Diffusion–relaxation correlation spectrum imaging for predicting tumor consistency and gross total resection in patients with pituitary adenomas: a preliminary study

Abstract

Objective

To evaluate the ability of diffusion–relaxation correlation spectrum imaging (DR-CSI) to predict the consistency and extent of resection (EOR) of pituitary adenomas (PAs).

Methods

Forty-four patients with PAs were prospectively enrolled. Tumor consistency was evaluated at surgery as either soft or hard, followed by histological assessment. In vivo DR-CSI was performed and spectra were segmented following to a peak-based strategy into four compartments, designated A (low ADC), B (mediate ADC, short T2), C (mediate ADC, long T2), and D (high ADC). The corresponding volume fractions (${f}_{\mathrm{A}}$, ${f}_{\mathrm{B}}$, ${f}_{\mathrm{C}}$, ${f}_{\mathrm{D}}$) along with the ADC and T2 values were calculated and assessed using univariable analysis for discrimination between hard and soft PAs. Predictors of EOR > 95% were analyzed using logistic regression model and receiver-operating-characteristic analysis.

Results

Tumor consistency was classified as soft (n = 28) or hard (n = 16). Hard PAs presented higher ${f}_{\mathrm{B}}$ (p = 0.001) and lower ${f}_{\mathrm{C}}$ (p = 0.013) than soft PAs, while no significant difference was found in other parameters. ${f}_{\mathrm{B}}$ significantly correlated with the level of collagen content (r = 0.448, p = 0.002). Knosp grade (odds ratio [OR], 0.299; 95% confidence interval [CI], 0.124–0.716; p = 0.007) and ${f}_{\mathrm{B}}$ (OR, 0.834, per 1% increase; 95% CI, 0.731–0.951; p = 0.007) were independently associated with EOR > 95%. A prediction model based on these variables yielded an AUC of 0.934 (sensitivity, 90.9%; specificity, 90.9%), outperforming the Knosp grade alone (AUC, 0.785; p < 0.05).

Conclusion

DR-CSI may serve as a promising tool to predict the consistency and EOR of PAs.

Clinical relevance statement

DR-CSI provides an imaging dimension for characterizing tissue microstructure of PAs and may serve as a promising tool to predict the tumor consistency and extent of resection in patients with PAs.

Key Points

• DR-CSI provides an imaging dimension for characterizing tissue microstructure of PAs by visualizing the volume fraction and corresponding spatial distribution of four compartments ( ${f}_{A}$ , ${f}_{B}$ , ${f}_{C}$ , ${f}_{D}$ ).

• ${f}_{B}$ correlated with the level of collagen content and may be the best DR-CSI parameter for discrimination between hard and soft PAs.

• The combination of Knosp grade and ${f}_{B}$ achieved an AUC of 0.934 for predicting the total or near-total resection, outperforming the Knosp grade alone (AUC, 0.785).