Adaptive Radiation Therapy: A Review of CT-based Techniques

Simulation-Omitting and Using Library Patients for Pre-Planning Online Adaptive Radiotherapy (SUPPORT): A Feasibility Study for Spine Stereotactic Ablative Radiotherapy (SAbR) Patients

Treatment planning in the field of radiation therapy has evolved from three-dimensional (3D) planning to inverse planning and, most recently, to personalized adaptive radiotherapy (ART) [...].

Minimizing Bowel Gas Artifact in Computed Tomography Guided Online Adaptive Radiation Therapy With Prolonged Supine Positioning

PURPOSE

Abdominal computed tomography guided online adaptive SABR treatments can be complicated by bowel gas artifact. We investigated whether prolonged patient supine positioning decreases bowel gas artifact.

METHODS AND MATERIALS

Three radiation oncologists, a physicist, and an advanced practice radiation therapist scored bowel gas artifact in 1644 images from 104 cone beam computed tomography (CBCT) data sets acquired in 52 fractions delivered to 26 pancreatic adenocarcinoma patients with a Halcyon/Ethos online adaptive SABR platform. Bowel gas artifact scoring followed an ordinal rubric from 1 (best) to 4 (worst). Ten patients were imaged with HyperSight CBCTs and 16 with an earlier CBCT imager, Halcyon v3.0.

RESULTS

Four-hundred forty-four CBCT images (27%) had bowel gas artifact that at least minimally obscured organ-at-risk borders. Artifact was worse in initial CBCTs than subsequent verification CBCTs (mean scores 2.26 vs 2.15, P = .006). The proportion of images scored 4 was significantly greater in Halcyon initial CBCTs (0.25) compared to verification CBCTs (0.16; P < .001). For HyperSight, this proportion was low in initial CBCTs (0.03) and lower in verification CBCTs (0.01; P = .09).

CONCLUSIONS

Clinically impactful bowel gas artifact in online adaptive SABR CBCT data sets was better in verification CBCTs than in initial CBCTs, potentially due to bowel settling during prolonged supine positioning.

A dosimetric comparison of helical tomotherapy treatment delivery with real-time adaption and no motion correction

This study assesses the ability of a helical tomotherapy system equipped with kV imaging and optical surface guidance to adapt to motion traces in real-time. To assess the delivery accuracy with motion, a unified testing framework was used. The average 2 %/2 mm γ-fail rates across all lung traces were 0.1 % for motion adapted and 17.4 % for no motion correction. Average 2 %/2 mm γ-fail rates across all prostate traces were 0.4 % for motion adapted and 12.2 % for no motion correction. Real-time motion adaption was shown to improve the accuracy of dose delivered to a moving phantom compared with no motion adaption. MeSH Keywords: Radiotherapy, image-guided; Radiation therapy, targeted.

Diagnostic image-based treatment planning for online adaptive ultra-hypofractionated prostate cancer radiotherapy with MR-Linac

PURPOSE

A new workflow was investigated for Elekta Unity MR-Linac by removing the computed tomography (CT)-simulation step and using diagnostic CT (DCT) for reference plan generation.

MATERIALS AND METHODS

Ten patients with ultra-hypofractionated prostate cancer treated with magnetic resonance imaging (MRI)-guided adaptive radiotherapy were retrospectively enrolled. Targets and organs at risk (OARs) were recontoured on DCT, and Hounsfield unit conversions to relative electron density were calibrated for DCT. Reference plans were reoptimized and recalculated using DCT for Unity. Subsequent adaptive plans were designed through an adapt-to-shape workflow to edit targets and OARs via daily MRI to generate a new treatment plan. Bulk electron density information for Unity adaptive plan was compared between planning CT (PCT) and DCT for volumes of interest. Dosimetric parameters were evaluated between PCT- and DCT-based reference and adaptive plans for target coverage and OAR dose constraints.

RESULTS

Bulk relative electron density differences between PCT and DCT were within ±1% for targets and OARs, excepting the rectum. PCT and DCT-based reference plans did not significantly differ in mean target coverages or for OARs in dosimetric difference except for V36 Gy of the rectum. PCT- and DCT-based adaptive plans did not significantly differ for most dosimetric parameters of targets and OARs except for V29 Gy and V36 Gy of the rectum, V18.1 Gy of the bladder, and D50% of the urethra.

CONCLUSIONS

By removing the CT simulation step, it is feasible to use DCT for designing reference and adaptive plans in the Unity ATS workflow. The workflow increased adaptive radiotherapy efficiency and decreased patient waiting time and additional radiation dose.

Modeling dosimetric benefits from daily adaptive RT for gynecological cancer patients with and without knowledge-based dose prediction

PURPOSE

Daily online adaptive radiotherapy (ART) improves dose metrics for gynecological cancer patients, but the on-treatment process is resource-intensive requiring longer appointments and additional time from the entire adaptive team. To optimize resource allocation, we propose a model to identify high-priority patients.

