Overview of cardiac toxicity from radiation therapy.

Radiotherapy is an essential part of treatment for many patients with thoracic cancers. However, proximity of the heart to tumour targets can lead to cardiac side effects, with studies demonstrating link between cardiac radiation dose and adverse outcomes. Although reducing cardiac dose can reduce associated risks, most cardiac constraint recommendations in clinical use are generally based on dose to the whole heart, as dose assessment at cardiac substructure levels on individual patients has been limited historically. Furthermore, estimation of an individual's cardiac risk is complex and multifactorial, which includes radiation dose alongside baseline risk factors, and the impact of systemic therapies. This review gives an overview of the epidemiological impact of cancer and cardiac disease, risk factors contributing to radiation-related cardiotoxicity, the evidence for cardiac side effects and future directions in cardiotoxicity research. A better understanding of the interactions between risk factors, balancing treatment benefit versus toxicity and the ongoing management of cardiac risk is essential for optimal clinical care. The emerging field of cardio-oncology is thus a multidisciplinary collaborative effort to enable better understanding of cardiac risks and outcomes for better-informed patient management decisions.

Lung cancer (internet-based) Delphi (LUCiD): A modified eDelphi consensus process to establish Australasian clinical quality indicators for thoracic cancer.

BACKGROUND AND OBJECTIVE: Approximately 16,000 new cases of lung cancer are diagnosed each year in Australia and Aotearoa New Zealand, and it is the leading cause of cancer death in the region. Unwarranted variation in lung cancer care and outcomes has been described for many years, although clinical quality indicators to facilitate benchmarking across Australasia have not been established. The purpose of this study was to establish clinical quality indicators applicable to lung and other thoracic cancers across Australia and Aotearoa New Zealand. METHODS: Following a literature review, a modified three round eDelphi consensus process was completed between October 2022 and June 2023. Participants included clinicians from all relevant disciplines, patient advocates, researchers and other stakeholders, with representatives from all Australian states and territories and Aotearoa New Zealand. Consensus was set at a threshold of 70%, with the first two rounds conducted as online surveys, and the final round held as a hybrid in person and virtual consensus meeting. RESULTS: The literature review identified 422 international thoracic oncology indicators, and a total of 71 indicators were evaluated over the course of the Delphi consensus. Ultimately, 27 clinical quality indicators reached consensus, covering the continuum of thoracic oncologic care from diagnosis to first line treatment. Indicators benchmarking supportive care were poorly represented. Attendant numeric quality standards were developed to facilitate benchmarking. CONCLUSION: Twenty-seven clinical quality indicators relevant to thoracic oncology care in Australasia were developed. Real world implementation will now be explored utilizing a prospective dataset collected across Australia.

Radiotherapy protocol compliance in routine clinical practice for patients with stages I-III non-small-cell lung cancer.

INTRODUCTION: Despite the availability of radiotherapy treatment protocols for lung cancer, considerable treatment variation occurs in clinical practice. This study assessed compliance with a radiotherapy protocol for the treatment of patients with stages I-III non-small-cell lung cancer (NSCLC) in routine clinical practice and to identify factors that were associated with compliance. METHODS: The Cancer Institute New South Wales eviQ treatment protocol for external beam radiotherapy of stages I-III NSCLC was taken as the reference to measure compliance. All inoperable patients with stages I-III NSCLC and documented ECOG performance status treated with radiotherapy between 2007 and 2019 at two radiotherapy facilities were available for analysis. Protocol compliance rates were calculated. Univariate and multivariate logistic regression models with 23 input factors were used to determine factors significantly associated with compliance. Survival analysis was conducted for both compliant and non-compliant treatments. RESULTS: Overall, 656 patients met the inclusion criteria. Protocol compliance was 16%. Alternative dose/fractionation was responsible for 49% of non-compliant treatments with 30% receiving an alternative curative fractionation. Five of 23 factors (age at the start of radiotherapy, stage group, ECOG performance status, tumour location and alcoholism history) showed significant associations with protocol compliance on multivariate analysis. There was no significant difference in median survival between patients receiving protocol compliant treatment (15.1 months) and non-compliant treatment (15.6 months). CONCLUSION: Adherence to the eviQ curative radiotherapy protocol for stages I-III NSCLC was low. Alternative dose/fractionation schemes were the main reason for non-compliance. Protocol compliance was not associated with outcome.

