Utilizing Large Language Models for Enhanced Clinical Trial Matching: A Study on Automation in Patient Screening.

Background Clinical trial matching, essential for advancing medical research, involves detailed screening of potential participants to ensure alignment with specific trial requirements. Research staff face challenges due to the high volume of eligible patients and the complexity of varying eligibility criteria. The traditional manual process, both time-consuming and error-prone, often leads to missed opportunities. Recently, large language models (LLMs), specifically generative pre-trained transformers (GPTs), have become impressive and impactful tools. Utilizing such tools from artificial intelligence (AI) and natural language processing (NLP) may enhance the accuracy and efficiency of this process through automated patient screening against established criteria. Methods Utilizing data from the National NLP Clinical Challenges (n2c2) 2018 Challenge, we utilized 202 longitudinal patient records. These records were annotated by medical professionals and evaluated against 13 selection criteria encompassing various health assessments. Our approach involved embedding medical documents into a vector database to determine relevant document sections and then using an LLM (OpenAI's GPT-3.5 Turbo and GPT-4) in tandem with structured and chain-of-thought prompting techniques for systematic document assessment against the criteria. Misclassified criteria were also examined to identify classification challenges. Results This study achieved an accuracy of 0.81, sensitivity of 0.80, specificity of 0.82, and a micro F1 score of 0.79 using GPT-3.5 Turbo, and an accuracy of 0.87, sensitivity of 0.85, specificity of 0.89, and micro F1 score of 0.86 using GPT-4. Notably, some criteria in the ground truth appeared mislabeled, an issue we couldn't explore further due to insufficient label generation guidelines on the website. Conclusion Our findings underscore the potential of AI and NLP technologies, including LLMs, in the clinical trial matching process. The study demonstrated strong capabilities in identifying eligible patients and minimizing false inclusions. Such automated systems promise to alleviate the workload of research staff and improve clinical trial enrollment, thus accelerating the process and enhancing the overall feasibility of clinical research. Further work is needed to determine the potential of this approach when implemented on real clinical data.

Multimodal radiotherapy dose prediction using a multi-task deep learning model.

BACKGROUND: In radiation therapy (RT), accelerated partial breast irradiation (APBI) has emerged as an increasingly preferred treatment modality over conventional whole breast irradiation due to its targeted dose delivery and shorter course of treatment. APBI can be delivered through various modalities including Cobalt-60-based systems and linear accelerators with C-arm, O-ring, or robotic arm design. Each modality possesses distinct features, such as beam energy or the degrees of freedom in treatment planning, which influence their respective dose distributions. These modality-specific considerations emphasize the need for a quantitative approach in determining the optimal dose delivery modality on a patient-specific basis. However, manually generating treatment plans for each modality across every patient is time-consuming and clinically impractical. PURPOSE: We aim to develop an efficient and personalized approach for determining the optimal RT modality for APBI by training predictive models using two different deep learning-based convolutional neural networks. The baseline network performs a single-task (ST), predicting dose for a single modality. Our proposed multi-task (MT) network, which is capable of leveraging shared information among different tasks, can concurrently predict dose distributions for various RT modalities. Utilizing patient-specific input data, such as a patient's computed tomography (CT) scan and treatment protocol dosimetric goals, the MT model predicts patient-specific dose distributions across all trained modalities. These dose distributions provide patients and clinicians quantitative insights, facilitating informed and personalized modality comparison prior to treatment planning. METHODS: The dataset, comprising 28 APBI patients and their 92 treatment plans, was partitioned into training, validation, and test subsets. Eight patients were dedicated to the test subset, leaving 68 treatment plans across 20 patients to divide between the training and validation subsets. ST models were trained for each modality, and one MT model was trained to predict doses for all modalities simultaneously. Model performance was evaluated across the test dataset in terms of Mean Absolute Percent Error (MAPE). We conducted statistical analysis of model performance using the two-tailed Wilcoxon signed-rank test. RESULTS: Training times for five ST models ranged from 255 to 430 min per modality, totaling 1925 min, while the MT model required 2384 min. MT model prediction required an average of 1.82 s per patient, compared to ST model predictions at 0.93 s per modality. The MT model yielded MAPE of 1.1033 ± 0.3627% as opposed to the collective MAPE of 1.2386 ± 0.3872% from ST models, and the differences were statistically significant (p = 0.0003, 95% confidence interval = [-0.0865, -0.0712]). CONCLUSION: Our study highlights the potential benefits of a MT learning framework in predicting RT dose distributions across various modalities without notable compromises. This MT architecture approach offers several advantages, such as flexibility, scalability, and streamlined model management, making it an appealing solution for clinical deployment. With such a MT model, patients can make more informed treatment decisions, physicians gain more quantitative insight for pre-treatment decision-making, and clinics can better optimize resource allocation. With our proposed goal array and MT framework, we aim to expand this work to a site-agnostic dose prediction model, enhancing its generalizability and applicability.

