RPS 911 - Skull-based essentials: from the sellar region to the cerebellopontine angle

Author Block: S. Nalubolu, R. Hiregouja Eranna; Bengaluru/IN
Purpose: While most pituitary adenomas are indolent, a subset exhibit atypical morphology characterized by increased firmness, invasiveness, and proliferative histopathology. These features contribute to surgical complexity and early postoperative recurrence. Preoperative identification of such atypical adenomas using quantitative MRI biomarkers—particularly diffusion, T2 signal, and enhancement pattern—may aid surgical planning and risk stratification.
Methods or Background: A retrospective review was conducted on 128 surgically resected pituitary adenomas (2020–2023). Preoperative 3 T MRI (T1, T2, DWI/ADC, dynamic contrast-enhanced sequences) and CT were assessed for cavernous sinus invasion (Knosp grade), dural contact, margin irregularity, T2 signal, cystic/necrotic change, and osseous invasion. Mean ADC values (b = 0, 1000 s/mm²) were measured from tumour ROIs. Atypicality was defined by intraoperative firmness and/or proliferative indices (Ki-67 > 3%, p53 positivity). Interobserver agreement (κ) and ROC analyses were performed. Follow-up MRI and hormonal data (6–18 months) assessed persistence or progression.
Results or Findings: Atypical lesions had a substantially lower mean ADC than typical ones (0.72 ± 0.09 vs. 0.91 ± 0.11 × 10⁻³ mm²/s; p < 0.001).T2 hypointensity occurred in 63% versus 31% (p = 0.002), correlating with intraoperative firmness (κ = 0.81).Cavernous sinus invasion (sensitivity 85%, specificity 82%) and irregular margins (74%, 68%) were strong predictors. The combined model with T2 hypointensity + invasion and ADC < 0.80 × 10⁻³ mm²/s had an AUC of 0.91. Early progression was seen in 17% of atypical versus 4% of typical adenomas (p = 0.03).
Conclusion: Atypical pituitary adenomas exhibit a reproducible MRI signature—low ADC, T2 hypointensity, and invasive margins—reflecting fibrous, highly cellular tissue. These multiparametric MRI biomarkers predict surgical difficulty and early recurrence, supporting their integration into preoperative evaluation and biologically guided management.
Limitations: Single-center limitation.
Funding for this study: No funding was provided for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: D. A. Zambrano, A. M. Delgado Brito, C. D. Solano, M. R. López De La Torre Carretero, J. M. Rodríguez Ortega, M. Calvo Imirizaldu; Pamplona/ES
Purpose: Describe and quantify dynamic perfusion-curve patterns in pituitary lesions (mainly microadenomas) and compare them with normal pituitaries.
Methods or Background: Pituitary microadenomas (<10 mm) are benign, often asymptomatic tumors with low clinical risk, typically identified through MRI and hormonal evaluation. However, their small size and subtle enhancement can lead to diagnostic uncertainty. Dynamic contrast-enhanced (DCE) MRI curve patterns (types I–III) have been proposed as diagnostic aids.

