Local Radiology Research Articles

Comparison of Expert Reads vs. Local Clinical Reads of Amyloid PET Scans in IDEAS

AbstractBackgroundAmyloid‐directed therapies will increase the clinical use of amyloid PET. IDEAS captured >17,000 clinical scan interpretations from local radiologists and nuclear medicine physicians. This study aimed to establish accuracy of these interpretations by comparison to expert readers across identical scans.MethodRandomly selected amyloid PET/CT scans performed using one of three FDA‐approved agents [18F‐Florbetaben (FBB), 18F‐Florbetapir (FBP), 18F‐Flutemetamol (FMM)] and previously interpreted by local readers from the Imaging Dementia ‐ Evidence for Amyloid Scanning (IDEAS) study were assigned to vendor‐selected panels of 3 expert readers for interpretation; 500 scans per agent and 1 panel per agent. Scans were stratified to match frequencies observed in the overall study by 1) participant age; 2) MCI vs dementia; 3) positive/negative interpretation; and 4) type of PET facility. Respective vendors performed quality control on attenuation‐corrected PET/CT images. Visual interpretations (positive or negative for Aß cortical deposition) were performed according to agent‐specific validated interpretation criteria. Expert panel readers were blinded to clinical and demographic data. Cohen’s kappa statistic with two‐sided 95% confidence interval estimates examined agreement between local reader and majority expert panel visual interpretation.ResultExpert panel readers’ visual interpretations totaled 4489. Median age for scanned participants was 75 (range 65‐96), 51.4% were female and 60.5% had MCI. Local readers interpreted 60.9% of scans positive and 39.1% negative (Table 1). Across all radiotracers, agreement between majority expert read and local readers was excellent (k coefficient 0.76; 95% CI 0.73 – 0.80, p<.0001) with 86.6% (791/913) agreement for positive scans and 90.9% (532/585) agreement for negative scans (Figure 1). Agreement by individual radiopharmaceuticals was good to excellent: k coefficient 0.78 (95% CI 0.72 – 0.83, p<.001) for FBB, k coefficient 0.72 (95% CI 0.66 – 0.78, p<.001) for FBP, and k coefficient 0.78 (95% CI 0.73 – 0.84) for FMM (Figure 1). Agreement levels between local readers and individual experts varied from good to excellent (Figure 2).ConclusionWe found good to excellent agreement between local readers and majority assessment of three‐member expert panels for all combined scans and scans by individual agents. Visual interpretation of Aß PET can be performed in a clinical setting with high reliability.

Comparison of Expert Reads vs. Local Clinical Reads of Amyloid PET Scans in IDEAS

AbstractBackgroundAmyloid‐directed therapies will increase the clinical use of amyloid PET. IDEAS captured >17,000 clinical scan interpretations from local radiologists and nuclear medicine physicians. This study aimed to establish accuracy of these interpretations by comparison to expert readers across identical scans.MethodRandomly selected amyloid PET/CT scans performed using one of three FDA‐approved agents [18F‐Florbetaben (FBB), 18F‐Florbetapir (FBP), 18F‐Flutemetamol (FMM)] and previously interpreted by local readers from the Imaging Dementia ‐ Evidence for Amyloid Scanning (IDEAS) study were assigned to vendor‐selected panels of 3 expert readers for interpretation; 500 scans per agent and 1 panel per agent. Scans were stratified to match frequencies observed in the overall study by 1) participant age; 2) MCI vs dementia; 3) positive/negative interpretation; and 4) type of PET facility. Respective vendors performed quality control on attenuation‐corrected PET/CT images.Visual interpretations (positive or negative for Aβ cortical deposition) were performed according to agent‐specific validated interpretation criteria. Expert panel readers were blinded to clinical and demographic data. Cohen’s kappa statistic with two‐sided 95% confidence interval estimates examined agreement between local reader and majority expert panel visual interpretation.ResultExpert panel readers’ visual interpretations totaled 4489. Median age for scanned participants was 75 (range 65‐96), 51.4% were female and 60.5% had MCI. Local readers interpreted 60.9% of scans positive and 39.1% negative (Table 1).Across all radiotracers, agreement between majority expert read and local readers was excellent (k coefficient 0.76; 95% CI 0.73 – 0.80, p<.0001) with 86.6% (791/913) agreement for positive scans and 90.9% (532/585) agreement for negative scans (Figure 1). Agreement by individual radiopharmaceuticals was good to excellent: k coefficient 0.78 (95% CI 0.72 – 0.83, p<.001) for FBB, k coefficient 0.72 (95% CI 0.66 – 0.78, p<.001) for FBP, and k coefficient 0.78 (95% CI 0.73 – 0.84) for FMM (Figure 1). Agreement levels between local readers and individual experts varied from good to excellent (Figure 2).ConclusionWe found good to excellent agreement between local readers and majority assessment of three‐member expert panels for all combined scans and scans by individual agents. Visual interpretation of Aβ PET can be performed in a clinical setting with high reliability.

