UMIN-ICDS Clinical Trial

Public title

Effect of AI-based image reconstruction on quantitative values (ADC) and image quality in diffusion-weighted MRI

Acronym

DLR-ADC study in pelvic DWI

Scientific Title

Impact of deep learning reconstruction on apparent diffusion coefficient values and image quality in pelvic diffusion-weighted MRI: a prospective observational study

Scientific Title:Acronym

DLR-ADC study in pelvic DWI

Region

Japan

Condition

Gynecological diseases (endometrial cancer, cervical cancer, ovarian cancer, etc.)

Classification by specialty

Obstetrics and Gynecology

Radiology

Classification by malignancy

Malignancy

Genomic information

NO

Narrative objectives1

To compare ADC values and image-quality metrics between conventional reconstruction and multiple acquisition and reconstruction conditions of pelvic diffusion-weighted imaging obtained within the same examination. The aims are to identify the range of conditions in which ADC values remain equivalent to conventional reconstruction, and to determine whether classification based on an established diagnostic cutoff of 0.9 x 10-3 mm2/s changes across reconstruction conditions.

Basic objectives2

Others

Basic objectives -Others

Evaluation of the effect of image reconstruction conditions on apparent diffusion coefficient values and image-quality metrics derived from diffusion-weighted imaging

Trial characteristics_1

Confirmatory

Trial characteristics_2

Pragmatic

Developmental phase

Not applicable

Primary outcomes

Difference in tumor ADC between each acquisition and reconstruction condition and the conventional reconstruction of each scanner, assessed against an equivalence margin of +/-0.05 x 10-3 mm2/s.

Key secondary outcomes

- Reclassification rate across conditions based on the ADC cutoff of 0.9 x 10-3 mm2/s
- Dose-response relationship between denoising level and ADC change, by tissue type
- ADC histogram metrics (10th percentile, standard deviation) across conditions
- ADC measured with a small region of interest placed on the visually lowest-signal area
- Signal intensity ratio, signal-to-noise ratio, and contrast-to-noise ratio, measured on both ADC maps and high b-value diffusion-weighted images
- ADC changes by tissue type (tumor, myometrium, skeletal muscle, and normal endometrium or coexisting leiomyoma when measurable)
- Comparison of ADC values and image quality between acquisition methods (single-shot echo-planar imaging, compressed sensing, and others)
- Three-point qualitative assessment of lesion-to-background separation by three readers, on both ADC maps and high b-value images

Study type

Observational

Basic design

Randomization

Randomization unit

Blinding

Control

Stratification

Dynamic allocation

Institution consideration

Blocking

Concealment

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

Age-lower limit

18

years-old

<=

Age-upper limit

Not applicable

Gender

Female

Key inclusion criteria

(1) Patients who underwent pelvic MRI on a 1.5-T or 3-T scanner at our institution for the workup of gynecological disease between study approval and March 2028. The standard clinical protocol includes diffusion-weighted imaging with multiple acquisition and reconstruction conditions. (2) Age 18 years or older. (3) Reconstruction data available for analysis.

Key exclusion criteria

Image quality judged inadequate for quantitative analysis because of severe motion or other causes.

Target sample size

60

Name of lead principal investigator

1st name	TSUKASA
Middle name
Last name	SAIDA

Organization

University of Tsukuba

Division name

Institute of Medicine, Department of Diagnostic Radiology and IVR

Zip code

305-8575

Address

1-1-1 Tennodai, Tsukuba, Ibaraki 305-8575, Japan

TEL

029-853-3205

Email

saida_sasaki_tsukasa@md.tsukuba.ac.jp

Name of contact person

1st name	TSUKASA
Middle name
Last name	SAIDA

Organization

University of Tsukuba

Division name

Institute of Medicine, Department of Diagnostic Radiology and IVR

Zip code

305-8575

Address

1-1-1 Tennodai, Tsukuba, Ibaraki 305-8575, Japan

TEL

029-853-3205

Homepage URL

Email

saida_sasaki_tsukasa@md.tsukuba.ac.jp

Institute

University of Tsukuba

Institute

Department

Personal name

Organization

University of Tsukuba

Organization

Division

Category of Funding Organization

Self funding

Nationality of Funding Organization

Co-sponsor

Name of secondary funder(s)

Organization

Institutional Review Board, University of Tsukuba Hospital

Address

2-1-1 Amakubo, Tsukuba, Ibaraki 305-8576, Japan

Tel

029-853-3749

Email

rinshokenkyu@un.tsukuba.ac.jp

Secondary IDs

NO

Study ID_1

Org. issuing International ID_1

Study ID_2

Org. issuing International ID_2

IND to MHLW

Institutions

筑波大学附属病院

Date of disclosure of the study information

2026

Year

09

Month

22

Day

URL releasing protocol

Publication of results

Unpublished

URL related to results and publications

Number of participants that the trial has enrolled

Results

Results date posted

Results Delayed

Results Delay Reason

Date of the first journal publication of results

Baseline Characteristics

Participant flow

Adverse events

Outcome measures

Plan to share IPD

IPD sharing Plan description

Recruitment status

Preinitiation

Date of protocol fixation

2026

Year

08

Month

15

Day

Date of IRB

Anticipated trial start date

2026

Year

10

Month

01

Day

Last follow-up date

2028

Year

03

Month

31

Day

Date of closure to data entry

Date trial data considered complete

Date analysis concluded

Other related information

Study design: A single-center prospective observational study. Consecutive patients imaged for clinical purposes after study approval are enrolled. No intervention is performed.

Recruitment: All consecutive patients who meet the inclusion criteria and undergo pelvic MRI on a 1.5-T or 3-T scanner at our institution for the workup of gynecological disease between study approval and March 2028. No volunteers are recruited. Consent is obtained through an opt-out procedure.

Recruitment period: From study approval to March 2028.

Measurements: Diffusion-weighted images acquired under the standard clinical protocol with multiple acquisition and reconstruction conditions are analyzed. Regions of interest are placed on the tumor and on reference tissues, including myometrium, skeletal muscle, and, when measurable, normal endometrium and coexisting leiomyoma. Mean, standard deviation, 10th percentile, and signal intensity are measured on both ADC maps and high b-value images. Signal intensity ratio, signal-to-noise ratio, and contrast-to-noise ratio are calculated. Three readers also perform a three-point qualitative assessment of lesion-to-background separation. Regions of interest are placed on the reference images and copied to all reconstruction conditions at identical coordinates.

Analysis: A mixed effects model with patient as a random effect is used to estimate the difference in ADC between each condition and the conventional reconstruction, with 95% confidence intervals, which are compared with a prespecified equivalence margin of +/-0.05 x 10-3 mm2/s. Dunnett-type multiple comparison is applied. Descriptive statistics are reported as median and interquartile range.

Registered date

2026

Year

09

Month

22

Day

Last modified on

2026

Year

09

Month

22

Day

Link to view the page

Value
https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000072144