Prospective validation of a machine learning model for applicator and hybrid interstitial needle selection in high-dose-rate (HDR) cervical brachytherapy

PURPOSE

To Demonstrate the clinical validation of a machine learning (ML) model for applicator and interstitial needle prediction in gynecologic brachytherapy through a prospective clinical study in a single institution.

METHODS

The study included cervical cancer patients receiving high-dose-rate brachytherapy using intracavitary (IC) or hybrid interstitial (IC/IS) applicators. For each patient, the primary radiation oncologist contoured the high-risk clinical target volume on a pre-brachytherapy MRI, indicated the approximate applicator location, and made a clinical determination of the first fraction applicator. A pre-trained ML model predicted the applicator and IC/IS needle arrangement using tumor geometry. Following the first fraction, ML and radiation oncologist predictions were compared and a replanning study determined the applicator providing optimal organ-at-risk (OAR) dosimetry. The ML-predicted applicator and needle arrangement and the clinical determination were compared to this dosimetric ground truth.

RESULTS

Ten patients were accrued from December 2020 to October 2022. Compared to the dosimetrically optimal applicator, both the radiation oncologist and ML had an accuracy of 70%. ML demonstrated better identification of patients requiring IC/IS applicators and provided balanced IC and IC/IS predictions. The needle selection model achieved an average accuracy of 82.5%. ML-predicted needle arrangements matched or improved plan quality when compared to clinically selected arrangements. Overall, ML predictions led to an average total improvement of 2.0 Gy to OAR doses over three treatment fractions when compared to clinical predictions.

CONCLUSION

In the context of a single institution study, the presented ML model demonstrates valuable decision-support for the applicator and needle selection process with the potential to provide improved dosimetry. Future work will include a multi-center study to assess generalizability.

The impact of advancing the standard of care in radiotherapy on operational treatment resources

PURPOSE

To demonstrate the impact of implementing hypofractionated prescription regimens and advanced treatment techniques on institutional operational hours and radiotherapy personnel resources in a multi-institutional setting. The study may be used to describe the impact of advancing the standard of care with modern radiotherapy techniques on patient and staff resources.

METHODS

This study uses radiation therapy data extracted from the radiotherapy information system from two tertiary care, university-affiliated cancer centers from 2012 to 2021. Across all patients in the analysis, the average fraction number for curative and palliative patients was reported each year in the decade. Also, the institutional operational treatment hours are reported for both centers. A sub-analysis for curative intent breast and lung radiotherapy patients was performed to contextualize the impact of changes to imaging, motion management, and treatment technique.

RESULTS

From 2012 to 2021, Center 1 had 42 214 patient plans and Center 2 had 43 252 patient plans included in the analysis. Averaged over both centers across the decade, the average fraction number per patient decreased from 6.9 to 5.2 (25%) and 21.8 to 17.2 (21%) for palliative and curative patients, respectively. The operational treatment hours for both institutions increased from 8 h 15 min to 9 h 45 min (18%), despite a patient population increase of 45%.

CONCLUSION

The clinical implementation of hypofractionated treatment regimens has successfully reduced the radiotherapy workload and operational treatment hours required to treat patients. This analysis describes the impact of changes to the standard of care on institutional resources.

Accelerated Partial Breast Irradiation Using 5 Daily Fractions: A Prospective, Phase 2, Multicenter Trial of Cosmetic Outcomes and Toxicity-ACCEL Final Results

PURPOSE

This study investigated the cosmetic degradation and toxicity for an accelerated partial breast irradiation (APBI) prescription delivered in 5 fractions over 1 week and compared the outcomes with those of whole breast irradiation (WBI).

METHODS AND MATERIALS

The trial was a multicenter, single-arm, phase 2 prospective cohort study. Eligible women 50 years of age or older with estrogen receptor-positive and human epidermal growth factor receptor 2-negative invasive ductal carcinoma or ductal carcinoma in situ after breast-conserving surgery received 27 Gy in 5 daily fractions of APBI. The primary endpoint was noninferiority of 2-year cosmesis using the RAPID trial's WBI arm as the control arm. A global consensus cosmetic score using a European Organisation for Research and Treatment of Cancer rating scale score of excellent, good, fair, or poor for each patient at baseline and 2 years was generated by a panel of 5 radiation oncologists using photographs of treated and untreated breasts.

RESULTS

From 2016 to 2019, 298 eligible women were enrolled. By the 2-year follow-up, 76 patients had been lost or withdrawn and 3 had died, resulting in 219 patients available for complete, 2-year photographic cosmetic evaluation. The median follow-up for all participants was 4.7 years (IQR, 3.8-5.5 years). No patient had a fair or poor cosmetic score at the 2-year evaluation. Cosmesis was better or unchanged for 97% of patients and worse for 3% (excellent to good), and no cosmetic failures occurred. The confidence intervals were 0.88 (0.86-0.90) and 1.00 (0.99-1.00) for the RAPID and ACCEL trials, respectively.

CONCLUSIONS

Cosmetic degradation with 5 daily treatments of the ACCEL trial's APBI intervention is noninferior to the WBI arm of the RAPID trial.

Prospective deployment of an automated implementation solution for artificial intelligence translation to clinical radiation oncology

Introduction

Artificial intelligence (AI)-based technologies embody countless solutions in radiation oncology, yet translation of AI-assisted software tools to actual clinical environments remains unrealized. We present the Deep Learning On-Demand Assistant (DL-ODA), a fully automated, end-to-end clinical platform that enables AI interventions for any disease site featuring an automated model-training pipeline, auto-segmentations, and QA reporting.

Materials and methods

We developed, tested, and prospectively deployed the DL-ODA system at a large university affiliated hospital center. Medical professionals activate the DL-ODA via two pathways (1): On-Demand, used for immediate AI decision support for a patient-specific treatment plan, and (2) Ambient, in which QA is provided for all daily radiotherapy (RT) plans by comparing DL segmentations with manual delineations and calculating the dosimetric impact. To demonstrate the implementation of a new anatomy segmentation, we used the model-training pipeline to generate a breast segmentation model based on a large clinical dataset. Additionally, the contour QA functionality of existing models was assessed using a retrospective cohort of 3,399 lung and 885 spine RT cases. Ambient QA was performed for various disease sites including spine RT and heart for dosimetric sparing.