Uncertainty estimation using a 3D probabilistic U-Net for segmentation with small radiotherapy clinical trial datasets.

BACKGROUND AND OBJECTIVES: Bio-medical image segmentation models typically attempt to predict one segmentation that resembles a ground-truth structure as closely as possible. However, as medical images are not perfect representations of anatomy, obtaining this ground truth is not possible. A surrogate commonly used is to have multiple expert observers define the same structure for a dataset. When multiple observers define the same structure on the same image there can be significant differences depending on the structure, image quality/modality and the region being defined. It is often desirable to estimate this type of aleatoric uncertainty in a segmentation model to help understand the region in which the true structure is likely to be positioned. Furthermore, obtaining these datasets is resource intensive so training such models using limited data may be required. With a small dataset size, differing patient anatomy is likely not well represented causing epistemic uncertainty which should also be estimated so it can be determined for which cases the model is effective or not. METHODS: We use a 3D probabilistic U-Net to train a model from which several segmentations can be sampled to estimate the range of uncertainty seen between multiple observers. To ensure that regions where observers disagree most are emphasised in model training, we expand the Generalised Evidence Lower Bound (ELBO) with a Constrained Optimisation (GECO) loss function with an additional contour loss term to give attention to this region. Ensemble and Monte-Carlo dropout (MCDO) uncertainty quantification methods are used during inference to estimate model confidence on an unseen case. We apply our methodology to two radiotherapy clinical trial datasets, a gastric cancer trial (TOPGEAR, TROG 08.08) and a post-prostatectomy prostate cancer trial (RAVES, TROG 08.03). Each dataset contains only 10 cases each for model development to segment the clinical target volume (CTV) which was defined by multiple observers on each case. An additional 50 cases are available as a hold-out dataset for each trial which had only one observer define the CTV structure on each case. Up to 50 samples were generated using the probabilistic model for each case in the hold-out dataset. To assess performance, each manually defined structure was matched to the closest matching sampled segmentation based on commonly used metrics. RESULTS: The TOPGEAR CTV model achieved a Dice Similarity Coefficient (DSC) and Surface DSC (sDSC) of 0.7 and 0.43 respectively with the RAVES model achieving 0.75 and 0.71 respectively. Segmentation quality across cases in the hold-out datasets was variable however both the ensemble and MCDO uncertainty estimation approaches were able to accurately estimate model confidence with a p-value < 0.001 for both TOPGEAR and RAVES when comparing the DSC using the Pearson correlation coefficient. CONCLUSIONS: We demonstrated that training auto-segmentation models which can estimate aleatoric and epistemic uncertainty using limited datasets is possible. Having the model estimate prediction confidence is important to understand for which unseen cases a model is likely to be useful.

The histopathological spectrum and molecular changes associated with KRAS G12C mutation in non-small cell lung carcinoma.

KRAS G12C is the most common KRAS mutation in non-small cell lung carcinoma (NSCLC), for which targeted therapy has recently been developed. From the 732 cases of NSCLC that underwent next-generation sequencing at the Department of Anatomical Pathology, Liverpool Hospital, between July 2021 and May 2023, we retrieved 83 (11%) consecutive cases of KRAS G12C mutated NSCLC, and analysed their clinical, pathological, and molecular features. Of the 83 cases of KRAS G12C mutated NSCLC, there were 46 (55%) men and 37 (45%) women, with mean age of 72 years. Of the 49 cases with known clinical information, 94% were current or ex-smokers, and 49% were stage IV at diagnosis with median survival of 12 months. Sixty-three percent were histology cases and the remainder were cytology cases. Eighty-two percent were non-mucinous adenocarcinomas, with conventional histology including lepidic, acinar, solid, single cells and micropapillary patterns, and 62% were poorly differentiated. There were five (6%) cases of mucinous adenocarcinoma, one case of pleomorphic carcinoma and one case of high-grade fetal adenocarcinoma. TTF1 was positive in the majority (89%) of cases. Nineteen (23%) cases had TP53 co-mutation, and these cases had trends towards higher PD-L1 expression, poor differentiation, and presentation as stage IV disease, but the differences were not statistically significant. KRAS G12C mutated NSCLCs almost exclusively occurred in smokers and were mostly non-mucinous adenocarcinomas with conventional histological patterns which ranged from well to poorly differentiated. Around a quarter had TP53 co-mutation, the histological impacts and immune profile of which need to be assessed in a larger study.