METHODS

For 49 retrospective cervical and endometrial cancer patients, we calculated two initial plans: the treated standard-of-care (InitialSOC) and a reduced margin initial plan (InitialART) for adapting with the Ethos treatment planning system. Daily doses corresponding to standard and reduced margins (DailySOC and DailyART) were determined by re-segmenting the anatomy based on the treatment CBCT and calculating dose on a synthetic CT. These initial and daily doses were used to estimate the ART benefit (Δ D a i l y ${{\Delta}}Daily$ = DailySOC-DailyART) versus initial plan differences (Δ I n i t i a l ${{\Delta}}Initial$ = InitialSOC-InitialART) via multivariate linear regression. Dosimetric benefits were modeled with initial plan differences (Δ I n i t i a l ${{\Delta}}Initial$ ) ofB o w e l V 40 G y $Bowel\ {{V}_{40Gy}}$ (cc),B l a d d e r D 50 % $Bladder\ {{D}_{50{\mathrm{\% }}}}$ (Gy), andR e c t u m D 50 % $Rectum\ {{D}_{50{\mathrm{\% }}}}$ (Gy). Anatomy (intact uterus or post-hysterectomy), DoseType (simultaneous integrated boost [SIB] vs. single dose), and/or prescription value. To establish a logistic model, we classified the top 10% in each metric as high-benefit patients. We then built a logistic model to predict these patients from the previous predictors. Leave-one-out validation and ROC analysis were used to evaluate the accuracy. To improve the clinical efficiency of this predictive process, we also created knowledge-based plans for the ΔInitial plans (Δ I n i t i a l R P ${{\Delta}}Initia{{l}_{RP}}$ ) and repeated the analysis.

RESULTS

In bothΔ I n i t i a l O r i g ${{\Delta}}Initia{{l}_{Orig}}$ andΔ I n i t i a l R P ${{\Delta}}Initia{{l}_{RP}}$ our multivariate analysis showed low R2 values 0.34-0.52 versus 0.14-0.38. The most significant predictor in each multivariate model was the corresponding ∆Initial metric (e.g.,Δ I n i t i a l ${{\Delta}}Initial$ Bowel (V40 Gy), p < 1e-05). In the logistic model, the metrics with the strongest correlation to the high-benefit patients wereB o w e l V 40 G y $Bowel\ {{V}_{40Gy}}$ (cc),B l a d d e r D 50 % $Bladder\ {{D}_{50{\mathrm{\% }}}}$ (Gy),D o s e T y p e $DoseType$ , andS I B D o s e $SIBDose$ prescription. The models for original and knowledge-based plans had an AUC of 0.85 versus 0.78. The sensitivity and specificity were 0.92/0.72 and 0.69/0.80, respectively.

CONCLUSION

This methodology will allow clinics to prioritize patients for resource-intensive daily online ART.

Experience and uncertainty analysis of CT-based adaptive radiotherapy for abdominal treatments

BACKGROUND

Online adaptive radiotherapy (ART) allows for daily replanning of treatment plans with adjustments according to current day anatomy. The purpose of this work is to present our methodology for using CT-based ART applied to abdominal cases along with our experience with this treatment. We additionally aim to estimate some of the uncertainties associated with the adaptive process.

METHODS AND MATERIALS

Analysis was performed on patients with abdominal targets (N = 41, 205 fractions), treated on a CT-based adaptive treatment unit; treatment sites were divided into 3 categories: pancreas, liver, and other (e.g., lymph nodes). Statistics regarding contouring time, planning target volume (PTV) coverage, and organ-at-risk (OAR) sparing are presented. Contouring uncertainty was estimated by expanding critical OARs and recalculating dose, and auto-registration uncertainty was estimated by adjusting the registration between the cone beam computed tomography scan and the dose cloud and recalculating dose.

RESULTS

Coverage for the planning optimization PTV (PTV_Opt) for adaptive plans was on average 94.7 ± 0.4 %, while for scheduled plans it was on average 92.0 ± 0.6 %. The average decrease in OAR maximum dose by using the adaptive plans was 11.6 ± 1.0 %. Contouring time was on average 23 ± 0 min. Uncertainty estimates for PTV V100% were on average 0.6 ± 0.4 %; combined uncertainties for maximum OAR dose were on average 4.6 ± 0.4 %.

CONCLUSION

Adaptive therapy on average led to plans with improved PTV coverage or OAR sparing, and our workflow allowed for treatment to be completed within a reasonable timeframe. The benefit of adaptive therapy largely outweighed estimates of uncertainty.

Intricacies of human-AI interaction in dynamic decision-making for precision oncology

AI decision support systems can assist clinicians in planning adaptive treatment strategies that can dynamically react to individuals' cancer progression for effective personalized care. However, AI's imperfections can lead to suboptimal therapeutics if clinicians over or under rely on AI. To investigate such collaborative decision-making process, we conducted a Human-AI interaction study on response-adaptive radiotherapy for non-small cell lung cancer and hepatocellular carcinoma. We investigated two levels of collaborative behavior: model-agnostic and model-specific; and found that Human-AI interaction is multifactorial and depends on the complex interrelationship between prior knowledge and preferences, patient's state, disease site, treatment modality, model transparency, and AI's learned behavior and biases. In summary, some clinicians may disregard AI recommendations due to skepticism; others will critically analyze AI recommendations on a case-by-case basis; clinicians will adjust their decisions if they find AI recommendations beneficial to patients; and clinician will disregard AI recommendations if deemed harmful or suboptimal and seek alternatives.