Results

Successful training of the breast model was completed in less than a day and resulted in clinically viable whole breast contours. For the retrospective analysis, we evaluated manual-versus-AI similarity for the ten most common structures. The DL-ODA detected high similarities in heart, lung, liver, and kidney delineations but lower for esophagus, trachea, stomach, and small bowel due largely to incomplete manual contouring. The deployed Ambient QAs for heart and spine sites have prospectively processed over 2,500 cases and 230 cases over 9 months and 5 months, respectively, automatically alerting the RT personnel.

Discussion

The DL-ODA capabilities in providing universal AI interventions were demonstrated for On-Demand contour QA, DL segmentations, and automated model training, and confirmed successful integration of the system into a large academic radiotherapy department. The novelty of deploying the DL-ODA as a multi-modal, fully automated end-to-end AI clinical implementation solution marks a significant step towards a generalizable framework that leverages AI to improve the efficiency and reliability of RT systems.

An Analysis of Clinical and Systemic Factors Associated with Palliative Radiotherapy Delivery and Completion at the End of Life in Alberta, Canada

Radiotherapy (RT) is often utilized for symptom control at the end of life. Palliative RT (pRT) may not be taken to completion by patients, thus decreasing clinical benefits and adversely impacting resource allocation. We determined rates of incomplete pRT and examined predictors of non-completion using an electronic questionnaire. Methods: A questionnaire was embedded within the RT electronic prescribing system for all five cancer centers of Alberta, Canada, between 2017 and 2020. Prescribing radiation oncologists (ROs) were tasked with completing the questionnaire. Treatment variables were collected for 2040 patients prescribed pRT. Details on pRT courses delivered and completed were used to determine rates of incomplete RT. Electronic medical records of a subset of 367 patients randomly selected from the 2040 patients were then analyzed to examine for association of non-completion of RT with patient, disease, and therapy-related factors. Results: Overall, 10% of patients did not complete pRT. The rate of single fractions prescribed as a proportion of all RT fractions increased from 18% (pre-2017: pre-study era) to 29% (2017-2020: study era) (p < 0.0001). After conducting multivariate analysis on the overall group, multiple lifetime malignancies (OR:0.64) or increasing the number of pRT fractions (OR:0.08-0.17) were associated with non-completion. Being selected for stereotactic RT (OR:3.75) or survival > 30 days post-RT prescription (OR:2.20-5.02) were associated with greater rates of RT completion. The ROs' estimates of life expectancy at the time of RT prescription were not predictive of RT completion. In the multivariate analysis of the 367-patient subset, the presence of hepatic metastases (OR 2.59), survival 30-59 days (OR 6.61) and survival 90+ days (OR 8.18) post-RT prescription were associated with pRT completion. Only increasing pRT fractionation (OR:0.05-0.2) was associated with non-completion. Conclusion: One in ten patients prescribed pRT did not complete their treatment course. Decreasing pRT fractionation and improving prognostication in patients near the end of life may decrease rates of incomplete RT courses.

Clinical implementation of 3D transvaginal ultrasound for intraoperative guidance of needle implant in template interstitial gynecologic high-dose-rate brachytherapy

PURPOSE

To demonstrate novel clinical implementation of a 3D transvaginal ultrasound (3DTVUS) system for intraoperative needle insertion guidance in perineal template interstitial gynecologic high-dose-rate brachytherapy and assess its impact on implant quality.

METHODS AND MATERIALS

Interstitial implants began with preimplant 3DTVUS to visualize the tumor and anatomy, with intermittent 3DTVUS to assess the implant and guide needle adjustment. Analysis includes visualization of the implant relative to anatomy, identification of cases where 3DTVUS is beneficial, dosimetry, and a survey distributed to 3DTVUS clinicians.

RESULTS

Seven patients treated between November 2021 and October 2022 were included in this study. Twenty needles were inserted under 3DTVUS guidance. The tumor and vaginal wall were well-differentiated in four and all seven patients, respectively. Patients with tumours below the superior aspect of the vagina are suited for 3DTVUS. Four radiation oncologists responded to the survey. There was general agreement that 3DTVUS improves implant and anatomy visualization and is preferred over standard 2D ultrasound guidance techniques.

CONCLUSIONS

Based on qualitative feedback from primary users and a small preliminary patient cohort, 3DTVUS imaging improves tumor and vaginal wall visualization during gynecologic perineal template interstitial needle implant and is a powerful tool for implant assessment in an intraoperative setting.

Microwave imaging for monitoring breast cancer treatment: A pilot study

BACKGROUND

Microwave imaging has been proposed for medical applications, creating maps related to water content of tissues. Breast imaging has emerged as a key application because the signals can be coupled directly into the breast and experience limited attenuation in fatty tissues. While the literature contains reports of tumor detection with microwave approaches, there is limited exploration of treatment monitoring.

PURPOSE

This study aims to detect treatment-related changes in breast tissue with a low-resolution microwave scanner.

METHODS

Microwave scans of 15 patients undergoing treatment for early-stage breast cancer are collected at up to 4 time points: after surgery (baseline), 6 weeks after accelerated partial breast radiation, as well as 1 and 2 years post-treatment. Both the treated and untreated breast are scanned at each time point. The microwave scanner consists of planar transmit and receive arrays and uses signals from 0.1 to 10 GHz. The average microwave frequency properties (permittivity) are calculated for each scan to enable quantitative comparison. Baseline and 6-week results are analyzed with a two-way ANOVA with blocking.

RESULTS

Consistent properties are observed for the untreated breast over time, similar to a previous study. Comparison of the scans of the treated and untreated breast suggests increased properties related to treatment, particularly at baseline and 6-weeks following radiotherapy. Analysis of the average properties of the scans with ANOVA indicates statistically significant differences (p < 0.05 $p < 0.05$ ) in the treated and untreated breast at these time points.

CONCLUSIONS

Microwave imaging has the potential to track treatment-related changes in breast tissues.