Nasal obstruction in a 14 year old girl caused by a huge middle turbinate mucocele appearing radiologically as an inverted papilloma

Introduction: We present a case of a fourteen year old girl who presented with a large intra-nasal mass to the ENT team at a district general hospital in the UK. Presentation of case: The girl presented predominantly with nasal obstruction and some symptoms of allergic rhinitis. Imaging revealed a large lesion abutting the skull base and causing bony remodelling with marked septal deviation. Based both on CT and MRI imaging, the reporting (non-head and neck) radiologist suggested inverted papilloma as a differential diagnosis. Intra-operative exploration in fact revealed a very large left middle turbinate mucocele extending to the left frontal sinus. The mass was excised endoscopically without complications. Discussion: Although concha bullosa of the middle turbinate of the nose are common, development of a mucocele within them is far less common and for such a mucocele to develop to this size in a child is extremely rare. The egg shell lining of the lesion can be a tell-tale sign of their aetiology when taken alongside other radiological factors. This case highlights challenges in radiological diagnosis of intra-nasal masses in children, which can lead to delays and increased anxiety. Conclusion: When assessing nasal masses in children it is important to keep a wide differential due to the challenges of diagnosis. A close conversation should be had with local head and neck radiologists and, of course, where there is a unilateral nasal mass tissue sampling is essential and may be taken as part of a full excision where clinically indicated.

Indeterminate pulmonary nodules in non-rhabdomyosarcoma soft tissue sarcoma: A study of the European paediatric Soft Tissue Sarcoma Study Group.

The aim of this study was to assess the clinical impact of indeterminate pulmonary nodules (no more than four pulmonary nodules of less than 5mm or one nodule measuring between 5 and less than 10mm by computed tomography [CT]) in children and adolescents with adult-type non-rhabdomyosarcoma soft tissue sarcoma (NRSTS) at diagnosis. Patients with NRSTS treated in 11 centers as part of the European paediatric Soft Tissue Sarcoma Study Group (EpSSG) were retrospectively assessed. Local radiologists, blinded to clinical information except for patients' age and tumor histotype, reviewed the chest CT at diagnosis and filled out a case report form. Because patients with or without indeterminate nodules in the EpSSG NRSTS 2005 study received the same type of treatment, event-free survival (EFS) and overall survival (OS) between groups by log-rank test were compared. Overall, 206 patients were examined: 109 (52.9%) were without any nodules, 78 (38%) had at least one indeterminate nodule, and 19 (9.2%) had nodules meeting the definition of metastases, which were then considered to be misclassified and were excluded from further analyses. Five-year EFS was 78.5% (95% CI, 69.4%-85.1%) for patients without nodules and 69.6% (95% CI, 57.9%-78.7%) for patients with indeterminate nodules (p=.135); 5-year OS was 87.4% (95% CI, 79.3%-92.5%) and 79.0% (95% CI, 67.5%-86.8%), respectively (p=.086). This study suggests that survival does not differ in otherwise nonmetastatic patients with indeterminate pulmonary nodules compared to nonmetastatic patients without pulmonary nodules. Radiologists should be aware of the classification of indeterminate pulmonary nodules in non-rhabdomyosarcoma soft tissue sarcomas and use it in their reports. More than a third of patients with non-rhabdomyosarcoma soft tissue sarcoma can be affected by indeterminate pulmonary nodules. Indeterminate pulmonary nodules do not significantly affect the overall survival of pediatric patients with non-rhabdomyosarcoma soft tissue sarcoma.

