Combining the radiomics signature and HPV status for the risk stratification of patients with OPC

Study of prediction model for high-grade meningioma using fractal geometry combined with radiological features

PURPOSE

To establish a prediction model combining fractal geometry and radiological features, which consider the complexity of tumour morphology advancing beyond the limitations of previous models.

METHODS

A total of 227 patients at the First Affiliated Hospital of Harbin Medical University from July 2021 to November 2023 were included. Fractal geometry was calculated and the radiomics features were extracted from regions of interest (ROIs). Weighted Gene Co-Expression Network Analysis (WGCNA) was employed for preliminary screening to identify those that were significantly associated with high-grade meningioma. In the training cohort, the least absolute shrinkage and selection operator (LASSO) regression was employed for further screening the radiomics features. Area under curve (AUC) was to evaluate models' performance.

RESULTS

In entire patient cohort, low-grade meningiomas had significantly lower fractal dimensions (P = 0.01), while high-grade meningiomas had higher lacunarity (P = 0.049). Fractal dimension (OR 6.8, 95% CI 1.49-36.51, P = 0.017), lacunarity (OR 3.7, 95% CI 1.36-11.75, P = 0.014), and Rscore (OR 2.8, 95% CI 1.55-5.75, P = 0.002) were independent risk factors for high-grade meningiomas. The final results demonstrated that the "fractal geometry + radiological features (semantic features + radiomics features)" model exhibited the most optimal performance in predicting high-grade meningioma, with an AUC of 0.854 in the training cohort and 0.757 in the validation cohort.

CONCLUSION

Significant differences in fractal dimension and lacunarity exist between high-grade and low-grade meningiomas, which can be potential predictive factors. The developed predictive model demonstrated good performance in predicting high-grade meningiomas.

Deep learning-based multimodal segmentation of oropharyngeal squamous cell carcinoma on CT and MRI using self-configuring nnU-Net

PURPOSE

To evaluate deep learning-based segmentation models for oropharyngeal squamous cell carcinoma (OPSCC) using CT and MRI with nnU-Net.

METHODS

This single-center retrospective study included 91 patients with OPSCC. The patients were grouped into the development (n = 56), test 1 (n = 13), and test 2 (n = 22) cohorts. In the development cohort, OPSCC was manually segmented on CT, MR, and co-registered CT-MR images, which served as the ground truth. The multimodal and multichannel input images were then trained using a self-configuring nnU-Net. For evaluation metrics, dice similarity coefficient (DSC) and mean Hausdorff distance (HD) were calculated for test cohorts. Pearson's correlation and Bland-Altman analyses were performed between ground truth and prediction volumes. Intraclass correlation coefficients (ICCs) of radiomic features were calculated for reproducibility assessment.

RESULTS

All models achieved robust segmentation performances with DSC of 0.64 ± 0.33 (CT), 0.67 ± 0.27 (MR), and 0.65 ± 0.29 (CT-MR) in test cohort 1 and 0.57 ± 0.31 (CT), 0.77 ± 0.08 (MR), and 0.73 ± 0.18 (CT-MR) in test cohort 2. No significant differences were found in DSC among the models. HD of CT-MR (1.57 ± 1.06 mm) and MR models (1.36 ± 0.61 mm) were significantly lower than that of the CT model (3.48 ± 5.0 mm) (p = 0.037 and p = 0.014, respectively). The correlation coefficients between the ground truth and prediction volumes for CT, MR, and CT-MR models were 0.88, 0.93, and 0.9, respectively. MR models demonstrated excellent mean ICCs of radiomic features (0.91-0.93).

CONCLUSION

The self-configuring nnU-Net demonstrated reliable and accurate segmentation of OPSCC on CT and MRI. The multimodal CT-MR model showed promising results for the simultaneous segmentation on CT and MRI.

CLINICAL RELEVANCE STATEMENT

Deep learning-based automatic detection and segmentation of oropharyngeal squamous cell carcinoma on pre-treatment CT and MRI would facilitate radiologic response assessment and radiotherapy planning.

KEY POINTS

• The nnU-Net framework produced a reliable and accurate segmentation of OPSCC on CT and MRI. • MR and CT-MR models showed higher DSC and lower Hausdorff distance than the CT model. • Correlation coefficients between the ground truth and predicted segmentation volumes were high in all the three models.

Prognosis Prediction in Head and Neck Squamous Cell Carcinoma by Radiomics and Clinical Information

(1) Background: head and neck squamous cell carcinoma (HNSCC) is a common cancer whose prognosis is affected by its heterogeneous nature. We aim to predict 5-year overall survival in HNSCC radiotherapy (RT) patients by integrating radiomic and clinical information in machine-learning models; (2) Methods: HNSCC radiotherapy planning computed tomography (CT) images with RT structures were obtained from The Cancer Imaging Archive. Radiomic features and clinical data were independently analyzed by five machine-learning algorithms. The results were enhanced through a voted ensembled approach. Subsequently, a probability-weighted enhanced model (PWEM) was generated by incorporating both models; (3) Results: a total of 299 cases were included in the analysis. By receiver operating characteristic (ROC) curve analysis, PWEM achieved an area under the curve (AUC) of 0.86, which outperformed both radiomic and clinical factor models. Mean decrease accuracy, mean decrease Gini, and a chi-square test identified T stage, age, and disease site as the most important clinical factors in prognosis prediction; (4) Conclusions: our radiomic-clinical combined model revealed superior performance when compared to radiomic and clinical factor models alone. Further prospective research with a larger sample size is warranted to implement the model for clinical use.

Risk assessment of venous thromboembolism in head and neck cancer patients and its establishment of a prediction model

IMPORTANCE

Venous thromboembolism (VTE) is closely relevant to head and neck cancer (HNC) prognosis, but little data exist on the risk prediction of VTE in patients with HNC.

OBJECTIVE

To study the risk factors regarding VTE in HNC patients and construct a nomogram model for its prediction.

DESIGN, SETTING, AND PARTICIPANTS

A cross-sectional retrospective study was implemented to comparatively analyze 220 HNC patients from January 2018 to December 2021. The Lasso algorithm was used to optimize the selection of variables. A nomogram model for predicting HNC-associated VTE was established using multivariate logistic regression analysis. Internal validation of the model was performed by bootstrap resampling (1000 times). Calibration plot and decision curve analysis (DCA) were applied to evaluate the calibration capability of the prediction model.

MAIN OUTCOME AND MEASURE

The demographics, medical history, blood biochemical indicators, and modalities of treatment were included for analysis.

RESULTS

The incidence of HNC-associated VTE was 2.8% (55/1967) in authors' affiliation. Five variables of risk factors, including surgery, radiochemotherapy, D-dimer, aspartate transaminase, and globulin, were screened and selected as predictors by Lasso algorithm. A prediction model that incorporated these independent predictors was developed and presented as the nomogram. The model showed good discrimination with a C-index of 0.972 (95% CI: 0.934-0.997), and had an area under the receiver operating characteristic curve value of 0.981 (p < 0.001, 95% CI: 0.964-0.998). The calibration curve displayed good agreement of the predicted probability with the actual observed probability for HNC-associated VTE. The DCA plot showed that the application of this nomogram was associated with net benefit gains in clinical practice.

CONCLUSIONS AND RELEVANCE

The high-performance nomogram model developed in this study may help early diagnose the risk of VTE in HNC patients and to guide individualized decision-making on thromboprophylaxis.