MRI-Guided Real-Time Online Adaptive Gated Stereotactic Body Radiation Therapy for Liver Tumors

Abstract

<h3>Purpose/Objective(s)</h3> Liver tumors (primary or metastases) are commonly encountered in Oncology. Single breathing phase Magnetic resonance imaging (MRI)-guided gated and possible adaptive stereotactic body radiation therapy (SBRT) for liver tumors offer dosimetric advantages and facilitate safe dose escalation compared to non- MRI-guided SBRT (MRgSBRT). We aimed to evaluate the impact of MRgSBRT on treatment outcomes and report treatment related adverse events. <h3>Materials/Methods</h3> This IRB approved retrospective study included liver tumors patients treated consecutively with MRgSBRT between 2019 and 2021. All patients received Gadoxetate disodium contrast followed by single phase post-exhale CT and MRI simulations utilizing 0.35 T MRI-guided linear accelerator. MRI images (hybrid T1-T2 sequence) were used as the primary set of images for contouring and planning. Real-time online adaptive SBRT plans were generated to account for daily anatomical changes of nearby organs at risk (OARs) when indicated. Patients were followed up every 3 to 6 months with a brief history, physical examination and imaging as indicated. Patients' data were analyzed based on age, sex, histology, tumor volume and delivered biologically equivalent dose (BED <100 Gy vs BED > 100 Gy). Kaplan Meier curves were generated for freedom from local progression (FFLP) and overall survival (OS). <h3>Results</h3> The study included 35 eligible patients with 49 liver tumors (hepatocellular carcinoma 21 lesions, liver metastases 28 lesions). The median age was 69.5 years (range 25 - 95). The median planning target volume (PTV) was 43.5 cm³ (4.9 - 344.8 cm³). Most patients (80%) were prescribed 50 Gy in 5 fractions (median 50 Gy, range 35 - 50Gy) to PTV. A total of 255 fractions were delivered for patients who completed treatment, 22 of which were adapted (8.6%). The median follow up was 19.4 months (range 1-37 months). The median OS was 25.9 months with 95% confidence intervals (CI) (22.1 - 29.6). The 1-year OS rate was 77.7%, 3-year OS rate was 47.9%. Local control was a driver for OS with patients with durable local control having a median OS of 27.8 months [95%CI (23.8-31.6)] vs 13.5 months [95%CI (5.6-21.3), p=0.007] in patients with local progression. The median FFLP was 29.8 months [95%CI (27.2-32.3]. The estimated 1-year FFLP was 95.6 % and 3-year FFLP was 87.1%. The FFLP in patients with BED>100Gy was 30.9 months [95%CI (28.7-33.1)] vs 13.3 months [95%CI (5.3-21.3), p value=0.004] in patients who received <100Gy BED. Most patients tolerated treatment well, no acute grade 3 toxicity. Two patients developed late grade 3 toxicity (gastric ulcer and non-classic radiation induced liver disease). <h3>Conclusion</h3> In our study, Gated MRgSBRT with real-time online adaptive planning is feasible, provides durable local control and impacts overall survival in a heavily pretreated older group of liver tumor patients, with remarkably good safety profile. BED>100Gy improved local control. Improved local control seems to drive overall survival benefit in liver tumors.