Katharin Deschesne Burkhardt

Quantitative assessment of workload and stressors in clinical radiation oncology

PURPOSE Workload level and sources of stressors have been implicated as sources of error in multiple settings. We assessed workload levels and sources of stressors among radiation oncology professionals. Furthermore, we explored the potential association between workload and the frequency of reported radiotherapy incidents by the World Health Organization (WHO). METHODS AND MATERIALS Data collection was aimed at various tasks performed by 21 study participants from different radiation oncology professional subgroups (simulation therapists, radiation therapists, physicists, dosimetrists, and physicians). Workload was assessed using National Aeronautics and Space Administration Task-Load Index (NASA TLX). Sources of stressors were quantified using observational methods and segregated using a standard taxonomy. Comparisons between professional subgroups and tasks were made using analysis of variance ANOVA, multivariate ANOVA, and Duncan test. An association between workload levels (NASA TLX) and the frequency of radiotherapy incidents (WHO incidents) was explored (Pearson correlation test). RESULTS A total of 173 workload assessments were obtained. Overall, simulation therapists had relatively low workloads (NASA TLX range, 30-36), and physicists had relatively high workloads (NASA TLX range, 51-63). NASA TLX scores for physicians, radiation therapists, and dosimetrists ranged from 40-52. There was marked intertask/professional subgroup variation (P<.0001). Mental demand (P<.001), physical demand (P=.001), and effort (P=.006) significantly differed among professional subgroups. Typically, there were 3-5 stressors per cycle of analyzed tasks with the following distribution: interruptions (41.4%), time factors (17%), technical factors (13.6%), teamwork issues (11.6%), patient factors (9.0%), and environmental factors (7.4%). A positive association between workload and frequency of reported radiotherapy incidents by the WHO was found (r = 0.87, P value=.045). CONCLUSIONS Workload level and sources of stressors vary among professional subgroups. Understanding the factors that influence these findings can guide adjustments to the workflow procedures, physical layout, and/or communication protocols to enhance safety. Additional evaluations are needed in order to better understand if these findings are systemic.

Empirical Evaluation of Workload of the Radiation Oncology Physicist During Radiation Treatment Planning and Delivery

In recent years, the practice of radiation oncology has changed due to several technological advances. As such, there is growing interest in the evolving nature of safety and operational challenges faced by radiation oncology professionals. This research focuses on physicists who play an important role in the radiation therapy treatment planning and delivery process. Specifically, the purpose of our research is to assess their workload levels using the NASA TLX method in order to identify tasks that might compromise patient safety. Based on empirical observations, this study provides practical suggestions for lowering workload levels that ultimately can reduce the probability of errors.

SU‐E‐T‐485: Comparison of the Oncogenic Potential for Radiation‐Associated, Second Malignant Neoplasms for Several Prostate Radiotherapy Modalities as a Function of Relative OAR & PTV Volumes

Purpose: To compare plan quality for a comprehensive set of OAR & PTV geometries considering relative risk estimates for the induction of secondary bladder and rectal cancers due to prostate radiotherapy using several modern delivery techniques. Methods: IMRT / 3D‐conformal, volumetric‐arc (VMAT), and Tomotherapy (TT) plans were generated using PlanUNC, Pinnacle3 ver. 9.2, and TomoTherapy treatment planning systems, respectively, for 12 prostate cases using, RTOG 0815‐derived planning goals. Selected cases were representative of relative (i.e. small (Svol), medium (Mvol), large (Lvol)) PTV, rectal & bladder structure volumes encountered clinically. PlanUNC was used to calculate the elevated relative‐risk (ERR) of secondary cancer induction in normal tissues, the bladder and rectal ERR, PTV conformity index (CI), dose heterogeneity index (DHI), and bladder and rectal V5, V15, V35, V50 (Bvx / Rvx) for all cases. All metrics were averaged for each relative tissue / target volume, and normalized to corresponding 3D‐conformal cases for comparison. Results: Considering relative rectal volumes, all metrics fell within the standard error for each modality excluding bladder ERR (TT 13%ave, IMRT −9% ave & VMAT −18% ave) and BV15 (VMAT −24% ave) for Lvol rectum cases. Considering relative bladder volumes, bladder ERR (TT 4%, IMRT −10%, VMAT −18%) for Lvol and (VMAT −20%) for Mvol; rectal ERR (TT −40%) for Mvol were observed to be significant in addition to bladder ERR (TT 13%) for Svol PTV cases. Generally, all modalities performed better when compared to 3D conformal delivery excluding nominal (<5%) increases in normal tissue dose and increased bladder ERR for several TT cases. Conclusion: We compared three prostate EBRT modalities considering OAR ERR as well as traditional plan quality metrics for a class of cases defined by relative structure size. Preliminary results indicate comparable improvements in PTV conformity & OAR dosing for each modality but differences in OAR ERR.

