patient population is needed to assess the generalizability of our algorithm.

Download Full-text