Measurement that matters: A systematic review and modified Delphi of multidisciplinary colorectal cancer quality indicators.

AIM: To develop a priority set of quality indicators (QIs) for use by colorectal cancer (CRC) multidisciplinary teams (MDTs). METHODS: The review search strategy was executed in four databases from 2009-August 2019. Two reviewers screened abstracts/manuscripts. Candidate QIs and characteristics were extracted using a tailored abstraction tool and assessed for scientific soundness. To prioritize candidate indicators, a modified Delphi consensus process was conducted. Consensus was sought over two rounds; (1) multidisciplinary expert workshops to identify relevance to Australian CRC MDTs, and (2) an online survey to prioritize QIs by clinical importance. RESULTS: A total of 93 unique QIs were extracted from 118 studies and categorized into domains of care within the CRC patient pathway. Approximately half the QIs involved more than one discipline (52.7%). One-third of QIs related to surgery of primary CRC (31.2%). QIs on supportive care (6%) and neoadjuvant therapy (6%) were limited. In the Delphi Round 1, workshop participants (n = 12) assessed 93 QIs and produced consensus on retaining 49 QIs including six new QIs. In Round 2, survey participants (n = 44) rated QIs and prioritized a final 26 QIs across all domains of care and disciplines with a concordance level > 80%. Participants represented all MDT disciplines, predominantly surgical (32%), radiation (23%) and medical (20%) oncology, and nursing (18%), across six Australian states, with an even spread of experience level. CONCLUSION: This study identified a large number of existing CRC QIs and prioritized the most clinically relevant QIs for use by Australian MDTs to measure and monitor their performance.

Faster and lower dose imaging: evaluating adaptive, constant gantry velocity and angular separation in fast low-dose 4D cone beam CT imaging.

BACKGROUND: The adoption of four-dimensional cone beam computed tomography (4DCBCT) for image-guided lung cancer radiotherapy is increasing, especially for hypofractionated treatments. However, the drawbacks of 4DCBCT include long scan times (∼240 s), inconsistent image quality, higher imaging dose than necessary, and streaking artifacts. With the emergence of linear accelerators that can acquire 4DCBCT scans in a short period of time (9.2 s) there is a need to examine the impact that these very fast gantry rotations have on 4DCBCT image quality. PURPOSE: This study investigates the impact of gantry velocity and angular separation between x-ray projections on image quality and its implication for fast low-dose 4DCBCT with emerging systems, such as the Varian Halcyon that provide fast gantry rotation and imaging. Large and uneven angular separation between x-ray projections is known to reduce 4DCBCT image quality through increased streaking artifacts. However, it is not known when angular separation starts degrading image quality. The study assesses the impact of constant and adaptive gantry velocity and determines the level when angular gaps impair image quality using state-of-the-art reconstruction methods. METHODS: This study considers fast low-dose 4DCBCT acquisitions (60-80 s, 200-projection scans). To assess the impact of adaptive gantry rotations, the angular position of x-ray projections from adaptive 4DCBCT acquisitions from a 30-patient clinical trial were analyzed (referred to as patient angular gaps). To assess the impact of angular gaps, variable and static angular gaps (20°, 30°, 40°) were introduced into evenly separated 200 projections (ideal angular separation). To simulate fast gantry rotations, which are on emerging linacs, constant gantry velocity acquisitions (9.2 s, 60 s, 120 s, 240 s) were simulated by sampling x-ray projections at constant intervals using the patient breathing traces from the ADAPT clinical trial (ACTRN12618001440213). The 4D Extended Cardiac-Torso (XCAT) digital phantom was used to simulate projections to remove patient-specific image quality variables. Image reconstruction was performed using Feldkamp-Davis-Kress (FDK), McKinnon-Bates (MKB), and Motion-Compensated-MKB (MCMKB) algorithms. Image quality was assessed using Structural Similarity-Index-Measure (SSIM), Contrast-to-Noise-Ratio (CNR), Signal-to-Noise-Ratio (SNR), Tissue-Interface-Width-Diaphragm (TIW-D), and Tissue-Interface-Width-Tumor (TIW-T). RESULTS: Patient angular gaps and variable angular gap reconstructions produced similar results to ideal angular separation reconstructions, while static angular gap reconstructions produced lower image quality metrics. For MCMKB-reconstructions, average patient angular gaps produced SSIM-0.98, CNR-13.6, SNR-34.8, TIW-D-1.5 mm, and TIW-T-2.0 mm, static angular gap 40° produced SSIM-0.92, CNR-6.8, SNR-6.7, TIW-D-5.7 mm, and TIW-T-5.9 mm and ideal produced SSIM-1.00, CNR-13.6, SNR-34.8, TIW-D-1.5 mm, and TIW-T-2.0 mm. All constant gantry velocity reconstructions produced lower image quality metrics than ideal angular separation reconstructions regardless of the acquisition time. Motion compensated reconstruction (MCMKB) produced the highest contrast images with low streaking artifacts. CONCLUSION: Very fast 4DCBCT scans can be acquired provided that the entire scan range is adaptively sampled, and motion-compensated reconstruction is performed. Importantly, the angular separation between x-ray projections within each individual respiratory bin had minimal effect on the image quality of fast low-dose 4DCBCT imaging. The results will assist the development of future 4DCBCT acquisition protocols that can now be achieved in very short time frames with emerging linear accelerators.