Deploying Artificial Intelligence for Thoracic Imaging Around the World

PurposeArtificial intelligence (AI) thoracic imaging applications are increasingly being deployed in low- and middle-income countries (LMICs). Radiologists have a critical gatekeeping role to ensure the effective and ethical implementation of AI solutions. RAD-AID International uses a three-pronged implementation strategy to overcome challenges pervasive in LMICs. MethodsDuring a similar period, an AI software for chest radiography (CXR) interpretation was deployed at two tertiary hospitals located in Guyana and Nigeria. The three-pronged implementation strategy of clinical education, infrastructure implementation, and phased AI introduction was used. A PACS with a cloud component was installed at each institution. Radiology residents and attending physicians at these institutions completed an introduction-to-AI course to prime them for the use of AI solutions. A phased introduction of the AI software was performed to allow local validation as well as trust building and workflow integration. Local validation processes were used at each site by comparing AI outputs with standardized prospectively generated reports by local radiologists and study team members, allowing for slight differences in the goals of AI software use between sites. ResultsThe PACS was successfully installed at both institutions. Thirty participants completed the introduction-to-AI course with an average pre–knowledge test score of 75% and an average posttest score of 95%. The focus of the validation process at various sites was reflective of the intended use of the AI software. In Guyana, it revealed an 87% concordance rate between radiologists and the AI model for triaging normal versus abnormal findings on CXR. In Nigeria, an 85% concordance rate between radiologists and the AI model for reporting tuberculosis on CXR was noted. The AI software was successfully deployed and is being used as intended at both institutions. ConclusionsThere are unique barriers to the adoption of AI in LMICs requiring an implementation strategy in collaboration with local institutions and industry partners to ensure success.