Optimal timing of organs-at-risk-sparing adaptive radiation therapy for head-and-neck cancer under re-planning resource constraints

Background and purpose

Prior work on adaptive organ-at-risk (OAR)-sparing radiation therapy has typically reported outcomes based on fixed-number or fixed-interval re-planning, which represent one-size-fits-all approaches and do not account for the variable progression of individual patients' toxicities. The purpose of this study was to determine the personalized optimal timing of re-planning in adaptive OAR-sparing radiation therapy, considering limited re-planning resources, for patients with head and neck cancer (HNC).

Materials and methods

A novel Markov decision process (MDP) model was developed to determine optimal timing of re-planning based on the patient's expected toxicity, characterized by normal tissue complication probability (NTCP), for four toxicities. The MDP parameters were derived from a dataset comprising 52 HNC patients treated between 2007 and 2013. Kernel density estimation was used to smooth the sample distributions. Optimal re-planning strategies were obtained when the permissible number of re-plans throughout the treatment was limited to 1, 2, and 3, respectively.

Results

The MDP (optimal) solution recommended re-planning when the difference between planned and actual NTCPs (ΔNTCP) was greater than or equal to 1%, 2%, 2%, and 4% at treatment fractions 10, 15, 20, and 25, respectively, exhibiting a temporally increasing pattern. The ΔNTCP thresholds remained constant across the number of re-planning allowances (1, 2, and 3).

Conclusion

In limited-resource settings that impeded high-frequency adaptations, ΔNTCP thresholds obtained from an MDP model could derive optimal timing of re-planning to minimize the likelihood of treatment toxicities.

An End-to-end Quality Assurance Procedure for Ethos Online Adaptive Radiotherapy

Purpose

Online adaptive radiation therapy (OART) poses unique challenges for quality assurance (QA), requiring innovative methodologies beyond traditional techniques. This study introduced an end-to-end (E2E) QA test for the Ethos OART system.

Materials and Methods

Initial treatment plans were developed using deformed computed tomography (CT) images of standard phantoms. During treatment sessions, adaptive plans were created and delivered using undistorted physical QA phantoms equipped with measuring detectors. Our approach was demonstrated using standard QA phantoms - OCTAVIUS-four-dimensional (PTW, Freiburg, Germany), ArcCHECK (Sun Nuclear Corp., FL, USA), and the RUBY (PTW, Freiburg, Germany) - to evaluate the accuracy of contouring, synthetic CT (sCT), and dosimetry of adaptive plans in the Ethos OART system.

Results

Our findings demonstrated the superior performance of the Ethos OART system, with a gamma pass rate exceeding 96% (2% local/2 mm) and point dose deviations below 0.5%. The Dice coefficients for body contours between the sCT and reference CT were above 0.9, and the sCT accuracy was confirmed by mean absolute errors of <27 Hounsfield unit.

Conclusion

This approach establishes a straightforward E2E test to assess the workflow accuracies essential for preclinical validation/monthly QA of OART systems.

Descriptive overview of AI applications in x-ray imaging and radiotherapy

Artificial intelligence (AI) is transforming medical radiation applications by handling complex data, learning patterns, and making accurate predictions, leading to improved patient outcomes. This article examines the use of AI in optimising radiation doses for x-ray imaging, improving radiotherapy outcomes, and briefly addresses the benefits, challenges, and limitations of AI integration into clinical workflows. In diagnostic radiology, AI plays a pivotal role in optimising radiation exposure, reducing noise, enhancing image contrast, and lowering radiation doses, especially in high-dose procedures like computed tomography (CT). Deep learning (DL)-powered CT reconstruction methods have already been incorporated into clinical routine. Moreover, AI-powered methodologies have been developed to provide real-time, patient-specific radiation dose estimates. These AI-driven tools have the potential to streamline workflows and potentially become integral parts of imaging practices. In radiotherapy, AI's ability to automate and enhance the precision of treatment planning is emphasised. Traditional methods, such as manual contouring, are time-consuming and prone to variability. AI-driven techniques, particularly DL models, are automating the segmentation of organs and tumours, improving the accuracy of radiation delivery, and minimising damage to healthy tissues. Moreover, AI supports adaptive radiotherapy, allowing continuous optimisation of treatment plans based on changes in a patient's anatomy over time, ensuring the highest accuracy in radiation delivery and better therapeutic outcomes. Some of these methods have been validated and integrated into radiation treatment systems, while others are not yet ready for routine clinical use mainly due to challenges in validation, particularly ensuring reliability across diverse patient populations and clinical settings. Despite the potential of AI, there are challenges in fully integrating these technologies into clinical practice. Issues such as data protection, privacy, data quality, model validation, and the need for large and diverse datasets are crucial to ensuring the reliability of AI systems.