Erectile Dysfunction Pharmacotherapy Utilization after 60Gy in 20 Fractions Volumetric Modulated Arc Therapy to the Prostate

PURPOSE/OBJECTIVE(S)

To determine which factors predict for worsening erectile function after highly conformal, modestly hypofractionated radiotherapy to the prostate.

MATERIALS/METHODS

All patients who received 60Gy in 20 fractions, volumetric modulated arc therapy to the prostate across 4 centers over 9 years were included in this study. The provincial electronic medical record was interrogated to identify any new prescriptions for erectile dysfunction (ED) medication, any change in prescription of ED medication or any permanent discontinuance of ED medication persisting beyond 6 months post completion of any androgen deprivation therapy. The penile bulb, penile crux and penile shaft structures were retrospectively contoured. A Youden receiver-operator-curve analysis, logistic regression, and neural network based interpretable machine learning analysis were then used to determine dependencies between worsening ED and clinical factors including mean doses to these structures.

RESULTS

Two-hundred-twelve patients with median (inter-quartile-range) follow-up of 3.6 (3.2-4.4) years were identified. Median age was 72 (67-76) years. 104 (49%) patients received androgen deprivation therapy. Prior to treatment, 52 (25%) patients were on ED medication: 20 (9%) on sildenafil, 28 (13%) on tadalafil and 4 (8%) on vardenafil. Median PTV volume was 158.9 (129.8-192.1) cc. Median penile bulb, penile crux and penile shaft volumes were 4.7 (3.6-6.2) cc, 6.5 (5.1-8.5) cc and 93.3 (80.6-106.2) cc, respectively. PTV V95 was 99.8 (99.5-99.9)%. Mean doses to penile bulb, penile crux and penile shaft were 2094.8 (1306.2-3036.3) cGy, 2094.8 (1306.2-3036.3) cGy and 444.4 (313.2-650.5), respectively. Fifty-nine (28%) patients had a worsening of ED after treatment: 25 (12%) started a new ED medication, 6 (3%) had a prescription change and 28 (13%) stopped ED medication. On univariate analyses pretreatment use of ED medication predicted for worsening ED: odds ratio (OR) yes vs no: 10.2 (5.0 - 20.8; p<0.001). A trend towards mean dose to penile bulb [OR ≤2343.9 vs >2343.9: 1.7 (0.9-3.2; p = 0.08)] predicting for worsening ED was observed. Mean doses to penile crux [OR <1725.8 vs > 1725.8: 2.6 (1.3-5.2; p = 0.005)] and penile shaft [OR ≤344.9 vs >344.9: 5.2 (2.2-12.2; p<0.001)] predicted for worsening ED. Use of androgen deprivation therapy, and age at time of radiotherapy were not predictive of worsening ED. On multivariate analysis, only mean dose to penile shaft [OR ≤344.9 vs >344.9: 6.3 (1.9-20.3; p = 0.002)] and pretreatment use of ED medication [OR yes vs no: 11.1 (5.3-23.2; p<0.001)] predicted for worsening ED. A neural network analysis suggested that penile shaft mean dose and pre-treatment ED medication use are the most important factors in predicting worsening ED.

CONCLUSION

In this limited analysis, pre-treatment use of ED medication and mean dose to penile shaft predicted for worsening ED after treatment with modestly hypofractionated radiotherapy for prostate cancer.

Assessment of dose to vaginal mucosa for gynecologic template interstitial high-dose-rate brachytherapy using Monte Carlo simulation

Purpose

This study investigated reliable vaginal mucosa dose-volume histogram (DVH) metrics in gynecologic template interstitial high-dose-rate brachytherapy (HDR-BT) for the purpose of standardized dose reporting.

Material and methods

Gynecologic template (Syed/Neblett) interstitial HDR-BT patients treated from September 2016 to November 2022 at the study institute were included in the cohort. Each patient implant included a vaginal mucosa contour defined by a 5 mm expansion from vaginal cylinder, then another volume with clinical target volume subtracted. DVH metrics were investigated between D0.1cc to D4cc. Clinical plans were re-calculated using Monte Carlo (MC) simulations both in heterogeneous material and in water.

Results

The patient cohort included 61 patients with clinical plans using conventional homogeneous dose calculation (TG43). Heterogeneous vs. water MC dose differences were between -1.1% and -1.4% for all metrics investigated. DVH metrics D1cc and smaller resulted in > 5% discrepancies between TG43 and MC dose (to water) calculation due to the proximity of source positions in/nearby the vaginal mucosa. Reliability improved when DVH metric volume was larger (D2cc and D4cc). Both D2cc and D4cc presented very high linear correlation between TG43 and MC reported doses for the vagina, and average ± standard deviation dose difference was 4.6 ±2.9% and -3.0 ±1.9%, respectively. Dose differences decreased when the clinical target volume was removed: -1.5 ±3.5% and -0.8 ±2.1% for D2cc and D4cc, respectively.

Conclusions

For perineal template gynecologic HDR-BT procedures, the 2 cc volume is the smallest representative volume that reliably reports vaginal dose and at minimum should be reported to establish dose and outcome evaluation.

The Impact of Cardiac Sparing Techniques on Major Cardiac Events for Patients Undergoing Breast Radiotherapy

PURPOSE/OBJECTIVE(S)

In radiotherapy, studies with historical estimates have shown a linear relationship between cardiac dose and risk of major cardiac events with no safe dose threshold. This study investigates the relationship between cardiac dose and event rates for patients undergoing breast radiotherapy with modern heart-sparing techniques, where substantially lower cardiac doses are achieved.

MATERIALS/METHODS

From 2008 to 2018, all patients undergoing curative whole breast or chest wall radiotherapy at a tertiary care center were included. In this study, the deep-inspiration breath hold (DIBH) technique was implemented in 2013 as the standard of care for left-sided patients. Cardiac dosimetric data from the clinical treatment plan was extracted and matched to patients from the ongoing, prospectively collected Alberta Provincial Project for Outcome Assessment in Coronary Heart disease (APPROACH) database that experienced a major adverse cardiac event (MACE). MACE was defined as myocardial infarction, hospitalization for unstable angina, hospitalization for heart failure, or clinically indicated coronary revascularization. For all the left- and right-sided patients, the population event rates, the median time to event, and the average (± standard deviation) of the mean heart physical dose is reported. The proportion of left- and right-sided patients experiencing a major cardiac event were tested for independence using a two-sample z-test for proportions. A student's t-test was used to measure significance for mean heart dose. An alpha of 0.05 was used to indicate significance for both tests.