Central airway squamous metaplasia following radiation therapy mimicking local tumour recurrence.

Radiation therapy can result in injury to the lung parenchyma and central airways; the latter is less well documented in the literature. Here, we describe a 65-year-old Caucasian male, who developed focal endobronchial nodules and right main bronchial stenosis suggesting tumour recurrence, 32 months following curative intent concurrent chemoradiation therapy for Stage 3B squamous cell carcinoma of the lung. Computed tomography and positron emission tomography results are detailed. Flexible bronchoscopy with bronchial biopsies revealed squamous metaplasia rather than malignant tumour recurrence, with ongoing observation planned.

Central airway squamous metaplasia following radiation therapy mimicking local tumour recurrence.

Radiation therapy can result in injury to the lung parenchyma and central airways; the latter is less well documented in the literature. Here, we describe a 65-year-old Caucasian male, who developed focal endobronchial nodules and right main bronchial stenosis suggesting tumour recurrence, 32 months following curative intent concurrent chemoradiation therapy for Stage 3B squamous cell carcinoma of the lung. Computed tomography and positron emission tomography results are detailed. Flexible bronchoscopy with bronchial biopsies revealed squamous metaplasia rather than malignant tumour recurrence, with ongoing observation planned.

SABR for Early Non-Small Cell Lung Cancer: Changes in Pulmonary Function, Dyspnea, and Quality of Life.

PURPOSE: The aim of this study was to report pulmonary function tests (PFTs) and clinician-reported and patient-reported quality-of-life (QoL) outcomes on a cohort of patients with non-small cell lung cancer (NSCLC) treated with SABR. METHODS AND MATERIALS: A total of 119 patients with NSCLC were treated with SABR in the prospective cohort SSBROC study of patients with T1-T2N0M0 NSCLC. PFTs and QoL measures were obtained at baseline pretreatment and at 6-month intervals. Here we report on the 6- to 18-month time points. Analysis of covariance (ANCOVA) methods adjusting for baseline analyzed potential predictors on outcomes of PFTs and patient-reported dyspnea at 18 months. RESULTS: The only statistically significant decline in PFTs was seen in forced expiratory volume in 1 second (FEV1) at 18 months post-SABR, with a decline of -0.11 L (P = .0087; 95% CI, -0.18 to -0.02). Of potential predictors of decline, only a 1-unit increase in smoking pack-years resulted in a -0.12 change in diffusing capacity for carbon monoxide (P = .026; 95% CI, -0.02 to -0.23) and a 0.003 decrease in FEV1 (P = .026; 95% CI, -0.006 to -0.0004). For patient-reported outcomes, statistically significant worsening in both the European Organisation for Research and Treatment of Cancer Quality of Life Core Questionnaire (QLQ-C30 Version 3) and the lung module (QLQ-LC13) dyspnea scores occurred at the 18-month time point, but not earlier. No potential predictors of worsening dyspnea were statistically significant. There was no statistically significant decline in clinician-reported outcomes or global QoL scores. CONCLUSIONS: We found a statistically significant decline in FEV1 at 18 months posttreatment. Smoking pack-years was a predictor for decline in diffusing capacity for carbon monoxide and FEV1 at 18 months. Worsening of patient-reported dyspnea scores was observed, consistent with the expected progression of lung comorbid disease.