Navigating the Complexities of Radiation Injuries: Therapeutic Principles and Reconstructive Strategies

Radiation injuries, particularly those resulting from therapeutic or accidental exposure, present complex challenges for medical management. These injuries can manifest localized skin damage or extend to deeper tissues, presenting as various clinical entities that require treatment strategies, ranging from conservative management to complex surgical interventions. Radiation treatment constitutes a fundamental component of neoplastic management, with nearly two out of three oncological instances undergoing it as an element of their therapeutic strategy. The therapeutic approach to radiation injury consists of expanding prophylactic measures while maintaining the efficacy of treatment, such as conservative treatment or local debridement followed by reconstruction. The armamentarium of reconstructive methods available for plastic surgeons, from secondary healing to free tissue transfer, can be successfully applied to radiation injuries. However, the unique pathophysiological changes induced by radiation necessitate a careful and specialized approach for their application, considering the altered tissue characteristics and healing dynamics. The therapeutic strategy is guided by both the severity and progression of the injury, with the primary aim of restoring functionality and aesthetic aspects while simultaneously minimizing the risk of complications. This paper explores the various conditions encompassed by the term "radiation injury," reviews both non-surgical and surgical therapeutic strategies for managing these injuries, and highlights the unique challenges associated with treating irradiated tissues within specific oncological contexts.

CBCT-based online adaptive radiotherapy of the prostate bed: first clinical experience and comparison to nonadaptive conventional IGRT

PURPOSE

Conventional image-guided radiotherapy (IGRT) of the prostate bed is challenged by the varying anatomy due to dynamic changes of surrounding organs such as the bladder and rectum. This leads to changed dose coverage of target and surrounding tissue. The novel online adaptive radiotherapy (oART) aims to improve target coverage as well as reduce dose exposure to surrounding healthy tissues by daily reoptimization of treatment plans. Here we set out to quantify the resulting changes of this adaptation for patients and treatment team.

METHODS

A total of 198 fractions of radiotherapy of the prostate bed (6 patients) were treated using oART with the Ethos accelerator (Varian Medical Systems, Palo Alto, CA, USA). For each fraction, volumes and several dose-volume parameters of target volumes and organs at risk were recorded for the scheduled plan (initial plan, recalculated based on daily cone beam computed tomography [CBCT]), the adapted plan, and the verification plan, which is the dose distribution of the applied plan recalculated on the closing CBCT after the adaptation process. Clinical acceptability for all plans was determined using given dose-volume parameters of target volumes. Additionally, the time needed for the adaptation process was registered and compared to the time required for the daily treatment of five conventional IGRT patients.

RESULTS

Volumes of target and organs at risk (OAR) exhibited broad variation from day to day. The differences in dose coverage D98% of the clinical target volume (CTV) were significant through adaptation (p < 0.0001; median D98% 97.1-98.0%) and further after verification CBCT (p < 0.001; median D98% 98.1%). Similarly, differences in D98% of the planning target volume (PTV) were significant with adaptation (p < 0.0001; median D98% 91.8-96.5%) and after verification CBCT (p < 0.001; median D98% 96.4%) with decreasing interquartile ranges (IQR). Dose to OAR varied extensively and did not show a consistent benefit from oART but decreased in IQR. Clinical acceptability increased significantly from 19.2% for scheduled plans to 76.8% for adapted plans and decreased to 70.7% for verification plans. The scheduled plan was never chosen for treatment. The median time needed for oART was 25 min compared to 8 min for IGRT.

CONCLUSION

Target dose coverage was significantly improved using oART. IQR decreased for target coverage as well as OAR doses indicating higher repeatability of dose delivery using oART. Differences in doses after verification CBCT for targets as well as OAR were significant compared to adapted plans but did not offset the overall dosimetric gain of oART. The median time required is three times higher for oART compared to IGRT.

Optimal Timing of Organs-at-Risk-Sparing Adaptive Radiation Therapy for Head- and-Neck Cancer under Re-planning Resource Constraints

Background and Purpose

Prior work on adaptive organ-at-risk (OAR)-sparing radiation therapy has typically reported outcomes based on fixed-number or fixed-interval re-planning, which represent one-size-fits-all approaches and do not account for the variable progression of individual patients' toxicities. The purpose of this study was to determine the personalized optimal timing for re-planning in adaptive OAR-sparing radiation therapy, considering limited re-planning resources, for patients with head and neck cancer (HNC).

Materials and Methods

A novel Markov decision process (MDP) model was developed to determine optimal timing of re-planning based on the patient's expected toxicity, characterized by normal tissue complication probability (NTCP), for four toxicities. The MDP parameters were derived from a dataset comprising 52 HNC patients treated at the University of Texas MD Anderson Cancer Center between 2007 and 2013. Kernel density estimation was used to smooth the sample distributions. Optimal re-planning strategies were obtained when the permissible number of re-plans throughout the treatment was limited to 1, 2, and 3, respectively.

