Evaluation of preoperative magnetic resonance imaging features and diagnostic effectiveness of grades II and III intracranial solitary fibroma

Yuncai Ran, Xiao Wang, Yong Zhang, Rui Chen, Chenchen Liu, Yunwei Ran, Weijian Wang, Xiaoyue Ma, Mengzhu Wang, Jingliang ChengEuropean 2024 · DOI

Abstract

Grade III intracranial solitary fibrous tumor (ISFT) patients typically present with more heterogeneous T2-FLAIR and Diffusion-Weighted Imaging (DWI) signal characteristics compared to Grade II patients. Key diagnostic indicators include higher rates of skull invasion, midline tumor mass collapse, and significantly lower Apparent Diffusion Coefficient (ADC) values ($ADC_{max}$, $ADC_{mean}$, and $ADC_{min}$). Among these, the $ADC_{min}$ value serves as a critical biomarker for the preoperative differentiation of Grade II and Grade III ISFT, substantially improving imaging-based grading accuracy.

Introduction

Intracranial solitary fibrous tumors (ISFT) are rare neoplasms originating from the meningeal stroma. The 2021 WHO classification categorizes these tumors into Grades I, II, and III. While Grade I tumors are generally benign, Grades II and III are malignant, characterized by aggressive growth and a high propensity for recurrence and metastasis.

The clinical management of ISFT is complicated by the significant difference in aggressiveness between Grade II and Grade III tumors. Grade III tumors are associated with markedly shorter survival times. Currently, definitive grading is predominantly performed postoperatively through histopathological analysis. Establishing reliable preoperative MRI "signatures" is essential for tailoring surgical margins and optimizing post-operative treatment strategies non-invasively.

Methods and Materials

This study involved a retrospective analysis of patients treated at the First Affiliated Hospital of Zhengzhou University between June 2012 and December 2022. The cohort was derived from an initial pool of 104 cases, ultimately focusing on 55 patients with confirmed Grade II or Grade III ISFT.

Flowchart of the patient selection process, detailing the inclusion and exclusion criteria that resulted in the final cohort of 55 ISFT patients.

Researchers evaluated a comprehensive set of preoperative MRI features using high-field 3T scanners. These included:

  • Signal Characteristics: T1-weighted, T2-weighted, T2-FLAIR, and DWI.
  • Morphological Features: Tumor location, maximum diameter, midline crossing, peritumoral edema, skull invasion, and the "meningeal tail" sign.
  • Quantitative Metrics: Apparent Diffusion Coefficient (ADC) values ($ADC_{max}$, $ADC_{mean}$, and $ADC_{min}$) were calculated within the solid portions of the tumors.

Statistical analysis utilized bivariate logistic regression to develop a predictive model, with Receiver Operating Characteristic (ROC) curves used to assess diagnostic efficacy.

Results

The study identified several key imaging differences between the two grades. Grade III tumors were significantly more likely to exhibit skull invasion (35.7% vs. 11.1% in Grade II) and displayed greater signal heterogeneity on T2-FLAIR and DWI sequences. This heterogeneity likely reflects the chaotic internal environment associated with rapid malignant growth and focal necrosis.

Violin plot illustrating the distribution of tumor maximum diameters (mm) for Grade II and Grade III ISFT patients.

The most significant finding was the predictive power of ADC values. $ADC_{max}$, $ADC_{mean}$, and $ADC_{min}$ were all significantly lower in Grade III patients ($P < 0.001$), correlating with the higher cellular density and nucleus-to-cytoplasm ratio typical of more aggressive tumors.

$ADC_{min}$ emerged as the most robust single predictor, capturing the densest, most malignant regions of the tumor. The diagnostic model achieved:

  • Overall Accuracy: 89.1%
  • Optimal $ADC_{min}$ Cutoff: 672 $mm^2/s$ (values below this threshold indicate Grade III).
  • Diagnostic Efficacy: An Area Under the Curve (AUC) of 0.805, with 85.2% sensitivity and 71.4% specificity.

Discussion

While this study provides a valuable roadmap for non-invasive grading, several limitations exist. The sample size of 55, though substantial for a rare disease, limits broad generalization. The retrospective, single-center design and manual selection of regions of interest (ROI) for ADC calculation may also introduce bias or variability.

Future research should focus on multi-center prospective studies. Additionally, the integration of radiomics and machine learning could automate feature extraction and potentially uncover high-dimensional imaging biomarkers that surpass the diagnostic precision of current manual methods.

Summary updated

Abstract

Grade III ISFT patients displayed more mixed T2-FLAIR signal characteristics and DWI signal characteristics than grade II patients, as shown by higher skull invasion and tumor mass collapse midline distribution and lower ADC<sub>max</sub>, ADC<sub>mean</sub>, and ADC<sub>min</sub> values. The ADC<sub>min</sub> value was significant in the preoperative assignment of grades II and III ISFT, thereby contributing to enhanced accuracy in the imaging grading diagnosis of the disease.
BibTeX
@inproceedings{ran2024evaluation,
  author = {Yuncai Ran and Xiao Wang and Yong Zhang and Rui Chen and Chenchen Liu and Yunwei Ran and Weijian Wang and Xiaoyue Ma and Mengzhu Wang and Jingliang Cheng},
  title = {{Evaluation of preoperative magnetic resonance imaging features and diagnostic effectiveness of grades II and III intracranial solitary fibroma}},
  booktitle = {European},
  year = {2024},
  doi = {10.1186/s40001-024-01959-5},
}
Figure