Pain ASsessment in CAncer Patients by Machine LEarning (PASCALE)

Официальное название

Home-Based Telemedicine for Automatic Pain Assessment in Cancer Patients: Dataset Creation and Development of Machine Learning Algorithms

Обзор

In cancer patients, the integration between anticancer therapies and palliative care is of fundamental importance. In this context, telemedicine can improve the quality of life (QoL) of chronic patients through self-management and remote monitoring solutions. This approach can favor the effectiveness of the treatment and therapeutic adherence. Of note, telemedicine can also be applied to the management of cancer pain. In the advanced stages of cancer disease, pain is one of the most obvious and most disabling symptoms. Consequently, proper pain management has a significant impact on the QoL, the ability to withstand treatment, and the recovery of patients. On the other hand, given the complexity of cancer pain, the main obstacle to its proper management is the lack of adequate measurement methods. Although in recent years a great deal of effort has been made in the direction of automatic pain assessment, both concerning the creation of datasets and the development of classification algorithms, the literature is lacking regarding the automatic measurement of pain in the setting of cancer patients. Observation by experienced clinical staff and self-assessment by patients could be useful for obtaining the ground truth and, in turn, for training automatic pain recognition systems.

Подробное описание

For the entire duration of the study, patients will remain under the care of the Early Palliative Care and Simultaneous Care Outpatient team of the Istituto Nazionale Tumori, Fondazione Pascale, at home. Pain and other symptoms will be managed according to the good clinical practice and patients will receive assistance in agreement to the routine medical care.

The following devices will be used:

1. Software 2. Instrumentation 3. Clinical Assessment Tools: European Organisation for Research and Treatment of Cancer Quality-of-life Questionnaire Core 30 (EORTC QLQ-C30), Daily Pain Diary, 0-10 numeric rating scale (NRS).

The project will be divided into three main Work Packages (WPs), dedicated respectively to the creation of the IT infrastructure to support acquisitions (WP1), the patient data collection campaign (WP2), and the development of machine learning algorithms for automatic pain recognition (WP3). The application of the devices and verification of correct functioning will be carried out at the patient's home by the IT staff involved in the study.

WP1 - The system consists of three main components: the server, with the attached database, the application for mobile devices, also responsible for managing data acquisition from physiological signal acquisition devices, and the desktop application, used by the clinical staff to monitor the progress of data collection.

The mobile application will have the role of interfacing directly with the patient and acquiring biometric data from wearable devices. Specifically, the following signals will be acquired: heart rate, body temperature, non-invasive blood pressure, and galvanic skin response (GSR). The heart rate will be obtained through a wearable device (Garmin Vivosmart 4) while the body temperature, the non-invasive blood pressure, and the GSR will be acquired by an external device (a BITalino platform).To further validate the accuracy of the algorithm that will deal with pain detection, patients will also be given a QoL questionnaire (EORTC QLQ-C30).

In order to acquire the ground truth of the data, the patient will be asked to provide feedback on the level of pain, both at certain intervals of time during the day, and in case of acute pain episodes. This feedback can be based on NRS and multimedia strategies (e.g., videos). Patients will fill out a daily pain diary.

WP2 - The campaign will include a preliminary acquisition phase aimed at testing the IT infrastructure. For obtaining an adequate inter-subject and intra-subject variability, it will be necessary to enroll at least 40 patients, acquiring data for 10-14 days. Thus, the data collection campaign will be conducted for about 6 months. Each subject will use the mobile application and sensors for 2 weeks. Data will be acquired using simultaneously data collection bundles (application, sensors, and any mobile device). Upon enrolment and at the end, EORTC QLQ-C30 will be administered.

WP3 - The objective is the development of algorithms able to predict the level of pain perceived by the patient. Having a considerable amount of labelled data available, the system will learn from the examples.