Results

The MDP (optimal) solution recommended re-planning when the difference between planned and actual NTCPs (ΔNTCP) was greater than or equal to 1%, 2%, 2%, and 4% at treatment fractions 10, 15, 20, and 25, respectively, exhibiting a temporally increasing pattern. The ΔNTCP thresholds remained constant across the number of re-planning allowances (1, 2, and 3).

Conclusion

In limited-resource settings that impeded high-frequency adaptations, ΔNTCP thresholds obtained from an MDP model could derive optimal timing of re-planning to minimize the likelihood of treatment toxicities.

Adaptive Approach to Treating Cervical Cancer in a Patient With Dramatic Uterine Movement

Adaptive radiation therapy is a modern technological advancement that allows radiation treatments to be adjusted daily to account for changes in the patient's anatomy, such as bladder and rectal filling, as well as changes in the tumor volume and position. In this case report, we present a patient with locally advanced cervical cancer who received definitive radiation therapy of 4500 cGy in 25 fractions using the Varian's Ethos system. We observed substantial daily uterine movement, which required re-optimization of each treatment fraction. Without the daily plan adaptation, the treatment would have resulted in markedly suboptimal dose coverage to the tumor. This case report highlights the importance of adaptive radiotherapy in managing anatomical changes in cervical cancer treatment and improving outcomes.

Autodelineation methods in a simulated fully automated proton therapy workflow for esophageal cancer

Background and purpose

Proton Online Adaptive RadioTherapy (ProtOnART) harnesses the dosimetric advantage of protons and immediately acts upon anatomical changes. Here, we simulate the clinical application of delineation and planning within a ProtOnART-workflow for esophageal cancer. We aim to identify the most appropriate technique for autodelineation and evaluate full automation by replanning on autodelineated contours.

Materials and methods

We evaluated 15 patients who started treatment between 11-2022 and 01-2024, undergoing baseline and three repeat computed tomography (CT) scans in treatment position. Quantitative and qualitative evaluations compared different autodelineation methods. For Organs-at-risk (OAR) deep learning segmentation (DLS), rigid and deformable propagation from baseline to repeat CT-scans were considered. For the clinical target volume (CTV), rigid and three deformable propagation methods (default, heart as controlling structure and with focus region) were evaluated. Adaptive treatment plans with 7 mm (ATP7mm) and 3 mm (ATP3mm) setup robustness were generated using best-performing autodelineated contours. Clinical acceptance of ATPs was evaluated using goals encompassing ground-truth CTV-coverage and OAR-dose.

Results

Deformation was preferred for autodelineation of heart, lungs and spinal cord. DLS was preferred for all other OARs. For CTV, deformation with focus region was the preferred method although the difference with other deformation methods was small. Nominal ATPs passed evaluation goals for 87 % of ATP7mm and 67 % of ATP3mm. This dropped to respectively 2 % and 29 % after robust evaluation. Insufficient CTV-coverage was the main reason for ATP-rejection.

Conclusion

Autodelineation aids a ProtOnART-workflow for esophageal cancer. Currently available tools regularly require manual annotations to generate clinically acceptable ATPs.

Colorectal cancer: Recent advances in management and treatment

Colorectal cancer (CRC) is the third most common cancer worldwide, and the second most common cause of cancer-related death. In 2020, the estimated number of deaths due to CRC was approximately 930000, accounting for 10% of all cancer deaths worldwide. Accordingly, there is a vast amount of ongoing research aiming to find new and improved treatment modalities for CRC that can potentially increase survival and decrease overall morbidity and mortality. Current management strategies for CRC include surgical procedures for resectable cases, and radiotherapy, chemotherapy, and immunotherapy, in addition to their combination, for non-resectable tumors. Despite these options, CRC remains incurable in 50% of cases. Nonetheless, significant improvements in research techniques have allowed for treatment approaches for CRC to be frequently updated, leading to the availability of new drugs and therapeutic strategies. This review summarizes the most recent therapeutic approaches for CRC, with special emphasis on new strategies that are currently being studied and have great potential to improve the prognosis and lifespan of patients with CRC.

