Large Language Model-Augmented Auto-Delineation of Treatment Target Volume in Radiation Therapy.

Radiation therapy (RT) is one of the most effective treatments for cancer, and its success relies on the accurate delineation of targets. However, target delineation is a comprehensive medical decision that currently relies purely on manual processes by human experts. Manual delineation is time-consuming, laborious, and subject to interobserver variations. Although the advancements in artificial intelligence (AI) techniques have significantly enhanced the auto-contouring of normal tissues, accurate delineation of RT target volumes remains a challenge. In this study, we propose a visual language model-based RT target volume auto-delineation network termed Radformer. The Radformer utilizes a hierarchical vision transformer as the backbone and incorporates large language models to extract text-rich features from clinical data. We introduce a visual language attention module (VLAM) for integrating visual and linguistic features for language-aware visual encoding (LAVE). The Radformer has been evaluated on a dataset comprising 2985 patients with head-and-neck cancer who underwent RT. Metrics, including the Dice similarity coefficient (DSC), intersection over union (IOU), and 95th percentile Hausdorff distance (HD95), were used to evaluate the performance of the model quantitatively. Our results demonstrate that the Radformer has superior segmentation performance compared to other state-of-the-art models, validating its potential for adoption in RT practice.

Optimizing efficiency and safety in external beam radiotherapy using automated plan check (APC) tool and six sigma methodology.

PURPOSE: To develop and implement an automated plan check (APC) tool using a Six Sigma methodology with the aim of improving safety and efficiency in external beam radiotherapy. METHODS: The Six Sigma define-measure-analyze-improve-control (DMAIC) framework was used by measuring defects stemming from treatment planning that were reported to the departmental incidence learning system (ILS). The common error pathways observed in the reported data were combined with our departmental physics plan check list, and AAPM TG-275 identified items. Prioritized by risk priority number (RPN) and severity values, the check items were added to the APC tool developed using Varian Eclipse Scripting Application Programming Interface (ESAPI). At 9 months post-APC implementation, the tool encompassed 89 check items, and its effectiveness was evaluated by comparing RPN values and rates of reported errors. To test the efficiency gains, physics plan check time and reported error rate were prospectively compared for 20 treatment plans. RESULTS: The APC tool was successfully implemented for external beam plan checking. FMEA RPN ranking re-evaluation at 9 months post-APC demonstrated a statistically significant average decrease in RPN values from 129.2 to 83.7 (P < .05). After the introduction of APC, the average frequency of reported treatment-planning errors was reduced from 16.1% to 4.1%. For high-severity errors, the reduction was 82.7% for prescription/plan mismatches and 84.4% for incorrect shift note. The process shifted from 4σ to 5σ quality for isocenter-shift errors. The efficiency study showed a statistically significant decrease in plan check time (10.1 ± 7.3 min, P = .005) and decrease in errors propagating to physics plan check (80%). CONCLUSIONS: Incorporation of APC tool has significantly reduced the error rate. The DMAIC framework can provide an iterative and robust workflow to improve the efficiency and quality of treatment planning procedure enabling a safer radiotherapy process.

Variability in MRI vs. ultrasound measures of prostate volume and its impact on treatment recommendations for favorable-risk prostate cancer patients: a case series.

BACKGROUND: Prostate volume can affect whether patients qualify for brachytherapy (desired size ≥20 mL and ≤60 mL) and/or active surveillance (desired PSA density ≤0.15 for very low risk disease). This study examines variability in prostate volume measurements depending on imaging modality used (ultrasound versus MRI) and volume calculation technique (contouring versus ellipsoid) and quantifies the impact of this variability on treatment recommendations for men with favorable-risk prostate cancer. METHODS: We examined 70 patients who presented consecutively for consideration of brachytherapy for favorable-risk prostate cancer who had volume estimates by three methods: contoured axial ultrasound slices, ultrasound ellipsoid (height × width × length × 0.523) calculation, and endorectal coil MRI (erMRI) ellipsoid calculation. RESULTS: Average gland size by the contoured ultrasound, ellipsoid ultrasound, and erMRI methods were 33.99, 37.16, and 39.62 mLs, respectively. All pairwise comparisons between methods were statistically significant (all p < 0.015). Of the 66 patients who volumetrically qualified for brachytherapy on ellipsoid ultrasound measures, 22 (33.33%) did not qualify on ellipsoid erMRI or contoured ultrasound measures. 38 patients (54.28%) had PSA density ≤0.15 ng/dl as calculated using ellipsoid ultrasound volumes, compared to 34 (48.57%) and 38 patients (54.28%) using contoured ultrasound and ellipsoid erMRI volumes, respectively. CONCLUSIONS: The ultrasound ellipsoid and erMRI ellipsoid methods appeared to overestimate ultrasound contoured volume by an average of 9.34% and 16.57% respectively. 33.33% of those who qualified for brachytherapy based on ellipsoid ultrasound volume would be disqualified based on ultrasound contoured and/or erMRI ellipsoid volume. As treatment recommendations increasingly rely on estimates of prostate size, clinicians must consider method of volume estimation.

