Evaluating prostate cancer bone metastasis using accelerated whole-body isotropic 3D T1-weighted Dixon MRI with compressed SENSE: a feasibility study.

OBJECTIVES: The study aimed to assess the efficiency of whole-body high-resolution compressed sensing-sensitivity encoding isotropic T1-Weighted Dixon (CSI-T1W-Dixon) scans in evaluating bone metastasis. METHODS: Forty-five high-risk prostate cancer patients with bone metastases were enrolled prospectively and underwent whole-body MRI sequences, which included the following: pre- and post-contrast CSI-T1W-Dixon and conventional multi-planar T1-Weighted Dixon (CMP-T1W-Dixon) (coronal, sagittal, and axial scans), short tau inversion recovery (STIR), and DWI. Comparison between the CMP-T1W-Dixon and CSI-T1W-Dixon images was done for the subjective image quality, the quantitative contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR). Furthermore, the diagnostic performance based on per-lesion and per-patient basis utilizing non-contrast T1-weighted (T1)/T1+ contrasted T1-weighted (T1C)/T1 + T1C + STIR + DWI sequences was compared between the CSI-T1W-Dixon and CMP-T1W-Dixon methods using reference standards (combining biopsy data and 6-month imaging follow-up). RESULT: The CSI-T1W-Dixon images produced fewer image artifacts in the axial and coronal planes compared to the CMP-T1W-Dixon images. Also, the CSI-T1W-Dixon images provided better a CNR in fat-only images of all three planes and water-only images of the axial plane (p < 0.05). The CSI-T1W-Dixon showed a higher sensitivity than the CMP-T1W-Dixon techniques in analyzing T1-only images on a per-lesion basis (82.7% vs. 53.8% for sensitivity, p = 0.03). On a per-patient basis, no difference was found in the diagnostic capacity between the CSI-T1W-Dixon and CMP-T1W-Dixon sequences either alone or in combinations (p = 0.57-1). CONCLUSION: High-resolution CSI-T1W-Dixon with higher image quality and diagnostic capacity can replace the CMP-T1W-Dixon method in evaluating bone metastasis in clinical practice. KEY POINTS: • Compressed sensing isotropic acquisition for 3D T1-weighted Dixon images can improve the image quality with fewer artifacts compared to the anisotropic multiplanar acquisition. • Compressed sensing isotropic acquisition can save 67% of scanning time compared to anisotropic multiplanar acquisition. • Compressed sensing isotropic 3D T1-weighted Dixon images can offer better diagnostic performance with higher sensitivity compared to anisotropic multiplanar images.

Dosimetric analysis of radiation-induced brainstem necrosis for nasopharyngeal carcinoma treated with IMRT.

BACKGROUND: Radiation-induced brainstem necrosis (RIBN) is a late life-threatening complication that can appear after treatment in patients with nasopharyngeal carcinoma (NPC). However, the relationship between RIBN and radiation dose is not still well-defined. METHODS: During January 2013 and December 2017, a total of 1063 patients with NPC were treated at Sichuan cancer hospital with IMRT. A total of 479 patients were eligible for dosimetric analysis. Dosimetric parameters of the RIBN, Dmax(the maximum dose), D0.1c (maximum average dose delivered to a 0.1-cc volume), D1cc, D2cc, D3cc, D5cc, D10cc and Dmean (mean does) were evaluated and recorded. ROC curve was used to analyze the area under curve (AUC) and cutoff points. Logistic regression for screening dose-volume parameter and logistic dose response model were used to predict the incidence of brainstem necrosis. RESULTS: Among the 479 patients with NPC, 6 patients were diagnosed with RIBN, the incidence of RIBN was 1.25% (6/479), and the median time to RIBN after treatment was 28.5 months (range 18-48 months). The dose of the brainstem in patients with RIBN were higher than that in patients without necrosis. ROC curve showed that the area under the curve (AUC) of Dmax was the largest (0.987). Moreover, logistic stepwise regression indicated that Dmax was the most important dose factor. The RIBN incidence at 5% over 5 years (TD5/5) and 50% incidence over 5 years (TD50/5) was 69.59 Gy and76.45 Gy, respectively. CONCLUSIONS: Brainstem necrosis is associated with high dose irritation. Dmax is the most significant predictive dosimetric factor for RIBN. Dmax of brainstem should be considered as the dose limitation parameter. We suggest that the limitation dose for brainstem was Dmax < 69.59 Gy.

Prognostic variables for temporal lobe injury after intensity modulated-radiotherapy of nasopharyngeal carcinoma.

To determine predictive factors for temporal lobe injury (TLI) in nasopharyngeal carcinoma patient (NPC) treated with intensity-modulated radiation therapy (IMRT). A total of 695 NPC cases treated with IMRT were retrospectively analyzed. TLI was diagnosed on MRI images. Volume-dose histograms for 870 evaluable temporal lobes were analyzed, and the predictive factors for the occurrence of TLI was evaluated. Receiver operating characteristic curve (ROC) and Logistic regression analysis was used to determine volume-dose parameters that predict temporal lobe injury (TLI). Univariate and multivariate analysis were used to analyze the predictive factors for TLI. The radiation dose-tolerance model of temporal lobe was calculated by logistic dose-response model. The median follow-up time was 73 months. A total of 8.5% patients were diagnosed with TLI. Among all the volume-dose parameters, logistic regression model showed D2cc (the dose Gray delivered to 2 cubic centimeter volume) was an only independent predictive factor. Multivariate analysis showed D2cc of temporal lobe, fraction size of prescription, T stage, and chemotherapy were the independent predictive factors for TLI. Logistic dose-response model has indicated the TD5/5 and TD50/5 of D2cc are 60.3 Gy and 76.9 Gy, respectively. D2cc of temporal lobe, fraction size of prescription, T stage, and chemotherapy were the possible independent predictive factors for TLI after IMRT of NPC. Biologic effective doses (TD5/5 and TD50/5 ) of D2cc are considered to prevent TLI.