Deep learning based automatic segmentation of the Internal Pudendal Artery in definitive radiotherapy treatment planning of localized prostate cancer.

Background and purpose: Radiation-induced erectile dysfunction (RiED) commonly affects prostate cancer patients, prompting clinical trials across institutions to explore dose-sparing to internal-pudendal-arteries (IPA) for preserving sexual potency. IPA, challenging to segment, isn't conventionally considered an organ-at-risk (OAR). This study proposes a deep learning (DL) auto-segmentation model for IPA, using Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) or CT alone to accommodate varied clinical practices. Materials and methods: A total of 86 patients with CT and MRI images and noisy IPA labels were recruited in this study. We split the data into 42/14/30 for model training, testing, and a clinical observer study, respectively. There were three major innovations in this model: 1) we designed an architecture with squeeze-and-excite blocks and modality attention for effective feature extraction and production of accurate segmentation, 2) a novel loss function was used for training the model effectively with noisy labels, and 3) modality dropout strategy was used for making the model capable of segmentation in the absence of MRI. Results: Test dataset metrics were DSC 61.71 ± 7.7 %, ASD 2.5 ± .87 mm, and HD95 7.0 ± 2.3 mm. AI segmented contours showed dosimetric similarity to expert physician's contours. Observer study indicated higher scores for AI contours (mean = 3.7) compared to inexperienced physicians' contours (mean = 3.1). Inexperienced physicians improved scores to 3.7 when starting with AI contours. Conclusion: The proposed model achieved good quality IPA contours to improve uniformity of segmentation and to facilitate introduction of standardized IPA segmentation into clinical trials and practice.

Rethinking the potential role of dose painting in personalized ultra-fractionated stereotactic adaptive radiotherapy.

Fractionated radiotherapy was established in the 1920s based upon two principles: (1) delivering daily treatments of equal quantity, unless the clinical situation requires adjustment, and (2) defining a specific treatment period to deliver a total dosage. Modern fractionated radiotherapy continues to adhere to these century-old principles, despite significant advancements in our understanding of radiobiology. At UT Southwestern, we are exploring a novel treatment approach called PULSAR (Personalized Ultra-Fractionated Stereotactic Adaptive Radiotherapy). This method involves administering tumoricidal doses in a pulse mode with extended intervals, typically spanning weeks or even a month. Extended intervals permit substantial recovery of normal tissues and afford the tumor and tumor microenvironment ample time to undergo significant changes, enabling more meaningful adaptation in response to the evolving characteristics of the tumor. The notion of dose painting in the realm of radiation therapy has long been a subject of contention. The debate primarily revolves around its clinical effectiveness and optimal methods of implementation. In this perspective, we discuss two facets concerning the potential integration of dose painting with PULSAR, along with several practical considerations. If successful, the combination of the two may not only provide another level of personal adaptation ("adaptive dose painting"), but also contribute to the establishment of a timely feedback loop throughout the treatment process. To substantiate our perspective, we conducted a fundamental modeling study focusing on PET-guided dose painting, incorporating tumor heterogeneity and tumor control probability (TCP).

An AI-based approach for modeling the synergy between radiotherapy and immunotherapy.

Personalized, ultra-fractionated stereotactic adaptive radiotherapy (PULSAR) is designed to administer tumoricidal doses in a pulsed mode with extended intervals, spanning weeks or months. This approach leverages longer intervals to adapt the treatment plan based on tumor changes and enhance immune-modulated effects. In this investigation, we seek to elucidate the potential synergy between combined PULSAR and PD-L1 blockade immunotherapy using experimental data from a Lewis Lung Carcinoma (LLC) syngeneic murine cancer model. Employing a long short-term memory (LSTM) recurrent neural network (RNN) model, we simulated the treatment response by treating irradiation and anti-PD-L1 as external stimuli occurring in a temporal sequence. Our findings demonstrate that: (1) The model can simulate tumor growth by integrating various parameters such as timing and dose, and (2) The model provides mechanistic interpretations of a "causal relationship" in combined treatment, offering a completely novel perspective. The model can be utilized for in-silico modeling, facilitating exploration of innovative treatment combinations to optimize therapeutic outcomes. Advanced modeling techniques, coupled with additional efforts in biomarker identification, may deepen our understanding of the biological mechanisms underlying the combined treatment.