MP38-10 MULTI-INSTITUTIONAL VALIDATION OF THE PRECISE CRITERIA ON PROSTATE MRI DURING ACTIVE SURVEILLANCE

You have accessJournal of UrologyCME1 Apr 2023MP38-10 MULTI-INSTITUTIONAL VALIDATION OF THE PRECISE CRITERIA ON PROSTATE MRI DURING ACTIVE SURVEILLANCE Francesco Giganti, Armando Stabile, Riccardo Leni, Mark Emberton, Caroline M. Moore, Massimo Valerio, Veeru Kasivisvanathan, and Precise Study Group Francesco GigantiFrancesco Giganti More articles by this author , Armando StabileArmando Stabile More articles by this author , Riccardo LeniRiccardo Leni More articles by this author , Mark EmbertonMark Emberton More articles by this author , Caroline M. MooreCaroline M. Moore More articles by this author , Massimo ValerioMassimo Valerio More articles by this author , Veeru KasivisvanathanVeeru Kasivisvanathan More articles by this author , and Precise Study GroupPrecise Study Group More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003276.10AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Currently, AS programmes vary across the world, with different timings in terms of follow-up on MRI and biopsy. The PRECISE criteria provide a dedicated 1-to-5 scale to evaluate the radiological change on serial MRI: a PRECISE score of 1 or 2 denotes radiological regression, PRECISE 3 indicates stability and PRECISE 4 or 5 implies progression. We report the first validation of the PRECISE score in a multicentre international setting. METHODS: We collected data from 22 centres across the world and applied two entry criteria: i) at least 2 MR scans (baseline and follow-up); ii) at least 2 biopsies (baseline and follow-up, the latter after the second MR). Local radiologists reported the scans using PRECISE. Histological progression was defined as any increase in Gleason Score from baseline biopsy. Progression-free survival (PFS) was estimated using Kaplan-Meier curves with landmark time starting from date of first follow-up MRI (i.e., second MR), and multivariable Cox proportional hazards model tested the predictive role of PRECISE. RESULTS: A total of 1,556 patients were included, 1389 (89%) of which had Gleason 3+3 and 167 (11%) had Gleason ≥ 3+4 at baseline. Median follow-up was 48 months, and 513 (33%) patients experienced histological progression. The overall 5- and 10-year progression-free survival (PFS) were 66% and 41%, with time starting at diagnosis. For PRECISE 1-2 (n=158), PFS was 88% at 2 years from first follow-up MRI and 77% at 5 years, while for PRECISE 4-5 (n=413), PFS was 52% and 35%. In those with PRECISE 3 [stability] and visible lesions (n=597), 2- and 5-year PFS was 76% and 50%, while for PRECISE 3 [stability] and non-visible lesions (n=388), PFS was similar to PRECISE 1-2 (2- and 5-year PFS: 91% and 80%) (Log Rank p<0.001). At multivariable analysis, PRECISE 4-5 remained an independent predictor of histological progression (HR: 2.63; p<0.0001). CONCLUSIONS: Our work has validated the PRECISE score in a multicentre setting and has the potential to be included in future risk models that incorporate both clinical and radiological data to fully benefit from prostate MRI and targeted biopsies. Source of Funding: Prostate Cancer Foundation / CRIS Cancer Foundation (Francesco Giganti is a recipient of the 2020 Young Investigator Award). © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e527 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Francesco Giganti More articles by this author Armando Stabile More articles by this author Riccardo Leni More articles by this author Mark Emberton More articles by this author Caroline M. Moore More articles by this author Massimo Valerio More articles by this author Veeru Kasivisvanathan More articles by this author Precise Study Group More articles by this author Expand All Advertisement PDF downloadLoading ...

Abstract OT2-01-01: Trial in Progress: Clinical Utility of Fluoroestradiol F18 PET/CT in Metastatic Breast Cancer Patients with ER-Positive and HER2-Negative Primary Lesions after Progression on First Line Hormonal Therapy

