PURPOSE Improved survival prediction and risk stratification in non–small-cell lung cancer (NSCLC) would lead to better prognosis counseling, adjuvant therapy selection, and clinical trial design. We propose the persistent homology (PHOM) score, the radiomic quantification of solid tumor topology, as a solution. MATERIALS AND METHODS Patients diagnosed with stage I or II NSCLC primarily treated with stereotactic body radiation therapy (SBRT) were selected (N = 554). The PHOM score was calculated for each patient's pretreatment computed tomography scan (October 2008-November 2019). PHOM score, age, sex, stage, Karnofsky Performance Status, Charlson Comorbidity Index, and post-SBRT chemotherapy were predictors in the Cox proportional hazards models for OS and cancer-specific survival. Patients were split into high– and low–PHOM score groups and compared using Kaplan-Meier curves for overall survival (OS) and cumulative incidence curves for cause-specific death. Finally, we generated a validated nomogram to predict OS, which is publicly available at Eashwarsoma.Shinyapps. RESULTS PHOM score was a significant predictor for OS (hazard ratio [HR], 1.17; 95% CI, 1.07 to 1.28) and was the only significant predictor for cancer-specific survival (1.31; 95% CI, 1.11 to 1.56) in the multivariable Cox model. The median survival for the high-PHOM group was 29.2 months (95% CI, 23.6 to 34.3), which was significantly worse compared with the low-PHOM group (45.4 months; 95% CI, 40.1 to 51.8; P < .001). The high-PHOM group had a significantly greater chance of cancer-specific death at post-treatment month 65 (0.244; 95% CI, 0.192 to 0.296) compared with the low-PHOM group (0.171; 95% CI, 0.123 to 0.218; P = .029). CONCLUSION The PHOM score is associated with cancer-specific survival and predictive of OS. Our developed nomogram can be used to inform clinical prognosis and assist in making post-SBRT treatment considerations.

Evolution underpins the survival of a population under environmental pressure. Resistance to treatment commonly arises as a result of such evolution. We analytically examine the addition of frequency-dependent effects on evolutionary outcomes. Through the lens of experimental biology, we frame these interactions as cell-extrinsic, growth rate-modifying, ecological interactions. Additionally, we show the extent to which the presence of these ecological interactions can modify evolutionary trajectories predicted from cell-intrinsic properties alone and show that these interactions can modify evolution in such ways as to mask or mimic or maintain the results of cell-intrinsic fitness advantages. This work has implications for the interpretation and understanding of evolution, a result which may explain an abundance of apparently neutral evolution in cancer systems and similarly heterogeneous populations. In addition, the derivation of an analytical result for stochastic, ecologically dependent evolution paves the way for treatment approaches involving genetic and ecological control.

Introduction: Predicting survival in NSCLC is a goal of clinicians in guiding therapy, trialists in risk stratifying patients, and patients for prognosis counseling. Traditional risk calculators for survival use clinical and histopathologic data to predict survival. We propose incorporating the PHOM (persistent homology) score, the radiomic quantification of solid tumor topology, to predict overall survival. Methods: Patients diagnosed with stage I or II NSCLC and status post definitive SBRT treatment were selected (n = 554). The PHOM score was calculated on each patient's pre-treatment CT scan. PHOM score, age, sex, stage, KPS, CCI, and post-SBRT chemotherapy were predictors in the generated Cox proportional hazards models for overall and cancer-specific survival. Patients were split into high and low PHOM score groups by median value and compared using KM curves for overall survival and cumulative incidence curves for cause specific death. Optimized tertile PHOM risk groups were calculated and similarly compared for overall survival and cause-specific death. Internal validation was conducted with bootstrap resampling and calibration curves were checked. Finally, a nomogram to predict overall survival was generated and is publicly available at https://eashwarsoma.shinyapps.io/LungCancerTDATest/. Results: PHOM score was a significant predictor for overall survival (HR: 1.17, 95% CI: 1.07-1.28) and was the only significant predictor for cancer-specific survival (1.31, 95% CI: 1.11-1.56) in the multivariable Cox model. The median survival for the high PHOM group was 29.18 months (95% CI: 23.62-34.27), which was significantly worse compared to the low PHOM group (45.41 months, 95% CI: 40.08-51.78, p < 0.001). The high PHOM group had a significantly greater chance of cancer-specific death at post treatment month 65 (0.244, 95%CI: 0.192-0.296) compared to the low PHOM group (0.171, 95% CI: 0.123-0.218, p = 0.029). Nomograms for 1, 2, 5, and 8-year overall survival were successfully calibrated using the Cox model. Conclusions: The PHOM score is associated with cancer-specific survival and predictive of overall survival. Our developed nomogram can be used to inform clinical prognosis and assist in making post-SBRT treatment considerations.

The calculation and use of Haralick texture features has been traditionally limited to imaging data and gray-level co-occurrence matrices calculated from images. We generalize the calculation of texture to graphs and networks with node attributes, focusing on cancer biology contexts such as fitness landscapes and gene regulatory networks with simulated and publicly available experimental gene expression data. We demonstrate the potential to calculate texture over multiple data set types including complex cancer networks and illustrate the potential for texture to distinguish cancer types and topologies of evolutionary landscapes through the summary metrics derived.

Objective &#xD;Image texture features, such as those derived by Haralick et al., are a powerful metric for image classification and&#xD;are used across fields including cancer research. Our aim is to demonstrate how analogous texture features can be derived for graphs and networks. We also aim to illustrate how these new metrics summarize graphs, may aid comparative graph studies, may help classify biological graphs, and might assist in detecting dysregulation in cancer.&#xD;&#xD;Approach &#xD;We generate the first analogies of image texture for graphs and networks. Co-occurrence matrices for graphs are&#xD;generated by summing over all pairs of neighboring nodes in the graph. We generate metrics for fitness landscapes, gene&#xD;co-expression and regulatory networks, and protein interaction networks. To assess metric sensitivity we varied discretization&#xD;parameters and noise. To examine these metrics in the cancer context we compare metrics for both simulated and publicly&#xD;available experimental gene expression and build random forest classifiers for cancer cell lineage.&#xD;&#xD;Main Results &#xD;Our novel graph “texture” features are shown to be informative of graph structure and node label distributions.&#xD;The metrics are sensitive to discretization parameters and noise in node labels. We demonstrate that graph texture features&#xD;vary across different biological graph topologies and node labelings. We show how our texture metrics can be used to classify&#xD;cell line expression by lineage, demonstrating classifiers with 82% and 89% accuracy.&#xD;&#xD;Significance &#xD;New metrics provide opportunities for better comparative analyses and new models for classification. Our&#xD;texture features are novel second-order graph features for networks or graphs with ordered node labels. In the complex&#xD;cancer informatics setting, evolutionary analyses and drug response prediction are two examples where new network science&#xD;approaches like this may prove fruitful.