Radiomics Nomograms Based on Multi-Parametric MRI for Preoperative Differential Diagnosis of Malignant and Benign Sinonasal Tumors: A Two-Centre Study

Abstract

Objectives: To investigate the efficacy of multi-parametric MRI-based radiomics nomograms for preoperative distinction between benign and malignant sinonasal tumors.

Methods: Data of 244 patients with sinonasal tumor (training set, n=192; test set, n=52) who had undergone pre-contrast MRI, and 101 patients who underwent post-contrast MRI (training set, n=74; test set, n=27) were retrospectively analyzed. Independent predictors of malignancy were identified and their performance were evaluated. Seven radiomics signatures (RSs) using maximum relevance minimum redundancy (mRMR), and the least absolute shrinkage selection operator (LASSO) algorithm were established. The radiomics nomograms, comprising the clinical model and the RS algorithms were built: one based on pre-contrast MRI (RNWOC); the other based on pre-contrast and post-contrast MRI (RNWC). The performances of the models were evaluated with area under the curve (AUC), calibration, and decision curve analysis (DCA) respectively.

Results: The efficacy of the clinical model (AUC=0.81) of RNWC was higher than that of the model (AUC=0.76) of RNWOC in the test set. There was no significant difference in the AUC of radiomic algorithms in the test set. The RS-T1T2 (AUC=0.74) and RS-T1T2T1C (RSWC, AUC=0.81) achieved a good distinction efficacy in the test set. The RNWC and the RNWOC showed excellent distinction (AUC=0.89 and 0.82 respectively) in the test set. The DCA of the nomograms showed better clinical usefulness than the clinical models and radiomics signatures.

Conclusions: The radiomics nomograms combining the clinical model and RS can be accurately, safely and efficiently used to distinguish between benign and malignant sinonasal tumors.

Keywords: benign; differential diagnosis; magnetic resonance imaging; malignant; radiomics; sinonasal.

Figure 1

Flow chart of radiomics in this study.

Figure 2

Selection of MRI features and confirmation of the predictive accuracy of RS. (A) Selection of the tuning parameter (λ). An optimal λ value of 0.061(RNWC)/0.037(RNWOC) with ln(λ)=–2.80/–3.30 was selected. (B) The coef-ficients have been plotted vs. ln(λ). (C) The selection of features with non-zero coeffi-cients and their corresponding roles. (D) The differential diagnostic efficacy of rad-scores.

Figure 3

AUC of RS-T1 model (A, B), RS-T2 model (C, D), RS-T1C model (E, F), RS-T1T1C model (G, H), RS-T2T1C model (I, J), Clinical model, RS-T1T2, RNWOC model (K, L) and Clinical model, RSWC, RNWC model (M, N) for distinguishing be-tween benign and malignant sinonasal tumors in the train set and test set.

Figure 4

Radiomics nomograms (A). Calibration curves of the radiomics nomograms in the training set (B) and test set (C). The calibration curves showed that the nomograms had good agreement between the predictive risk of malignant status and the patho-logical outcome.

Figure 5

DCA of the radiomics nomograms. In the RNWC, the decision curves indicated that the radiomics nomograms were more beneficial than the clinical and RS model when the threshold probability is between 0.1 and 0.9. In the RNWOC, the threshold probability was between 0.2 and 1.0.