RESULTS

A total of 4893 patients received breast radiotherapy with 2429 left-sided, 2429 right-sided, and 35 synchronous bilateral. For all patients, Table 1 describes laterality, treatment characteristics, and prescription. The median (range) age at the time of first radiotherapy treatment was 59 (23-94). Of these patients with contours available, 96 left-sided (3.9%) and 69 right-sided (2.8%) patients experienced a MACE (p = 0.03) with a mean heart dose of 1.4 (±0.7) Gy and 0.5 (±0.2) Gy, in the event cohorts respectively (p < 0.001). The mean time to event was 4.7 (±3.2) years and 4.8 (±3.0) years for left- and right-sided patients, respectively (p = 0.83). For the left-sided cohort treated with DIBH, the mean heart dose is 1.0 (±0.8) Gy.

CONCLUSION

In the era of modern radiotherapy and cardiac-sparing techniques, mean heart dose and corresponding event rates are still significantly higher for left-sided patients than right-sided patients. Future work will aim to test whether there is a safe threshold for cardiac dose in the context of these results.

Radiation Treatment Preparation Safety Risk Prediction

PURPOSE/OBJECTIVE(S)

To determine the effect of delays in radiation treatment preparation task timeliness on reported safety events.

MATERIALS/METHODS

For 2022, all patients treated at four community-based radiation oncology clinics were included for analysis. Safety reports which occurred during treatment preparation tasks (i.e., between CT simulation and the first treatment) were included for this analysis. Patients with a reported safety event were the report cohort and those without were the control cohort. Treatment preparation tasks are completed sequentially and included contouring, treatment planning, Radiation Oncologist review, Medical Physicist review, Therapist quality check. At baseline, each task was scheduled a standard interval for completion based on treatment technique. The time taken to complete each task was captured using ARIA v16.0 Care Path module (Varian Medical Systems, Palo Alto, CA) and extracted through database query. For each task, two metrics were quantified: Task (1) Completion Timeliness: the time (hours) to complete each task, relative to the time allocated to each task. A negative value indicated more time taken than scheduled. (2) Overall Timeliness: specific task timeliness (hours) relative to overall timeliness for the whole Care Path. A negative value indicated the tasks were behind schedule. A student's t-test with an alpha of 0.05 was used to indicate significance.

RESULTS

Over the study period, 66 courses had a reported treatment preparation safety event (report) and 2690 did not (control). Techniques include 54% and 56% 3DCRT, 35% and 35% IMRT, 3% and 3% SRS, and 6% and 8% SBRT, for the report and control groups, respectively. Disease sites include breast, GU, GI, head and neck, CNS, thoracic, skin, secondary, gynecological, sarcoma, and heme. Table 1 displays the timeliness for each of the five tasks. Of the courses with safety events, overall timeliness was significantly behind the control group from the in contouring, treatment planning, Radiation Oncologist review, and Medical Physicist review. Courses with safety events took significantly more time than scheduled early in treatment preparation with a significant increase at treatment planning and significantly less time towards the end with the therapist quality check.

CONCLUSION

Patient courses with reported safety events had significantly longer treatment planning times, and as a result inadequate time to perform each downstream pre-treatment activity. The implication of this analysis is that a flag could be created to indicate risk of a safety event early in the treatment preparation process.

Prospective Application of an Artificial Intelligence Decision Support Tool for Applicator and Needle Selection in High-Dose-Rate Brachytherapy for Cervical Cancer Patients

PURPOSE/OBJECTIVE(S)

To assess the performance of an artificial intelligence (AI)-based decision support tool under differing thresholds to determine optimal applicator selection in a prospective clinical setting.

MATERIALS/METHODS

Cervical cancer patients scheduled for high-dose-rate brachytherapy implants with intracavitary tandem and ring (IC) or interstitial tandem and ring (IC/IS) implants in a single tertiary cancer center were eligible for enrollment. Prior to the first brachytherapy implant, a diagnostic MRI was acquired. The clinical target volume and the expected inserted position of the intrauterine tandem and ring were identified and contoured on the T2-MRI. An in-house artificial intelligence-based predicted the need for an IC or IC/IS implant, based on target volume geometric features. For IC/IS implants, an optimal needle arrangement for target coverage was also predicted. Blinded to the AI outcome, a clinical determination was made by the clinician reviewing the MR image. AI algorithm prediction provided a confidence level associated with each decision. The algorithm performance for different confidence thresholds using the IC applicator of 50%, 60%, and 80% was investigated. Performance metrics of the initial clinical determination and the AI prediction were calculated based on the consensus optimal applicator determined from an assessment of planning dosimetry in the first fraction and clinical use for the final brachytherapy implants and fractions. The performance metrics were accuracy, precision, and recall.

RESULTS

A total of 10 eligible patients were accrued between December 2020 and October 2022. Table 1 shows the initial clinical determination and consensus applicator AI predictions that were made with different confidence thresholds. The optimal confidence threshold (60%) yielded performance scores of 80%, 83.3%, and 83.3% for the accuracy, precision, and recall, respectively. The performance metrics were equivalent for the optimal confidence threshold and the initial clinical decision.

CONCLUSION

The AI-based decision support tool shows strong predictive results for a clinical brachytherapy application that is important to patient outcome. This prospective study demonstrates that the algorithm's utility is a critical step in using AI-based tools clinically. Further work to determine the optimal brachytherapy applicator, based on treatment planning dosimetry, is required.

Deep Learning-Based Automated Quality Assurance for Palliative Spinal Treatment Planning in Radiotherapy

PURPOSE/OBJECTIVE(S)

Robustquality assurance (QA) for palliative spine radiation therapy (RT) remains critical due to the risk of wrong anatomic level treatment on account of human error in enumerating vertebral bodies accurately based on morphology with incomplete imaging of the spine and prevalence of anatomic variants (10%). We propose a rapid, fully automated deep learning-based QA (DL-QA) tool for segmenting and enumerating vertebral structures from image data, capable of identifying misalignment based on discrepancies in calculated dose coverage.

