Engineering topics

Hope, Andrew J.

Publications and source records attributed to Hope, Andrew J..

4 records

DOE OSTI2015

Automatic learning-based beam angle selection for thoracic IMRT

Purpose: The treatment of thoracic cancer using external beam radiation requires an optimal selection of the radiation beam directions to ensure effective coverage of the target volume and to avoid unnecessary treatment of normal healthy tissues. Intensity modulated radiation therapy (IMRT) planning is a lengthy process, which requires the planner to iterate between choosing beam angles, specifying dose–volume objectives and executing IMRT optimization. In thorax treatment planning, where there are no class solutions for beam placement, beam angle selection is performed manually, based on the planner’s clinical experience. The purpose of this work is to propose and study a computationally efficient framework that utilizes machine learning to automatically select treatment beam angles. Such a framework may be helpful for reducing the overall planning workload. Methods: The authors introduce an automated beam selection method, based on learning the relationships between beam angles and anatomical features. Using a large set of clinically approved IMRT plans, a random forest regression algorithm is trained to map a multitude of anatomical features into an individual beam score. An optimization scheme is then built to select and adjust the beam angles, considering the learned interbeam dependencies. The validity and quality of the automatically selected beams evaluated using the manually selected beams from the corresponding clinical plans as the ground truth. Results: The analysis included 149 clinically approved thoracic IMRT plans. For a randomly selected test subset of 27 plans, IMRT plans were generated using automatically selected beams and compared to the clinical plans. The comparison of the predicted and the clinical beam angles demonstrated a good average correspondence between the two (angular distance 16.8° ± 10°, correlation 0.75 ± 0.2). The dose distributions of the semiautomatic and clinical plans were equivalent in terms of primary target volume coverage and organ at risk sparing and were superior over plans produced with fixed sets of common beam angles. The great majority of the automatic plans (93%) were approved as clinically acceptable by three radiation therapy specialists. Conclusions: The results demonstrated the feasibility of utilizing a learning-based approach for automatic selection of beam angles in thoracic IMRT planning. The proposed method may assist in reducing the manual planning workload, while sustaining plan quality.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2014

Incidental Prophylactic Nodal Irradiation and Patterns of Nodal Relapse in Inoperable Early Stage NSCLC Patients Treated With SBRT: A Case-Matched Analysis

Purpose: Reported rates of non-small cell lung cancer (NSCLC) nodal failure following stereotactic body radiation therapy (SBRT) are lower than those reported in the surgical series when matched for stage. We hypothesized that this effect was due to incidental prophylactic nodal irradiation. Methods and Materials: A prospectively collected group of medically inoperable early stage NSCLC patients from 2004 to 2010 was used to identify cases with nodal relapses. Controls were matched to cases, 2:1, controlling for tumor volume (ie, same or greater) and tumor location (ie, same lobe). Reference (normalized to equivalent dose for 2-Gy fractions [EQD2]) point doses at the ipsilateral hilum and carina, demographic data, and clinical outcomes were extracted from the medical records. Univariate conditional logistical regression analyses were performed with variables of interest. Results: Cases and controls were well matched except for size. The controls, as expected, had larger gross tumor volumes (P=.02). The mean ipsilateral hilar doses were 9.6 Gy and 22.4 Gy for cases and controls, respectively (P=.014). The mean carinal doses were 7.0 Gy and 9.2 Gy, respectively (P=.13). Mediastinal nodal relapses, with and without ipsilateral hilar relapse, were associated with mean ipsilateral hilar doses of 3.6 Gy and 19.8 Gy, respectively (P=.01). The conditional density plot appears to demonstrate an inverse dose-effect relationship between ipsilateral hilar normalized total dose and risk of ipsilateral hilar relapse. Conclusions: Incidental hilar dose greater than 20 Gy is significantly associated with fewer ipsilateral hilar relapses in inoperable early stage NSCLC patients treated with SBRT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2014

