Computer-Aided Detection of Colorectal Polyps at CT Colonography: Prospective Clinical Performance and Third-Party Reimbursement

Ziemlewicz, Timothy J.; Kim, David H.; Hinshaw, J. Louis; Lubner, Meghan G.; Robbins, Jessica B.; Pickhardt, Perry J.

doi:10.2214/ajr.16.17499

Cited by 7 publications

(3 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Studies of three-dimensional computed tomography air-contrast enema (CT enema) simulating BE have been reported [ 28 , 29 , 33 – 35 ]. Also, computer-aided diagnosis systems including artificial intelligence (AI) have been developed [ 36 , 37 ]. Several studies have measured the wall rigidity under the profile view of the CT enema in CRC.…”

Section: Discussionmentioning

confidence: 99%

Role of barium enema examination for the diagnosis of submucosal invasion depth in T1 colorectal cancers

et al. 2021

View full text Add to dashboard Cite

Background The indication for endoscopic resection for submucosally invasive colorectal cancer (T1-CRC) depends on the preoperative diagnosis of invasion depth. The aim of this investigation was to evaluate the association between barium enema examination (BE) profile views and depth of submucosal (SM) invasion in CRCs. Methods We reviewed the radiographic and endoscopic findings of 145 T1-CRCs diagnosed from 2008 to 2019. We measured the widths of horizontal and vertical rigidity under a BE profile view corresponding to CRC and compared the values with SM invasion depth. Horizontal rigidity was defined as the horizontal length and vertical rigidity as the vertical width of the barium defect corresponding to each target lesion. The most appropriate cut-off values for predicting SM invasion ≥1.8 mm were calculated by receiver operating characteristic curve analysis. Results Values of horizontal rigidity (r = 0.626, P < 0.05) and vertical rigidity (r = 0.482, P < 0.05) correlated significantly with SM invasion depth. The most appropriate cut-off values for the prediction of SM invasion depth ≥ 1.8 mm were 4.5 mm for horizontal rigidity, with an accuracy of 80.7%; and 0.7 mm for vertical rigidity, with an accuracy of 77.9%. The prevalence of lympho-vascular invasion was significantly different when those cut-off values were applied (43.2% vs. 17.5% for horizontal rigidity, P < 0.005). Conclusions In T1-CRC, values of horizontal and vertical rigidities under a BE profile view were correlated with SM invasion depth. While the accuracy of the rigidities for the prediction of SM invasion depth ≥ 1.8 mm was not high, horizontal rigidity may be predictive of lympho-vascular invasion, thus aiding in therapeutic decision-making.

show abstract

Section: Discussionmentioning

confidence: 99%

Role of barium enema examination for the diagnosis of submucosal invasion depth in T1 colorectal cancers

et al. 2021

View full text Add to dashboard Cite

show abstract

“…This technology can boost multiple functions in an objective manner, which can assist in both image diagnostics and image enhancement. These models have well‐documented use in imaging modalities such as computed tomography (CT) and radiography …”

Section: List Of Common Terms and Definitions In MLmentioning

confidence: 99%

“…The application of CNNs in the field of image and video processing has increased considerably in the past 6 years. This has been facilitated by easily accessible, free, open‐source (software code that is publicly available) software packages such as TensorFlow (TensorFlow 1.4; Google, LLC, Mountain View, CA), Cognitive Toolkit (Microsoft Corporation, Redmond, WA), PyTorch, Caffe (University of California, Berkeley, CA), and MXNet (Apache Software Foundation, Forest Hill, MD) for rapidly implementing and fine‐tuning DL models . Within medical imaging, a large variety of use cases have been reported from almost every aspect of medical image analysis, including detection of pleural effusion and cardiomegaly on chest radiography, mediastinal lymph nodes on CT, lung nodules on CT, and detection of tuberculosis on chest radiographs …”

Section: Deep‐learning Technologymentioning

confidence: 99%

Enhanced Point‐of‐Care Ultrasound Applications by Integrating Automated Feature‐Learning Systems Using Deep Learning

Shokoohi

LeSaux

Roohani

et al. 2018

J of Ultrasound Medicine

View full text Add to dashboard Cite

Recent applications of artificial intelligence (AI) and deep learning (DL) in health care include enhanced diagnostic imaging modalities to support clinical decisions and improve patients' outcomes. Focused on using automated DLbased systems to improve point-of-care ultrasound (POCUS), we look at DL-based automation as a key field in expanding and improving POCUS applications in various clinical settings. A promising additional value would be the ability to automate training model selections for teaching POCUS to medical trainees and novice sonologists. The diversity of POCUS applications and ultrasound equipment, each requiring specialized AI models and domain expertise, limits the use of DL as a generic solution. In this article, we highlight the most advanced potential applications of AI in POCUS tailored to high-yield models in automated image interpretations, with the premise of improving the accuracy and efficacy of POCUS scans.

show abstract

Visual Image Annotation for Bowel Obstruction: Repeatability and Agreement with Manual Annotation and Neural Networks

Murphy

2023

J Digit Imaging

View full text Add to dashboard Cite

Bowel obstruction is a common cause of acute abdominal pain. The development of algorithms for automated detection and characterization of bowel obstruction on CT has been limited by the effort required for manual annotation. Visual image annotation with an eye tracking device may mitigate that limitation. The purpose of this study is to assess the agreement between visual and manual annotations for bowel segmentation and diameter measurement, and to assess agreement with convolutional neural networks (CNNs) trained using that data. Sixty CT scans of 50 patients with bowel obstruction from March to June 2022 were retrospectively included and partitioned into training and test data sets. An eye tracking device was used to record 3-dimensional coordinates within the scans, while a radiologist cast their gaze at the centerline of the bowel, and adjusted the size of a superimposed ROI to approximate the diameter of the bowel. For each scan, 59.4 ± 15.1 segments, 847.9 ± 228.1 gaze locations, and 5.8 ± 1.2 m of bowel were recorded. 2d and 3d CNNs were trained using this data to predict bowel segmentation and diameter maps from the CT scans. For comparisons between two repetitions of visual annotation, CNN predictions, and manual annotations, Dice scores for bowel segmentation ranged from 0.69 ± 0.17 to 0.81 ± 0.04 and intraclass correlations [95% CI] for diameter measurement ranged from 0.672 [0.490–0.782] to 0.940 [0.933–0.947]. Thus, visual image annotation is a promising technique for training CNNs to perform bowel segmentation and diameter measurement in CT scans of patients with bowel obstruction.

show abstract

Computer-Aided Detection of Colorectal Polyps at CT Colonography: Prospective Clinical Performance and Third-Party Reimbursement

Abstract: In our routine clinical practice, CAD showed good sensitivity for detecting colorectal polyps 6 mm or larger, with an acceptable number of false-positive marks. Importantly, CAD is already being reimbursed by some third-party payers in our clinical CTC practice.

Cited by 7 publications

References 30 publications

Role of barium enema examination for the diagnosis of submucosal invasion depth in T1 colorectal cancers

Role of barium enema examination for the diagnosis of submucosal invasion depth in T1 colorectal cancers

Enhanced Point‐of‐Care Ultrasound Applications by Integrating Automated Feature‐Learning Systems Using Deep Learning

Visual Image Annotation for Bowel Obstruction: Repeatability and Agreement with Manual Annotation and Neural Networks

Contact Info

Product

Resources

About