Adaptive Stereotactic Body Radiation Therapy in the Management of Oligometastatic Uterine Leiomyosarcoma: A Clinical Case Report

Safe delivery of stereotactic body radiation therapy (SBRT) to large (>5 cm) oligometastatic abdominopelvic tumors can often be challenging, especially in tumors that require a higher biologically effective dose (BED) for tumor control. Adaptive stereotactic body radiation therapy (A-SBRT) involves inter-fraction and real-time replanning while the patient is on the treatment table, potentially allowing for improved dose coverage and greater sparing of critical structures. Our case report illustrates the benefit of CT-based A-SBRT in the treatment and management of an oligometastatic uterine leiomyosarcoma patient with a rapidly enlarging pelvic recurrence. A 60-year-old female presented to the radiation oncology clinic for treatment of an enlarging, right pelvic oligometastatic leiomyosarcoma. She was prescribed 35 Gy in five fractions. Planning prioritized the sparing of nearby small bowels while maximizing coverage of the planning target volume (PTV). On treatment day, two plans were calculated, the initial plan recalculated on the current CBCT (scheduled plan) and a plan reoptimized using current contours (adapted plan), and the more appropriate one was chosen for delivery. The adapted plan was chosen for all five fractions, with the adapted plan offering better small bowel sparing in five fractions and better target coverage in four fractions, delivering a total of 34 Gy to 95% of the PTV while limiting the small bowel to a maximum point dose of 37 Gy. At approximately six months out from treatment, the patient showed continued radiographic response and resolved acute Grade 1 gastrointestinal toxicity. This case study therefore demonstrates the successful treatment of a large oligometastatic abdominopelvic tumor using CT-based A-SBRT and builds on previous experience treating abdominal cases adaptively.

Custom-Trained Deep Learning-Based Auto-Segmentation for Male Pelvic Iterative CBCT on C-Arm Linear Accelerators

PURPOSE

The purpose of this investigation was to evaluate the clinical applicability of a commercial artificial intelligence-driven deep learning auto-segmentation (DLAS) tool on enhanced iterative cone beam computed tomography (iCBCT) acquisitions for intact prostate and prostate bed treatments.

METHODS AND MATERIALS

DLAS models were trained using 116 iCBCT data sets with manually delineated organs at risk (bladder, femoral heads, and rectum) and target volumes (intact prostate and prostate bed) adhering to institution-specific contouring guidelines. An additional 25 intact prostate and prostate bed iCBCT data sets were used for model testing. Segmentation accuracy relative to a reference structure set was quantified using various geometric comparison metrics and qualitatively evaluated by trained physicists and physicians. These results were compared with those obtained for an additional DLAS-based model trained on planning computed tomography (pCT) data sets and for a deformable image registration (DIR)-based automatic contour propagation method.

RESULTS

In most instances, statistically significant differences in the Dice similarity coefficient (DSC), 95% directed Hausdorff distance, and mean surface distance metrics were observed between the models, as the iCBCT-trained DLAS model outperformed the pCT-trained DLAS model and DIR-based method for all organs at risk and the intact prostate target volume. Mean DSC values for the proposed method were ≥0.90 for these volumes of interest. The iCBCT-trained DLAS model demonstrated a relatively suboptimal performance for the prostate bed segmentation, as the mean DSC value was <0.75 for this target contour. Overall, 90% of bladder, 93% of femoral head, 67% of rectum, and 92% of intact prostate contours generated by the proposed method were deemed clinically acceptable based on qualitative scoring, and approximately 63% of prostate bed contours required moderate or major manual editing to adhere to institutional contouring guidelines.

CONCLUSIONS

The proposed method presents the potential for improved segmentation accuracy and efficiency compared with the DIR-based automatic contour propagation method as commonly applied in CBCT-based dose evaluation and calculation studies.

CT-Based Adaptive Short Course Radiotherapy for Rectal Cancer: A Case Report

Rectal cancer is a highly prevalent malignancy with an increasing number of treatment strategies and sequencing. Short-course radiotherapy (SCRT) combined with chemotherapy has shown to be a feasible treatment option for non-metastatic disease within the total neoadjuvant therapy paradigm. In contrast to conventionally fractionated chemoradiation, SCRT carries less logistical burden and reduces time off systemic therapy, which is particularly beneficial for patients with metastatic disease. We present a case of a 42-year-old male with metastatic rectal cancer who received SCRT using CT-based adaptive radiation, delivering 25 Gy to the pelvis with a simultaneous integrated boost of 30 Gy to the rectal tumor over five treatments. CT-based adaptive radiation was effective and well-tolerated and offers a pathway towards isotoxic radiotherapy dose escalation for patients considered to be inoperable. This case demonstrates the integration of adaptive SCRT as a promising strategy for rectal cancer.

Implementation of Cone Beam Computed Tomography-Guided Online Adaptive Radiotherapy for Challenging Trimodal Therapy in Bladder Preservation: A Report of Two Cases

Muscle invasive bladder cancer (MIBC) is an aggressive disease with a high risk of metastasis. Bladder preservation with trimodality therapy (TMT) is an option for well-selected patients or poor cystectomy candidates. Cone beam computed tomography (CBCT)-guided online adaptive radiotherapy (oART) shows promise in improving the dose to treatment targets while better sparing organs at risk (OARs). The following series presents two cases in which the capabilities of a CBCT-guided oART platform were leveraged to meet clinical challenges. The first case describes a patient with synchronous MIBC and high-risk prostate cancer with challenging target-OAR interfaces. The second recounts the case of a patient with a history of low dose rate (LDR) brachytherapy to the prostate who was later diagnosed with MIBC and successfully treated with CBCT-guided oART with reduced high-dose volume bladder targeting. To date, both patients report minimal side effects and are without disease recurrence. These cases illustrate how CBCT-guided online adaptive systems may efficiently aid radiation oncologists in treating patients with more complex clinical scenarios who desire bladder-sparing therapy.

