T2-Weighted Image-Based Radiomics Signature for Discriminating Between Seminomas and Nonseminoma

Peipei Zhang, Zhaoyan Feng, Wei Cai, Huijuan You, Chanyuan Fan, Wenzhi Lv, Xiangde Min, Liang WangFrontiers in Oncology 2019 · DOI

Abstract

Objective: To evaluate the performance of a T2-weighted image (T2WI)-based radiomics signature for differentiating between seminomas and nonseminomas.
Materials and Methods: In this retrospective study, 39 patients with testicular germ-cell tumors (TGCTs) confirmed by radical orchiectomy were enrolled (19 seminomas, 20 nonseminomas). All patients underwent 3T magnetic resonance imaging (MRI) before surgery. Eight hundred fifty-one radiomics features were extracted from the T2WI of each patient. Feature selection was performed using intra- and interclass correlation coefficients (ICC), minimum-redundancy maximum-relevance (mRMR), and the least absolute shrinkage and selection operator (LASSO) algorithms. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis.
Results: Five features were selected to build the radiomics signature. The signature was significantly different between groups (p < 0.01), achieving an area under the curve (AUC) of 0.979, sensitivity of 90.00%, and specificity of 100.00%.
Conclusion: The T2WI-based radiomics signature has the potential to non-invasively discriminate between seminomas and nonseminomas.

Introduction

Testicular germ-cell tumors (TGCTs) are the most common malignancy in young men. Clinically, distinguishing between seminomas and nonseminomas is vital because their management strategies differ significantly; seminomas are highly sensitive to radiotherapy, whereas nonseminomas often require more complex chemotherapy or surgical interventions.

Currently, the "gold standard" for diagnosis is radical orchiectomy followed by pathological examination. Traditional biopsies are strictly discouraged due to the high risk of tumor seeding. While ultrasound and standard MRI are used for initial screening, they often rely on subjective interpretations. There is a clear need for a non-invasive, objective method to characterize these tumors before surgical intervention.

Materials and Methods

This retrospective study analyzed 39 patients (median age 29) treated at a single institution between 2014 and 2019. The cohort included 19 seminomas and 20 nonseminomas, all pathologically confirmed.

Flow chart of patient inclusion and exclusion criteria.

Radiomics Workflow

  1. Image Acquisition & Segmentation: Patients underwent 3T MRI. Radiologists manually outlined the three-dimensional volume of interest (VOI) on T2-weighted images (T2WI).
  2. Feature Extraction: Using the PyRadiomics library, 851 features were extracted, including shape, intensity, and complex textures (e.g., Gray Level Co-occurrence Matrix).
  3. Feature Selection: To prevent overfitting, the team used a multi-step reduction process: ICC for stability, mRMR to rank unique information, and LASSO to identify the most predictive variables.

Results

The final radiomics signature was built using five core features. The process of narrowing down 851 features was necessary due to the high degree of redundancy and correlation between raw quantitative metrics.

Correlation matrices showing the high dimensionality (A) and subsequent filtering (B) of radiomics features.

The radiomics signature demonstrated exceptional performance:

  • AUC: 0.979 (95% CI: 0.873–1.000)
  • Sensitivity: 90.00%
  • Specificity: 100.00%
  • Statistical Significance: Differences between groups were highly significant (p < 0.01).

Notably, four of the five final features were derived from wavelet filters. Wavelet transforms decompose images into different scales, allowing the model to detect subtle tissue heterogeneity invisible to the naked eye.

Discussion and Conclusion

This study demonstrates that high-quality T2WI alone, when processed through a radiomics pipeline, contains sufficient information for high-accuracy discrimination between TGCT subtypes. By using objective algorithms like LASSO, the researchers moved away from subjective "cherry-picking" of features.

However, several limitations remain. The sample size is small (n=39), and the model lacks external validation on datasets from different institutions or scanners. Future work should focus on multi-center trials and the potential integration of functional MRI sequences (like DWI) to create an even more robust multiparametric signature. Despite these hurdles, this T2WI-based radiomics signature represents a promising step toward non-invasive, data-driven clinical decision-making in testicular cancer.

Summary updated

Abstract

<b>Objective:</b> To evaluate the performance of a T2-weighted image (T2WI)-based radiomics signature for differentiating between seminomas and nonseminomas. <b>Materials and Methods:</b> In this retrospective study, 39 patients with testicular germ-cell tumors (TGCTs) confirmed by radical orchiectomy were enrolled, including 19 cases of seminomas and 20 cases of nonseminomas. All patients underwent 3T magnetic resonance imaging (MRI) before radical orchiectomy. Eight hundred fifty-one radiomics features were extracted from the T2WI of each patient. Intra- and interclass correlation coefficients were used to select the features with excellent stability and repeatability. Then, we used the minimum-redundancy maximum-relevance (mRMR) and the least absolute shrinkage and selection operator (LASSO) algorithms for feature selection and radiomics signature development. Receiver operating characteristic curve analysis was used to evaluate the diagnostic performance of the radiomics signature. <b>Results:</b> Five features were selected to build the radiomics signature. The radiomics signature was significantly different between the seminomas and nonseminomas (<i>p</i> < 0.01). The area under the curve (AUC), sensitivity, and specificity of the radiomics signature for discriminating between seminomas and nonseminomas were 0.979 (95% CI: 0.873-1.000), 90.00 (95% CI: 68.3-98.8), and 100.00 (95% CI: 82.4-100.0), respectively. <b>Conclusion:</b> The T2WI-based radiomics signature has the potential to non-invasively discriminate between seminomas and nonseminomas.
BibTeX
@inproceedings{zhang2019t,
  author = {Peipei Zhang and Zhaoyan Feng and Wei Cai and Huijuan You and Chanyuan Fan and Wenzhi Lv and Xiangde Min and Liang Wang},
  title = {{T2-Weighted Image-Based Radiomics Signature for Discriminating Between Seminomas and Nonseminoma}},
  booktitle = {Frontiers in Oncology},
  year = {2019},
  doi = {10.3389/fonc.2019.01330},
}
Figure