Optimal biomechanical parameters for measuring sclerotic chronic graft-versus-host disease

Abstract

Skin biomechanical parameters (dynamic stiffness, frequency, relaxation time, creep, and decrement) measured using a myotonometer (MyotonPRO) could inform management of sclerotic disease. To determine which biomechanical parameter(s) can accurately differentiate sclerotic chronic graft-versus-host disease (cGVHD) patients from post-hematopoietic cell transplant (post-HCT) controls, 15 sclerotic cGVHD patients and 11 post-HCT controls were measured with the myotonometer on 18 anatomic sites. Logistic regression and two machine learning algorithms, LASSO regression and random forest, were developed to classify subjects. In univariable analysis, frequency had the highest overfit-corrected area under the receiver operating characteristic curve (AUC 0.91). Backward stepwise selection and random forest machine learning identified frequency and relaxation time as the optimal parameters for differentiating sclerotic cGVHD patients from post-HCT controls. LASSO regression selected the combination of frequency and relaxation time (overfit-corrected AUC 0.87). Discriminatory ability was maintained when only the sites accessible while the patient is supine (12 sites) were used. We report the distribution of values for these highly discriminative biomechanical parameters, which could inform assessment of disease severity in future quantitative biomechanical studies of sclerotic cGVHD.

Figure 1

The 18 measured sites. Green circles represent the 12 sites that are measured in the supine position.

Figure 2

Box and whisker plots of the parameter values by sites in patients with sclerotic cGVHD(red)and in post-HCT controls(blue). Sites are displayed in descending order of the frequency AUCs shown in Figure 4. Bilateral sites are displayed next to each other. The bottom and top boundaries of each box represent the 25th and 75th quartiles, respectively; the horizontal line within the box represents the median; and the whiskers represent the minimum and maximum values. AUC, area under the curve; cGVHD, chronic graft-versus-host disease; HCT, hematopoietic cell transplantation; L, left; R, right.

Figure 3

Distributions of biomechanical parameter values for sclerotic cGVHD patients and post-HCT controls. Patients with sclerotic cGVHD (red boxes) had increased stiffness and frequency and decreased relaxation time and creep compared with post-HCT controls (blue boxes). Patients with sclerotic cGVHD and post-HCT controls were compared by Wilcoxon rank-sum test using averages across the 18 sites (solid boxes, n for patients with cGVHD = 13, n for post-HCT control = 11) and averages across the 12 supine sites (diagonal pattern, n for patients with cGVHD = 15, n for post-HCT control = 11); ∗P < 0.05. Frequency and stiffness (top row) values are expected to be higher for sclerotic skin than for healthy skin, whereas relaxation time and creep (bottom row) values are expected to be lower for sclerotic skin. Logarithmic decrement (overall: P = 0.209, supine: P = 0.209) is not shown. cGVHD, chronic graft-versus-host disease; HCT, hematopoietic cell transplantation; Q, quartile.

Figure 4

AUC of univariable analyses of frequency and relaxation time for individual anatomic sites. Frequency = open circles, and relaxation time = open triangles. Averages across the 12 supine sites (red-filled markers, n for patients with cGVHD = 15; n for controls = 11) and across all the 18 measured sites (black-filled markers, n for patients with cGVHD = 13, n for controls = 11) are shown. Markers represent the averages of the left and right anatomic sites’ AUCs. Error bars represent the ranges between the left and right AUCs. AUC, area under the curve.