Abstract BACKGROUND: Cyclin-dependent kinase 4/6 (CDK4/6) inhibitors combined with endocrine therapy are the recommended first-line standard-of-care (SOC) treatment for estrogen receptor (ER)-positive, human epidermal growth factor receptor 2 (HER2)-negative metastatic breast cancer (MBC) patients. After progression on first-line treatment, current clinical practice guidelines (NCCN, ESMO) recommend to sequence endocrine therapy (in absence of visceral crisis) until there is no clinical benefit after up to 3 regimens, and subsequently initiate chemotherapy regimens. However, clinical benefit for endocrine monotherapy after progression on first-line standard-of-care is limited, with median progression-free survival (PFS) of 2-5 months. Multiple factors can contribute to endocrine therapy resistance, including changes in ER expression, (epi)genetic mutations, and/or clonal selection under therapeutic pressure. ER heterogeneity can occur over time (temporal) as well as between lesions in a patient (spatial). As confirmed by tissue biopsy, greater than 20% of MBC patients have discordance in ER expression between their primary tumor and a metastatic lesion. Measuring total ER heterogeneity in a patient – the variation in ER expression at whole body level across all MBC lesions – is not clinically feasible by tissue biopsy alone. Fluoroestradiol F18 (Cerianna), also called 18F-FES, is a radioactive diagnostic agent that can be used with PET/CT for in vivo detection of ER-positive lesions in MBC with high accuracy. Results from 18F-FES PET/CT studies have shown that ER expression across MBC lesions at whole body level is heterogeneous in approximately half of ER-positive, HER2-negative MBC patients. In this study, we aim to evaluate the clinical utility of 18F-FES PET/CT to guide second-line treatment decision in ER-positive, HER2-negative MBC patients with progressive disease on first-line SOC hormonal therapy. METHODS: Fluoroestradiol F18 (Cerianna) is being evaluated in a phase 4, open-label, prospective cohort study enrolling 206 patients at 20-30 centers in the United States. Patients diagnosed with ER-positive, HER2-negative MBC who have progressive disease on first-line aromatase inhibitor therapy with or without CDK4/6 inhibitor will be included in the study; patients with isolated hepatic metastases will be excluded. All patients will undergo an 18F-FES PET/CT scan. The treating physician will complete a standardized questionnaire to indicate the second-line therapeutic management plan before the scheduled 18F-FES PET/CT. After interpretation of 18F-FES PET/CT by local radiologist or nuclear medicine physician, imaging results will be provided to the treating physician who will then fill out a similar questionnaire to specify the final therapeutic decision. The proportion of patients with a change in therapeutic management plan based on incorporation of 18F-FES PET/CT results will be the primary endpoint. During the study follow-up period of 18 months, data on SOC imaging, treatments/procedures received, and clinical outcomes will be collected. Secondary endpoints include visual and quantitative heterogeneity assessment of tumor 18F-FES uptake, and PFS rates at 6 months and 18 months after 18F-FES PET/CT, which will be assessed between patients with and without a change in therapeutic management plan. The trial is planned to begin enrollment in Q3 2022. Clinical trial information: NCT05068726 Citation Format: Stephanie van de Ven, Feng Luo, Nicholas DiGregorio, Adolfo Fuentes-Alburo, Francois Tranquart. Trial in Progress: Clinical Utility of Fluoroestradiol F18 PET/CT in Metastatic Breast Cancer Patients with ER-Positive and HER2-Negative Primary Lesions after Progression on First Line Hormonal Therapy [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr OT2-01-01.

127 Years of New Orleans Radiology: The Evolution of Practice and Those Who Led It.

The history of New Orleans radiology helps us understand how innovative medical ideas are transmitted and adopted and how the specialty evolved to incorporate these new ideas. As radiology became a required component of medical practice, additional expertise was needed. This resulted in the need for standardized specialized training and the necessity for some form of certification. New Orleans radiologists have always been involved in these American Board of Radiology efforts and also have held leadership positions in other national radiology organizations. The practice of radiology evolved from photographers being employees to physicians becoming members of a unique specialty with practice ownership interests. Radiologists united to form large practice groups and joined professional associations to share technological innovations as well as ensure a culture of professional collegiality. In New Orleans, organizations and events such as the New Orleans Radiology Society, the Spring Roentgen Conference, and the New Orleans Ultrasound Association were organized. Local radiologists who directed these efforts also participated on boards of many national radiologic organizations and helped advance the science of radiology. As the financial complexity and political stresses of radiology practice increased, groups merged, with a resultant decrease in individual autonomy and an increase in radiologist burnout. This has directly and indirectly resulted in many radiologists becoming employees as they were in the early days of New Orleans radiology.

Look how far we have come: BREAST cancer detection education on the international stage.

