Risk Prediction in Cancer Imaging Using Enriched Radiomics Features

TL;DR

The proposed framework leverages enhancement pattern mapping images to provide an automated and robust radiomics representation that captures intratumoral heterogeneity through pixel-level functional information and can potentially replace classical radiomics analysis and be used for imaging biomarkers in cross-sectional and in longitudinal cancer imaging studies.

Abstract

Background: We aim to develop enriched radiomics features that integrate classical structural radiomics with novel functional radiomics derived from liver MRI for diagnosis and risk stratification in liver cancer. The proposed framework leverages enhancement pattern mapping (EPM) images to provide an automated and robust radiomics representation that captures intratumoral heterogeneity through pixel-level functional information. Methods: Pixel-wise EPM data reflecting blood perfusion were extracted from T1-weighted MRI scans. Classical structural radiomics features were extracted via existing software such as PyRadiomics. In addition, empirical quantiles of EPM values over all pixels within the image, and then smoothed using suitable basis. The smoothed quantiles, along with the classical structural quantiles, are used as functional radiomics features for diagnostic classification and tumor grade stratification, using L1-penalized logistic model that automatically downweights the contribution of the irrelevant features. Further, we conducted longitudinal analyses using Bayesian tensor response regression, which enables spatial smoothing and parsimonious modeling of temporally evolving imaging patterns. Results: The enriched radiomics features illustrate higher diagnostic classification performance (AUC=0.96, sensitivity> 0.8) and superior tumor grade stratification accuracy (AUC=0.87, sensitivity=0.8) compared to alternate radiomics features. Moreover, we find that the proportion of lesion pixels with significant reduction in EPM values over time is considerably higher (median = 0.12) in aggressive lesions versus stable or mildly aggressive lesions (median = 0.025). Conclusion: The enriched novel radiomics features can potentially replace classical radiomics analysis and be used for imaging biomarkers in cross-sectional and in longitudinal cancer imaging studies.

Figures (4)

• Figure S1: A schematic of the proposed analytical pipeline. The leftmost panel captures the 5 different phases of the T1-MRI liver scan that are all registered to a common template for a given subject. Subsequently the EPM images are computed from these T1-MRI scans, and then empirical quantile distributions followed by smoothed quantile distributions (based on quantlet basis expansions) were computed. The arterial phase of T1-MRI is used for segmentation of tumor and also used to compute the structural radiomics features. These two types of features, along with demographics are used for three classification tasks that include diagnosis and lesion score stratification.
• Figure S2: Visualizations of lesion EPM images with segmentations for LIRADS scores 2 to 5 and healthy control. Liver masks are shown in yellow; lesion masks in red and surrounding peri-lesional areas are shaded in blue. Mean empirical quantiles are shown along with 95% confidence bands for each LIRADS score and controls. For cases, quantiles are shown across all lesional (red) and peri-lesional (blue) areas, and for controls, quantiles are shown across all liver masks (yellow). Distributions of log-lesion volumes are also shown for each LIRADS score. Observed lesion volumes ranged from 0.17-0.950 cm$^3$, with approximate diameters of 0.70-12.2 cm.
• Figure S3: Example of subject-specific liver map showing regions of significant longitudinal EPM changes around lesion, as determined via longitudinal BTRR. White indicates presence of lesion, red indicates positive (+) EPM change, and green indicates negative (-) EPM change.
• Figure S4: Boxplots displaying proportions of significant positive (left) and negative (right) EPM changes within lesion ROI, stratified by change in lesion score. For right-hand plot, the medians are $0.0240$ for no change, $0.00278$ for change of $+1$, and $0.113$ for change of $+2$; the proportions of lesions with $<5\%$ negative significant changes are $0.667$ for no change, $0.727$ for $+1$ change, and $0.333$ for $+2$ change.