Bone texture analysis for prediction of incident radiographic hip osteoarthritis using machine learning: data from the Cohort Hip and Cohort Knee (CHECK) study

Abstract

Our aim was to assess the ability of radiography-based bone texture parameters in proximal femur and acetabulum to predict incident radiographic hip osteoarthritis (rHOA) over a 10 years period. Pelvic radiographs from CHECK (Cohort Hip and Cohort Knee) at baseline (987 hips) were analyzed for bone texture using fractal signature analysis in proximal femur and acetabulum. Elastic net (machine learning) was used to predict the incidence of rHOA (Kellgren-Lawrence grade (KL) ≥ 2 or total hip replacement (THR)), joint space narrowing score (JSN, range 0-3), and osteophyte score (OST, range 0-3) after 10 years. Performance of prediction models was assessed using the area under the receiver operating characteristic curve (ROC AUC). Of the 987 hips without rHOA at baseline, 435 (44%) had rHOA at 10-year follow-up. Of the 667 hips with JSN grade 0 at baseline, 471 (71%) had JSN grade ≥ 1 at 10-year follow-up. Of the 613 hips with OST grade 0 at baseline, 526 (86%) had OST grade ≥ 1 at 10-year follow-up. AUCs for the models including age, gender, and body mass index to predict incident rHOA, JSN, and OST were 0.59, 0.54, and 0.51, respectively. The inclusion of bone texture parameters in the models improved the prediction of incident rHOA (ROC AUC 0.66 and 0.71 when baseline KL was also included in the model) and JSN (ROC AUC 0.62), but not incident OST (ROC AUC 0.53). Bone texture analysis provides additional information for predicting incident rHOA or THR over 10 years.

Constitutional morphological features and risk of hip osteoarthritis: a case–control study using standard radiographs

“…This suggests that local morphological risk factors may contribute more than systemic factors to development of hip OA. The results align with the literature for incidence and progression of hip OA[42,43] and may be explained by shared single nucleotide polymorphisms (SNPs) between OA and hip shape [44,45]. There are several caveats to this study.…”

Radiomics for classification of bone mineral loss: A machine learning study

“…In addition, as stated by Bolotin et al ''the DXA-derived BMD value does not correctly represent the areal density of bone mineral material, as it is contaminated by sizable, unavoidable, inextricable, independent soft tissue contributions'' [28]. On the other hand, radiomic texture features extracted from medical bone images contain meaningful information, which could be used to predict bone mineral disorders more accurately [29]. In addition, previous studies indicated that the combination of texture analysis and bone mineral density improves the prediction of fracture load in human femurs [30].…”

Osteoarthritis year in review 2019: imaging

“…Hivasniemi et al created risk models for predicting hip OA incidence over 10 years using clinical findings and bone texture analysis of pelvic X-rays in 1243 subjects from the Cohort Hip and Knee (CHECK) study. A clinical model including age, gender, and BMI had an AUC of 0.59, while a combined clinical and X-ray model which also included KL grade and bone texture variables had an AUC of 0.71 53 .…”