MATERIALS/METHODS

Ina retrospective cohort of 514 patients who received palliative spine radiation treatment at a single institution for spinal metastases, vertebral volumes for each individual spine level were automatically segmented on RT planning computed tomography scans using a publicly available deep learning algorithm, Total Segmentator (TS) deployed in the treatment planning system (Wasserthal et al, 2022). Departmental policy requires that the prescription/plan name include all spinal levels that receive a prescribed dosimetric threshold of V50% > 50%. By comparing the intended spine level target in the prescription and plan name against the TS volumes, the DL-QA flagged all cases for which any target vertebrae did not receive this threshold dose and/or any non-targeted vertebrae that received V50% > 50%. To detect spine anomalies, cases were also flagged if any vertebrae volume was not within ±1σ of the entire population of vertebrae volumes. Flagged cases were either categorized as: (1) wrong spine level RT error; (2) documentation error, in which treatment was correct but the prescription/plan name did not follow Departmental policy; or (3) potential spinal geometric error. All flagged cases were verified manually by checking the original images and treatment planning data.

RESULTS

Outof 514 patients, 29 cases were flagged as potential errors. Manual review revealed that one of these was a previously discovered true treatment error (due to anatomic variant with 4 lumbar bodies) while 10 were treated as intended but showed documentation errors due to variants in the number of vertebral bodies, kyphosis of the spine causing non-targeted vertebrae to appear in the treatment field, or improper observation of the Departmental plan naming policy. The remaining 18 cases were associated with flagged vertebrae volumes. Reviewing those patients, we identified spinal anomalies where TS attempted to account for extra or missing vertebrae (N = 9) and cases where TS made segmentation errors (N = 9).

CONCLUSION

Theproposed automated DL-QA system successfully identified patients with spine anomalies, flagged documentation errors, verified the correct target levels of spine RT treatments, and detected a known misadministration. The next phase will involve prospective testing of the system in a clinical setting upstream of treatment delivery.

The transition in practice to reduce bolus use in post-mastectomy radiotherapy: A dosimetric study of skin and subcutaneous tissue

To inform clinical practice for women receiving post-mastectomy radiotherapy (PMRT), this study demonstrates the dosimetric impact of removing daily bolus on skin and subcutaneous tissue. Two planning strategies were used: clinical field-based (n = 30) and volume-based planning (n = 10). The clinical field-based plans were created with bolus and recalculated without bolus for comparison. The volume-based plans were created with bolus to ensure a minimum target coverage of the chest wall PTV and recalculated without bolus. In each scenario, the dose to superficial structures, including skin (3 mm and 5 mm) and subcutaneous tissue (a 2 mm layer, 3 mm deep from surface) were reported. Additionally, the difference in the clinically evaluated dosimetry to skin and subcutaneous tissue in volume-based plans were recalculated using Acuros (AXB) and compared to the Anisotropic Analytical Algorithm (AAA) algorithm. For all treatment planning strategies, chest wall coverage (V90%) was maintained. As expected, superficial structures demonstrate significant loss in coverage. The largest difference observed in the most superficial 3 mm where V90% coverage is reduced from a mean (± standard deviation) of 95.1% (± 2.8) to 18.9% (± 5.6) for clinical field-based treatments with and without bolus, respectively. For volume-based planning, the subcutaneous tissue maintains a V90% of 90.5% (± 7.0) compared to the clinical field-based planning coverage of 84.4% (± 8.0). In all skin and subcutaneous tissue, the AAA algorithm underestimates the volume of the 90% isodose. Removing bolus results in minimal dosimetric differences in the chest wall and significantly lower skin dose while dose to the subcutaneous tissue is maintained. Unless the skin has disease involvement, the most superficial 3 mm is not considered part of the target volume. The continued use of the AAA algorithm is supported for the PMRT setting.

Technical note: Investigating the suitability of existing facilities for a new Lu-177 prostate-specific membrane antigen therapy program

BACKGROUND

¹⁷⁷ Lu prostate-specific membrane antigen (PSMA) therapy prolongs survival for some prostate cancer patients. To adopt this technique, institutions may need to evaluate the suitability of existing infrastructure.

PURPOSE

Develop a methodology to determine whether existing facilities can accommodate a ¹⁷⁷ Lu-PSMA therapy program.

METHODS

Room suitability is defined by both the ability to accommodate ¹⁷⁷ Lu-PSMA therapy workflow and to provide appropriate radiation shielding. Two methods of shielding calculation were performed: (1) National Council on Radiation Protection and Measurements report 151 (NCRP-151), with workload defined in terms of the activity of ¹⁷⁷ Lu administered, and (2) using the RadPro shielding calculator. This methodology was applied to ¹³¹ I therapy, PET-CT uptake, PET-SPECT injection, and orthovoltage therapy rooms.

RESULTS

¹³¹ I therapy rooms were found to meet both shielding and workflow requirements. The shielding was found to be adequate for orthovoltage and PET-SPECT facilities, neglecting patient transit between external washrooms. The workflow was the limiting factor for these rooms due to the requirement of dedicated washrooms that shield the patient and contain possible contamination. The PET-CT facility did not meet either criteria. The NCRP-151 method generally predicted a higher dose rate on the other side of shielding than did the RadPro calculator. The dose rate on the other side of concrete shielding as predicted by the NCRP-151 method increased relative to the dose rate predicted by the RadPro calculator as shielding thickness increased. For lead shielding, the dose rate predicted by the NCRP-151 method decreased relative to the result predicted by the RadPro calculator with increasing material thickness.

CONCLUSIONS

¹³¹ I therapy, PET-CT uptake, PET-SPECT injection, and orthovoltage therapy rooms were considered. The ¹³¹ I treatment rooms were the best candidate for ¹⁷⁷ Lu-PSMA therapy, due to their shielding and capability to accommodate the necessary workflow.