This retrospective descriptive study included 30 patients (20 with pituitary microadenomas and 10 controls) evaluated between January 2022 and December 2024. Imaging was performed on 3T and 1.5T scanners using 3D T1-weighted DCE sequences, followed by motion correction and temporal signal-intensity curve analysis. Regions of interest (ROIs) were placed in the lesion and contralateral pituitary tissue (or bilaterally in controls). Perfusion curves were classified as type I, II, or III for descriptive comparison between microadenomas and normal pituitaries.
Results or Findings: In this study of 30 patients (20 with microadenomas, 10 controls), 95% of microadenomas exhibited a type II DCE curve, whereas all controls showed normal physiological curves. These results indicate a strong association between microadenomas and type II perfusion patterns, suggesting that DCE-curve analysis may support visual diagnosis in uncertain cases. However, due to the retrospective design and small sample size, findings are exploratory. Further research is needed to standardize metrics, evaluate reproducibility, and validate diagnostic performance before adopting type II curves as a formal criterion.
Conclusion: Most pituitary microadenomas exhibited a type IIb DCE pattern, suggesting that perfusion-curve analysis may aid in their detection. Further studies are needed to validate these findings.
Limitations: Retrospective design and small sample size render findings exploratory. Standardization of quantitative metrics, interobserver reproducibility testing, and validation in larger prospective studies are required before clinical adoption.
Funding for this study: The authors did not receive any funding for this study
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: M. O. Sousa, R. Silva; São Luís/BR
Purpose: Evaluate magnetic resonance imaging (MRI) craniopharyngioma diagnosis patients, correlating them with histological subtypes, age and gender. Furthermore, observe differential diagnosis, and the disease evolution based on MRI aspect.
Methods or Background: Retrospective cohort analysis of 111 MRIs from 34 craniopharyngioma diagnosis patients performed between 2018 and 2025, comparing histological diagnosis, age, gender and progression. Descriptive statistical analysis, Kruskal-Wallis test and Fisher's exact test were performed using Stata software.
Results or Findings: Most females patients ranging from 2 to 71 years old, median of 13, interquartile range (IQR) = 8-24 years old. Twenty-nine were craniopharyngioma confirmed, two pituitary macroadenoma, two suprasellar pilocytic astrocytoma, and one Rathke's pouch cyst. Adamatinomatous craniopharyngiomas (86,67%), papillary (12,33%). Kruskal-Wallis test being statistically significant for the association between the histological type and age group (P = 0,27) but not for progression (P = 0,53). Fisher's exact test demonstrated no association between progression and histological type (P = 1,0).
Conclusion: In this study sample, we observed a predominance of adamatinomatous subtype craniopharyngiomas in children and young adults, with a small percentage of elderly individuals affected by the papillary subtype, predominantly female.
These lesions presented challenging imaging diagnoses and was confused with macroadenomas, Rathke's pouch cysts, and pilocytic astrocytomas. Although most presented reduction, many progressed despite treatment, and some died, characterizing the aggressive aspect of the tumor.
Limitations: The study was limited by the small sample size given the rarity of the disease. Other limitations were the high number of missing information on histological subtype and the lack of follow-up in some cases.
Funding for this study: No funding
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: A. Mangili¹, A. Arrigoni¹, G. Pezzetti², B. Frigeni², R. Bivona², S. Capelli¹, G. Danesi², A. Caroli¹, S. Gerevini²; ¹Ranica/IT, ²Bergamo/IT
Purpose: Cerebellopontine angle (CPA) tumors are often benign but can compress adjacent structures, including the facial nerve (FN). Surgical resection is standard treatment but carries a risk of iatrogenic injury and facial palsy. This study investigates the role of anatomical and diffusion-weighted MRI (DW-MRI) in presurgical planning, focusing on FN reconstruction. Radiomic features and diffusion tensor imaging (DTI) biomarkers were extracted and integrated with clinical variables to train machine learning models predicting FN integrity, postoperative outcomes, and long-term function.
Methods or Background: Forty-seven CPA patients who underwent preoperative MRI and surgery were analyzed. MRI protocols included DW-MRI AP (b0 and 1500; 50 directions; voxel size 2×2×2 mm), b0 PA, post-contrast volumetric T1-w, and volumetric T2-w scans. A custom pipeline, implemented in Python with MRtrix3Tissue and FSL, corrected noise and artifacts, segmented lesions, and performed a SS3T-CSD-based tractography using the iFOD2 algorithm with anatomically guided seeding.
Results or Findings: The FN was successfully reconstructed in all patients. The tumor-affected side showed higher fractional anisotropy (FA) and lower mean diffusivity (MD) than the contralateral side. Diffusion metrics correlated with conventional prognostic markers, including blink reflex and Koos grade. Compound Muscle Action Potential (CMAP) correlated inversely with diffusion values and FN tract length. Postoperative and 1-year House-Brackmann (HB) grades were significantly associated with lesion volume, FN length, and CMAP. Machine learning models achieved accuracies of 0.76 for predicting FN adherence, 0.82 for postoperative HB, and 0.90 for 12-month HB outcomes.
Conclusion: Diffusion tractography enhances CPA presurgical planning. Combined radiomic, clinical, and neurophysiological data improve prediction of FN integrity, short-term function, and long-term outcomes after resection.
Limitations: The study is limited by its single-center design, modest sample size, and lack of external validation.
Funding for this study: The study didn’t receive any funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Protocol title: Facial nerve Assessment in patients with Cerebellopontine angle tumours: a quantitative Evaluation from pre-surgical brain MRI Scan (FACES)

Author Block: C. Chen, J. Zhou; LanZhou/CN
Purpose: Although most pituitary neuroendocrine tumors (PitNETs) are benign with good postoperative outcomes, they frequently recur despite multidisciplinary treatment. Early prediction of PitNET recurrence is clinically vital. This study aims to explore the value of preoperative assessment of PitNET recurrence based on MRI fractal analysis and radiomic features.
Methods or Background: A retrospective study included 123 PitNET patients who underwent MRI, divided into recurrence (n=50) and non-recurrence (n=73) groups based on follow-up. Clinical, pathological, and conventional MRI data were collected. Fractal dimension was calculated from sagittal CE-T1WI using ImageJ. Radiomic features were extracted from whole-tumor images on axial T2, sagittal CE-T1WI, and coronal CE-T1WI sequences via Darwin platform. Group differences in fractal parameters, radiomic features, clinical/MRI/pathological traits were analyzed to construct an optimal predictive model.
Results or Findings: Statistically significant differences in fractal dimension were observed between the two groups (P < 0.001). Among clinical features, preoperative ACTH levels, tumor texture, extent of tumor resection, presence of tumor apoplexy, and CSF leakage all differed significantly (all P < 0.05). For conventional MRI signs, statistically significant differences were found in maximum tumor diameter/height, tumor hemorrhage, carotid artery encasement, and cavernous sinus invasion (all P < 0.05). Pathologically, P53 positivity showed a significant difference (P = 0.03). The SVM model based on radiomic features outperformed others. The SVM model constructed using statistically significant fused features achieved the highest performance, with AUC values of 0.951 in the test set and 0.778 in the validation set .
Conclusion: Combining MRI fractal analysis, radiomics, and clinical/pathological data enhances clinical decision-making to improve treatment and patient quality of life.
Limitations: This is a retrospective study.
Multi-dimensional parameters such as molecular omics and genomics will be integrated to refine the model.
Funding for this study: National Natural Science Foundation Project (82371914)
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: