The Image Biomarker Standardization Initiative: Standardized Convolutional Filters for Reproducible Radiomics and Enhanced Clinical Insights.

Filters are commonly used to enhance specific structures and patterns in images, such as vessels or peritumoral regions, to enable clinical insights beyond the visible image using radiomics. However, their lack of standardization restricts reproducibility and clinical translation of radiomics decision support tools. In this special report, teams of researchers who developed radiomics software participated in a three-phase study (September 2020 to December 2022) to establish a standardized set of filters. The first two phases focused on finding reference filtered images and reference feature values for commonly used convolutional filters: mean, Laplacian of Gaussian, Laws and Gabor kernels, separable and nonseparable wavelets (including decomposed forms), and Riesz transformations. In the first phase, 15 teams used digital phantoms to establish 33 reference filtered images of 36 filter configurations. In phase 2, 11 teams used a chest CT image to derive reference values for 323 of 396 features computed from filtered images using 22 filter and image processing configurations. Reference filtered images and feature values for Riesz transformations were not established. Reproducibility of standardized convolutional filters was validated on a public data set of multimodal imaging (CT, fluorodeoxyglucose PET, and T1-weighted MRI) in 51 patients with soft-tissue sarcoma. At validation, reproducibility of 486 features computed from filtered images using nine configurations × three imaging modalities was assessed using the lower bounds of 95% CIs of intraclass correlation coefficients. Out of 486 features, 458 were found to be reproducible across nine teams with lower bounds of 95% CIs of intraclass correlation coefficients greater than 0.75. In conclusion, eight filter types were standardized with reference filtered images and reference feature values for verifying and calibrating radiomics software packages. A web-based tool is available for compliance checking.

Multicentre validation of CT grey-level co-occurrence matrix features for overall survival in primary oesophageal adenocarcinoma.

BACKGROUND: Personalising management of primary oesophageal adenocarcinoma requires better risk stratification. Lack of independent validation of proposed imaging biomarkers has hampered clinical translation. We aimed to prospectively validate previously identified prognostic grey-level co-occurrence matrix (GLCM) CT features for 3-year overall survival. METHODS: Following ethical approval, clinical and contrast-enhanced CT data were acquired from participants from five institutions. Data from three institutions were used for training and two for testing. Survival classifiers were modelled on prespecified variables ('Clinical' model: age, clinical T-stage, clinical N-stage; 'ClinVol' model: clinical features + CT tumour volume; 'ClinRad' model: ClinVol features + GLCM_Correlation and GLCM_Contrast). To reflect current clinical practice, baseline stage was also modelled as a univariate predictor ('Stage'). Discrimination was assessed by area under the receiver operating curve (AUC) analysis; calibration by Brier scores; and clinical relevance by thresholding risk scores to achieve 90% sensitivity for 3-year mortality. RESULTS: A total of 162 participants were included (144 male; median 67 years [IQR 59, 72]; training, 95 participants; testing, 67 participants). Median survival was 998 days [IQR 486, 1594]. The ClinRad model yielded the greatest test discrimination (AUC, 0.68 [95% CI 0.54, 0.81]) that outperformed Stage (ΔAUC, 0.12 [95% CI 0.01, 0.23]; p = .04). The Clinical and ClinVol models yielded comparable test discrimination (AUC, 0.66 [95% CI 0.51, 0.80] vs. 0.65 [95% CI 0.50, 0.79]; p > .05). Test sensitivity of 90% was achieved by ClinRad and Stage models only. CONCLUSIONS: Compared to Stage, multivariable models of prespecified clinical and radiomic variables yielded improved prediction of 3-year overall survival. CLINICAL RELEVANCE STATEMENT: Previously identified radiomic features are prognostic but may not substantially improve risk stratification on their own. KEY POINTS: • Better risk stratification is needed in primary oesophageal cancer to personalise management. • Previously identified CT features-GLCM_Correlation and GLCM_Contrast-contain incremental prognostic information to age and clinical stage. • Compared to staging, multivariable clinicoradiomic models improve discrimination of 3-year overall survival.

Predicting programmed death-ligand 1 (PD-L1) expression with fluorine-18 fluorodeoxyglucose ([18F]FDG) positron emission tomography/computed tomography (PET/CT) metabolic parameters in resectable non-small cell lung cancer.

BACKGROUND: Programmed death-ligand 1 (PD-L1) expression is a predictive biomarker for immunotherapy in non-small cell lung cancer (NSCLC). PD-L1 and glucose transporter 1 expression are closely associated, and studies demonstrate correlation of PD-L1 with glucose metabolism. AIM: The aim of this study was to investigate the association of fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography ([18F]FDG-PET/CT) metabolic parameters with PD-L1 expression in primary lung tumour and lymph node metastases in resected NSCLC. METHODS: We conducted a retrospective analysis of 210 patients with node-positive resectable stage IIB-IIIB NSCLC. PD-L1 tumour proportion score (TPS) was determined using the DAKO 22C3 immunohistochemical assay. Semi-automated techniques were used to analyse pre-operative [18F]FDG-PET/CT images to determine primary and nodal metabolic parameter scores (including max, mean, peak and peak adjusted for lean body mass standardised uptake values (SUV), metabolic tumour volume (MTV), total lesional glycolysis (TLG) and SUV heterogeneity index (HISUV)). RESULTS: Patients were predominantly male (57%), median age 70 years with non-squamous NSCLC (68%). A majority had negative primary tumour PD-L1 (TPS < 1%; 53%). Mean SUVmax, SUVmean, SUVpeak and SULpeak values were significantly higher (p < 0.05) in those with TPS ≥ 1% in primary tumour (n = 210) or lymph nodes (n = 91). However, ROC analysis demonstrated only moderate separability at the 1% PD-L1 TPS threshold (AUCs 0.58-0.73). There was no association of MTV, TLG and HISUV with PD-L1 TPS. CONCLUSION: This study demonstrated the association of SUV-based [18F]FDG-PET/CT metabolic parameters with PD-L1 expression in primary tumour or lymph node metastasis in resectable NSCLC, but with poor sensitivity and specificity for predicting PD-L1 positivity ≥ 1%. CLINICAL RELEVANCE STATEMENT: Whilst SUV-based fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography metabolic parameters may not predict programmed death-ligand 1 positivity ≥ 1% in the primary tumour and lymph nodes of resectable non-small cell lung cancer independently, there is a clear association which warrants further investigation in prospective studies. TRIAL REGISTRATION: Non-applicable KEY POINTS: • Programmed death-ligand 1 immunohistochemistry has a predictive role in non-small cell lung cancer immunotherapy; however, it is both heterogenous and dynamic. • SUV-based fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography ([18F]FDG-PET/CT) metabolic parameters were significantly higher in primary tumour or lymph node metastases with positive programmed death-ligand 1 expression. • These SUV-based parameters could potentially play an additive role along with other multi-modal biomarkers in selecting patients within a predictive nomogram.

Perceptions of radiologists on structured reporting for cancer imaging-a survey by the European Society of Oncologic Imaging (ESOI).

OBJECTIVES: To assess radiologists' current use of, and opinions on, structured reporting (SR) in oncologic imaging, and to provide recommendations for a structured report template. MATERIALS AND METHODS: An online survey with 28 questions was sent to European Society of Oncologic Imaging (ESOI) members. The questionnaire had four main parts: (1) participant information, e.g., country, workplace, experience, and current SR use; (2) SR design, e.g., numbers of sections and fields, and template use; (3) clinical impact of SR, e.g., on report quality and length, workload, and communication with clinicians; and (4) preferences for an oncology-focused structured CT report. Data analysis comprised descriptive statistics, chi-square tests, and Spearman correlation coefficients. RESULTS: A total of 200 radiologists from 51 countries completed the survey: 57.0% currently utilized SR (57%), with a lower proportion within than outside of Europe (51.0 vs. 72.7%; p = 0.006). Among SR users, the majority observed markedly increased report quality (62.3%) and easier comparison to previous exams (53.5%), a slightly lower error rate (50.9%), and fewer calls/emails by clinicians (78.9%) due to SR. The perceived impact of SR on communication with clinicians (i.e., frequency of calls/emails) differed with radiologists' experience (p < 0.001), and experience also showed low but significant correlations with communication with clinicians (r = - 0.27, p = 0.003), report quality (r = 0.19, p = 0.043), and error rate (r = - 0.22, p = 0.016). Template use also affected the perceived impact of SR on report quality (p = 0.036). CONCLUSION: Radiologists regard SR in oncologic imaging favorably, with perceived positive effects on report quality, error rate, comparison of serial exams, and communication with clinicians. CLINICAL RELEVANCE STATEMENT: Radiologists believe that structured reporting in oncologic imaging improves report quality, decreases the error rate, and enables better communication with clinicians. Implementation of structured reporting in Europe is currently below the international level and needs society endorsement. KEY POINTS: • The majority of oncologic imaging specialists (57% overall; 51% in Europe) use structured reporting in clinical practice. • The vast majority of oncologic imaging specialists use templates (92.1%), which are typically cancer-specific (76.2%). • Structured reporting is perceived to markedly improve report quality, communication with clinicians, and comparison to prior scans.

Multivariable prognostic modelling to improve prediction of colorectal cancer recurrence: the PROSPeCT trial.

OBJECTIVE: Improving prognostication to direct personalised therapy remains an unmet need. This study prospectively investigated promising CT, genetic, and immunohistochemical markers to improve the prediction of colorectal cancer recurrence. MATERIAL AND METHODS: This multicentre trial (ISRCTN 95037515) recruited patients with primary colorectal cancer undergoing CT staging from 13 hospitals. Follow-up identified cancer recurrence and death. A baseline model for cancer recurrence at 3 years was developed from pre-specified clinicopathological variables (age, sex, tumour-node stage, tumour size, location, extramural venous invasion, and treatment). Then, CT perfusion (blood flow, blood volume, transit time and permeability), genetic (RAS, RAF, and DNA mismatch repair), and immunohistochemical markers of angiogenesis and hypoxia (CD105, vascular endothelial growth factor, glucose transporter protein, and hypoxia-inducible factor) were added to assess whether prediction improved over tumour-node staging alone as the main outcome measure. RESULTS: Three hundred twenty-six of 448 participants formed the final cohort (226 male; mean 66 ± 10 years. 227 (70%) had ≥ T3 stage cancers; 151 (46%) were node-positive; 81 (25%) developed subsequent recurrence. The sensitivity and specificity of staging alone for recurrence were 0.56 [95% CI: 0.44, 0.67] and 0.58 [0.51, 0.64], respectively. The baseline clinicopathologic model improved specificity (0.74 [0.68, 0.79], with equivalent sensitivity of 0.57 [0.45, 0.68] for high vs medium/low-risk participants. The addition of prespecified CT perfusion, genetic, and immunohistochemical markers did not improve prediction over and above the clinicopathologic model (sensitivity, 0.58-0.68; specificity, 0.75-0.76). CONCLUSION: A multivariable clinicopathological model outperformed staging in identifying patients at high risk of recurrence. Promising CT, genetic, and immunohistochemical markers investigated did not further improve prognostication in rigorous prospective evaluation. CLINICAL RELEVANCE STATEMENT: A prognostic model based on clinicopathological variables including age, sex, tumour-node stage, size, location, and extramural venous invasion better identifies colorectal cancer patients at high risk of recurrence for neoadjuvant/adjuvant therapy than stage alone. KEY POINTS: Identification of colorectal cancer patients at high risk of recurrence is an unmet need for treatment personalisation. This model for recurrence, incorporating many patient variables, had higher specificity than staging alone. Continued optimisation of risk stratification schema will help individualise treatment plans and follow-up schedules.

Temporary treatment cessation versus continuation of first-line tyrosine kinase inhibitor in patients with advanced clear cell renal cell carcinoma (STAR): an open-label, non-inferiority, randomised, controlled, phase 2/3 trial.

BACKGROUND: Temporary drug treatment cessation might alleviate toxicity without substantially compromising efficacy in patients with cancer. We aimed to determine if a tyrosine kinase inhibitor drug-free interval strategy was non-inferior to a conventional continuation strategy for first-line treatment of advanced clear cell renal cell carcinoma. METHODS: This open-label, non-inferiority, randomised, controlled, phase 2/3 trial was done at 60 hospital sites in the UK. Eligible patients (aged ≥18 years) had histologically confirmed clear cell renal cell carcinoma, inoperable loco-regional or metastatic disease, no previous systemic therapy for advanced disease, uni-dimensionally assessed Response Evaluation Criteria in Solid Tumours-defined measurable disease, and an Eastern Cooperative Oncology Group performance status of 0-1. Patients were randomly assigned (1:1) at baseline to a conventional continuation strategy or drug-free interval strategy using a central computer-generated minimisation programme incorporating a random element. Stratification factors were Memorial Sloan Kettering Cancer Center prognostic group risk factor, sex, trial site, age, disease status, tyrosine kinase inhibitor, and previous nephrectomy. All patients received standard dosing schedules of oral sunitinib (50 mg per day) or oral pazopanib (800 mg per day) for 24 weeks before moving into their randomly allocated group. Patients allocated to the drug-free interval strategy group then had a treatment break until disease progression, when treatment was re-instated. Patients in the conventional continuation strategy group continued treatment. Patients, treating clinicians, and the study team were aware of treatment allocation. The co-primary endpoints were overall survival and quality-adjusted life-years (QALYs); non-inferiority was shown if the lower limit of the two-sided 95% CI for the overall survival hazard ratio (HR) was 0·812 or higher and if the lower limit of the two-sided 95% CI of the marginal difference in mean QALYs was -0·156 or higher. The co-primary endpoints were assessed in the intention-to-treat (ITT) population, which included all randomly assigned patients, and the per-protocol population, which excluded patients in the ITT population with major protocol violations and who did not begin their randomisation allocation as per the protocol. Non-inferiority was to be concluded if it was met for both endpoints in both analysis populations. Safety was assessed in all participants who received a tyrosine kinase inhibitor. The trial was registered with ISRCTN, 06473203, and EudraCT, 2011-001098-16. FINDINGS: Between Jan 13, 2012, and Sept 12, 2017, 2197 patients were screened for eligibility, of whom 920 were randomly assigned to the conventional continuation strategy (n=461) or the drug-free interval strategy (n=459; 668 [73%] male and 251 [27%] female; 885 [96%] White and 23 [3%] non-White). The median follow-up time was 58 months (IQR 46-73 months) in the ITT population and 58 months (46-72) in the per-protocol population. 488 patients continued on the trial after week 24. For overall survival, non-inferiority was demonstrated in the ITT population only (adjusted HR 0·97 [95% CI 0·83 to 1·12] in the ITT population; 0·94 [0·80 to 1·09] in the per-protocol population). Non-inferiority was demonstrated for QALYs in the ITT population (n=919) and per-protocol (n=871) population (marginal effect difference 0·06 [95% CI -0·11 to 0·23] for the ITT population; 0·04 [-0·14 to 0·21] for the per-protocol population). The most common grade 3 or worse adverse events were hypertension (124 [26%] of 485 patients in the conventional continuation strategy group vs 127 [29%] of 431 patients in the drug-free interval strategy group); hepatotoxicity (55 [11%] vs 48 [11%]); and fatigue (39 [8%] vs 63 [15%]). 192 (21%) of 920 participants had a serious adverse reaction. 12 treatment-related deaths were reported (three patients in the conventional continuation strategy group; nine patients in the drug-free interval strategy group) due to vascular (n=3), cardiac (n=3), hepatobiliary (n=3), gastrointestinal (n=1), or nervous system (n=1) disorders, and from infections and infestations (n=1). INTERPRETATION: Overall, non-inferiority between groups could not be concluded. However, there seemed to be no clinically meaningful reduction in life expectancy between the drug-free interval strategy and conventional continuation strategy groups and treatment breaks might be a feasible and cost-effective option with lifestyle benefits for patients during tyrosine kinase inhibitor therapy in patients with renal cell carcinoma. FUNDING: UK National Institute for Health and Care Research.

Diffusion-weighted imaging complements T2-weighted MRI for tumour response assessment in squamous anal carcinoma.

OBJECTIVES: A published tumour regression grade (TRG) score for squamous anal carcinoma treated with definitive chemoradiotherapy based on T2-weighted MRI yields a high proportion of indeterminate responses (TRG-3). We investigate whether the addition of diffusion-weighted imaging (DWI) improves tumour response assessment in the early post treatment period. MATERIALS AND METHODS: This retrospective observational study included squamous anal carcinoma patients undergoing MRI before and within 3 months of completing chemoradiotherapy from 2009 to 2020. Four independent radiologists (1-20 years' experience) scored MRI studies using a 5-point TRG system (1 = complete response; 5 = no response) based on T2-weighted sequences alone, and then after a 12-week washout period, using a 5-point DWI-TRG system based on T2-weighted and DWI. Scoring confidence was recorded on a 5-point scale (1 = low; 5 = high) for each reading and compared using the Wilcoxon test. Indeterminate scores (TRG-3) from each reading session were compared using the McNemar test. Interobserver agreement was assessed using kappa statistics. RESULTS: Eighty-five patients were included (mean age, 59 years ± 12 [SD]; 55 women). T2-weighted TRG-3 scores from all readers combined halved from 24% (82/340) to 12% (41/340) with DWI (p < 0.001). TRG-3 scores changed most frequently (41%, 34/82) to DWI-TRG-2 (excellent response). Complete tumour response was recorded clinically in 77/85 patients (91%). Scoring confidence increased using DWI (p < 0.001), with scores of 4 or 5 in 84% (287/340). Interobserver agreement remained fair to moderate (kappa range, 0.28-0.58). CONCLUSION: DWI complements T2-weighted MRI by reducing the number of indeterminate tumour responses (TRG-3). DWI increases radiologist's scoring confidence. CLINICAL RELEVANCE STATEMENT: Diffusion-weighted imaging improves T2-weighted tumour response assessment in squamous anal cancer, halving the number of indeterminate responses in the early post treatment period, and increases radiologists' confidence. KEY POINTS: Tumour response based on T2-weighted MRI is often indeterminate in squamous anal carcinoma. Diffusion-weighted imaging alongside T2-weighted MRI halved indeterminate tumour regression grade scores assigned by four radiologists from 24 to 12%. Scoring confidence of expert and non-expert radiologists increased with the inclusion of diffusion-weighted imaging.

18F FDG PET/CT and Novel Molecular Imaging for Directing Immunotherapy in Cancer.

Immunotherapy has transformed the treatment landscape of many cancers, with durable responses in disease previously associated with a poor prognosis. Patient selection remains a challenge, with predictive biomarkers an urgent unmet clinical need. Current predictive biomarkers, including programmed death-ligand 1 (PD-L1) (measured with immunohistochemistry), are imperfect. Promising biomarkers, including tumor mutation burden and tumor infiltrating lymphocyte density, fail to consistently predict response and have yet to translate to routine clinical practice. Heterogeneity of immune response within and between lesions presents a further challenge where fluorine 18 fluorodeoxyglucose PET/CT has a potential role in assessing response, stratifying treatment, and detecting and monitoring immune-related toxicities. Novel radiopharmaceuticals also present a unique opportunity to define the immune tumor microenvironment to better predict which patients may respond to therapy, for example by means of in vivo whole-body PD-L1 and CD8+ T cell expression imaging. In addition, longitudinal molecular imaging may help further define dynamic changes, particularly in cases of immunotherapy resistance, helping to direct a more personalized therapeutic approach. This review highlights current and emerging applications of molecular imaging to stratify, predict, and monitor molecular dynamics and treatment response in areas of clinical need.

Exercise prehabilitation during neoadjuvant chemotherapy may enhance tumour regression in oesophageal cancer: results from a prospective non-randomised trial.

BACKGROUND: There is increasing evidence for the use of exercise in cancer patients and data supporting enhanced tumour volume reduction following chemotherapy in animal models. To date, there is no reported histopathological evidence of a similar oncological benefit in oesophageal cancer. METHODS: A prospective non-randomised trial compared a structured prehabilitation exercise intervention during neoadjuvant chemotherapy and surgery versus conventional best-practice for oesophageal cancer patients. Biochemical and body composition analyses were performed at multiple time points. Outcome measures included radiological and pathological markers of disease regression. Logistic regression calculated ORs with 95% CI for the likelihood of pathological response adjusting for chemotherapy regimen and chemotherapy delivery. RESULTS: Comparison of the Intervention (n=21) and Control (n=19) groups indicated the Intervention group had higher rates of tumour regression (Mandard TRG 1-3 Intervention n=15/20 (75%) vs Control n=7/19 (36.8%) p=0.025) including adjusted analyses (OR 6.57; 95% CI 1.52 to 28.30). Combined tumour and node downstaging (Intervention n=9 (42.9%) vs Control n=3 (15.8%) p=0.089) and Fat Free Mass index were also improved (Intervention 17.8 vs 18.7 kg/m2; Control 16.3 vs 14.7 kg/m2, p=0.026). Differences in markers of immunity (CD-3 and CD-8) and inflammation (IL-6, VEGF, INF-y, TNFa, MCP-1 and EGF) were observed. CONCLUSION: The results suggest improved tumour regression and downstaging in the exercise intervention group and should prompt larger studies on this topic. TRIAL REGISTRATION NUMBER: NCT03626610.

Distortion-free 3D diffusion imaging of the prostate using a multishot diffusion-prepared phase-cycled acquisition and dictionary matching.

PURPOSE: To achieve three-dimensional (3D) distortion-free apparent diffusion coefficient (ADC) maps for prostate imaging using a multishot diffusion prepared-gradient echo (msDP-GRE) sequence and ADC dictionary matching. METHODS: The msDP-GRE sequence is combined with a 3D Cartesian, centric k-space trajectory with center oversampling. Oversampled k-space center averaging and phase cycling are used to address motion- and eddy current-induced magnitude corruption. Extended-phase-graph (EPG) simulations and ADC dictionary matching are used to compensate for T1 effects. To shorten the acquisition time, each volume is undersampled by a factor of two and reconstructed using iterative sensitivity encoding. The proposed approach is characterized using simulations and validated in a kiwifruit phantom, comparing the msDP-GRE ADC maps obtained using both standard monoexponential fitting and dictionary matching with the clinical standard single-shot diffusion weighted-echo planar imaging (ssDW-EPI) ADC. Initial in vivo feasibility is tested in three healthy subjects, and geometric distortion is compared with anatomical T2 -weighted-turbo spin echo. RESULTS: In the kiwifruit phantom experiment, the signal magnitude could be recovered using k-space center averaging and phase cycling. No statistically significant difference was observed in the ADC values estimated using msDP-GRE with dictionary matching and clinical standard DW-EPI (P < .05). The in vivo prostate msDP-GRE scans were free of geometric distortion caused by off-resonance susceptibility, and the ADC values in the prostate were in agreement with values found in the published literature. CONCLUSION: Nondistorted 3D ADC maps of the prostate can be achieved using a msDP sequence and dictionary matching.

Comparison of the diagnostic performance and impact on management of 18F-FDG PET/CT and whole-body MRI in multiple myeloma.

PURPOSE: Comparative data on the impact of imaging on management is lacking for multiple myeloma. This study compared the diagnostic performance and impact on management of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) and whole-body magnetic resonance imaging (WBMRI) in treatment-naive myeloma. METHODS: Forty-six patients undergoing 18F-FDG PET/CT and WBMRI were reviewed by a nuclear medicine physician and radiologist, respectively, for the presence of myeloma bone disease. Blinded clinical and imaging data were reviewed by two haematologists in consensus and management recorded following clinical data ± 18F-FDG PET/CT or WBMRI. Bone disease was defined using International Myeloma Working Group (IMWG) criteria and a clinical reference standard. Per-patient sensitivity for lesion detection was established. McNemar test compared management based on clinical assessment ± 18F-FDG PET/CT or WBMRI. RESULTS: Sensitivity for bone lesions was 69.6% (32/46) for 18F-FDG PET/CT (54.3% (25/46) for PET component alone) and 91.3% (42/46) for WBMRI. 27/46 (58.7%) of cases were concordant. In 19/46 patients (41.3%) WBMRI detected more focal bone lesions than 18F-FDG PET/CT. Based on clinical data alone, 32/46 (69.6%) patients would have been treated. Addition of 18F-FDG PET/CT to clinical data increased this to 40/46 (87.0%) patients (p = 0.02); and WBMRI to clinical data to 43/46 (93.5%) patients (p = 0.002). The difference in treatment decisions was not statistically significant between 18F-FDG PET/CT and WBMRI (p = 0.08). CONCLUSION: Compared to 18F-FDG PET/CT, WBMRI had a higher per patient sensitivity for bone disease. However, treatment decisions were not statistically different and either modality would be appropriate in initial staging, depending on local availability and expertise.

Systematic review of research design and reporting of imaging studies applying convolutional neural networks for radiological cancer diagnosis.

OBJECTIVES: To perform a systematic review of design and reporting of imaging studies applying convolutional neural network models for radiological cancer diagnosis. METHODS: A comprehensive search of PUBMED, EMBASE, MEDLINE and SCOPUS was performed for published studies applying convolutional neural network models to radiological cancer diagnosis from January 1, 2016, to August 1, 2020. Two independent reviewers measured compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM). Compliance was defined as the proportion of applicable CLAIM items satisfied. RESULTS: One hundred eighty-six of 655 screened studies were included. Many studies did not meet the criteria for current design and reporting guidelines. Twenty-seven percent of studies documented eligibility criteria for their data (50/186, 95% CI 21-34%), 31% reported demographics for their study population (58/186, 95% CI 25-39%) and 49% of studies assessed model performance on test data partitions (91/186, 95% CI 42-57%). Median CLAIM compliance was 0.40 (IQR 0.33-0.49). Compliance correlated positively with publication year (ρ = 0.15, p = .04) and journal H-index (ρ = 0.27, p < .001). Clinical journals demonstrated higher mean compliance than technical journals (0.44 vs. 0.37, p < .001). CONCLUSIONS: Our findings highlight opportunities for improved design and reporting of convolutional neural network research for radiological cancer diagnosis. KEY POINTS: • Imaging studies applying convolutional neural networks (CNNs) for cancer diagnosis frequently omit key clinical information including eligibility criteria and population demographics. • Fewer than half of imaging studies assessed model performance on explicitly unobserved test data partitions. • Design and reporting standards have improved in CNN research for radiological cancer diagnosis, though many opportunities remain for further progress.

Integrated slice-specific dynamic shimming for whole-body diffusion-weighted MR imaging at 1.5 T.

OBJECTIVE: To compare integrated slice-specific dynamic shim (iShim) with distortion correction post-processing to conventional 3D volume shim for the reduction of artefacts and signal loss in 1.5 T whole-body diffusion-weighted imaging (WB-DWI). METHODS: Ten volunteers underwent WB-DWI using conventional 3D volume shim and iShim. Forty-eight consecutive patients underwent WB-DWI with either volume shim (n = 24) or iShim (n = 24) only. For all subjects, displacement of the spinal cord at imaging station interfaces was measured on composed b = 900 s/mm2 images. The signal intensity ratios, computed as the average signal intensity in a region of high susceptibility gradient (sternum) divided by the average signal intensity in a region of low susceptibility gradient (vertebral body), were compared in volunteers. For patients, image quality was graded from 1 to 5 (1 = Poor, 5 = Excellent). Signal intensity discontinuity scores were recorded from 1 to 4 (1 = 2 + steps, 4 = 0 steps). A p value of < 0.05 was considered significant. RESULTS: Spinal cord displacement artefacts were lower with iShim (p < 0.05) at the thoracic junction in volunteers and at the cervical and thoracic junctions in patients (p < 0.05). The sternum/vertebra signal intensity ratio in healthy volunteers was higher with iShim compared with the volume shim sequence (p < 0.05). There were no significant differences between the volume shim and iShim patient groups in terms of image quality and signal intensity discontinuity scores. CONCLUSION: iShim reduced the degree of spinal cord displacement artefact between imaging stations and susceptibility-gradient-induced signal loss.

The Impact of the COVID-19 Pandemic on the Radiology Research Enterprise: Radiology Scientific Expert Panel.

The current coronavirus disease 2019 (COVID-19) crisis continues to grow and has resulted in marked changes to clinical operations. In parallel with clinical preparedness, universities have shut down most scientific research activities. Radiology researchers are currently grappling with these challenges that will continue to affect current and future imaging research. The purpose of this article is to describe the collective experiences of a diverse international group of academic radiology research programs in managing their response to the COVID-19 pandemic. The acute response at six distinct institutions will be described first, exploring common themes, challenges, priorities, and practices. This will be followed by reflections about the future of radiology research in the wake of the COVID-19 pandemic.

The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping.

Background Radiomic features may quantify characteristics present in medical imaging. However, the lack of standardized definitions and validated reference values have hampered clinical use. Purpose To standardize a set of 174 radiomic features. Materials and Methods Radiomic features were assessed in three phases. In phase I, 487 features were derived from the basic set of 174 features. Twenty-five research teams with unique radiomics software implementations computed feature values directly from a digital phantom, without any additional image processing. In phase II, 15 teams computed values for 1347 derived features using a CT image of a patient with lung cancer and predefined image processing configurations. In both phases, consensus among the teams on the validity of tentative reference values was measured through the frequency of the modal value and classified as follows: less than three matches, weak; three to five matches, moderate; six to nine matches, strong; 10 or more matches, very strong. In the final phase (phase III), a public data set of multimodality images (CT, fluorine 18 fluorodeoxyglucose PET, and T1-weighted MRI) from 51 patients with soft-tissue sarcoma was used to prospectively assess reproducibility of standardized features. Results Consensus on reference values was initially weak for 232 of 302 features (76.8%) at phase I and 703 of 1075 features (65.4%) at phase II. At the final iteration, weak consensus remained for only two of 487 features (0.4%) at phase I and 19 of 1347 features (1.4%) at phase II. Strong or better consensus was achieved for 463 of 487 features (95.1%) at phase I and 1220 of 1347 features (90.6%) at phase II. Overall, 169 of 174 features were standardized in the first two phases. In the final validation phase (phase III), most of the 169 standardized features could be excellently reproduced (166 with CT; 164 with PET; and 164 with MRI). Conclusion A set of 169 radiomics features was standardized, which enabled verification and calibration of different radiomics software. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Kuhl and Truhn in this issue.

Cost-Effectiveness of Immediate Magnetic Resonance Imaging In the Management of Patients With Suspected Scaphoid Fracture: Results From a Randomized Clinical Trial.

OBJECTIVES: Given the limited diagnostic accuracy of radiographs on presentation to the emergency department (ED), the management of suspected scaphoid fractures remains clinically challenging and poses an unknown economic burden to healthcare systems. We aimed to evaluate the cost-effectiveness of immediate magnetic resonance imaging (MRI) in the management of patients presenting with suspected scaphoid fracture to an ED in England. METHODS: A pragmatic, randomized, single-center trial compared the use of immediate MRI in the ED against standard care with radiographs only. Participants' use of healthcare services and costs were estimated from primary care and secondary care databases and questionnaires at baseline, 1, 3, and 6 months postrecruitment. Costs were compared using generalized linear models and combined with quality-adjusted life years (QALYs, based on the EQ-5D-5L) to estimate cost-effectiveness at 6 months postrecruitment. Cost-effectiveness acceptability curves and bootstrapping techniques were used to estimate the probability of cost-effectiveness at different willingness-to-pay (WTP) thresholds. Four deterministic sensitivity scenarios were considered around key parameters. RESULTS: The MRI intervention dominated standard care in the base case and all 4 deterministic sensitivity scenarios, costing less and achieving more QALY gains, with a probability of 100% of being cost-effective at 6 months using the conventional United Kingdom WTP thresholds of £20 000 to £30 000 per QALY. CONCLUSION: The use of immediate MRI is a cost-effective intervention in the management of suspected scaphoid fractures in a Central Hospital in London. Routine clinical practice at our institution has been changed to include the intervention.

Automated Triaging of Adult Chest Radiographs with Deep Artificial Neural Networks.

Purpose To develop and test an artificial intelligence (AI) system, based on deep convolutional neural networks (CNNs), for automated real-time triaging of adult chest radiographs on the basis of the urgency of imaging appearances. Materials and Methods An AI system was developed by using 470 388 fully anonymized institutional adult chest radiographs acquired from 2007 to 2017. The free-text radiology reports were preprocessed by using an in-house natural language processing (NLP) system modeling radiologic language. The NLP system analyzed the free-text report to prioritize each radiograph as critical, urgent, nonurgent, or normal. An AI system for computer vision using an ensemble of two deep CNNs was then trained by using labeled radiographs to predict the clinical priority from radiologic appearances only. The system's performance in radiograph prioritization was tested in a simulation by using an independent set of 15 887 radiographs. Prediction performance was assessed with the area under the receiver operating characteristic curve; sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were also determined. Nonparametric testing of the improvement in time to final report was determined at a nominal significance level of 5%. Results Normal chest radiographs were detected by our AI system with a sensitivity of 71%, specificity of 95%, PPV of 73%, and NPV of 94%. The average reporting delay was reduced from 11.2 to 2.7 days for critical imaging findings (P < .001) and from 7.6 to 4.1 days for urgent imaging findings (P < .001) in the simulation compared with historical data. Conclusion Automated real-time triaging of adult chest radiographs with use of an artificial intelligence system is feasible, with clinically acceptable performance. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Auffermann in this issue.

Accelerated 3D T2 w-imaging of the prostate with 1-millimeter isotropic resolution in less than 3 minutes.

PURPOSE: To achieve 3D T2 w imaging of the prostate with 1-mm isotropic resolution in less than 3 min. METHODS: We devised and implemented a 3D T2 -prepared multishot balanced steady state free precession (T2 prep-bSSFP) acquisition sequence with a variable density undersampled trajectory combined with a total variation regularized iterative SENSE (TV-SENSE) reconstruction. Prospectively undersampled images of the prostate (acceleration factor R = 3) were acquired in 11 healthy subjects in an institutional review board-approved study. Image quality metrics (subjective signal-to-noise ratio, contrast, sharpness, and overall prostate image quality) were evaluated by 2 radiologists. Scores of the proposed accelerated sequence were compared using the Wilcoxon signed-rank and Kruskal-Wallis non-parametric tests to prostate images acquired using a fully sampled 3D T2 prep-bSSFP acquisition, and with clinical standard 2D and 3D turbo spin echo (TSE) T2 w acquisitions. A P-value < 0.05 was considered significant. RESULTS: The 3× accelerated 3D T2 prep-bSSFP images required a scan time (min:s) of 2:45, while the fully sampled 3D T2 prep-bSSFP and clinical standard 3D TSE images were acquired in 8:23 and 7:29, respectively. Image quality scores (contrast, sharpness, and overall prostate image quality) of the accelerated 3D T2 prep-bSSFP, fully sampled T2 prep-bSSFP, and clinical standard 3D TSE acquisitions along all 3 spatial dimensions were not significantly different (P > 0.05). CONCLUSION: 3D T2 w images of the prostate with 1-mm isotropic resolution can be acquired in less than 3 min, with image quality that is comparable to a clinical standard 3D TSE sequence but only takes a third of the acquisition time.

Accelerated 3D T2 mapping with dictionary-based matching for prostate imaging.

PURPOSE: To develop a fast and accurate method for 3D T2 mapping of prostate cancer using undersampled acquisition and dictionary-based fitting. METHODS: 3D high-resolution T2 -weighted images (0.9 × 0.9 × 3 mm3 ) were obtained with a multishot T2 -prepared balanced steady-state free precession (T2 -prep-bSSFP) acquisition sequence using a 3D variable density undersampled Cartesian trajectory. Each T2 -weighted image was reconstructed using total variation regularized sensitivity encoding. A flexible simulation framework based on extended phase graphs generated a dictionary of magnetization signals, which was customized to the proposed sequence. The dictionary was matched to the acquired T2 -weighted images to retrieve quantitative T2 values, which were then compared to gold-standard spin echo acquisition values using monoexponential fitting. The proposed approach was validated in simulations and a T1 /T2 phantom, and feasibility was tested in 8 healthy subjects. RESULTS: The simulation analysis showed that the proposed T2 mapping approach is robust to noise and typically observed T1 variations. T2 values obtained in the phantom with T2 prep-bSSFP and the acquisition-specific, dictionary-based matching were highly correlated with the gold-standard spin echo method (r = 0.99). Furthermore, no differences were observed with the accelerated acquisition compared to the fully sampled acquisition (r = 0.99). T2 values obtained in prostate peripheral zone, central gland, and muscle in healthy subjects (age, 26 ± 6 years) were 97 ± 14, 76 ± 7, and 36 ± 3 ms, respectively. CONCLUSION: 3D quantitative T2 mapping of the whole prostate can be achieved in 3 minutes.

What can artificial intelligence teach us about the molecular mechanisms underlying disease?

While molecular imaging with positron emission tomography or single-photon emission computed tomography already reports on tumour molecular mechanisms on a macroscopic scale, there is increasing evidence that there are multiple additional features within medical images that can further improve tumour characterization, treatment prediction and prognostication. Early reports have already revealed the power of radiomics to personalize and improve patient management and outcomes. What remains unclear is how these additional metrics relate to underlying molecular mechanisms of disease. Furthermore, the ability to deal with increasingly large amounts of data from medical images and beyond in a rapid, reproducible and transparent manner is essential for future clinical practice. Here, artificial intelligence (AI) may have an impact. AI encompasses a broad range of 'intelligent' functions performed by computers, including language processing, knowledge representation, problem solving and planning. While rule-based algorithms, e.g. computer-aided diagnosis, have been in use for medical imaging since the 1990s, the resurgent interest in AI is related to improvements in computing power and advances in machine learning (ML). In this review we consider why molecular and cellular processes are of interest and which processes have already been exposed to AI and ML methods as reported in the literature. Non-small-cell lung cancer is used as an exemplar and the focus of this review as the most common tumour type in which AI and ML approaches have been tested and to illustrate some of the concepts.