Первичные конечные точки

• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]
• To develop a machine learning algorithm useful for predicting the level of pain in cancer patients. A database containing clinical data and pain features will be obtained. [Срок оценки: Whenever the patient has a worsening of his/her pain, up to 2 weeks]

Вторичные конечные точки (1)

• Patients' quality of life assessed by the EORTC QLQ-C30 questionnaire. [Срок оценки: At the beginning and at the end of the observation, up to 2 weeks]

Критерии участия

Критерии включения

• Patients aged > 18 years
• Home care patients diagnosed with advanced cancer disease and life expectancy ≤ 1 year
• Patients receiving treatment for cancer pain
• Patients who have given their consent

Критерии исключения

• Patients aged < 18 years
• Willingness to sign the informed consent form (unable to read or write)
• Cognitive deficit (e.g. Alzheimer disease or senile dementia)

Критерии приведены из реестра в оригинале (на английском). Окончательную оценку соответствия проводит исследовательский центр.

Здоровые добровольцы: Нет

Дизайн исследования

Модель наблюдения

Когортное

Центры проведения

Италия · 2 центра

• National Cancer Institute of Naples — Naples
• A.O.U. Federico II — Naples

Публикации

• Reis-Pina P, Lawlor PG, Barbosa A. Adequacy of cancer-related pain management and predictors of undertreatment at referral to a pain clinic. J Pain Res. 2017 Aug 31;10:2097-2107. doi: 10.2147/JPR.S139715. eCollection 2017. PMID 28919809
• Adamse C, Dekker-Van Weering MG, van Etten-Jamaludin FS, Stuiver MM. The effectiveness of exercise-based telemedicine on pain, physical activity and quality of life in the treatment of chronic pain: A systematic review. J Telemed Telecare. 2018 Sep;24(8):511-526. doi: 10.1177/1357633X17716576. Epub 2017 Jul 11. PMID 28696152
• Sirintrapun SJ, Lopez AM. Telemedicine in Cancer Care. Am Soc Clin Oncol Educ Book. 2018 May 23;38:540-545. doi: 10.1200/EDBK_200141. PMID 30231354
• Aung MSH, Kaltwang S, Romera-Paredes B, Martinez B, Singh A, Cella M, Valstar M, Meng H, Kemp A, Shafizadeh M, Elkins AC, Kanakam N, de Rothschild A, Tyler N, Watson PJ, de C Williams AC, Pantic M, Bianchi-Berthouze N. The Automatic Detection of Chronic Pain-Related Expression: Requirements, Challenges and the Multimodal EmoPain Dataset. IEEE Trans Affect Comput. 2016 Oct-Dec;7(4):435-451. doi: 10 PMID 30906508
• Gruss S, Geiger M, Werner P, Wilhelm O, Traue HC, Al-Hamadi A, Walter S. Multi-Modal Signals for Analyzing Pain Responses to Thermal and Electrical Stimuli. J Vis Exp. 2019 Apr 5;(146). doi: 10.3791/59057. PMID 31009005
• Pfeifer AC, Uddin R, Schroder-Pfeifer P, Holl F, Swoboda W, Schiltenwolf M. Mobile Application-Based Interventions for Chronic Pain Patients: A Systematic Review and Meta-Analysis of Effectiveness. J Clin Med. 2020 Nov 5;9(11):3557. doi: 10.3390/jcm9113557. PMID 33167300
• Cuomo A, Bimonte S, Forte CA, Botti G, Cascella M. Multimodal approaches and tailored therapies for pain management: the trolley analgesic model. J Pain Res. 2019 Feb 19;12:711-714. doi: 10.2147/JPR.S178910. eCollection 2019. PMID 30863143
• Rashidi P, Edwards DA, Tighe PJ. Primer on machine learning: utilization of large data set analyses to individualize pain management. Curr Opin Anaesthesiol. 2019 Oct;32(5):653-660. doi: 10.1097/ACO.0000000000000779. PMID 31408024

Идентификаторы

NCT: NCT04726228 · 41/20 oss

Первоисточники (государственные реестры)

• ClinicalTrials.gov (NIH, США)

Открыть это исследование на ClinicalTrials.gov ↗