UMIN-CTR Clinical Trial
| Unique ID issued by UMIN | UMIN000037053 |
|---|---|
| Receipt number | R000042207 |
| Scientific Title | Development of Computer-Aid Diagnosis System for T1b Colorectal Cancer on Non-magnified Plain Endoscopic Images: A Multicenter Study |
| Date of disclosure of the study information | 2019/07/01 |
| Last modified on | 2022/12/16 06:17:54 |
* This page includes information on clinical trials registered in UMIN clinical trial registed system.
* We don't aim to advertise certain products or treatments
Basic information
Public title
Development of AI Diagnosis System for T1b Colorectal Cancer on Non-magnified Plain Endoscopic Images
Acronym
AI Dx for T1b CRC
Scientific Title
Development of Computer-Aid Diagnosis System for T1b Colorectal Cancer on Non-magnified Plain Endoscopic Images: A Multicenter Study
Scientific Title:Acronym
CAD for T1b CRC by Plain Images
Region
| Japan |
Condition
Condition
Early stage colorectal cancer (Tis/T1a/T1b) receiving colonoscoppy between 2016 Jan. and 2018 Dec.
Classification by specialty
| Gastroenterology |
Classification by malignancy
Malignancy
Genomic information
NO
Objectives
Narrative objectives1
To develop and evaluate Computer-aid Diagnosis System to differentiate Tis/T1a stage from T1b stage colorectal cancer.
Basic objectives2
Others
Basic objectives -Others
Comparison with human endoscopists
Trial characteristics_1
Confirmatory
Trial characteristics_2
Others
Developmental phase
Not applicable
Assessment
Primary outcomes
Specificity for T1b colorectal cancer
Key secondary outcomes
Sensitivity, PPV, NPV and overall accuracy for T1b colorectal cancer
Base
Study type
Others,meta-analysis etc
Study design
Basic design
Randomization
Randomization unit
Blinding
Control
Stratification
Dynamic allocation
Institution consideration
Blocking
Concealment
Intervention
No. of arms
Purpose of intervention
Type of intervention
Interventions/Control_1
Interventions/Control_2
Interventions/Control_3
Interventions/Control_4
Interventions/Control_5
Interventions/Control_6
Interventions/Control_7
Interventions/Control_8
Interventions/Control_9
Interventions/Control_10
Eligibility
Age-lower limit
| 20 | years-old | <= |
Age-upper limit
| 100 | years-old | >= |
Gender
Male and Female
Key inclusion criteria
1) Consecutive series of Tis/T1a/T1b stage colorectal cancer endoscopically observed between 2016 Jan and 2018 December.
2) Non-magnified plain still images (3-5 shots)
Key exclusion criteria
1) Pedunculated morphology
2) Huge lesion, >5cm in size
3) Chromoendoscopic images
4) Images enhanced by NBI, BLI etc.
Target sample size
1500
Research contact person
Name of lead principal investigator
| 1st name | Kazutomo |
| Middle name | |
| Last name | Togashi |
Organization
Aizu Medical Center Fukushima Medical University
Division name
Dept. of Coloproctology
Zip code
969-3492
Address
21-2Maeda, Tanisawa, Kawahigashi-machi, Aizuwakamatsu, Fukushima
TEL
+81-242752100
togashik@fmu.ac.jp
Public contact
Name of contact person
| 1st name | Masato |
| Middle name | |
| Last name | Aizawa |
Organization
Aizu Medical Center Fukushima Medical University
Division name
Dept. of Coloproctology
Zip code
969-3492
Address
21-2Maeda, Tanisawa, Kawahigashi-machi, Aizuwakamatsu, Fukushima
TEL
+81-242752100
Homepage URL
aizawa-m@fmu.ac.jp
Sponsor or person
Institute
Fukushima Medical University
Institute
Department
Personal name
Funding Source
Organization
Fukushima Medical University
Organization
Division
Category of Funding Organization
Self funding
Nationality of Funding Organization
Other related organizations
Co-sponsor
Name of secondary funder(s)
IRB Contact (For public release)
Organization
