Artificial Intelligence (AI), Cancer, Cancer-related Pain, Database, Facial Expression, Oncology, Oncology Pain, Pain Assessment, Voice
Conditions
Keywords
Cancer-related Pain, Pain assessment, Artificial Intelligence (AI), Oncology, Oncology Pain, Database, Facial Action Unit, Vocalizations, Acoustic, Automatic assessment, Machine learning, Audiovisual data, Paralinguistics
Brief summary
The goal of this observational study is to collect short video and sound recordings of people with cancer to create a secure database that can be used in future research to develop an artificial intelligence (AI) tool for pain assessment. The main aim is to build a large, high-quality collection of audiovisual data showing how people with cancer express themselves when they do and do not have pain. Participants will include adults with cancer who are admitted to the oncology ward for pain treatment and a control group admitted for chemotherapy who have no pain. After giving consent, participants will: * Be recorded on video (from the shoulders up) for up to 60 seconds while reading a short sentence and describing their pain or daily experience. * Complete a short questionnaire about their mood and pain expression. * Allow researchers to collect some information from their medical record, such as their pain score, medications, and cancer type. These recordings will be securely stored and used to create a database for future AI research. No medical tests, new treatments, or extra hospital visits are involved. This study will provide the foundation for developing future AI-based tools that could support doctors and patients in monitoring and managing pain more accurately and easily.
Detailed description
This is a prospective, single-centre database development study conducted at the Erasmus MC Cancer Institute. The purpose of the study is to create a high-quality audiovisual database for future research on automated pain assessment in people with cancer. The database will include short (up to 60 seconds) video and sound recordings of participants' facial and vocal expressions, alongside relevant clinical and demographic information. The study will include two groups: (1) adults admitted to the oncology ward for cancer-related pain (pain group) and (2) adults admitted for chemotherapy who report no pain (control group). Participants will provide written informed consent before any recording takes place. Each participant will be recorded from the shoulders up while reading a neutral prompt and, if applicable, describing their pain experience. The recordings will be securely stored in compliance with the General Data Protection Regulation (GDPR). In addition, researchers will collect secondary parameters from the participants' electronic medical record, such as pain scores, analgesic use, tumour type, and clinical status. Participants in the pain group may be recorded on multiple days during admission, while controls will be recorded once. The audiovisual recordings will be used to extract facial and vocal features (e.g., Facial Action Units, voice parameters) with open-source software such as OpenFace and OpenSmile. These features will form the foundation for future artificial intelligence (AI) model development aimed at automatic pain assessment. No experimental interventions, additional clinical procedures, or diagnostic tests are part of this study. The study carries minimal risk to participants, primarily related to the handling of identifiable audiovisual data. All data will be stored on Erasmus MC's secure Research Storage & Compute infrastructure, accessible only to authorised researchers. The resulting dataset will be used to develop and validate AI models that can objectively estimate pain intensity from audiovisual data, supporting more accurate and continuous pain monitoring in clinical care.
Interventions
OTHERAudiovisual and questionnaire-based data collection (once)
Participants will be video recorded for up to 60 seconds from the shoulders up while reading a short sentence and describing their pain or daily experience. They will also complete a brief questionnaire about their mood and how they express pain. No drugs, medical devices, or clinical treatments are used. This non-invasive data collection is performed once for participants without pain.
OTHERAudiovisual and questionnaire-based data collection (consecutive)
Participants will be video recorded for up to 60 seconds from the shoulders up while reading a short sentence and describing their pain or daily experience. They will also complete a brief questionnaire about their mood and how they express pain. No drugs, medical devices, or clinical treatments are used. This non-invasive data collection is performed on several consecutive days during admission for participants with pain.
Sponsors
Dr. Mark Mulder
Delft University of Technology
Study design
Observational model
CASE_CONTROL
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Adult patients (≥18 years) * Diagnosed with cancer, active * Able to communicate verbally in Dutch or English * Able to provide written informed consent. * Pain group specific: * Experiencing pain related to cancer * Admitted to the hospital due to pain * Control group specific: * Not experiencing pain (NRS = 0) * Admitted to the hospital for chemotherapy
Exclusion criteria
* Cognitive, physical, or medical limitations that prevent participation in the audiovisual recording sessions or affect facial expressions or voice (e.g. facial paralysis, tracheostomy, severe speech impairment). * Critical illness or end-of-life care where participation would impose an additional burden. * Experiencing pain not associated with cancer * Infectious isolation precautions that prevent safe data collection
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Audiovisual recordings of facial and vocal pain expression | Up to 2 weeks | The main endpoint consists of audiovisual recordings of each patient's face (from the shoulders up) and voice while the patient reads a short standardised sentence and describes their pain experience. Recordings (maximum 60 seconds after filtering for relevant parts) will be stored digitally and used as input data for the training and validation of an AI-based pain assessment algorithm. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Clinical and patient-reported parameters from electronic health records | At each recording session | Secondary data used to label the audiovisual recordings and provide context for model development, including patient-reported pain intensity (NRS 0-10), patient-reported change in pain (increasing, decreasing, no change), patient description of pain (short annotations), quality of life (if reported), observer-reported pain intensity (NRS 0-10), analgesics in use (drug name, dose, frequency), cancer tumour type, stage, Karnofsky score, and current treatment, demographics: age, sex, height, weight |
| Patient-reported mood and pain questionnaire outcomes | During each recording session | Assessment of patients' self-reported mood and perceived pain expression using a short questionnaire administered alongside audiovisual recordings. The questionnaire evaluates: current mood state, perceived intensity and expressiveness of pain, ease or difficulty in expressing pain verbally and non-verbally. Responses will be used to explore associations between subjective reports and multimodal AI-derived pain indicators. |
Countries
Netherlands
Contacts
CONTACTMarsha Kamsteeg, MSc
PRINCIPAL_INVESTIGATORMark Mulder, dr.ir.
Erasmus University Medical Centre
Outcome results
None listed