2020
DOI: 10.1007/s00261-020-02741-x
|View full text |Cite
|
Sign up to set email alerts
|

Development of a volumetric pancreas segmentation CT dataset for AI applications through trained technologists: a study during the COVID 19 containment phase

Abstract: Purpose To evaluate the performance of trained technologists vis-à-vis radiologists for volumetric pancreas segmentation and to assess the impact of supplementary training on their performance. Methods In this IRB-approved study, 22 technologists were trained in pancreas segmentation on portal venous phase CT through radiologist-led interactive videoconferencing sessions based on an image-rich curriculum. Technologists segmented pancreas in 188 CTs using freehand tools on custom image-viewing software. Subsequ… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
13
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
7

Relationship

1
6

Authors

Journals

citations
Cited by 19 publications
(13 citation statements)
references
References 24 publications
0
13
0
Order By: Relevance
“…These regions of pancreas are difficult to delineate even for expert radiologists and often contribute to intra-reader and inter-reader variability. 44,45 The model-predicted mean (SD) pancreatic volumes were not different from the volumes derived from GT segmentations either in the internal test set (P = 0.08) or in the TCIA dataset (P = 0.3). Only one prior study reports results from an AI model-predicted pancreatic volumes.…”
Section: Discussionmentioning
confidence: 90%
See 1 more Smart Citation
“…These regions of pancreas are difficult to delineate even for expert radiologists and often contribute to intra-reader and inter-reader variability. 44,45 The model-predicted mean (SD) pancreatic volumes were not different from the volumes derived from GT segmentations either in the internal test set (P = 0.08) or in the TCIA dataset (P = 0.3). Only one prior study reports results from an AI model-predicted pancreatic volumes.…”
Section: Discussionmentioning
confidence: 90%
“…Segmentation errors were commonly seen at the interface of pancreatic head and duodenum and pancreatic tail and jejunum or splenic vessels. These regions of pancreas are difficult to delineate even for expert radiologists and often contribute to intra‐reader and inter‐reader variability 44,45 …”
Section: Discussionmentioning
confidence: 99%
“…Hence, the preparation of CT scans is time-consuming. Future research needs to address the potential benefit of artificial intelligence for this part [24][25][26][27]. Once the G-code from CT data is deduced, the replica can easily be re-printed in a larger quantity, thereby opening the possibility for extended teaching and training purposes [28,29].…”
Section: Discussionmentioning
confidence: 99%
“… 25 [ 146 ] X-ray Data set based on previous literature 26 [ 59 ] X-ray JSRT/SCR dataset NLM( MC) Cohen dataset CoronaHack 27 [ 94 ] X-ray X-ray images from GitHub 28 [ 103 ] X-ray Shenzhen set NIH (National Institutes of Health, US) Clinical Center Hospital dataset COVID-Chest Xray-Dataset 29 [ 74 ] CT images Data set based on previous literature 30 [ 88 ] X-ray, CT images Hitherto dataset 31 [ 99 ] X-ray Github Kaggle 32 [ 134 ] US Nine COVID-19 patients 33 [ 102 ] CT images COVID-19 patient dataset. CT datasets of patients with different medical histories other than COVID-19 34 [ 148 ] CT images Radiologists’ review [ 128 ]. 35 [ 131 ] X-ray COVID-CT-Dataset [ 155 ] <...>…”
Section: Datasets and Imaging Modalities In The Surveyed Studiesmentioning
confidence: 99%