Proteome-assisted multi-feature discriminant modelling of chronic post-surgical pain in female patients: a proof-of-concept study

Pain after surgery is of major perioperative concern because it is associated with substantial complications, including chronic post-surgical pain (CPSP) development. CPSP incidence is high, and its accurate prediction or prevention have been so far unsuccessful. Based on experimental studies in healthy volunteers, we hypothesised that pre-surgical plasma proteome data and multi-feature discriminant modelling can improve the accuracy of predicting susceptibility versus resilience to CPSP. To test this, we conducted a proof-of-concept case-control study: Thirty-two female surgical patients undergoing either open hysterectomy or thoracotomy were stratified by CPSP presence three months post-surgery. Preoperative blood samples were analysed by unbiased deep proteomics to identify plasma proteins associated with phenotypes of CPSP susceptibility vs resilience. These were then integrated with pre-surgical psycho-social factors to develop discriminative models. Among 684 identified plasma proteins, 106 turned out to be discriminatory for the CPSP-susceptible and 104 for the CPSP-resilient phenotype. At postoperative month 3, physical dysfunction, anxiety, and depression were significantly higher in CPSP-susceptible patients. The addition of proteomic data to the model improved the accuracy of phenotype discrimination in internal cross-validation when compared to psychosocial and neurocognitive factors alone. Protein network analysis was consistent with the hypothesis that pre-surgical immune and complement activation may be associated with CPSP risk. Furthermore, computational drug repositioning suggested candidate molecular targets potentially relevant to modulating the CPSP risk profile. Overall, our results illustrate the feasibility and potential utility of multimodal datasets to discriminate between CPSP phenotypes. When combined with network-based analysis and drug-repositioning this approach may open new avenues for identifying drug targets and personalized mitigation of CPSP in the future.