Dosimetric predictors of toxicity in a randomized study of short-course vs conventional radiotherapy for glioblastoma

• Published in: Radiotherapy and oncology Vol. 177; pp. 152 - 157
• Main Authors: Yang, Fan, Dinakaran, Deepak, Heikal, Amr A., Yaghoobpour Tari, Shima, Ghosh, Sunita, Amanie, John, Murtha, Albert, Rowe, Lindsay S., Roa, Wilson H., Patel, Samir
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.12.2022
• Subjects: Adolescent; Adult; Aged; Glioblastoma; Glioblastoma - radiotherapy; Hippocampus; Humans; Middle Aged; Organ at Risk; Organs at Risk; Prospective Studies; Radiometry; Radiotherapy Dosage; Radiotherapy Planning, Computer-Assisted - methods; Radiotherapy, Intensity-Modulated - methods; Short-course; Toxicity; Young Adult; CTCAE; OAR; RT; Organ at Risk; Toxicity; Glioblastoma; PTV; ECOG; MMSE; MCO; Short-course; EQD2; RTOG; QUANTEC; Hippocampus
• ISSN: 0167-8140; 1879-0887
• DOI: 10.1016/j.radonc.2022.10.016

•Critical structure constraints used in our randomized controlled trial of short-course versus conventional radiotherapy for patients with glioblastoma were appropriate for limiting grade 2 or higher toxicities.•Higher left-hippocampal mean doses were the most predictive for neuro-cognitive decline post-radiotherapy.•Routine contouring and use of dose constraints for the hippocampus is recommended.•Our left hippocampus sparing treatment model may help minimize neuro-cognitive decline in patients with glioblastoma treated with radiotherapy. There is no consensus on appropriate organ at risk (OAR) constraints for short-course radiotherapy for patients with glioblastoma. Using dosimetry and prospectively-collected toxicity data from a trial of short-course radiotherapy for glioblastoma, this study aims to empirically examine the OAR constraints, with particular attention to left hippocampus dosimetry and impact on neuro-cognitive decline. Data was taken from a randomized control trial of 133 adults (age 18–70 years; ECOG performance score 0–2) with newly diagnosed glioblastoma treated with 60 Gy in 30 (conventional arm) versus 20 (short-course arm) fractions of adjuvant chemoradiotherapy (ClinicalTrials.gov Identifier: NCT02206230). The delivered plan’s dosimetry to the OARs was correlated to prospective-collected toxicity and Mini-Mental State Examination (MMSE) data. Toxicity events were not significantly increased in the short-course arm versus the conventional arm. Across all OARs, delivered radiation doses within protocol-allowable maximum doses correlated with lack of grade ≥ 2 toxicities in both arms (p < 0.001), while patients with OAR doses at or above protocol limits correlated with increased grade ≥ 2 toxicities across all examined OARs in both arms (p-values 0.063–0.250). Mean left hippocampus dose was significantly associated with post-radiotherapy decline in MMSE scores (p = 0.005), while the right hippocampus mean dose did not reach statistical significance (p = 0.277). Compared to the original clinical plan, RapidPlan left hippocampus sparing model decreased left hippocampus mean dose by 43 % (p < 0.001), without compromising planning target volume coverage. In this trial, protocol OAR constraints were appropriate for limiting grade ≥ 2 toxicities in conventional and short-course adjuvant chemoradiotherapy for glioblastoma. Higher left hippocampal mean doses were predictive for neuro-cognitive decline post-radiotherapy. Routine contouring and use of dose constraints to limit hippocampal dose is recommended to minimize neuro-cognitive decline in patients with glioblastoma treated with chemoradiotherapy.