Persistent Homology of Tumor CT Scans is Associated with Survival In Lung Cancer

Abstract

Radiomics, the objective study of non-visual features in clinical imaging, has been useful in informing decisions in clinical oncology. However, radiomics currently lacks the ability to characterize the overall structure of the data. This field may benefit by incorporating persistent homology, a popular new algorithm that analyzes whole data structure. We hypothesized that persistent homology could be applied to lung tumor scans and predict clinical variables. We obtained computed tomography lung scans (n = 565) from the NSCLC-Radiomics and NSCLC-Radiogenomics datasets in The Cancer Imaging Archive. Segmentation data was used to create a cubical region centered on the primary tumor in each scan. For each scan, a cubical complex filtration based on Hounsfield units was generated. We created a feature curve that plotted the number of 0 dimensional topological features against each Hounsfield unit. The curve’s first moment of the distribution was utilized as a summary statistic to predict survival in a Cox proportional hazards model. The first moment of the distribution is equivalent to the area under the curve of our topological feature curves (AUC). The Kruskal-Wallis H Test and a post-hoc Dunn’s test with Bonferroni correction were used to test AUC differences among survival quartiles. After controlling for tumor image size, age, and stage, AUC, was associated with poorer survival (HR = 1.118; 95% CI = 1.026-1.218; p = 0.01). AUC was significantly higher for patients in the lowest survival quartile compared to the highest survival quartile (p < 0.001). We have shown that persistent homology can generate useful clinical correlates from tumor CT scans. Our 0-dimensional topological feature curve statistic predicts survival in lung cancer patients. This novel statistic may be used in tandem with standard radiomics variables to better inform clinical oncology decisions.

Visualization of exhaled breath metabolites reveals distinct diagnostic signatures for acute cardiorespiratory breathlessness

“…TDA is a well-established data analytic technique for unbiased data-driven discovery–based phenotyping ( 7 ). TDA has proven to be a powerful tool, yielding critical insights in the prognostic phenotyping ( 8 ), cancer imaging biomarker stratification ( 9 ), disease classification using pathology biomarkers ( 10 ), and omics-based cancer phenotyping ( 11 ). Several publications have reported the use of TDA in the metabolomics field, for example, unbiased lipid phenotyping of lung epithelial lining fluid ( 12 ).…”

A Topological Data Analytic Approach for Discovering Biophysical Signatures in Protein Dynamics

“…Another natural extension would be to apply the SINATRA Pro pipeline to other data types used to study variation in 3D protein structures such as cryogenic electron microscopy (cryo-EM), nuclear magnetic resonance (NMR) ensembles, and X-ray crystallography (i.e., electron density) data. Previous work has already shown that topological characteristics computed on tumors from magnetic resonance images (MRIs) have the potential to be powerful predictors of survival times for patients with glioblastoma multiforme (GBM) [17,51] and other cancer subtypes [52][53][54]; however, it has also been noted that the efficacy of current topological summaries decreases when heterogeneity between two phenotypic classes is driven by minute differences [13]. For example, cryo-EM images can look quick similar even for two proteins harboring different mutations.…”

gtexture: Haralick texture analysis for graphs and its application to biological networks

“…al., describing the co-occurrence of nuclear features in physical cell neighborhoods (Saito et al 2016). Recent interdisciplinary work has successfully extended different graph-based topological analyses to image derived point clouds and more recently to images themselves, including the use of cubical complexes to derive prognostic topological features from medical images (Lawson et al 2019;Hajij, Zamzmi, and Batayneh 2021;Somasundaram, Litzler, et al 2021;Somasundaram, Wadhwa, et al 2022).…”