A Prospective Study of Machine Learning-Assisted Radiotherapy Planning for Patients Receiving 54 Gy to the Brain

PURPOSE/OBJECTIVE(S)

Radiotherapy (RT) planning is presently a semi-manual, iterative, labor-intensive process which may result in unnecessary variation in plan quality. To improve treatment plan quality and decrease RT planning time, we conducted a prospective, blinded study to compare machine learning-assisted planning with conventional manual planning for patients receiving 54 Gy in 30 fractions for a primary brain tumor.

MATERIALS/METHODS

From January 31, 2022 to January 10, 2023, 40 patients receiving 54 Gy for primary CNS tumors were prospectively enrolled (median age 50 years, range 4-78 years). Patients underwent standard CT/MR simulation and target/OAR delineation by the treating radiation oncologist. Each patient had one ML plan and 1-2 manual RT plans created by different planners. The reviewing oncologist was blinded to planning method by removing optimization and IMRT/VMAT beam arrangement details from all plans, which were then rated based on clinical acceptability, target coverage, OAR sparing, conformity, and dose-fall off. One preferred plan was chosen and used for clinical treatment.

RESULTS

A total of 115 plans for 40 patients were evaluated: 40 ML plans (35% of all plans), and 75 manual plans (65% of all plans; 5 and 35 patients had 1 and 2 manual plans created, respectively). ML plans required a mean planning time of 65 min as compared to 107 min for manual plans, with a mean time savings of 41 min per patient (paired t-test p = 0.002). 97% of ML plans (95% confidence interval [CI] 85-100) and 96% of manual plans (95% CI 87-99) were designated clinically acceptable by the treating radiation oncologist. While ML-assisted plans represented 35% of plans evaluated, they were chosen as preferred for clinical treatment in 43% of cases (17/40, 95% CI 29-58, p = 0.32). Median doses to the brain (10.8 Gy vs. 11.3 Gy, Wilcoxon rank-sum p = 0.012) and brain minus PTV (9.2 Gy vs 10.0 Gy, Wilcoxon rank-sum p = 0.009) were lower with ML planning versus manual planning, respectively. Doses to other structures, including hippocampi, cochlea, pituitary and hypothalamus were not statistically different.

CONCLUSION

In this prospective study with blinded oncologist evaluation, ML-assisted RT planning for primary CNS tumors was faster than manual planning, and produced a very high rate of acceptable plans with similar or superior OAR sparing. Future work will be undertaken to iteratively refine the ML model using the preferred cases from this study.

Dose Reductions to Critical Brain Organs-at-Risk and Better Cognition in Children with Medulloblastoma Receiving Proton Therapy

PURPOSE/OBJECTIVE(S)

Our group previously demonstrated an improvement in cognition among children with medulloblastoma treated with proton therapy, as compared to photon therapy. However, the reason for this cognitive improvement was unclear. In this study, our aim was to determine whether dose to critical brain structures acted as a mediator of improved cognition in patients treated with proton therapy.

MATERIALS/METHODS

In this retrospective study, a cohort of 75 children with medulloblastoma from two institutions was assembled (39 photon, 36 proton). Included patients were treated with similar radiation and cognitive follow-up protocols. Study endpoints were verbal comprehension (VCI), perceptual reasoning (PRI), working memory (WMI), processing speed (PSI) indices and full-scale IQ (FSIQ). Brain structures were segmented and dose comparisons by RT modality were compared using independent t-tests. Linear mixed effects models with random intercepts were created to evaluate cognitive endpoints using R version 4.2.2.

RESULTS

Median follow-up from RT to last cognitive assessment was 4.8 years. Total dose, including RT boost, was slightly lower in the proton cohort than the photon cohort (mean, 54.6 Gy vs. 56.1 Gy, respectively, p < 0.001). Eleven children (31%) treated with proton therapy received 36 Gy CSI, while 6 children (15%) treated with photon therapy received 36 Gy CSI (p = 0.07). Children treated with proton therapy had reduced total doses to the brain (mean, D40), left and right temporal lobes (mean, D40, D50), and left and right hippocampi (mean, D40, D50 - see Table). After adjustment for age at RT and posterior fossa syndrome, higher whole brain mean dose and time since RT were associated with greater decrease in VCI (p = 0.033), higher left temporal D50 and time since RT with greater decrease in PRI (p = 0.031), higher whole brain D40 and time since RT with greater decrease in PSI (p < 0.001) and FSIQ (p = 0.030).

CONCLUSION

Our study demonstrates that proton therapy for patients with medulloblastoma reduces dose to normal brain tissues, which is associated with better intellectual outcomes. Children with medulloblastoma who undergo RT should be treated with proton therapy, if available.