Our research explores noninvasive pulmonary nodule assessment. Through multi-varietal spectrum analysis of CT images, we aim to answer why the multi-varietal characteristics can reliably differentiate between benign and the malignant nodules, potentially reducing the need for invasive diagnostic procedures. Recent advances include AI-based radial mix approaches for nodule characterization, but this typically focused on either pathological image or morphological analysis alone.
Our multifractal spectrum analysis bridges imaging and the pathological findings. We demonstrated that pulmonary nodule exhibit distinct multifractal spectrum at different stage, with later-stage nodules showing wider scale range and higher extreme points, enabling quantitative differentiation of malignancy. Our protocol offers noninvasive quantitative assessment of nodule malignancy through simultaneous analysis of morphological characteristics and tissue heterogeneity, reducing reliance on invasive biopsies.
These findings enable more precise staging of pulmonary nodules and early detection of malignancy, potentially improving clinical decision making and patient outcomes in pulmonary oncology. To begin, obtain the patient's computed tomography or CT scan data as DICOM files, and generate a 3D volume matrix in MATLAB. Visualize the image sequence using MATLAB slice viewer function.
Use the scroll bar at the bottom of the graphical user interface to browse through different slices in the CT sequence. Identify the presence of malignant pulmonary nodules in the lungs at the appropriate frame. Locate and use the icons for zooming in, zooming out, and returning to the global view.
Activate the data tip icon to mark the pixel coordinates of the selected region. Right click the color bar to use the common gray color map, and select the appropriate option. If the filter effect is not satisfactory, use the left mouse button to drag up and down in the middle of the figure to adjust the window level.
Drag left and right to adjust the window width, and the corresponding accurate filtering range will be displayed on the color bar. Identify the pixel coordinates of the nodule after marking its edges, and note the X and Y coordinates displayed in the data tip popup. After defining the region of interest, or ROI, use the MATLAB command to create a 3D surface plot.
Observe the 3D grayscale intensity of the pulmonary nodule. Locate and use the zooming, unzooming, rotating, and restore view tools for a detailed inspection. Call the function pic_size, fractal_dimension equals PN_fractal_feature M, with the previously obtained M matrix as input.
This will calculate the fractal dimensions at different scales. Visualize the multifractal spectrum of the pulmonary nodule using the code. Calculate the multifractal spectrum for another benign pulmonary nodule using the same steps.
Plot the results in the same coordinate system using a different color for comparison, generating the multifractal spectrum comparison between different pulmonary nodules. Finally, use the data tip tool to mark the coordinates of key extreme points for precise comparison of different benign and malignant pulmonary nodules. The fractal dimension of malignant nodules displayed a multifractal spectrum with higher extreme points and broader scale ranges compared to benign nodules.
Early-stage lung adenocarcinoma nodules exhibited a narrower fractal scale range and lower extreme points in the multifractal spectrum than late-stage nodules.
