PySERA

Radiomics analyses extract quantitative biomarkers from medical images for precision modeling, yet reproducibility and scalability remain limited by heterogeneous and limited implementations. Existing tools support only partial standards and lack integration with deep learning (DL) radiomics. To address these gaps, we developed PySERA, an open-source, Python-native, standardized radiomics framework designed for automation, reproducibility, and AI integration.MethodsPySERA re-implements MATLAB-based SERA (standardized environment for radiomics analysis) in a modular, object-oriented Python architecture. It computes 557 features, including 487 features compliant with the Image Biomarker Standardization Initiative (IBSI) and 10 moment-invariant descriptors, as well as 60 additional diagnostic features, along with DL radiomics embeddings from pre-trained DL: ResNet50 (2,048 features) DL radiomics features), DenseNet121 (1,024), and VGG16 (512). It includes standardized preprocessing (resampling, discretization, normalization), multi-format I/O (DICOM, NIfTI, NRRD), adaptive memory handling, and a parallel multi-core engine for scalable feature extraction. PySERA integrates directly with libraries: scikit-learn/PyTorch/TensorFlow/MONAI, and others for downstream machine learning applications.ResultsPySERA demonstrated >94% IBSI reproducibility, closely matching MITK and substantially outperforming PyRadiomics against the 487 IBSI-compliant feature reference set. Across 8 public datasets, PySERA achieved accuracies of 0.43–0.84, exceeding PyRadiomics for outcome prediction tasks. Benchmarking showed efficient processing (including added higher-order features not implemented in other software): 583 seconds (305 MB) for 166 features, and 2,325 seconds (491 MB) for full extraction, with deterministic outputs across platforms.ConclusionsBy uniting standardized handcrafted/DL radiomics in a scalable, transparent, and Python-integrable framework, PySERA establishes a reproducible and extensible foundation for next-generation, AI-ready precision imaging research.

Radiomics analysis transforms medical imaging into a quantitative resource by extracting features that capture tumor morphology, intensity, heterogeneity, shape, and texture, enabling precision diagnosis, treatment response prediction, and outcome modeling. However, existing radiomics solutions face major limitations, including restricted feature sets, limited interpretability, lack of higher‑order/multiscale descriptors, absence of moment‑based features, and/or poor compatibility with large‑scale, automated, reproducible AI workflows. These constraints reduce model robustness, hinder small‑lesion analysis, and slow clinical translation. To overcome these limitations, we introduce PySERA, an open-source, IBSIcompliant Python library for automated, scalable, and reproducible radiomics feature extraction. PySERA supports multiple image formats—including multi-slice and single-slice DICOM, NIfTI, NRRD, DICOM RT Struct, and NumPy arrays—and integrates seamlessly into both research and clinical workflows. It computes 497 radiomics features (487 IBSI-compliant, 10 moment-invariant), supports batch processing, multi-core parallelization, and deterministic logging, and uniquely enables reliable feature extraction from very small lesions through synthetic augmentation to facilitate early detection in precision oncology. PySERA was evaluated on two independent cancer imaging datasets: 999 CT lung cancer scans and 883 PET head and neck cancer cases, using binary survival as the outcome. Machine learning pipelines combining Support Vector Machine, Logistic Regression, and K-Nearest Neighbors—with and without Principal Component Analysis—achieved 0.80±0.05 accuracy, significantly outperforming results using PyRadiomics (0.60±0.05, p<0.05). By enabling richer feature sets, robust small‑lesion analysis, and seamless AI integration, PySERA bridges research and clinical translation through scalable, reproducible, and clinically explainable radiomics, ultimately driving precision medicine and accelerating multi‑center imaging studies.