Interrater Reliability of the Categorization of Late Radiographic Changes After Lung Stereotactic Body Radiation Therapy

Purpose: Radiographic changes after lung stereotactic body radiation therapy (SBRT) have been categorized into 4 groups: modified conventional pattern (A), mass-like fibrosis; (B), scar-like fibrosis (C), and no evidence of increased density (D). The purpose of this study was to assess the interrater reliability of this categorization system in patients with early-stage non-small cell lung cancer (NSCLC). Methods and Materials: Seventy-seven patients were included in this study, all treated with SBRT for early-stage (T1/2) NSCLC at a single institution, with a minimum follow-up of 6 months. Six experienced clinicians familiar with post-SBRT radiographic changes scored the serial posttreatment CT images independently in a blinded fashion. The proportion of patients categorized as A, B, C, or D at each interval was determined. Krippendorff's alpha (KA), Multirater kappa (M-kappa), and Gwet's AC1 (AC1) scores were used to establish interrater reliability. A leave-one-out analysis was performed to demonstrate the variability among raters. Interrater agreement of the first and last 20 patients scored was calculated to explore whether a training effect existed. Results: The number of ratings ranged from 450 at 6 months to 84 at 48 months of follow-up. The proportion of patients in each category was as follows: A, 45%; B, 16%; C, 13%; and D, 26%. KA and M-kappa ranged from 0.17 to 0.34. AC1 measure range was 0.22 to 0.48. KA increased from 0.24 to 0.36 at 12 months with training. The percent agreement for pattern A peaked at 12 month with a 54% chance of having >50% raters in agreement and decreased over time, whereas that for patterns B and C increased over time to a maximum of 20% and 22%, respectively. Conclusion: This post-SBRT radiographic change categorization system has modest interrater agreement, and there is a suggestion of a training effect. Patterns of fibrosis evolve after SBRT and alternative categorization systems should be evaluated.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2013

The Impact of Tumor Size on Outcomes After Stereotactic Body Radiation Therapy for Medically Inoperable Early-Stage Non-Small Cell Lung Cancer

Purpose: Stereotactic body radiation therapy for medically inoperable early-stage non-small cell lung cancer (NSCLC) offers excellent control rates. Most published series deal mainly with small (usually <4 cm), peripheral, solitary tumors. Larger tumors are associated with poorer outcomes (ie, lower control rates, higher toxicity) when treated with conventional RT. It is unclear whether SBRT is sufficiently potent to control these larger tumors. We therefore evaluated and examined the influence of tumor size on treatment outcomes after SBRT. Methods and Materials: Between October 2004 and October 2010, 185 medically inoperable patients with early (T1-T2N0M0) NSCLC were treated on a prospective research ethics board-approved single-institution protocol. Prescription doses were risk-adapted based on tumor size and location. Follow-up included prospective assessment of toxicity (as per Common Terminology Criteria for Adverse Events, version 3.0) and serial computed tomography scans. Patterns of failure, toxicity, and survival outcomes were calculated using Kaplan-Meier method, and the significance of tumor size (diameter, volume) with respect to patient, treatment, and tumor factors was tested. Results: Median follow-up was 15.2 months. Tumor size was not associated with local failure but was associated with regional failure (P=.011) and distant failure (P=.021). Poorer overall survival (P=.001), disease-free survival (P=.001), and cause-specific survival (P=.005) were also significantly associated with tumor size (with tumor volume more significant than diameter). Gross tumor volume and planning target volume were significantly associated with grade 2 or worse radiation pneumonitis. However, overall rates of grade ≥3 pneumonitis were low and not significantly affected by tumor or target size. Conclusions: Currently employed stereotactic body radiation therapy dose regimens can provide safe effective local therapy even for larger solitary NSCLC tumors (up to 5.7 cm in tumor diameter or 100 cm{sup 3} in tumor volume) but are associated with more nonlocal failures as well as poorer survival. These observations suggest these patients may benefit from more extensive staging or consideration of adjuvant therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