Do you need help?

The Cancer Imaging Biobank (CAIB) is a collaborative project aimed to make health data readily available for AI applications. It was conceived in view of the limited availability of comprehensive, complete, timeline data in the public domain, representative of patients from India. The project collects, pre-processes and generates anonymized data centred around pathology, and radiology images, linked to their clinical information including treatment details, outcomes in addition to patient level, study level and image level labels.

The data custodian is TMC-ACTREC and a copy of the data is maintained at Indian Biological Data Centre at Faridabad.

Eligibile users can access this data by registering on CAIB. Potential users would be required to validate their researcher status and affiliation to a research/academic/government/not-for-profit organization in the profile section. Once the validation is successful, users would be able to browse through the data, select the data/files for requesting downloads. The data will also be available through DBT at the IBDC data portal (https://ibdc.dbtindia.gov.in/).

Complete registration steps are as follows:

1. Fill registration form at the sign up link. Enter valid organizational & academic email id only.
2. Verify your email by clicking on the verification link sent to your email.
3. Once your email is verified, you can login into the platform and complete your profile. Complete the mandatory fields at the minimum before submitting/updating your profile.
4. Your profile will be reviewed by the administrator and access granted.
5. You can then browse through the data, filter data and select appropriate data for requesting download access.

The data access is governed by the BIOTECH-PRIDE guidelines and FeED protocols from DBT, Govt. of India. The data is not licensed for commercial usage in any form. Private/for-profit/start-up entitites/individuals involved in research activities would need to partner with an eligible organization/researcher, who can in-turn submit application for data-access.

Currently, there is no cost to register on CAIB or for data access. However, this policy may change depending upon guidance from DBT, and the CAIB Steering committee, at a future date.

Yes. You need to create a separate account if you move to a different institute or your academic email address changes. You will need to complete all procedures related to a new account registration before you can access the data.

Currently, the data is accessible to researchers and academics working in academic and research organizations, in medical AI-related fields or other medical related fields where the data collected in this project would help solve real-world problems. The access policy aligns with the Biotech-PRIDE Guidelines and FeED protocols published by DBT (https://dbtindia.gov.in/).

No. The email address used to register must be your personal academic institutional email address. PI/teachers/faculty registering on behalf of their student or any third party, take full responsibility for their actions.

The CAIB project focuses on Pathology and Radiology images along with linked laboratory and clinical information. Accordingly, the following types of files are available on CAIB:

1. Timeline Pathology whole slide images
2. Timeline Radiology DICOM files
3. Annotations, segmentation masks
4. Pathology and radiology report data
5. Slide-level, study level, patient level labels
6. Key timeline patient information, treatment and outcome data.

Care has been taken to ensure that all data is anonymized. Accordingly the following steps have been taken:

1. All personal identifiers are erased/replaced with CAIB identifiers.
2. All dates have been replaced by "Days from Date of Birth".
3. All reports and documents have been passed through a custom anonymization pipeline to redact any identifiable information or dates.

If any identifiable information is found on any of the data on CAIB, users are required to report this finding with details to CAIB management team at caib@actrec.gov.in.

The data submitters take the responsibility for the veracity and accuracy of the data they submit. The submission system helps in ensuring accuracy by limiting free text entry and having a pre-defined data entry options. The data once entered is verified and then finalised/committed. This data is verified by pathologists, radiologists or oncologists. Only finalised/committed records are visible to the users.

As of December 2025, about 100 TB of data from roughly 2700 patients is available on CAIB for public access. CAIB envisages to have minimum 1000 patient data for each cancer type initially and gradually build up the repository to have minimum 10000 patient data for each cancer type. Current number of patients and files is reflected on the data portal's landing page.

CAIB aims to provide access to data and files from all types of cancers. Currently, data related to the following types of cancers are available for public access:

1. Head Neck Cancer
2. Lung Cancer.

In addition, data related to the following types of cancer are in different stages of processing in preparation for public access:

1. Breast Cancer
2. Ovarian Cancer
3. GI Cancer
4. Haematological Cancers
5. others.

Data within CAIB is available strictly for non-commercial academic and research usage only. Accordingly, the data is made available under a CC BY-NC-SA 4.0 license, unless a specific dataset states otherwise. In plain terms, this means you may:

1. Use the data for academic and non-commercial research only as approved in the Data Use Agreement, including training and validating AI/ML models.
2. Publish results, share code, and present findings, provided CAIB is credited as the data source.
3. Create & further share derived datasets for non-commercial academic or research usage, provided you do not share the raw data received from CAIB as is, and provided CAIB is credited as the data source.