Simulation and pre-planning omitted radiotherapy (SPORT): a feasibility study for prostate cancer.

This study explored the feasibility of on-couch intensity modulated radiotherapy (IMRT) planning for prostate cancer (PCa) on a cone-beam CT (CBCT)-based online adaptive RT platform without an individualized pre-treatment plan and contours. Ten patients with PCa previously treated with image-guided IMRT (60 Gy/20 fractions) were selected. In contrast to the routine online adaptive RT workflow, a novel approach was employed in which the same preplan that was optimized on one reference patient was adapted to generate individual on-couch/initial plans for the other nine test patients using Ethos emulator. Simulation CTs of the test patients were used as simulated online CBCT (sCBCT) for emulation. Quality assessments were conducted on synthetic CTs (sCT). Dosimetric comparisons were performed between on-couch plans, on-couch plans recomputed on the sCBCT and individually optimized plans for test patients. The median value of mean absolute difference between sCT and sCBCT was 74.7 HU (range 69.5-91.5 HU). The average CTV/PTV coverage by prescription dose was 100.0%/94.7%, and normal tissue constraints were met for the nine test patients in on-couch plans on sCT. Recalculating on-couch plans on the sCBCT showed about 0.7% reduction of PTV coverage and a 0.6% increasing of hotspot, and the dose difference of the OARs was negligible (<0.5 Gy). Hence, initial IMRT plans for new patients can be generated by adapting a reference patient's preplan with online contours, which had similar qualities to the conventional approach of individually optimized plan on the simulation CT. Further study is needed to identify selection criteria for patient anatomy most amenable to this workflow.

Zero-Shot Medical Image Translation via Frequency-Guided Diffusion Models.

Recently, the diffusion model has emerged as a superior generative model that can produce high quality and realistic images. However, for medical image translation, the existing diffusion models are deficient in accurately retaining structural information since the structure details of source domain images are lost during the forward diffusion process and cannot be fully recovered through learned reverse diffusion, while the integrity of anatomical structures is extremely important in medical images. For instance, errors in image translation may distort, shift, or even remove structures and tumors, leading to incorrect diagnosis and inadequate treatments. Training and conditioning diffusion models using paired source and target images with matching anatomy can help. However, such paired data are very difficult and costly to obtain, and may also reduce the robustness of the developed model to out-of-distribution testing data. We propose a frequency-guided diffusion model (FGDM) that employs frequency-domain filters to guide the diffusion model for structure-preserving image translation. Based on its design, FGDM allows zero-shot learning, as it can be trained solely on the data from the target domain, and used directly for source-to-target domain translation without any exposure to the source-domain data during training. We evaluated it on three cone-beam CT (CBCT)-to-CT translation tasks for different anatomical sites, and a cross-institutional MR imaging translation task. FGDM outperformed the state-of-the-art methods (GAN-based, VAE-based, and diffusion-based) in metrics of Fréchet Inception Distance (FID), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index Measure (SSIM), showing its significant advantages in zero-shot medical image translation.

ART2Dose: A comprehensive dose verification platform for online adaptive radiotherapy.

BACKGROUND: Online adaptive radiotherapy (ART) involves the development of adaptable treatment plans that consider patient anatomical data obtained right prior to treatment administration, facilitated by cone-beam computed tomography guided adaptive radiotherapy (CTgART) and magnetic resonance image-guided adaptive radiotherapy (MRgART). To ensure accuracy of these adaptive plans, it is crucial to conduct calculation-based checks and independent verification of volumetric dose distribution, as measurement-based checks are not practical within online workflows. However, the absence of comprehensive, efficient, and highly integrated commercial software for secondary dose verification can impede the time-sensitive nature of online ART procedures. PURPOSE: The main aim of this study is to introduce an efficient online quality assurance (QA) platform for online ART, and subsequently evaluate it on Ethos and Unity treatment delivery systems in our clinic. METHODS: To enhance efficiency and ensure compliance with safety standards in online ART, ART2Dose, a secondary dose verification software, has been developed and integrated into our online QA workflow. This implementation spans all online ART treatments at our institution. The ART2Dose infrastructure comprises four key components: an SQLite database, a dose calculation server, a report generator, and a web portal. Through this infrastructure, file transfer, dose calculation, report generation, and report approval/archival are seamlessly managed, minimizing the need for user input when exporting RT DICOM files and approving the generated QA report. ART2Dose was compared with Mobius3D in pre-clinical evaluations on secondary dose verification for 40 adaptive plans. Additionally, a retrospective investigation was conducted utilizing 1302 CTgART fractions from ten treatment sites and 1278 MRgART fractions from seven treatment sites to evaluate the practical accuracy and efficiency of ART2Dose in routine clinical use. RESULTS: With dedicated infrastructure and an integrated workflow, ART2Dose achieved gamma passing rates that were comparable to or higher than those of Mobius3D. Additionally, it significantly reduced the time required to complete pre-treatment checks by 3-4 min for each plan. In the retrospective analysis of clinical CTgART and MRgART fractions, ART2Dose demonstrated average gamma passing rates of 99.61 ± 0.83% and 97.75 ± 2.54%, respectively, using the 3%/2 mm criteria for region greater than 10% of prescription dose. The average calculation times for CTgART and MRgART were approximately 1 and 2 min, respectively. CONCLUSION: Overall, the streamlined implementation of ART2Dose notably enhances the online ART workflow, offering reliable and efficient online QA while reducing time pressure in the clinic and minimizing labor-intensive work.