Radiotherapy with Cyclin-Dependent Kinase 4/6 Inhibitors: A Multi-institutional Safety and Toxicity Study

PURPOSE

To investigate radiotherapy (RT) toxicity when given with Cyclin-Dependent Kinase 4/6 inhibitors (CDK4/6i) compared to RT alone.

METHODS AND MATERIALS

We conducted a retrospective cohort study of patients with hormonal receptor-positive and human epidermal growth factor-2 negative metastatic breast cancer treated with RT at four cancer centers in Alberta, Canada between 2016 and 2020. Toxicity in patients treated with RT within 30 days of initiating to discontinuing CDK4/6i (RT+CDK4/6i) was compared to toxicity of RT in CDK4/6i naïve patients (RT alone). The primary outcome was acute grade (G) II or higher, non-hematological toxicity within 30 days of RT. We also explored toxicity based on the timing of RT (prior, concurrent, post) in relation to CDK4/6i. Propensity score matching was applied to create comparable cohorts. A generalized linear mixed model was used to evaluate factors associated with acute toxicity.

RESULTS

132 patients (220 RT sites) in the RT+CDK4/6i and 53 patients (93 RT sites) in RT alone were eligible. The rate of acute GII or higher non-hematological toxicity was 11.5% vs. 7%, respectively (p=0.439), and acute GIII or higher non-hematological toxicity was 3.7% vs. 0%, respectively (p=0.151). Acute toxicity in RT+CDK4/6i group was mainly observed when RT was given concurrently (67%) with most of the GIII toxicity recorded. After propensity score matching, the association of acute toxicity with RT+CDK4/6i vs. RT alone was not significant on multivariable analysis, Odds Ratio 3.13 (95% confidence interval: 0.74 - 13.2; p=0.121).

CONCLUSIONS

We did not observe a significant association between CDK4/6i use and acute, GII or higher, non-hematological toxicity, in women with metastatic breast cancer receiving palliative RT. Given the findings of GIII toxicity, caution is advised whenever CDK4/6i is given concurrently with RT.

Assessment of Tissue Toxicity Risk in Breast Radiotherapy using Bayesian Networks

PURPOSE

The purpose of this analysis is to predict worsening post-treatment normal tissue toxicity in patients undergoing accelerated partial breast irradiation (APBI) therapy and to quantitatively identify which diagnostic, anatomical, and dosimetric features are contributing to these outcomes.

METHODS

A retrospective study of APBI treatments was performed using 32 features pertaining to various stages of the patient's treatment journey. These features were used to inform and construct a Bayesian network (BN) based on both statistical analysis of feature distributions and relative clinical importance. The target feature for prediction was defined as a measurable worsening of telangiectasia, subcutaneous tissue induration, or fibrosis when compared against the observed baseline. Parameter learning for the network was performed using data from the 299 patients included in the ACCEL trial and predictive performance was measured. Feature importance for the BN was quantified using a novel information-theoretic approach.

RESULTS

Cross validated performance of the BN for predicting toxicity was consistently higher when compared against conventional machine learning (ML) techniques. The measured BN receiver operating characteristic area under the curve was 0.960±0.013 against the best ML result of 0.942±0.021 using 5-fold cross validation with separate test data across 100 trials. The volume of the clinical target volume, gross target volume, and baseline toxicity measurements were found to have the highest feature importance and mutual dependence with normal tissue toxicity in the network, representing the strongest contribution to patient outcomes.

CONCLUSIONS

The BN outperformed conventional ML techniques in predicting tissue toxicity outcomes and provided deeper insight into which features are contributing to these outcomes. This article is protected by copyright. All rights reserved.

An updated approach for deriving PTV margins using image guidance and deformable dose accumulation

Objective. To demonstrate an updated approach for deriving planning target volume (PTV) margins for a patient population treated with volumetric image-guided radiotherapy. Approach. The approach uses a semi-automated workflow within commercial radiotherapy applications that combines dose accumulation with the bidirectional local distance (BLD) metric. The patient cohort is divided into derivation and validation datasets. For each patient in the derivation dataset, a treatment plan is generated with a 0 mm PTV margin (the idealized treatment scenario without the influence of the standard margin). Deformable image registration enabled dose accumulation of these zero-margin plans. PTV margins are derived by using the BLD to calculate the geometric extent of underdosed regions of the clinical target volume (CTV). The PTV margin is validated by ensuring the specified CTV coverage criterion is met when the margin is applied to the validation dataset. Main results. The methodology was applied to two cohorts: 40 oropharyngeal cancer patients and 50 early-stage breast cancer patients. Ten patients from each cohort were used for validation. PTV margins derived for the oropharyngeal and early-stage breast cancer patient cohorts were 3 and 5 mm, respectively, and ensure that 95% of the prescription dose is delivered to 98% of the CTV for 90% of patients. Dose accumulation showed that the CTV coverage criterion was achieved for at least 90% of patients when the margins were applied. Significance. This methodology can be used to derive appropriate PTV margins for realistic treatment scenarios and any disease site, which will improve our understanding of patient outcomes.

Machine learning generated decision boundaries for prediction and exploration of patient-specific quality assurance failures in Stereotactic Radiosurgery plans

INTRODUCTION

Stereotactic Radiosurgery (SRS) is a form of radiotherapy treatment during which high radiation dose is delivered in a single or few fractions. These treatments require highly conformal plans with steep dose gradients which can result in an increase in plan complexity prompting the need for stringent pre-treatment patient specific quality assurance (QA) measurements to ensure the planned and measured dose distributions agree within clinical standards. Complexity scores and machine learning (ML) techniques may help with prediction of QA outcomes however interpretability and usability of those results continues to be an area of study. This study investigates the use of plan complexity metrics as input for an ML model to allow for prediction of QA outcomes for SRS plans as measured via 3D phantom dose verification. Explorations into interpretability and predictive performance changes as model dimensionality increases, as well as a prospective in-clinic implementation using the resulting model were also performed.