Morphometric and clinical comparison of MRI-based synthetic CT to conventional CT of the hip in children

BACKGROUND

MRI-based synthetic CT (sCT) generates CT-like images from MRI data.

OBJECTIVE

To evaluate equivalence, inter- and intraobserver reliability, and image quality of sCT compared to conventional (cCT) for assessing hip morphology and maturity in pediatric patients.

MATERIALS AND METHODS

We prospectively enrolled patients <21 years old with cCT and 3T MRI of the hips/pelvis. A dual-echo gradient-echo sequence was used to generate sCT via a commercially available post-processing software (BoneMRI v1.5 research version, MRIguidance BV, Utrecht, NL). Two pediatric musculoskeletal radiologists measured seven morphologic hip parameters. 3D surface distances between cCT and sCT were computed. Physeal status was established at seven locations with cCT as reference standard. Images were qualitatively scored on a 5-point Likert scale regarding diagnostic quality, signal-to-noise ratio, clarity of bony margin, corticomedullary differentiation, and presence and severity of artifacts. Quantitative evaluation of Hounsfield units (HU) was performed in bone, muscle, and fat tissue. Inter- and intraobserver reliability were measured by intraclass correlation coefficients. The cCT-to-sCT intermodal agreement was assessed via Bland-Altman analysis. The equivalence between modalities was tested using paired two one-sided tests. The quality parameter scores of each imaging modality were compared via Wilcoxon signed-rank test. For tissue-specific HU measurements, mean absolute error and mean percentage error values were calculated using the cCT as the reference standard.

RESULTS

Thirty-eight hips in 19 patients were included (16.6 ± 3 years, range 9.9-20.9; male = 5). cCT- and sCT-based morphologic measurements demonstrated good to excellent inter- and intraobserver correlation (0.77

CONCLUSION

sCT is equivalent to cCT for the assessment of hip morphology, physeal status, and radiodensity assessment in pediatric patients.

Collapse

Key Words

• Hip imaging
• Hip morphology
• Hounsfield units
• MRI based
• Quality
• Synthetic CT

Collapse

MESH Headings

• Humans
• Male
• Child
• Female
• Magnetic Resonance Imaging/methods
• Prospective Studies
• Reproducibility of Results
• Adolescent
• Tomography, X-Ray Computed/methods
• Hip Joint/diagnostic imaging
• Child, Preschool
• Imaging, Three-Dimensional/methods

Collapse

Grants Collapse

Affiliation(s)

Jade Iwasaka-Neder Department of Radiology, Boston Children's Hospital, 300 Longwood Ave, Boston, MA, 02115, USA.
M Alejandra Bedoya Department of Radiology, Boston Children's Hospital, 300 Longwood Ave, Boston, MA, 02115, USA
James Connors Department of Radiology, Boston Children's Hospital, 300 Longwood Ave, Boston, MA, 02115, USA
Simon Warfield Computational Radiology Laboratory, Boston Children's Hospital, 401 Park Drive, Boston, MA, 02215, USA
Sarah D Bixby Department of Radiology, Boston Children's Hospital, 300 Longwood Ave, Boston, MA, 02115, USA

Collapse

Novel Technology Allowing Cone Beam Computed Tomography in 6 Seconds: A Patient Study of Comparative Image Quality

PURPOSE

The goal of this study was to evaluate the image quality provided by a novel cone beam computed tomography (CBCT) platform (HyperSight, Varian Medical Systems), a platform with enhanced reconstruction algorithms as well as rapid acquisition times. Image quality was compared with both status quo CBCT for image guidance, and to fan beam CT (FBCT) acquired on a CT simulator (CTsim).

METHODS AND MATERIALS

In a clinical study, 30 individuals were recruited for whom either deep inspiration (DIBH) or deep exhalation breath hold (DEBH) was used during imaging and radiation treatment of tumors involving liver, lung, breast, abdomen, chest wall, and pancreatic sites. All subjects were imaged during breath hold with CBCT on a standard image guidance platform (TrueBeam 2.7, Varian Medical Systems) and FBCT CT (CTsim, GE Optima). HyperSight imaging with both breath hold (HSBH) and free breathing (HSFB) was performed in a single session. The 4 image sets thus acquired were registered and compared using metrics quantifying artifact index, image nonuniformity, contrast, contrast-to-noise ratio, and difference of Hounsfield unit (HU) from CTsim.

RESULTS

HSBH provided less severe artifacts compared with both HSFB and TrueBeam. The severity of artifacts in HSBH images was similar to that in CTsim images, with statistically similar artifact index values. CTsim provided the best image uniformity; however, HSBH provided improved uniformity compared with both HSFB and TrueBeam. CTsim demonstrated elevated contrast compared with HyperSight imaging, but both HSBH and HSFB imaging showed superior contrast-to-noise ratio characteristics compared with TrueBeam. The median HU difference of HSBH from CTsim was within 1 HU for muscle/fat tissue, 12 HU for bone, and 14 HU for lung.

CONCLUSIONS

The HyperSight system provides 6-second CBCT acquisition with image artifacts that are significantly reduced compared with TrueBeam and comparable to those in CTsim FBCT imaging. HyperSight breath hold imaging was of higher quality compared with free breathing imaging on the same system. The median HU value in HyperSight breath hold imaging is within 15 HU of that in CTsim imaging for muscle, fat, bone, and lung tissue types, indicating the utility of image data for direct dose calculation in adaptive workflows.

Intricacies of Human-AI Interaction in Dynamic Decision-Making for Precision Oncology: A Case Study in Response-Adaptive Radiotherapy

Background

Adaptive treatment strategies that can dynamically react to individual cancer progression can provide effective personalized care. Longitudinal multi-omics information, paired with an artificially intelligent clinical decision support system (AI-CDSS) can assist clinicians in determining optimal therapeutic options and treatment adaptations. However, AI-CDSS is not perfectly accurate, as such, clinicians' over/under reliance on AI may lead to unintended consequences, ultimately failing to develop optimal strategies. To investigate such collaborative decision-making process, we conducted a Human-AI interaction case study on response-adaptive radiotherapy (RT).

Methods

We designed and conducted a two-phase study for two disease sites and two treatment modalities-adaptive RT for non-small cell lung cancer (NSCLC) and adaptive stereotactic body RT for hepatocellular carcinoma (HCC)-in which clinicians were asked to consider mid-treatment modification of the dose per fraction for a number of retrospective cancer patients without AI-support (Unassisted Phase) and with AI-assistance (AI-assisted Phase). The AI-CDSS graphically presented trade-offs in tumor control and the likelihood of toxicity to organs at risk, provided an optimal recommendation, and associated model uncertainties. In addition, we asked for clinicians' decision confidence level and trust level in individual AI recommendations and encouraged them to provide written remarks. We enrolled 13 evaluators (radiation oncology physicians and residents) from two medical institutions located in two different states, out of which, 4 evaluators volunteered in both NSCLC and HCC studies, resulting in a total of 17 completed evaluations (9 NSCLC, and 8 HCC). To limit the evaluation time to under an hour, we selected 8 treated patients for NSCLC and 9 for HCC, resulting in a total of 144 sets of evaluations (72 from NSCLC and 72 from HCC). Evaluation for each patient consisted of 8 required inputs and 2 optional remarks, resulting in up to a total of 1440 data points.

Results

AI-assistance did not homogeneously influence all experts and clinical decisions. From NSCLC cohort, 41 (57%) decisions and from HCC cohort, 34 (47%) decisions were adjusted after AI assistance. Two evaluations (12%) from the NSCLC cohort had zero decision adjustments, while the remaining 15 (88%) evaluations resulted in at least two decision adjustments. Decision adjustment level positively correlated with dissimilarity in decision-making with AI [NSCLC: ρ = 0.53 ( p < 0.001); HCC: ρ = 0.60 ( p < 0.001)] indicating that evaluators adjusted their decision closer towards AI recommendation. Agreement with AI-recommendation positively correlated with AI Trust Level [NSCLC: ρ = 0.59 ( p < 0.001); HCC: ρ = 0.7 ( p < 0.001)] indicating that evaluators followed AI's recommendation if they agreed with that recommendation. The correlation between decision confidence changes and decision adjustment level showed an opposite trend [NSCLC: ρ = -0.24 ( p = 0.045), HCC: ρ = 0.28 ( p = 0.017)] reflecting the difference in behavior due to underlying differences in disease type and treatment modality. Decision confidence positively correlated with the closeness of decisions to the standard of care (NSCLC: 2 Gy/fx; HCC: 10 Gy/fx) indicating that evaluators were generally more confident in prescribing dose fractionations more similar to those used in standard clinical practice. Inter-evaluator agreement increased with AI-assistance indicating that AI-assistance can decrease inter-physician variability. The majority of decisions were adjusted to achieve higher tumor control in NSCLC and lower normal tissue complications in HCC. Analysis of evaluators' remarks indicated concerns for organs at risk and RT outcome estimates as important decision-making factors.

Conclusions

Human-AI interaction depends on the complex interrelationship between expert's prior knowledge and preferences, patient's state, disease site, treatment modality, model transparency, and AI's learned behavior and biases. The collaborative decision-making process can be summarized as follows: (i) some clinicians may not believe in an AI system, completely disregarding its recommendation, (ii) some clinicians may believe in the AI system but will critically analyze its recommendations on a case-by-case basis; (iii) when a clinician finds that the AI recommendation indicates the possibility for better outcomes they will adjust their decisions accordingly; and (iv) When a clinician finds that the AI recommendation indicate a worse possible outcome they will disregard it and seek their own alternative approach.