Cancer Informatics for Cancer Centers: Sharing Ideas on How to Build an Artificial Intelligence-Ready Informatics Ecosystem for Radiation Oncology.

In August 2022, the Cancer Informatics for Cancer Centers brought together cancer informatics leaders for its biannual symposium, Precision Medicine Applications in Radiation Oncology, co-chaired by Quynh-Thu Le, MD (Stanford University), and Walter J. Curran, MD (GenesisCare). Over the course of 3 days, presenters discussed a range of topics relevant to radiation oncology and the cancer informatics community more broadly, including biomarker development, decision support algorithms, novel imaging tools, theranostics, and artificial intelligence (AI) for the radiotherapy workflow. Since the symposium, there has been an impressive shift in the promise and potential for integration of AI in clinical care, accelerated in large part by major advances in generative AI. AI is now poised more than ever to revolutionize cancer care. Radiation oncology is a field that uses and generates a large amount of digital data and is therefore likely to be one of the first fields to be transformed by AI. As experts in the collection, management, and analysis of these data, the informatics community will take a leading role in ensuring that radiation oncology is prepared to take full advantage of these technological advances. In this report, we provide highlights from the symposium, which took place in Santa Barbara, California, from August 29 to 31, 2022. We discuss lessons learned from the symposium for data acquisition, management, representation, and sharing, and put these themes into context to prepare radiation oncology for the successful and safe integration of AI and informatics technologies.

Real-time liver motion estimation via deep learning-based angle-agnostic X-ray imaging.

BACKGROUND: Real-time liver imaging is challenged by the short imaging time (within hundreds of milliseconds) to meet the temporal constraint posted by rapid patient breathing, resulting in extreme under-sampling for desired 3D imaging. Deep learning (DL)-based real-time imaging/motion estimation techniques are emerging as promising solutions, which can use a single X-ray projection to estimate 3D moving liver volumes by solved deformable motion. However, such techniques were mostly developed for a specific, fixed X-ray projection angle, thereby impractical to verify and guide arc-based radiotherapy with continuous gantry rotation. PURPOSE: To enable deformable motion estimation and 3D liver imaging from individual X-ray projections acquired at arbitrary X-ray scan angles, and to further improve the accuracy of single X-ray-driven motion estimation. METHODS: We developed a DL-based method, X360, to estimate the deformable motion of the liver boundary using an X-ray projection acquired at an arbitrary gantry angle (angle-agnostic). X360 incorporated patient-specific prior information from planning 4D-CTs to address the under-sampling issue, and adopted a deformation-driven approach to deform a prior liver surface mesh to new meshes that reflect real-time motion. The liver mesh motion is solved via motion-related image features encoded in the arbitrary-angle X-ray projection, and through a sequential combination of rigid and deformable registration modules. To achieve the angle agnosticism, a geometry-informed X-ray feature pooling layer was developed to allow X360 to extract angle-dependent image features for motion estimation. As a liver boundary motion solver, X360 was also combined with priorly-developed, DL-based optical surface imaging and biomechanical modeling techniques for intra-liver motion estimation and tumor localization. RESULTS: With geometry-aware feature pooling, X360 can solve the liver boundary motion from an arbitrary-angle X-ray projection. Evaluated on a set of 10 liver patient cases, the mean (± s.d.) 95-percentile Hausdorff distance between the solved liver boundary and the "ground-truth" decreased from 10.9 (±4.5) mm (before motion estimation) to 5.5 (±1.9) mm (X360). When X360 was further integrated with surface imaging and biomechanical modeling for liver tumor localization, the mean (± s.d.) center-of-mass localization error of the liver tumors decreased from 9.4 (± 5.1) mm to 2.2 (± 1.7) mm. CONCLUSION: X360 can achieve fast and robust liver boundary motion estimation from arbitrary-angle X-ray projections for real-time imaging guidance. Serving as a surface motion solver, X360 can be integrated into a combined framework to achieve accurate, real-time, and marker-less liver tumor localization.

Single patient learning for adaptive radiotherapy dose prediction.

BACKGROUND: Throughout a patient's course of radiation therapy, maintaining accuracy of their initial treatment plan over time is challenging due to anatomical changes-for example, stemming from patient weight loss or tumor shrinkage. Online adaptation of their RT plan to these changes is crucial, but hindered by manual and time-consuming processes. While deep learning (DL) based solutions have shown promise in streamlining adaptive radiation therapy (ART) workflows, they often require large and extensive datasets to train population-based models. PURPOSE: This study extends our prior research by introducing a minimalist approach to patient-specific adaptive dose prediction. In contrast to our prior method, which involved fine-tuning a pre-trained population model, this new method trains a model from scratch using only a patient's initial treatment data. This patient-specific dose predictor aims to enhance clinical accessibility, thereby empowering physicians and treatment planners to make more informed, quantitative decisions in ART. We hypothesize that patient-specific DL models will provide more accurate adaptive dose predictions for their respective patients compared to a population-based DL model. METHODS: We selected 33 patients to train an adaptive population-based (AP) model. Ten additional patients were selected, and their respective initial RT data served as single samples for training patient-specific (PS) models. These 10 patients contained an additional 26 ART plans that were withheld as the test dataset to evaluate AP versus PS model dose prediction performance. We assessed model performance using Mean Absolute Percent Error (MAPE) by comparing predicted doses to the originally delivered ground truth doses. We used the Wilcoxon signed-rank test to determine statistically significant differences in terms of MAPE between the AP and PS model results across the test dataset. Furthermore, we calculated differences between predicted and ground truth mean doses for segmented structures and determined statistical significance in the differences for each of them. RESULTS: The average MAPE across AP and PS model dose predictions was 5.759% and 4.069%, respectively. The Wilcoxon signed-rank test yielded two-tailed p-value =  2.9802 × 10 - 8 $2.9802\ \times \ {10}^{ - 8}$ , indicating that the MAPE differences between the AP and PS model dose predictions are statistically significant, and 95% confidence interval = [-2.1610, -1.0130], indicating 95% confidence that the MAPE difference between the AP and PS models for a population lies in this range. Out of 24 total segmented structures, the comparison of mean dose differences for 12 structures indicated statistical significance with two-tailed p-values < 0.05. CONCLUSION: Our study demonstrates the potential of patient-specific deep learning models in application to ART. Notably, our method streamlines the training process by minimizing the size of the required training dataset, as only a single patient's initial treatment data is required. External institutions considering the implementation of such a technology could package such a model so that it only requires the upload of a reference treatment plan for model training and deployment. Our single patient learning strategy demonstrates promise in ART due to its minimal dataset requirement and its utility in personalization of cancer treatment.

Intentional deep overfit learning for patient-specific dose predictions in adaptive radiotherapy.

BACKGROUND: The framework of adaptive radiation therapy (ART) was crafted to address the underlying sources of intra-patient variation that were observed throughout numerous patients' radiation sessions. ART seeks to minimize the consequential dosimetric uncertainty resulting from this daily variation, commonly through treatment planning re-optimization. Re-optimization typically consists of manual evaluation and modification of previously utilized optimization criteria. Ideally, frequent treatment plan adaptation through re-optimization on each day's computed tomography (CT) scan may improve dosimetric accuracy and minimize dose delivered to organs at risk (OARs) as the planning target volume (PTV) changes throughout the course of treatment. PURPOSE: Re-optimization in its current form is time-consuming and inefficient. In response to this ART bottleneck, we propose a deep learning based adaptive dose prediction model that utilizes a head and neck (H&N) patient's initial planning data to fine-tune a previously trained population model towards a patient-specific model. Our fine-tuned, patient-specific (FT-PS) model, which is trained using the intentional deep overfit learning (IDOL) method, may enable clinicians and treatment planners to rapidly evaluate relevant dosimetric changes daily and re-optimize accordingly. METHODS: An adaptive population (AP) model was trained using adaptive data from 33 patients. Separately, 10 patients were selected for training FT-PS models. The previously trained AP model was utilized as the base model weights prior to re-initializing model training for each FT-PS model. Ten FT-PS models were separately trained by fine-tuning the previous model weights based on each respective patient's initial treatment plan. From these 10 patients, 26 ART treatment plans were withheld from training as the test dataset for retrospective evaluation of dose prediction performance between the AP and FT-PS models. Each AP and FT-PS dose prediction was compared against the ground truth dose distribution as originally generated during the patient's course of treatment. Mean absolute percent error (MAPE) evaluated the dose differences between a model's prediction and the ground truth. RESULTS: MAPE was calculated within the 10% isodose volume region of interest for each of the AP and FT-PS models dose predictions and averaged across all test adaptive sessions, yielding 5.759% and 3.747% respectively. MAPE differences were compared between AP and FT-PS models across each test session in a test of statistical significance. The differences were statistically significant in a paired t-test with two-tailed p-value equal to 3.851 × 10 - 9 $3.851 \times {10}^{ - 9}$ and 95% confidence interval (CI) equal to [-2.483, -1.542]. Furthermore, MAPE was calculated using each individually segmented structure as an ROI. Nineteen of 24 structures demonstrated statistically significant differences between the AP and FT-PS models. CONCLUSION: We utilized the IDOL method to fine-tune a population-based dose prediction model into an adaptive, patient-specific model. The averaged MAPE across the test dataset was 5.759% for the population-based model versus 3.747% for the fine-tuned, patient-specific model, and the difference in MAPE between models was found to be statistically significant. Our work demonstrates the feasibility of patient-specific models in adaptive radiotherapy, and offers unique clinical benefit by utilizing initial planning data that contains the physician's treatment intent.

ChatDoctor: A Medical Chat Model Fine-Tuned on a Large Language Model Meta-AI (LLaMA) Using Medical Domain Knowledge.

Objective The primary aim of this research was to address the limitations observed in the medical knowledge of prevalent large language models (LLMs) such as ChatGPT, by creating a specialized language model with enhanced accuracy in medical advice. Methods We achieved this by adapting and refining the large language model meta-AI (LLaMA) using a large dataset of 100,000 patient-doctor dialogues sourced from a widely used online medical consultation platform. These conversations were cleaned and anonymized to respect privacy concerns. In addition to the model refinement, we incorporated a self-directed information retrieval mechanism, allowing the model to access and utilize real-time information from online sources like Wikipedia and data from curated offline medical databases. Results The fine-tuning of the model with real-world patient-doctor interactions significantly improved the model's ability to understand patient needs and provide informed advice. By equipping the model with self-directed information retrieval from reliable online and offline sources, we observed substantial improvements in the accuracy of its responses. Conclusion Our proposed ChatDoctor, represents a significant advancement in medical LLMs, demonstrating a significant improvement in understanding patient inquiries and providing accurate advice. Given the high stakes and low error tolerance in the medical field, such enhancements in providing accurate and reliable information are not only beneficial but essential.

Evaluation of fan-beam kilovoltage computed tomography image quality on a novel biological-guided radiotherapy platform.

Background and Purpose: A recently developed biology-guided radiotherapy platform, equipped with positron emission tomography (PET) and computed tomography (CT), provides both anatomical and functional image guidance for radiotherapy. This study aimed to characterize performance of the kilovoltage CT (kVCT) system on this platform using standard quality metrics measured on phantom and patient images, using CT simulator images as reference. Materials and Methods: Image quality metrics, including spatial resolution/modular transfer function (MTF), slice sensitivity profile (SSP), noise performance and image uniformity, contrast-noise ratio (CNR) and low-contrast resolution, geometric accuracy, and CT number (HU) accuracy, were evaluated on phantom images. Patient images were evaluated mainly qualitatively. Results: On phantom images the MTF10% is about 0.68 lp/mm for kVCT in PET/CT Linac. The SSP agreed with nominal slice thickness within 0.7 mm. The diameter of the smallest visible target (1% contrast) is about 5 mm using medium dose mode. The image uniformity is within 2.0 HU. The geometric accuracy tests passed within 0.5 mm. Relative to CT simulator images, the noise is generally higher and the CNR is lower in PET/CT Linac kVCT images. The CT number accuracy is comparable between the two systems with maximum deviation from the phantom manufacturer range within 25 HU. On patient images, higher spatial resolution and image noise are observed on PET/CT Linac kVCT images. Conclusions: Major image quality metrics of the PET/CT Linac kVCT were within vendor-recommended tolerances. Better spatial resolution but higher noise and better/comparable low contrast visibility were observed as compared to a CT simulator when images were acquired with clinical protocols.