METHODS

498 plans (1571 VMAT arcs) were processed via in-house script to generate several complexity scores. 3D phantom dose verification measurement results were extracted and classified as pass or failure (with failures defined as below 95% voxel agreement passing 3%/1mm gamma criteria with 10% threshold,) and 1472 of the arcs were split into training and testing sets, with 99 arcs as a sequential holdout set. A z-score scaler was trained on the training set and used to scale all other sets. Variations of MLC leaf movement variability, aperture complexity, and leaf size and MU at control point weighted target area scores were used as input to a Support Vector Classifier to generate a series of 1-D, 2-D, and 5-D decision boundaries. The best performing 5D model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements for 38 patient plans (112 arcs). The decision to order 3D phantom dose verification measurements was recorded before and after prediction.

RESULTS

Best performing 1-D threshold, and 2-D prediction models with best performance produced a QA failure recall and QA passing recall of 1.00 and 0.55, and 0.82 and 0.82 respectively. Best performing 5-D prediction model produced a QA failure recall (sensitivity) of 1.00, and QA passing recall (specificity) of 0.72. This model was then used within a prospective in-clinic study providing predictions to physicists prior to ordering 3D phantom dose verification measurements and achieved a QA failure recall of 1.00 and QA passing recall of 0.58. The decision to order 3D phantom dose verification measurements was recorded before and after measurement. A single initially unidentified failing plan of the prospective cohort was successfully predicted to fail by the model.

CONCLUSION

Implementation of complexity score based prediction models for SRS would allow for support of a clinician's decision to reduce time spent performing QA measurements, and avoid patient treatment delays (i.e. in case of QA failure). This article is protected by copyright. All rights reserved.

Performance of a knowledge-based planning model for optimizing intensity-modulated radiotherapy plans for partial breast irradiation

Purpose

To evaluate a knowledge‐based (KB) planning model for RapidPlan, generated using a five‐field intensity‐modulated radiotherapy (IMRT) class solution beam strategy and rigorous dosimetric constraints for accelerated partial breast irradiation (APBI).

Materials and methods

The RapidPlan model was configured using 64 APBI treatment plans and validated for 120 APBI patients who were not included in the training dataset. KB plan dosimetry was compared to clinical plan dosimetry, the clinical planning constraints, and the constraints used in phase III APBI trials. Dosimetric differences between clinical and KB plans were evaluated using paired two‐tailed Wilcoxon signed‐rank tests.

Results

KB planning was able to produce IMRT‐based APBI plans in a single optimization without manual intervention that are comparable or better than the conventionally optimized, clinical plans. Comparing KB plans to clinical plans, differences in PTV, heart, contralateral breast, and ipsilateral lung dose–volume metrics were not clinically significant. The ipsilateral breast volume receiving at least 50% of the prescription dose was statistically and clinically significantly lower in the KB plans.

Conclusion

KB planning for IMRT‐based APBI provides equivalent or better dosimetry compared to conventional inverse planning. This model may be reliably applied in clinical practice and could be used to transfer planning expertise to ensure consistency in APBI plan quality.

Clinical Outcomes from Dose-Reduced Radiotherapy to the Prostate in Elderly Patients with Localized Prostate Cancer

Radical treatment of localized prostate cancer in elderly patients may lead to unacceptable treatment-associated toxicities that adversely impact quality of life without improving survival outcomes. This study reports on a cohort of 54 elderly (>70 years) patients that received 4000–5000 cGy of palliative external beam radiotherapy (EBRT) as an alternative to androgen deprivation therapy (ADT). The primary outcome of interest was the period of ADT-free survival, and secondary outcomes included overall survival (OS) and metastases-free survival (MFS). Kaplan–Meier regression was used to estimate survival outcomes. Thirty-six (67%) patients achieved a break in ADT post-radiotherapy, with a median time to ADT reinitiation of 20 months. Common Terminology Criteria for Adverse Events (CTCAE) were limited to low-grade gastrointestinal (GI) or genitourinary (GU) toxicities, with no skin toxicities observed. Grade 1 GI toxicity was observed in 9 (17%) patients, and grades 1 and 2 GU toxicities were observed in 13 (24%) and 3 (6%) patients, respectively, with no higher-grade toxicities reported. Five-year MFS and OS were 56% and 78%, respectively. In summary, the treatment regimen was well-tolerated and achieved durable ADT-free survival in most patients. Dose-reduced EBRT appears to be a viable alternative to ADT in elderly patients with localized prostate cancer.

Increasing Demand on Human Capital and Resource Utilization in Radiation Therapy: The Past Decade

PURPOSE

To quantify the change resource utilization in radiation therapy in the context of advancing technologies and techniques over the last decade.

METHODS AND MATERIALS

Prospectively, the time to complete radiation therapy workflow tasks was captured between January 1, 2020, and December 31, 2020. The institutional task workflows are specific to each technique and broadly organized into 4 categories: 3-dimenstional conformal radiation therapy, intensity modulated radiation therapy, volumetric modulated arc therapy simple, and volumetric modulated arc therapy complex. These discipline-specific task times were used to quantify a resource utilization factor, which is the median time taken to complete all tasks for each category divided by the median time for 3-dimensional conformal radiation therapy treatments. Retrospectively, all plans treated between January 1, 2012, and December 31, 2019, were quantified and categorized. The resource factor was applied to determine resource utilization. For context, institutional staffing levels were captured across the same decade for medical dosimetrists, medical physicists, and radiation oncologists.

RESULTS

This analysis includes 30,229 patient plans in the retrospective data set and 4747 patient plans in the prospective data set. This analysis demonstrates that over this period, patient numbers increased by approximately 45%, whereas time-based human resources increased by almost 150%. The resource allocation factors for 3-dimenstional conformal radiation therapy, intensity modulated radiation therapy, volumetric modulated arc therapy simple, and volumetric arc therapy complex were 1.0, 2.4, 2.9, and 4.3, respectively. Across the 3 disciplines, staffing levels increased from 15 to 17 (13%) for medical dosimetrists, from 10 to 13 (30%) for medical physicists, and from 16 to 23 (44%) for radiation oncologists.

CONCLUSIONS

This work demonstrates the increase in resource utilization due to the introduction of advanced technologies and changes in radiation therapy techniques over the past decade. Human resource utilization is the predominant factor and should be considered with increasing patient volume for operational planning.