Clinical target volume delineation quality assurance for MRI-guided prostate radiotherapy using deep learning with uncertainty estimation.

BACKGROUND AND PURPOSE: Previous studies on automatic delineation quality assurance (QA) have mostly focused on CT-based planning. As MRI-guided radiotherapy is increasingly utilized in prostate cancer treatment, there is a need for more research on MRI-specific automatic QA. This work proposes a clinical target volume (CTV) delineation QA framework based on deep learning (DL) for MRI-guided prostate radiotherapy. MATERIALS AND METHODS: The proposed workflow utilized a 3D dropblock ResUnet++ (DB-ResUnet++) to generate multiple segmentation predictions via Monte Carlo dropout which were used to compute an average delineation and area of uncertainty. A logistic regression (LR) classifier was employed to classify the manual delineation as pass or discrepancy based on the spatial association between the manual delineation and the network's outputs. This approach was evaluated on a multicentre MRI-only prostate radiotherapy dataset and compared with our previously published QA framework based on AN-AG Unet. RESULTS: The proposed framework achieved an area under the receiver operating curve (AUROC) of 0.92, a true positive rate (TPR) of 0.92 and a false positive rate of 0.09 with an average processing time per delineation of 1.3 min. Compared with our previous work using AN-AG Unet, this method generated fewer false positive detections at the same TPR with a much faster processing speed. CONCLUSION: To the best of our knowledge, this is the first study to propose an automatic delineation QA tool using DL with uncertainty estimation for MRI-guided prostate radiotherapy, which can potentially be used for reviewing prostate CTV delineation in multicentre clinical trials.

Qualitative study of international key informants' perspectives on the current and future state of healthcare quality measurement and feedback.

OBJECTIVES: The aim of this study is to explore the current and future state of quality measurement and feedback and identify factors influencing measurement feedback systems, including the barriers and enablers to their effective design, implementation, use and translation into quality improvement. DESIGN: This qualitative study used semistructured interviews with key informants. A deductive framework analysis was conducted to code transcripts to the Theoretical Domains Framework (TDF). An inductive analysis was used to produce subthemes and belief statements within each TDF domain. SETTING: All interviews were conducted by videoconference and audio-recorded. PARTICIPANTS: Key informants were purposively sampled experts in quality measurement and feedback, including clinical (n=5), government (n=5), research (n=4) and health service leaders (n=3) from Australia (n=7), the USA (n=4), the UK (n=2), Canada (n=2) and Sweden (n=2). RESULTS: A total of 17 key informants participated in the study. The interview length ranged from 48 to 66 min. 12 theoretical domains populated by 38 subthemes were identified as relevant to measurement feedback systems. The most populous domains included environmental context and resources, memory, attention and decision-making, and social influences. The most populous subthemes included 'quality improvement culture', 'financial and human resource support' and 'patient-centred measurement'. There were minimal conflicting beliefs outside of 'data quality and completeness'. Conflicting beliefs in these subthemes were predominantly between government and clinical leaders. CONCLUSIONS: Multiple factors were found to influence measurement feedback systems and future considerations are presented within this manuscript. The barriers and enablers that impact these systems are complex. While there are some clear modifiable factors in the design of measurement and feedback processes, influential factors described by key informants were largely socioenvironmental. Evidence-based design and implementation, coupled with a deeper understanding of the implementation context, may lead to enhanced quality measurement feedback systems and ultimately improved care delivery and patient outcomes.

Long-Term Outcomes of TROG 13.01 SAFRON II Randomized Trial of Single- Versus Multifraction Stereotactic Ablative Body Radiotherapy for Pulmonary Oligometastases.