Bony structure enhanced synthetic CT generation using Dixon sequences for pelvis MR-only radiotherapy.

BACKGROUND: MRI-only radiotherapy planning (MROP) is beneficial to patients by avoiding MRI/CT registration errors, simplifying the radiation treatment simulation workflow and reducing exposure to ionizing radiation. MRI is the primary imaging modality for soft tissue delineation. Treatment planning CTs (i.e., CT simulation scan) are redundant if a synthetic CT (sCT) can be generated from the MRI to provide the patient positioning and electron density information. Unsupervised deep learning (DL) models like CycleGAN are widely used in MR-to-sCT conversion, when paired patient CT and MR image datasets are not available for model training. However, compared to supervised DL models, they cannot guarantee anatomic consistency, especially around bone. PURPOSE: The purpose of this work was to improve the sCT accuracy generated from MRI around bone for MROP. METHODS: To generate more reliable bony structures on sCT images, we proposed to add bony structure constraints in the unsupervised CycleGAN model's loss function and leverage Dixon constructed fat and in-phase (IP) MR images. Dixon images provide better bone contrast than T2-weighted images as inputs to a modified multi-channel CycleGAN. A private dataset with a total of 31 prostate cancer patients were used for training (20) and testing (11). RESULTS: We compared model performance with and without bony structure constraints using single- and multi-channel inputs. Among all the models, multi-channel CycleGAN with bony structure constraints had the lowest mean absolute error, both inside the bone and whole body (50.7 and 145.2 HU). This approach also resulted in the highest Dice similarity coefficient (0.88) of all bony structures compared with the planning CT. CONCLUSION: Modified multi-channel CycleGAN with bony structure constraints, taking Dixon-constructed fat and IP images as inputs, can generate clinically suitable sCT images in both bone and soft tissue. The generated sCT images have the potential to be used for accurate dose calculation and patient positioning in MROP radiation therapy.

Evaluating machine learning enhanced intelligent-optimization-engine (IOE) performance for ethos head-and-neck (HN) plan generation.

PURPOSE: Varian Ethos utilizes novel intelligent-optimization-engine (IOE) designed to automate the planning. However, this introduced a black box approach to plan optimization and challenge for planners to improve plan quality. This study aims to evaluate machine-learning-guided initial reference plan generation approaches for head & neck (H&N) adaptive radiotherapy (ART). METHODS: Twenty previously treated patients treated on C-arm/Ring-mounted were retroactively re-planned in the Ethos planning system using a fixed 18-beam intensity-modulated radiotherapy (IMRT) template. Clinical goals for IOE input were generated using (1) in-house deep-learning 3D-dose predictor (AI-Guided) (2) commercial knowledge-based planning (KBP) model with universal RTOG-based population criteria (KBP-RTOG) and (3) an RTOG-based constraint template only (RTOG) for in-depth analysis of IOE sensitivity. Similar training data was utilized for both models. Plans were optimized until their respective criteria were achieved or DVH-estimation band was satisfied. Plans were normalized such that the highest PTV dose level received 95% coverage. Target coverage, high-impact organs-at-risk (OAR) and plan deliverability was assessed in comparison to clinical (benchmark) plans. Statistical significance was evaluated using a paired two-tailed student t-test. RESULTS: AI-guided plans were superior to both KBP-RTOG and RTOG-only plans with respect to clinical benchmark cases. Overall, OAR doses were comparable or improved with AI-guided plans versus benchmark, while they increased with KBP-RTOG and RTOG plans. However, all plans generally satisfied the RTOG criteria. Heterogeneity Index (HI) was on average <1.07 for all plans. Average modulation factor was 12.2 ± 1.9 (p = n.s), 13.1 ± 1.4 (p = <0.001), 11.5 ± 1.3 (p = n.s.) and 12.2 ± 1.9 for KBP-RTOG, AI-Guided, RTOG and benchmark plans, respectively. CONCLUSION: AI-guided plans were the highest quality. Both KBP-enabled and RTOG-only plans are feasible approaches as clinics adopt ART workflows. Similar to constrained optimization, the IOE is sensitive to clinical input goals and we recommend comparable input to an institution's planning directive dosimetric criteria.

Artificial intelligence guided physician directive improves head and neck planning quality and practice Uniformity: A prospective study.

•AI dose predictor was fully integrated with treatment planning system and used as a physicain decision support tool to improve uniformity of practice.•Model was trained based on our standard of practice, but implemented at the time of expansion with 3 new physicians join the practice.•Phase 1 retrospective evaluation demonstrated the non-uniform practice among 3 MDs and only 52.9% frequency planner can achieve physicians' directives.•Significant improvement in practice uniformity of practice was observed after utilizing AI as DST and 80.4% frequency clinical plan can achieve AI-guided physician directives.