SU‐E‐T‐717: An Automatic Self‐Customizable Tool to Analyze the Dosimetric Quality of a Radiation Therapy Plan

Purpose: The review of radiation therapy plan is a crucial step for the quality control. It can be time consuming when many anatomic structures are involved and plan needs multiple revisions. The purpose of this work is to develop a real time, automatic and self‐customizable tool to evaluate the dosimetric quality against the protocol. This will improve the efficiency of plan design and review and reduce the human error. Methods: We developed a software tool to import DVH, evaluate the dosimetric quality and export the QA report. Current version handles the DVH formats of three main treatment planning softwares (Eclipse, Varian; Pinnacle, Philips; TomoTherapy, Accuray). A rule language is created to describe the type of dosimetric constraints such as the percentage volume of percentage dose and etc. The protocol is written in ASCII format which allows the easy modification and the creation of a new one. The name of center, protocol, radiation oncologist and planner are in an editable configuration file. All of these design features make the tool easily customized to fulfill the protocols and report format of individual treatment center. Structure association between the protocol and clinic plan is programmed to be automatic as well. 12 patients and 3 protocols were used in the test of the software. Results: The dosimetric report generated with the new software tool agrees with the report manually produced. Three protocols, head and neck, lung and prostate, were used by the software with no error. Program associated 70% structures correctly between the standard names in the protocol and operator defined‐names in the clinical plan. Conclusion: We developed an automatic dosimetric quality evaluation tool to improve the efficiency of plan design and review. It allows the users to customize their dosimetric protocol and report format without the need of changing the software.

SU-E-T-188: An Effective Quality Assurance Method for TomoTherapy Craniospinal Irradiation Patients with A 3D Semiconductor Phantom

PURPOSE This study aims to develop an effective and efficient approach to perform delivery quality assurance (DQA) to any large length spine in Tomo craniospinal irradiation (CSI) with a 3D semiconductor phantom. METHODS It is challenging to use conventional method to perform DQA to an entire large length of spine because the Tomo machine limits the maximum longitudinal distance between red and green laser (=18 cm). In order to overcome this limit, three DQA plans for each CSI patient were generated for covering the entire treatment site, in which we intentionally overlapped the red and green laser in DQA calculation, then we measured the distance (d) between the centers of the phantom and lasers. When using SunNucelar QA software to compute the plannar dose, we set the center of the cylindrical surface by applying d to the IEC Z coordinate. In phantom setup, we first align the center of phantom with green laser then manually shift the phantom superiorly or inferiorly d. With this new approach, we measured five CSI patients using a commercial 3D semiconductor phantom, the ArcCHECK (SunNuclear). The measured and computed dose is compared with gamma analysis and (3%, 3 mm) agreement criteria. RESULTS For all CSI patients and sections measured, the acquired dose distribution covered the complete treatment length and matched the computed data very well. All the DQA passed the gamma criteria with an average passing rate of 99.1%. The total delivery time of DQA using the new method was 66.7% of the time using the conventional method. CONCLUSION The measured DQA results from the five CSI patients demonstrate that this new method overcomes the distance limitation of 18 cm between the red and green laser in the conventional DQA method. The DQA delivery becomes more efficient with 33.3% reduction of beam-on time.

SU‐E‐T‐635: Effect of Planning Parameters on Tomotherapy Dosimetric Quality and Treatment Efficiency

PURPOSE To investigate how the setting of optimization parameters, fractional dose and tuning structure in tomotherapy treatment planning affects plan dosimetric quality and treatment efficiency. METHODS A digital phantom to simulate head and neck radiotherapy was constructed for this study. Tumor was 10cm long C-shaped with two surrounding parallel normal structures (P-NS) and one serial normal structure (S-NS). Dose prescription was 54 Gy in total. Fourteen treatment plans were generated with varied parameters in five categories: a) jaw size (1 to 5cm), b) pitch (0.215 to 0.43), c) modulation factor (1.5 to 4), d) dose per fraction (100 to 600cGy) and e) whether to use tuning structure. Plans were compared with multiple dosimetric endpoints including target minimum/maximum/mean dose, V100%, conformity, heterogeneity, S-NS maximum dose, P-NS and body mean dose, and treatment times. The reference plan was defined for the plan with conventional parameters: jaw 2.5cm, pitch 0.287, modulation factor 3.0, 200cGy per fraction and use of a 2cm ring structure in optimization. RESULTS Compared with 2.5cm jaw reference plan, 1cm jaw plan decreased the mean body dose 10.7% while 5 cm jaw plan increased the dose 17.9%. Smaller pitch (p=0.215) made the plan more conform than reference plan, and bigger pitch (p=0.43) had opposite effect. A small modulation factor (M=1.5) failed to spare critical structures. A medium modulation factor (M=2) resulted in similar plan to the reference but with 29% less treatment time. A low fractional dose (100 cGy) planned with similar parameter as reference had much inferior target coverage (V100%=85.6% vs V100%=96.4). Lastly, the use of tuning structure improved the conformity of target. CONCLUSIONS Selection of optimization parameters in tomotherapy treatment planning affects target coverage, critical structure sparing, body dose, and treatment time. Target coverage is compromised if fractional dose is low to the range of 100 cGy.