Technical Note: A standardized automation framework for monitoring institutional radiotherapy protocol compliance

PURPOSE

To establish a framework for the standardization of monitoring radiotherapy protocol compliance.

METHODS

An automated protocol compliance tool was developed using best practice in software design and a flexible framework to easily adapt to changing institutional standards. The Eclipse scripting environment was used to develop the application with the scripting application programing interface (API) and direct data extraction from ARIA. For each institutional protocol, external validation was specified in a JavaScript Object Notation (JSON) file that stores protocol specific constraints and evaluates compliance of the data from Eclipse and Aria. This tool was applied prospectively to a cohort of prostate cancer patients undergoing radiotherapy with a prescription regimen of 60 Gy in 20 fractions.

RESULTS

The prospective evaluation was performed on 58 prostate cancer patients. For this cohort, the mean (standard deviation) pass rate is 92.3% (6.1%). The overall fail rate is 6.0% (5.8%); the percentage of these failures is in 2.6% in Patient Assessment, 0% in Simulation, and 97.4% in Treatment Planning.

CONCLUSIONS

A protocol compliance application is developed and implemented in a standard radiotherapy information system. The application functionality is demonstrated on a cohort of 58 patients undergoing prostate radiotherapy, which highlights the utility of assessing adherence to institutional protocols. A unified method must be available for the community to ensure consistency in compliance reporting.

Dosimetric consequences of seed placement accuracy in permanent breast seed implant brachytherapy

PURPOSE

This study quantifies the dosimetric impact of implant accuracy and derives a quantitative relationship relating implant accuracy to target dosimetry.

METHODS AND MATERIALS

A framework was developed to simulate multiple implants for error combinations. Spherical clinical target volumes (CTVs) were modeled with volumes 1.4 cm³, 9.2 cm³, and 20.6 cm³, representing the range seen clinically. Each CTV was expanded 10 mm isotropically to create a planning target volume (PTV).. Random and systematic seed placement errors were simulated by shifting needles from their planned positions. Implant errors were simulated over the range of clinically practical errors in permanent breast seed implant. The relative effect on target coverage was evaluated. Regression analysis was performed to derive relationships between CTV dosimetry and the magnitude of implant accuracy. The validity of the clinically used 10 mm PTV margin for each of the CTVs was assessed.

RESULTS

Introducing practical implant errors resulted in CTV V_90% median (10th and 90th percentile) of 97.7% (85.9% and 100%), 96.2% (86.8% and 99.7%), and 100% (77.8% and 100%) for the typical, large, and small CTV, respectively. All CTVs show similar trends in target coverage. Polynomials were derived relating seed placement accuracy to median (R² = 0.82) and 10th percentile (R² = 0.78) CTV V_90%._. CONCLUSIONS: This work quantitatively describes the relationship between implant accuracy and CTV coverage. Based on simulations, the 10 mm PTV margin is adequate to maintain target coverage. These equations can be used with institutional seed placement accuracy to estimate coverage.

Establishing a simulation-based education program for radiation oncology learners in permanent seed implant brachytherapy: Building validation evidence

PURPOSE

The purpose of this study was to establish a simulation-based education program for radiation oncology learners in permanent seed implant brachytherapy. The first step in formalizing any education program is a validation process that builds evidence-based verification that the learning environment is appropriate.

METHODS AND MATERIALS

The primary education task allowed practitioners to use an anthropomorphic breast phantom to simulate a permanent seed implant brachytherapy delivery. Validation evidence is built by generating data to assess learner and expert cohorts according to their proficiency. Each practitioner's performance during the simulation was evaluated by seed placement accuracy, procedural time-to-complete, and two qualitative evaluation tools-a global rating scale and procedural checklist.

RESULTS

The average seed placement accuracy (±SD) was 8.1 ± 3.5 mm compared to 6.1 ± 2.6 mm for the learner and expert cohort, respectively. The median (range) procedural time-to-complete was 64 (60-77) minutes and 43 (41-50) minutes for the learner and expert cohort, respectively. Seed placement accuracy (student t-test, p < 0.05) and procedural time-to-complete (Mann-Whitney U-test, p < 0.05) were statistically different between the cohorts. In both the global rating scale and procedural checklist, the expert cohort demonstrated improved proficiency compared to the learner cohort.

CONCLUSIONS

This validation evidence supports the utilization of this simulation environment toward appropriately capturing the delivery experience of practitioners. The results demonstrate that, in all areas of evaluation, expert cohort proficiency was superior to learner cohort proficiency. This methodology will be used to establish a simulation-based education program for radiation oncology learners in permanent seed implant brachytherapy.

What Conditions Make Proton Beam Therapy Financially Viable in Western Canada?

Background

Proton beam therapy (PBT) is available in many western and Asian countries, but there is no clinical, gantry-based PBT facility in Canada.

Methods

A cost analysis was conducted from the Alberta Ministry of Health perspective with a 15-year horizon. Estimated costs were: PBT unit, facility development as part of an ongoing capital project, electricity, maintenance contract, and staffing. Revenues were: savings from stopping USA referrals, avoiding the costs of standard radiation therapy (RT) for Albertans receiving PBT instead, and cost-recovery charges for out-of-province patients.

Results

The Ministry of Health funded 15 Albertans for PBT in the USA in the 2014/15 fiscal year (mean CAD$ 237,348/patient). A single-vault, compact PBT unit operating 10 hours/day could treat 250 patients annually. A 100 Albertans, with accepted indications, such as the curative-intent treatment of chordomas, ocular melanomas, and selected pediatric cancers, would likely benefit annually from PBT’s improved conformality and/or reduced integral dose compared to RT. The estimated capital cost was $40 million for a single beamline built within an ongoing capital project. Operating costs were $4.8 million/year at capacity. With 50% capacity reserved for non-Albertans at a cost recovery of $45,000/patient, a Western Canadian PBT facility would achieve net positive cash flow by year eight of clinical operations, assuming Alberta-to-USA referrals reach 21 patients/year by 2024 and increase at 3%/year thereafter. Sensitivity analysis indicates the lifetime net savings is robust to the assumptions made.

Conclusion

This business case, based on Canadian costing data and estimates, demonstrates the potential for a financially viable PBT facility in Western Canada.