You may not:

1. Sell, sublicense, or otherwise commercially exploit the data or any derived datasets from this data itself.
2. Redistribute the raw data to a third party who has not separately been approved for access by CAIB.

If your intended use is commercial (for example, building a product for sale), contact individual data providing organizations separately to discuss a commercial data licence; this policy covers non-commercial research access only.

Once your profile is approved (after email & profile verification), researchers can browse and search through the data available.

1. Add the files and records, you would like to download to the cart.
2. On checkout/submitting the cart, you will be required to fill a short form detailing how you would be using the data.
3. Once your submit the form, a project ID will be assigned to the access request.
4. The CAIB management team will review the access request and approve/reject the request. This process is expected to take up to 30 working days.
5. Once data access/project is approved, the researcher would need to upload the signed data use/material transfer agreement.
6. The approved data will be made available for download along with access credentials. This download access will be available for 15 days. If the download is not completed within 15 days, you may request an extension for a further 15 days. Beyond this time, the temporary access provided will automatically expire and the researcher will be required to submit a fresh data access request.

You may view your previous requests for access, their current status and details for downloading the data in the 'My Downloads' section.

Yes. Users are required to submit intended usage details along with the data access request form.

In addition, all registered users would be required to submit an annual report of how they have been using the data access provided by CAIB. The users should submit details of any publications, presentations done, or any further modification/share created from the downloaded data. This submission is mandatory to maintain access to the CAIB database through the next year.

Students with an institutional email address are welcome to apply for access. If you do not have an institutional email address, you can request your supervisor/guide/teacher to apply for access. Supervisor/guide/teacher registering on behalf of a student/individual, take full responsibility for their actions.

Currently the registration process is only available for researchers/individuals working/registered with academic organizations. This is validated through your organizational email address. If you feel that you are eligible for access and system is erroneously blocking your application despite filling the form correctly, you may contact the CAIB management team on caib@actrec.gov.in. Be as elaborate as possible to justify your eligibility for access. The CAIB management team will revert to your query within 2 working weeks time.

No. No such collaboration is required for accessing the images and data stored in CAIB database.

The data in CAIB is sourced from multiple hospitals across India. Currently the data is sourced from Tata Memorial Hospital, Mumbai; ACTREC, Navi Mumbai; AIIMS-New Delhi; PGIMER-Chandigarh and RGCIRC- New Delhi. Additional institutions may contribute to this database in the future.

No. CAIB does not provide access to the tissues or any data not already present in the database. No such requests will be approved from CAIB. However, users are free to enter into collaboration agreements with the original data contributors if any tissue access/data access beyond that already available in CAIB.

Good Clinical Practice (GCP) certification emphasises ethical treatment and safety of human participants in pre-clinical and clinical studies, ensures the integrity and reliability of the data collected, and builds public trust by demonstrating compliance with international standards. It provides a common framework for conducting research, protects participant rights, safeguards well-being, and ensures that the resulting data is scientifically sound and credible for regulatory approval and medical advancement. 

Since data in CAIB is actual patient data (anonymized), it is recommended that data users are aware of their responsibilities, rights, ethics and duties while dealing with patient data (even anonymised), and that they ensure that human participants or their data are treated with respect and their confidentiality is maintained. Hence, it is recommended, though not mandatory, that researchers undergo GCP training.

Good clinical practice certification is provided by many organizations, including hospitals, academic organizations, clinical research organizations, etc. If your organization has an ethics committee, it is possible they provide some courses or can guide you how to acquire this certificaiton. If your organization does not provide this certification, researchers can get this certification from external agencies. The following is a non-comprehensive list of organizations and online programs offering such certification:

1. Institute of Good Clinical Practice (https://igmpi.ac.in/)
2. https://gcp.nidatraining.com/
3. https://coursera.org/learn/good-clinical-practice-novartis
4. https://about.citiprogram.org/series/good-clinical-practice-gcp/

The Technical Documentation contains all details related to the data structure, formats and other detailed information you may seek from this database. The Technical Documentation can be accessed after login. Please go through this documentation before contacting the team seeking any clarification.