Hemoglobin state-flux: A finite-state model representation of the hemoglobin signal for evaluation of the resting state and the influence of disease

Detection of breast cancer using machine learning on time-series diffuse optical transillumination data

Significance

Optical mammography as a promising tool for cancer diagnosis has largely fallen behind expectations. Modern machine learning (ML) methods offer ways to improve cancer detection in diffuse optical transmission data.

Aim

We aim to quantitatively evaluate the classification of cancer-positive versus cancer-negative patients using ML methods on raw transmission time series data from bilateral breast scans during subjects' rest.

Approach

We use a support vector machine (SVM) with hyperparameter optimization and cross-validation to systematically explore a range of data preprocessing and feature-generation strategies. We also apply an automated ML (AutoML) framework to validate our findings. We use receiver operating characteristics and the corresponding area under the curve (AUC) to quantify classification performance.

Results

For the sample group available ( N = 63 , 18 cancer patients), we demonstrate an AUC score of up to 93.3% for SVM classification and up to 95.0% for the AutoML classifier.

Conclusions

ML offers a viable strategy for clinically relevant breast cancer diagnosis using diffuse-optical transmission measurements. The diagnostic performance of ML on raw data can outperform traditional statistical biomarkers derived from reconstructed image time series. To achieve clinically relevant performance, our ML approach requires simultaneous bilateral scanning of the breasts with spatially dense channel coverage.

Hemoglobin signal network mapping reveals novel indicators for precision medicine

Precision medicine currently relies on a mix of deep phenotyping strategies to guide more individualized healthcare. Despite being widely available and information-rich, physiological time-series measures are often overlooked as a resource to extend insights gained from such measures. Here we have explored resting-state hemoglobin measures applied to intact whole breasts for two subject groups - women with confirmed breast cancer and control subjects - with the goal of achieving a more detailed assessment of the cancer phenotype from a non-invasive measure. Invoked is a novel ordinal partition network method applied to multivariate measures that generates a Markov chain, thereby providing access to quantitative descriptions of short-term dynamics in the form of several classes of adjacency matrices. Exploration of these and their associated co-dependent behaviors unexpectedly reveals features of structured dynamics, some of which are shown to exhibit enzyme-like behaviors and sensitivity to recognized molecular markers of disease. Thus, findings obtained strongly indicate that despite the use of a macroscale sensing method, features more typical of molecular-cellular processes can be identified. Discussed are factors unique to our approach that favor a deeper depiction of tissue phenotypes, its extension to other forms of physiological time-series measures, and its expected utility to advance goals of precision medicine.