Classification of painful or painless diabetic peripheral neuropathy and identification of the most powerful predictors using machine learning models in large cross-sectional cohorts.

BACKGROUND: To improve the treatment of painful Diabetic Peripheral Neuropathy (DPN) and associated co-morbidities, a better understanding of the pathophysiology and risk factors for painful DPN is required. Using harmonised cohorts (N = 1230) we have built models that classify painful versus painless DPN using quality of life (EQ5D), lifestyle (smoking, alcohol consumption), demographics (age, gender), personality and psychology traits (anxiety, depression, personality traits), biochemical (HbA1c) and clinical variables (BMI, hospital stay and trauma at young age) as predictors. METHODS: The Random Forest, Adaptive Regression Splines and Naive Bayes machine learning models were trained for classifying painful/painless DPN. Their performance was estimated using cross-validation in large cross-sectional cohorts (N = 935) and externally validated in a large population-based cohort (N = 295). Variables were ranked for importance using model specific metrics and marginal effects of predictors were aggregated and assessed at the global level. Model selection was carried out using the Mathews Correlation Coefficient (MCC) and model performance was quantified in the validation set using MCC, the area under the precision/recall curve (AUPRC) and accuracy. RESULTS: Random Forest (MCC = 0.28, AUPRC = 0.76) and Adaptive Regression Splines (MCC = 0.29, AUPRC = 0.77) were the best performing models and showed the smallest reduction in performance between the training and validation dataset. EQ5D index, the 10-item personality dimensions, HbA1c, Depression and Anxiety t-scores, age and Body Mass Index were consistently amongst the most powerful predictors in classifying painful vs painless DPN. CONCLUSIONS: Machine learning models trained on large cross-sectional cohorts were able to accurately classify painful or painless DPN on an independent population-based dataset. Painful DPN is associated with more depression, anxiety and certain personality traits. It is also associated with poorer self-reported quality of life, younger age, poor glucose control and high Body Mass Index (BMI). The models showed good performance in realistic conditions in the presence of missing values and noisy datasets. These models can be used either in the clinical context to assist patient stratification based on the risk of painful DPN or return broad risk categories based on user input. Model's performance and calibration suggest that in both cases they could potentially improve diagnosis and outcomes by changing modifiable factors like BMI and HbA1c control and institute earlier preventive or supportive measures like psychological interventions.

What is associated with painful polyneuropathy? A cross-sectional analysis of symptoms and signs in patients with painful and painless polyneuropathy.

ABSTRACT: It is still unclear how and why some patients develop painful and others painless polyneuropathy. The aim of this study was to identify multiple factors associated with painful polyneuropathies (NeuP). A total of 1181 patients of the multicenter DOLORISK database with painful (probable or definite NeuP) or painless (unlikely NeuP) probable or confirmed neuropathy were investigated clinically, with questionnaires and quantitative sensory testing. Multivariate logistic regression including all variables (demographics, medical history, psychological symptoms, personality items, pain-related worrying, life-style factors, as well as results from clinical examination and quantitative sensory testing) and machine learning was used for the identification of predictors and final risk prediction of painful neuropathy. Multivariate logistic regression demonstrated that severity and idiopathic etiology of neuropathy, presence of chronic pain in family, Patient-Reported Outcomes Measurement Information System Fatigue and Depression T-Score, as well as Pain Catastrophizing Scale total score are the most important features associated with the presence of pain in neuropathy. Machine learning (random forest) identified the same variables. Multivariate logistic regression archived an accuracy above 78%, random forest of 76%; thus, almost 4 out of 5 subjects can be classified correctly. This multicenter analysis shows that pain-related worrying, emotional well-being, and clinical phenotype are factors associated with painful (vs painless) neuropathy. Results may help in the future to identify patients at risk of developing painful neuropathy and identify consequences of pain in longitudinal studies.

Genetic associations of neuropathic pain and sensory profile in a deeply phenotyped neuropathy cohort.

ABSTRACT: We aimed to investigate the genetic associations of neuropathic pain in a deeply phenotyped cohort. Participants with neuropathic pain were cases and compared with those exposed to injury or disease but without neuropathic pain as control subjects. Diabetic polyneuropathy was the most common aetiology of neuropathic pain. A standardised quantitative sensory testing protocol was used to categorize participants based on sensory profile. We performed genome-wide association study, and in a subset of participants, we undertook whole-exome sequencing targeting analyses of 45 known pain-related genes. In the genome-wide association study of diabetic neuropathy (N = 1541), a top significant association was found at the KCNT2 locus linked with pain intensity (rs114159097, P = 3.55 × 10-8). Gene-based analysis revealed significant associations between LHX8 and TCF7L2 and neuropathic pain. Polygenic risk score for depression was associated with neuropathic pain in all participants. Polygenic risk score for C-reactive protein showed a positive association, while that for fasting insulin showed a negative association with neuropathic pain, in individuals with diabetic polyneuropathy. Gene burden analysis of candidate pain genes supported significant associations between rare variants in SCN9A and OPRM1 and neuropathic pain. Comparison of individuals with the "irritable" nociceptor profile to those with a "nonirritable" nociceptor profile identified a significantly associated variant (rs72669682, P = 4.39 × 10-8) within the ANK2 gene. Our study on a deeply phenotyped cohort with neuropathic pain has confirmed genetic associations with the known pain-related genes KCNT2, OPRM1, and SCN9A and identified novel associations with LHX8 and ANK2, genes not previously linked to pain and sensory profiles, respectively.