Cancer diagnostics using artificial intelligence

ISRCTN	ISRCTN16907234
DOI	https://doi.org/10.1186/ISRCTN16907234
ClinicalTrials.gov (NCT)	Nil known
Clinical Trials Information System (CTIS)	Nil known
Protocol serial number	480_21
Sponsor	Rigshospitalet
Funders	Aage og Johanne Louis-Hansens Fond, Region Hovedstaden, Hartmann Fonden

Submission date: 13/09/2022
Registration date: 16/09/2022
Last edited: 13/03/2025

Recruitment status: No longer recruiting
Overall study status: Completed
Condition category: Cancer

Prospectively registered

Protocol

Statistical analysis plan

Results

Individual participant data

Plain English summary of protocol

Background and study aims
The exact location, size and activity level of malignant tumors in the head and neck region can be difficult to determine accurately. A correct assessment these tumors on PET-CT images is essential in order to secure successful treatment. The aim of this study is to investigate whether or not artificial intelligence tools can assist in the assessment of head and neck tumors.

Who can participate?
Patients of all ages who were treated for head and neck cancer with radiotherapy at Rigshospitalet between 01/01/2014 and 01/01/2020, who were also scanned with PET-CT for treatment planning purposes

What does the study involve?
The study does not have an effect on patient treatment. Tumor delineations are used to train and validate artificial intelligence algorithms in order to determine if the quality of these is sufficient to be used in a clinical setting.

What are the possible benefits and risks of participating?
Participation involves no risks or benefits.

Where is the study run from?
Rigshospitalet (Denmark)

When is the study starting and how long is it expected to run for?
January 2020 to June 2022

Who is funding the study?
1. Louis-Hansen Fonden (Denmark)
2. Capital Region of Denmark (Denmark)
3. Hartmann Fonden (Denmark)

Who is the main contact?
David Gergely Kovacs Petersen, david.gergely.kovacs.petersen@regionh.dk

Contact information

Mrs Barbara Malene Fischer
Principal investigator

Blegdamsvej 9
Copenhagen
2100
Denmark

ORCID ID	0000-0002-6065-3375
Phone	+45 (0)35459824
Email	barbara.malene.fischer@regionh.dk

Mr David Gergely Kovacs Petersen
Scientific

Blegdamsvej 9
Copenhagen
2100
Denmark

ORCID ID	0000-0002-0383-1446
Phone	+45 (0)30470433
Email	david.gergely.kovacs.petersen@regionh.dk

Study information

Primary study design	Observational
Study design	Single-centre observational study
Secondary study design
Study type	Participant information sheet
Scientific title	Clinical comparison and validation of openly available deep learning methods for automated metabolic tumor volume delineation on PET-CT of head and neck cancer
Study objectives	Artificial intelligence can be used for automated tumor delineation of head and neck cancer in quality matching that of a nuclear medicine specialist.
Ethics approval(s)	Approved 18/06/2020, Danish Patient Safety Authority (Islands Brygge 67, 2300 København S, Denmark; +45 (0)7228 6600; stps@stps.dk), case no. 31-1521-340, ref: SMMO.
Health condition(s) or problem(s) studied	Head and neck cancer
Intervention	A single-centre study of automated tumor delineation accuracy with retrospectively registered head and neck cancer patients scanned with PET-CT for radiotherapy treatment planning between 01/01/2014 and 01/01/2020. Patients are not exposed to any new interventions as a part of this study. The researchers study PET-CT scans that were acquired as a part of routine clinical radiotherapy treatment.
Intervention type	Device
Phase	Not Applicable
Drug / device / biological / vaccine name(s)	Automated metabolic tumor volume delineation
Primary outcome measure(s)	Tumor delineation accuracy measured using the dice coefficient at a single timepoint
Key secondary outcome measure(s)	1. Tumor delineation accuracy measured using Hausdorff distance at a single timepoint 2. Lesion-level detection accuracy measured using f1 score (harmonic mean of positive predictive value and sensitivity) at a single timepoint
Completion date	06/06/2022

Eligibility

Participant type(s)	Patient
Age group	Adult
Sex	All
Target sample size at registration	1200
Total final enrolment	1184
Key inclusion criteria	Patients treated with radiotherapy for cancer of the head and neck who received a PET-CT scan for treatment planning purposes as a part of clinical routine
Key exclusion criteria	1. Patient PET or CT image not acquired according to required protocol 2. Clinical metabolic tumor volume delineation incomplete
Date of first enrolment	21/03/2021
Date of final enrolment	15/10/2021

Locations

Countries of recruitment

Denmark

Study participating centre

Rigshospitalet

Blegdamsvej 9
Copenhagen
2100
Denmark

Results and Publications

Individual participant data (IPD) Intention to share	Yes
IPD sharing plan summary	Stored in publicly available repository
IPD sharing plan	The researchers intend to publish the datasets generated and analysed at either https://www.cancerimagingarchive.net/ or using an in-house server. The data sharing is pending legal approval, and the platform used for sharing depends on this approval. The type of data shared will be PET images, CT images and tumor volumes delineated by nuclear medicine physicians. The data is expected to become available indefinitely for researchers upon publication at the end of 2022. The data will be accessible upon legal approval for healthcare researchers in an anonymized form without prior participant consent. Updated 13/03/2025: The data is shared and available at the website https://rigshospitalet-tumour-segmentation.regionh.dk/-

Study outputs

Output type	Details	Date created	Date added	Peer reviewed?	Patient-facing?
Results article		22/02/2024	23/02/2024	Yes	No
Dataset			13/03/2025	No	No
Participant information sheet	Participant information sheet	11/11/2025	11/11/2025	No	Yes
Protocol file	version 11	16/09/2021	16/09/2022	No	No
Statistical Analysis Plan	version 1		16/09/2022	No	No

Additional files

ISRCTN16907234_PROTOCOL_V11_16Sep21.pdf: Protocol file
ISRCTN16907234_SAP_V1.pdf: Statistical Analysis Plan

Editorial Notes

13/03/2025: IPD sharing plan updated, link to dataset added.
23/02/2024: Publication reference added.
25/07/2023: The intention to publish date has been changed from 30/04/2023 to 31/08/2023.
14/03/2023: The intention to publish date has been changed from 28/02/2023 to 30/04/2023.
28/12/2022: The following changes have been made:
1. The final enrolment number has been changed from 1095 to 1184.
2. The intention to publish date has been changed from 31/12/2022 to 28/02/2023.
16/09/2022: Protocol and statistical analysis plan uploaded.
13/09/2022: Trial's existence confirmed by the Danish Patient Safety Authority, the Danish Data Protection Agency and the Louis-Hansen Foundation.