Clinical experience on patient-specific quality assurance for CBCT-based online adaptive treatment plan.

PURPOSE: Ethos CBCT-based adaptive radiotherapy (ART) system can generate an online adaptive plan by re-optimizing the initial reference plan based on the patient anatomy at the treatment. The optimization process is fully automated without any room for human intervention. Due to the change in anatomy, the ART plan can be significantly different from the initial plan in terms of plan parameters such as the aperture shapes and number of monitor units (MUs). In this study, we investigated the feasibility of using calculation-based patient specific QA for ART plans in conjunction with measurement-based and calculation-based QA for initial plans to establish an action level for the online ART patient-specific QA. METHODS: A cohort of 98 cases treated on CBCT-based ART system were collected for this study. We performed measurement-based QA using ArcCheck and calculation-based QA using Mobius for both the initial plan and the ART plan for analysis. For online the ART plan, Mobius calculation was conducted prior to the delivery, while ArcCheck measurement was delivered on the same day after the treatment. We first investigated the modulation factors (MFs) and MU numbers of the initial plans and ART plans, respectively. The γ passing rates of initial and ART plan QA were analyzed. Then action limits were derived for QA calculation and measurement for both initial and online ART plans, respectively, from 30 randomly selected patient cases, and were evaluated using the other 68 patient cases. RESULTS: The difference in MF between initial plan and ART-plan was 12.9% ± 12.7% which demonstrates their significant difference in plan parameters. Based on the patient QA results, pre-treatment calculation and measurement results are generally well aligned with ArcCheck measurement results for online ART plans, illustrating their feasibility as an indicator of failure in online ART QA measurements. Furthermore, using 30 randomly selected patient cases, the γ analysis action limit derived for initial plans and ART plans are 89.6% and 90.4% in ArcCheck QA (2%/2 mm) and are 92.4% and 93.6% in Mobius QA(3%/2 mm), respectively. According to the calculated action limits, the ArcCheck measurements for all the initial and ART plans passed QA successfully while the Mobius calculation action limits flagged seven and four failure cases respectively for initial plans and ART plans, respectively. CONCLUSION: An ART plan can be substantially different from the initial plan, and therefore a separate session of ART plan QA is needed to ensure treatment safety and quality. The pre-treatment QA calculation via Mobius can serve as a reliable indicator of failure in online ART plan QA. However, given that Ethos ART system is still relatively new, ArcCheck measurement of initial plan is still in practice. It may be skipped as we gain more experience and have better understanding of the system.

Real-time liver tumor localization via combined surface imaging and a single x-ray projection.

Objective. Real-time imaging, a building block of real-time adaptive radiotherapy, provides instantaneous knowledge of anatomical motion to drive delivery adaptation to improve patient safety and treatment efficacy. The temporal constraint of real-time imaging (<500 milliseconds) significantly limits the imaging signals that can be acquired, rendering volumetric imaging and 3D tumor localization extremely challenging. Real-time liver imaging is particularly difficult, compounded by the low soft tissue contrast within the liver. We proposed a deep learning (DL)-based framework (Surf-X-Bio), to track 3D liver tumor motion in real-time from combined optical surface image and a single on-board x-ray projection.Approach. Surf-X-Bio performs mesh-based deformable registration to track/localize liver tumors volumetrically via three steps. First, a DL model was built to estimate liver boundary motion from an optical surface image, using learnt motion correlations between the respiratory-induced external body surface and liver boundary. Second, the residual liver boundary motion estimation error was further corrected by a graph neural network-based DL model, using information extracted from a single x-ray projection. Finally, a biomechanical modeling-driven DL model was applied to solve the intra-liver motion for tumor localization, using the liver boundary motion derived via prior steps.Main results. Surf-X-Bio demonstrated higher accuracy and better robustness in tumor localization, as compared to surface-image-only and x-ray-only models. By Surf-X-Bio, the mean (±s.d.) 95-percentile Hausdorff distance of the liver boundary from the 'ground-truth' decreased from 9.8 (±4.5) (before motion estimation) to 2.4 (±1.6) mm. The mean (±s.d.) center-of-mass localization error of the liver tumors decreased from 8.3 (±4.8) to 1.9 (±1.6) mm.Significance. Surf-X-Bio can accurately track liver tumors from combined surface imaging and x-ray imaging. The fast computational speed (<250 milliseconds per inference) allows it to be applied clinically for real-time motion management and adaptive radiotherapy.