Machine-learned identification of psychological subgroups with relation to pain interference in patients after breast cancer treatments

Abstract

Background: Persistent pain in breast cancer survivors is common. Psychological and sleep-related factors modulate perception, interpretation and coping with pain and may contribute to the clinical phenotype. The present analysis pursued the hypothesis that breast cancer survivors form subgroups, based on psychological and sleep-related parameters that are relevant to the impact of pain on the patients' life. Methods: We analysed 337 women treated for breast cancer, in whom psychological and sleep-related parameters as well as parameters related to pain intensity and interference had been acquired. Data were analysed by using supervised and unsupervised machine-learning techniques (i) to detect patient subgroups based on the pattern of psychological or sleep-related parameters, (ii) to interpret the detected cluster structure and (iii) to relate this data structure to pain interference and impact on life. Results: Artificial intelligence-based detection of data structure, implemented as self-organizing neuronal maps, identified two different clusters of patients. A smaller cluster (11.5% of the patients) had comparatively lower resilience, more depressive symptoms and lower extraversion than the other patients. In these patients, life-satisfaction, mood, and life in general were comparatively more impeded by persistent pain. Conclusions: The results support the initial hypothesis that psychological and sleep-related parameter patterns are meaningful for subgrouping patients with respect to how persistent pain after breast cancer treatments interferes with their life. This indicates that management of pain should address more complex features than just pain intensity. Artificial intelligence is a useful tool in the identification of subgroups of patients based on psychological factors.

Resilience, pain, and health-related quality of life in gynecological patients undergoing surgery for benign and malignant conditions: a 12-month follow-up study

“…Thus, in contrast to those before surgery, patients with malignant diseases had clinically importantly lower HRQoL than those with benign diseases at 12 months after surgery. Consistent with previous studies [ 7 , 11 , 12 , 15 , 28 ], higher resilience had a moderate to strong correlation with better HRQoL, higher life satisfaction, and lower anxiety and depression. Anxiety decreased in both groups during the follow-up, but depression was rare before and after surgery.…”

“…The novelty of this study was that we evaluated resilience for the first time in gynecology patients undergoing surgery for benign and malignant diagnoses. In the present study, resilience was moderately high, and patients with malignant diseases had comparable resilience scores to previous studies in gynecological cancer [ 12 , 15 ]. In the present study, HRQoL decreased in patients with malignant diseases and increased in those with benign diseases.…”

“…Most people are sufficiently resilient, but resilience skills in those with less adaptive capacities can be learned and enhanced [ 9 , 10 ]. Resilience has gained increasing research interest in the medical setting due to its associations with a better recovery, enhanced quality of life [ 11 , 12 ], and functional status [ 13 ] and fewer psychological symptoms [ 14 , 15 ] and less pain interference after surgery [ 15 , 16 ]. A meta-analysis showed a cautious positive effect at improving resilience and other health outcomes [ 17 ].…”

Phenomapping of Patients with Primary Breast Cancer Using Machine Learning-Based Unsupervised Cluster Analysis

“…Like K-means, it is designed for quantitative variables and the choice of (weighted) Euclidean distance as a dissimilarity measure. SOM methods have been used in a few medical studies [ 28 , 29 , 30 ].…”

“…The parameters of the distributions are generally unknown, and they have to be estimated. A maximum likelihood estimator (MLE) is usually employed to estimate the parameters of the distribution, using the expectation-maximisation (EM) algorithm [ 30 , 31 ]. Most applications assume that all components have the same parameters of the Gaussians, i.e., the same covariance matrix and the same mean for all the K distributions.…”