Use of a rectal spacer with low-dose-rate brachytherapy for treatment of prostate cancer in previously irradiated patients: Initial experience and short-term results.

BACKGROUND: Salvage brachytherapy in patients with prior pelvic radiation carries a risk of rectal injury. Herein, we report our initial experience using a hydrogel spacer between the prostate and the rectum during salvage brachytherapy. METHODS AND MATERIALS: A total of 11 patients with prostate cancer and prior radiotherapy (5 prostate brachytherapy, 2 prostate external beam radiation therapy [EBRT], and 4 rectal cancer EBRT) received (125)I brachytherapy after attempted placement of 10cc of a diluted hydrogel spacer between the prostate and rectum. RESULTS: Spacing was achieved in 8 of the 11 (73%) patients but was not possible in 3 (1 prior brachytherapy and 2 prior EBRT) owing to fibrosis and adhesions. For the 8 patients in whom spacing was accomplished, the median space between the prostate and rectum was 10.9mm (prior EBRT) vs. 7.7mm (prior brachytherapy), p=0.048. Median followup was 15.7 months. One patient developed a prostato-rectal fistula requiring a diverting colostomy. The 16-month estimate of late Grade 3 or 4 gastrointestinal or genitourinary toxicity was 26%. One patient developed lymph node-positive recurrence. The 16-month prostate-specific antigen failure-free survival rate was 89%. Compared with baseline, Expanded Prostate Cancer Index Composite for Clinical Practice urinary quality of life (QoL) was significantly worse at 3 and 6 months but not significantly worse by 1 year. There were no significant changes throughout the study period in bowel or sexual QoL. CONCLUSION: Hydrogel spacer placements may be feasible in most patients with prior pelvic radiation. Further followup is needed to determine whether spacer placement will produce long-term improvements in toxicity or QoL.

Placement of empty catheters for an HDR-emulating LDR prostate brachytherapy technique: comparison to standard intraoperative planning.

PURPOSE: We sought to determine whether placing empty catheters within the prostate and then inverse planning iodine-125 seed locations within those catheters (High Dose Rate-Emulating Low Dose Rate Prostate Brachytherapy [HELP] technique) would improve concordance between planned and achieved dosimetry compared with a standard intraoperative technique. METHODS AND MATERIALS: We examined 30 consecutive low dose rate prostate cases performed by standard intraoperative technique of planning followed by needle placement/seed deposition and compared them to 30 consecutive low dose rate prostate cases performed by the HELP technique. The primary endpoint was concordance between planned percentage of the clinical target volume that receives at least 100% of the prescribed dose/dose that covers 90% of the volume of the clinical target volume (V100/D90) and the actual V100/D90 achieved at Postoperative Day 1. RESULTS: The HELP technique had superior concordance between the planned target dosimetry and what was actually achieved at Day 1 and Day 30. Specifically, target D90 at Day 1 was on average 33.7 Gy less than planned for the standard intraoperative technique but was only 10.5 Gy less than planned for the HELP technique (p < 0.001). Day 30 values were 16.6 Gy less vs. 2.2 Gy more than planned, respectively (p = 0.028). Day 1 target V100 was 6.3% less than planned with standard vs. 2.8% less for HELP (p < 0.001). There was no significant difference between the urethral and rectal concordance (all p > 0.05). CONCLUSIONS: Placing empty needles first and optimizing the plan to the known positions of the needles resulted in improved concordance between the planned and the achieved dosimetry to the target, possibly because of elimination of errors in needle placement.