Clinical trials frequently include multiple end points that mature at different times. The initial report, typically based on the primary end point, may be published when key planned co-primary or secondary analyses are not yet available. Clinical Trial Updates provide an opportunity to disseminate additional results from studies, published in JCO or elsewhere, for which the primary end point has already been reported.In a randomized phase II clinical trial, the Trans Tasman Radiation Oncology Group compared single- versus multifraction stereotactic ablative body radiotherapy (SABR) in 90 patients with 133 oligometastases to the lung. The study found no differences in safety, efficacy, systemic immunogenicity, or survival between arms, with single-fraction SABR picked as the winner on the basis of cost-effectiveness. In this article, we report the final updated survival outcome analysis. The protocol mandated no concurrent or post-therapy systemic therapy until progression. Modified disease-free survival (mDFS) was defined as any progression not addressable by local therapy, or death. At a median follow-up of 5.4 years, the 3- and 5-year estimates for overall survival (OS) were 70% (95% CI, 59 to 78) and 51% (95% CI, 39 to 61). There were no significant differences between the multi- and single-fraction arms for OS (hazard ratio [HR], 1.1 [95% CI, 0.6 to 2.0]; P = .81). The 3- and 5-year estimates for disease-free survival were 24% (95% CI, 16 to 33) and 20% (95% CI, 13 to 29), with no differences between arms (HR, 1.0 [95% CI, 0.6 to 1.6]; P = .92). The 3- and 5-year estimates for mDFS were 39% (95% CI, 29 to 49) and 34% (95% CI, 24 to 44), with no differences between arms (HR, 1.0 [95% CI, 0.6 to 1.8]; P = .90). In this patient population, where patients receive SABR in lieu of systemic therapy, one-in-three patients are alive without disease in the long term. There were no differences in outcomes by fractionation schedule.

Implementing Smoking Cessation Interventions for Tobacco Users Within Oncology Settings: A Systematic Review.

Importance: Patients with cancer who continue to smoke tobacco experience greater treatment-related complications, higher risk of secondary cancers, and greater mortality. Despite research to improve smoking cessation care within clinical oncology, implementation of proposed interventions within routine care remains challenging. Objective: To identify and recommend implementation strategies for smoking cessation interventions associated with improved screening, advice-giving, and referral for tobacco users recently diagnosed with cancer, as well as shifting smoking behaviors and attitudes in this patient population. Evidence Review: MEDLINE, CINAHL, Embase, and PsycINFO databases, as well as Google Scholar, were searched for articles published before September 7, 2020, using terms related to cancer, smoking cessation, and implementation science. Outcomes of interest were study characteristics, implementation strategies, and outcome measures (screening, advice, referral, abstinence rates, and attitudes). The Cochrane Risk of Bias Tool for randomized and nonrandomized studies was used to assess bias. The review was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline and Synthesis Without Meta-analysis (SWiM) guideline. Implementation strategies were categorized according to Expert Recommendations for Implementing Change (ERIC) study taxonomy. A systematic analysis was conducted focusing on studies with low or moderate risk of bias due to high heterogeneity in outcome measurement. Findings: In total, 6047 records were screened, yielding 43 articles (10 randomized clinical trials and 33 nonrandomized studies). Four strategies were associated with improved screening, advice-giving, and referral: (1) supporting clinicians, (2) training implementation stakeholders (including clinicians), (3) changing the infrastructure, and (4) developing stakeholder interrelationships. Conclusions and Relevance: In this systematic review, supporting clinicians by providing cessation care through a trained tobacco specialist was identified as important for achieving short-term abstinence and changing attitudes among patients with cancer. Combined with a theoretical framework and stakeholder involvement, these strategies provide the basis for successful implementation of cessation support; this systematic review serves as an illustration of the methodological application and synthesis of implementation studies and other medical conditions more generally.

Outcome prediction models incorporating clinical variables for Head and Neck Squamous cell Carcinoma: A systematic review of methodological conduct and risk of bias.

