Initial clinical assessment of 'center-specific' automated treatment plans for low-dose-rate prostate brachytherapy
| Autor: | Ron S. Sloboda, John Doucette, John Amanie, Mustafa Ege Babadagli, Nawaid Usmani, Muhammad F. Jamaluddin, Don Yee, Albert Murtha |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Male
Organs at Risk medicine.medical_specialty Linear programming medicine.medical_treatment Brachytherapy Urinary Bladder Pilot Projects Plan (drawing) 030218 nuclear medicine & medical imaging Iodine Radioisotopes 03 medical and health sciences Consistency (database systems) 0302 clinical medicine Treatment plan medicine Humans Radiology Nuclear Medicine and imaging Medical physics Low dose rate Single institution Retrospective Studies business.industry Radiotherapy Planning Computer-Assisted Prostate implant Prostatic Neoplasms Radiotherapy Dosage Oncology 030220 oncology & carcinogenesis business Prostate brachytherapy |
| Zdroj: | Brachytherapy. 17(2) |
| ISSN: | 1873-1449 |
| Popis: | Purpose To report results of an initial pilot study assessing iodine-125 prostate implant treatment plans created automatically by a new seed-placement method. Methods and Materials A novel mixed-integer linear programming method incorporating spatial constraints on seed locations in addition to standard dose–volume constraints was used to place seeds. The approach, described in detail elsewhere, was used to create treatment plans fully automatically on a retrospective basis for 20 patients having a wide range of prostate sizes and shapes. Corresponding manual plans used for patient treatment at a single institution were combined with the automated plans, and all 40 plans were anonymized, randomized, and independently evaluated by five clinicians using a common scoring tool. Numerical and clinical features of the plans were extracted for comparison purposes. Results A full 51% of the automated plans were deemed clinically acceptable without any modification by the five practitioners collectively versus 90% of the manual plans. Automated plan seed distributions were for the most part not substantially different from those for the manual plans. Two observed shortcomings of the automated plans were seed strands not intersecting the prostate and strands extending into the bladder. Both are amenable to remediation by adjusting existing spatial constraints. Conclusions After spatial and dose–volume constraints are set, the mixed-integer linear programming method is capable of creating prostate implant treatment plans fully automatically, with clinical acceptability sufficient to warrant further investigation. These plans, intended to be reviewed and refined as necessary by an expert planner, have the potential to both save planner time and enhance treatment plan consistency. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|