Fukushima Medical University
Address
1 Hikarigaoka, Fukushima-city
Tel
+81-25471825
rs@@fmu.ac.jp
Secondary IDs
Secondary IDs
YES
Study ID_1
30301
Org. issuing International ID_1
Fukushima Medical University
Study ID_2
Org. issuing International ID_2
IND to MHLW
Institutions
Institutions
福島県立医科大学会津医療センター（福島県）、自治医科大学（栃木県）、群馬大学医学部（群馬県）、小樽掖済会病院（北海道）、佐賀大学医学部（佐賀県）、自治医科大学さいたま医療センター（埼玉県）、東京女子医科大学東医療センター（東京都）、岩手医科大学（岩手県）、東京女子医科大学消化器病センター（東京都）、京都府立医科大学（京都府）、会津大学コンピュータ理工学部（福島県）
Other administrative information
Date of disclosure of the study information
| 2019 | Year | 07 | Month | 01 | Day |
Related information
URL releasing protocol
https://upload.umin.ac.jp/cgi-open-bin/icdr_e/ctr_view.cgi?recptno=R000042207
Publication of results
Partially published
Result
URL related to results and publications
https://www.giejournal.org/article/S0016-5107(21)00682-9/fulltext
Number of participants that the trial has enrolled
1513
Results
At 90% cutoff for per-lesion score, the CADx system showed the highest specificity of 94.4% [95% confidence interval: 91.3-96.6], with 59.8% [48.3-70.4] sensitivity and 87.3% [83.7-90.4] accuracy. The area under the characteristic curve was 85.1% [79.9-90.4] for CADx, 88.2% [83.7-92.8] for expert 1, 85.9% [80.9-90.9] for expert 2, 77.0% [71.5-82.4] for trainee 1 (CADx vs. trainee 1: p=0.0076), and 66.2% [60.6-71.9] for trainee 2 (CADx vs. trainee 2: p<0.0001).
Results date posted
| 2021 | Year | 12 | Month | 19 | Day |
Results Delayed
Results Delay Reason
Date of the first journal publication of results
Baseline Characteristics
A total of 1513 lesions (Tis 1074, T1a 145, T1b 294) in 5108 images were collected from 1470 patients at ten academic hospitals.
Participant flow
The lesions were assigned to training and testing datasets (3:1). The ResNet-50 network was used as the backbone to extract features from images. Over-sampling and focal loss were used to compensate class imbalance of invasive stage. Diagnostic performance was assessed using the testing dataset including 403 CRCs (Tis 276, T1a 45, T1b 82) with 1392 images. The CADx system generated a score for T1b diagnosis for each lesion. Two experts and two trainees read the identical testing dataset.
Adverse events
None
Outcome measures
Per-lesion specificity for T1b colorectal cancers
Area under the characteristic curve
Other diagnostic values
Plan to share IPD
IPD sharing Plan description
Progress
Recruitment status
Completed
Date of protocol fixation
| 2019 | Year | 04 | Month | 01 | Day |
Date of IRB
| 2019 | Year | 05 | Month | 24 | Day |
Anticipated trial start date
| 2019 | Year | 10 | Month | 01 | Day |
Last follow-up date
| 2020 | Year | 06 | Month | 30 | Day |
Date of closure to data entry
| 2020 | Year | 10 | Month | 31 | Day |
Date trial data considered complete
| 2020 | Year | 12 | Month | 31 | Day |
Date analysis concluded
| 2022 | Year | 03 | Month | 31 | Day |
Other
Other related information
After deep learning, diagnostic performance of AI is measured in a new reading set. Also, human endoscopists read the same reading set, and the diagnostic performance is compared with AI.
Management information
Registered date
| 2019 | Year | 06 | Month | 13 | Day |
Last modified on
| 2022 | Year | 12 | Month | 16 | Day |
Link to view the page
Value
https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000042207
| Research Plan | |
|---|---|
| Registered date | File name |
| 2021/09/27 | Protocol_200714.docx |
| Research case data specifications | |
|---|---|
| Registered date | File name |
| 2021/12/19 | Individual case data specification.xlsx |
| Research case data | |
|---|---|
| Registered date | File name |
| 2021/12/19 | individual case data.xlsx |