Multiple outcome prediction models have been developed for Head and Neck Squamous Cell Carcinoma (HNSCC). This systematic review aimed to identify HNSCC outcome prediction model studies, assess their methodological quality and identify those with potential utility for clinical practice. Inclusion criteria were mucosal HNSCC prognostic prediction model studies (development or validation) incorporating clinically available variables accessible at time of treatment decision making and predicting tumour-related outcomes. Eligible publications were identified from PubMed and Embase. Methodological quality and risk of bias were assessed using the checklist for critical appraisal and data extraction for systematic reviews of prediction modelling studies (CHARMS) and prediction model risk of bias assessment tool (PROBAST). Eligible publications were categorised by study type for reporting. 64 eligible publications were identified; 55 reported model development, 37 external validations, with 28 reporting both. CHARMS checklist items relating to participants, predictors, outcomes, handling of missing data, and some model development and evaluation procedures were generally well-reported. Less well-reported were measures accounting for model overfitting and model performance measures, especially model calibration. Full model information was poorly reported (3/55 model developments), specifically model intercept, baseline survival or full model code. Most publications (54/55 model developments, 28/37 external validations) were found to have high risk of bias, predominantly due to methodological issues in the PROBAST analysis domain. The identified methodological issues may affect prediction model accuracy in heterogeneous populations. Independent external validation studies in the local population and demonstration of clinical impact are essential for the clinical implementation of outcome prediction models.

Open-source, fully-automated hybrid cardiac substructure segmentation: development and optimisation.

Radiotherapy for thoracic and breast tumours is associated with a range of cardiotoxicities. Emerging evidence suggests cardiac substructure doses may be more predictive of specific outcomes, however, quantitative data necessary to develop clinical planning constraints is lacking. Retrospective analysis of patient data is required, which relies on accurate segmentation of cardiac substructures. In this study, a novel model was designed to deliver reliable, accurate, and anatomically consistent segmentation of 18 cardiac substructures on computed tomography (CT) scans. Thirty manually contoured CT scans were included. The proposed multi-stage method leverages deep learning (DL), multi-atlas mapping, and geometric modelling to automatically segment the whole heart, cardiac chambers, great vessels, heart valves, coronary arteries, and conduction nodes. Segmentation performance was evaluated using the Dice similarity coefficient (DSC), mean distance to agreement (MDA), Hausdorff distance (HD), and volume ratio. Performance was reliable, with no errors observed and acceptable variation in accuracy between cases, including in challenging cases with imaging artefacts and atypical patient anatomy. The median DSC range was 0.81-0.93 for whole heart and cardiac chambers, 0.43-0.76 for great vessels and conduction nodes, and 0.22-0.53 for heart valves. For all structures the median MDA was below 6 mm, median HD ranged 7.7-19.7 mm, and median volume ratio was close to one (0.95-1.49) for all structures except the left main coronary artery (2.07). The fully automatic algorithm takes between 9 and 23 min per case. The proposed fully-automatic method accurately delineates cardiac substructures on radiotherapy planning CT scans. Robust and anatomically consistent segmentations, particularly for smaller structures, represents a major advantage of the proposed segmentation approach. The open-source software will facilitate more precise evaluation of cardiac doses and risks from available clinical datasets.

Standardising Breast Radiotherapy Structure Naming Conventions: A Machine Learning Approach.

In progressing the use of big data in health systems, standardised nomenclature is required to enable data pooling and analyses. In many radiotherapy planning systems and their data archives, target volumes (TV) and organ-at-risk (OAR) structure nomenclature has not been standardised. Machine learning (ML) has been utilised to standardise volumes nomenclature in retrospective datasets. However, only subsets of the structures have been targeted. Within this paper, we proposed a new approach for standardising all the structures nomenclature by using multi-modal artificial neural networks. A cohort consisting of 1613 breast cancer patients treated with radiotherapy was identified from Liverpool & Macarthur Cancer Therapy Centres, NSW, Australia. Four types of volume characteristics were generated to represent each target and OAR volume: textual features, geometric features, dosimetry features, and imaging data. Five datasets were created from the original cohort, the first four represented different subsets of volumes and the last one represented the whole list of volumes. For each dataset, 15 sets of combinations of features were generated to investigate the effect of using different characteristics on the standardisation performance. The best model reported 99.416% classification accuracy over the hold-out sample when used to standardise all the nomenclatures in a breast cancer radiotherapy plan into 21 classes. Our results showed that ML based automation methods can be used for standardising naming conventions in a radiotherapy plan taking into consideration the inclusion of multiple modalities to better represent each volume.