Fig 1.
Main features of dynamic hypothesis.
Schematic representation of initial infection (A, white arrow) lesions’ growth (A–F), endogenous reinfection (C–E, yellow arrows) and coalescence of neighbouring lesions (F).
Fig 2.
Summary of experimental results.
Left: CT image reconstruction of a minipig’s pulmonary surface. Right: 3D representation of location and size of all minipig lesions; each colour is for a different minipig (red, magenta, blue, yellow, and green).
Fig 3.
Normalized contour lines obtained from CT-scan images.
Contour lines were obtained from CT-scan images and used for a 3D computational reconstruction of the pulmonary surface. Trachea division point (carina) is marked for purposes of reconstruction. (A) Sagittal plane CT image. (B) Coronal plane CT image. (C) Axial plane CT image. (D) Sagittal plane outline reconstruction. (E) Coronal plane outline reconstruction. (F) Axial plane outline reconstruction.
Fig 4.
Bifurcation of 0 branch into two (1, 2) daughter branches. The cabal ratio for branch 1 is: q1 = 0.6. Length is 3 times the diameter of each branch as may be seen in Eq 2. Diameter relations are obtained from Eq 4, as d1 = 0.84·d0 and d2 = 0.74·d0. Angular values are computed using Eq 5, as ϕ1 = 32° and ϕ2 = 43°.
Table 1.
Summary of the model for building the computational bronchial tree of each minipig.
Fig 5.
Reinfection probability for a lesion with rmax = 1 mm and ρ = 0.10 day-1. In black, original Bubble model [14]; in blue, the updated model considering an exponential decrease during the control phase. The area under the two curves is equivalent.
Fig 6.
(A) Mean distance where lesions appear as a function of the dispersion parameter, β ∈ [0, 0.3] mm-1. Due to the quantization of the space, for β > 0.3 mm-1 the probability that the lesion appears at the nearest terminal is very high and the value is a constant. (B) Density of probability of new lesions spreading with β = 0.08 mm-1.
Table 2.
Outcome variables of the model.
Table 3.
Values of proportionality constant, A(OC,S), for different outcomes and sets of input parameters.
Table 4.
Sensitivity of the error functions (NLE, DE, and SE) to the three parameters explored (β, ρ and rmax).
Fig 7.
Computational minipig bronchial tree.
It is represented as a tubular structure. Each conducting airway is represented by an empty cylinder. Three axial planes are shown: (A) Sagittal plane, (B) Coronal plane, and (C) Axial plane.
Fig 8.
Computational bronchial tree analysis.
(A) Angular distribution of bronchial tree bifurcations. The angle is measured between the mother branch and the new one. (B) Generation distribution; in red, the terminal branches, in blue the non-terminal, and in pink the intersection between them. (C) Diameter distribution of the terminal branches. (D) Terminal distribution (blue) and density (black dots) along X coordinate. (E) Terminal distribution (blue) and density (black dots) along Y coordinate. (F) Terminal distribution (blue) and density (black dots) along Z coordinate.
Fig 9.
Comparison of experimental and computationally obtained distributions of lesion location and size.
Computational distributions were obtained considering as initial infection one lesion in the mass centre of the observed lesions. The set of parameters used is: ρ = 0.13 day-1, β = 0.08 mm-1, and rmax = 0.68 mm. (A) Coordinate X (Left—Right) histogram comparison. (B) Coordinate Y (Anterior–Posterior) histogram comparison. (C) Coordinate Z (Vertical) histogram comparison. (D) Diameter distribution histogram comparison.
Table 5.
Sensitivity analysis for the set of parameters: S = {ρ, β, rmax} = {0.12 day-1, 0.08 mm-1, 0.68 mm}.
Table 6.
Initial configurations explored with the model using experimental data.
Fig 10.
Active Disease Index for different sets of parameters.
Exploration of parameters space (rmax, β and ρ) to see the fraction of in silico experiments that present an active TB disease. The colour is proportional to this frequency; green colour means most of the cases remained latent, red colour means that most of the cases derived into an active disease, and intermediate colours mean that both dynamics are possible.