The development of screening mammography over 30 years has remarkedly reduced breast cancer-associated mortality by 20%-30% through detection of small cancer lesions at early stages. Yet breast screening programmes may function differently in each nation depending on the incidence rate, national legislation, local health infrastructure and training opportunities including feedback on performance. Mammography has been the frontline breast cancer screening tool for several decades; however, it is estimated that there are 15% to 35% of cancers missed on screening which are owing to perceptual and decision-making errors by radiologists and other readers. Furthermore, mammography screening is not available in all countries and the increased speed in the number of new breast cancer cases among less developed countries exceeds that of the developed world in recent decades. Studies conducted through the BreastScreen Reader Assessment Strategy (BREAST) training tools for breast screening readers have documented benchmarking and significant variation in diagnostic performances in screening mammogram test sets in different countries. The performance of the radiologists from less well-established breast screening countries such as China, Mongolia and Vietnam were significant lower in detecting early-stage cancers than radiologists from developed countries such as Australia, USA, Singapore, Italy. Differences in breast features and cancer presentations, discrepancies in the level of experiences in reading screening mammograms, the availability of high-quality national breast screening program and breast image interpretation training courses between developed and less developed countries are likely to have impact on the variation of readers' performances. Hence dedicated education training programs with the ability to tailor to different reader cohorts and different population presentations are suggested to ameliorate challenges in exposure to a range of cancer cases and improve the interpretation skills of local radiologists. Findings from this review provide a good understanding of the radiologist' performances and their improvement using the education interventions, primarily the BREAST program, which has been deployed in a large range of developing and developed countries in the last decade. Self-testing and immediate feedback loops have been shown to have important implications for benchmarking and improving the diagnostic accuracy in radiology worldwide for better breast cancer control.

Clinical Activity of Anti-PD-1 Therapy Following CD30 CAR-T Cell Therapy in Relapsed Hodgkin Lymphoma

Background: Anti-CD30 chimeric antigen receptor (CD30.CAR) T-cell therapy has been studied in heavily pretreated relapsed/refractory (r/r) classic Hodgkin lymphoma (cHL) patients with an objective response rate (ORR) of 62% and complete response (CR) rate of 51% (Ramos et al, JCO, 2020). Despite high response rates, disease progression has been observed, with a 1-year PFS of 36%. The anti-PD-1 antibodies nivolumab and pembrolizumab have both been studied in phase II studies in patients with r/r cHL with ORR ranging between 65-87% and median progression free survival (PFS) ranging from 13-15 months. Limited data is available regarding the clinical benefit of anti-PD-1 therapies following CD30.CAR-T cell therapy and the potential mechanism of response in patients with prior anti-PD-1 exposure. Methods: We followed and analyzed the clinical outcome of 10 patients with r/r HL who progressed after CD30.CAR-T cell therapy at the University of North Carolina at Chapel Hill and subsequently received anti-PD-1 therapy to treat relapsed disease. Clinical responses in follow up were determined per local radiologist assessment. Expansion and persistence of CD30.CAR-T cells in vivo was determined by quantitative PCR from peripheral blood samples. Results: Between October 2016 and November 2019, patients were treated with either CD30.CAR-T cells (9 patients) or CD30.CCR4.CAR-T cells (1 patient) on one of three clinical trials (NCT02690545, NCT02663297, NCT03602157). The median age of patients was 30 years (range, 24-51 years). The median number of prior lines of therapy was 8 (range, 5-16). Eight patients had a prior autologous stem cell transplant, and six patients had a prior allogeneic stem cell transplant. The best response to CAR-T cell therapy was a CR in 5 patients, partial response (PR) in 1 patient, and stable/progressive disease in 4 patients. At the time of progression following CD30.CAR-T cell therapy, patients were offered standard of care anti-PD-1 therapy. Eight patients were treated with nivolumab, and 2 patients were treated with pembrolizumab based on decision of the treating physician. The median time from CAR-T cell therapy to anti-PD-1 therapy was 171 days (range, 56-858 days). Seven of the 10 patients had prior anti-PD-1 exposure before CAR-T cell therapy. All 10 patients had an objective response to anti-PD-1 therapy after CAR-T cells, with a best response of CR in 7 patients and PR in 3 patients. Four of the 6 patients who had a PR to anti-PD-1 therapy prior to CAR-T cell treatment, achieved CR with anti-PD-1 re-challenge afterwards. At the time of the last follow up, 8 patients remain in an ongoing response, lasting over 1 year in 7 and over 3 years in 3 patients. The median PFS was not reached in this cohort (Fig 1A). Five patients stopped anti-PD-1 therapy after CAR-T cell therapy (4 related to toxicity of therapy, 1 for ongoing durable response). At the time of data collection, with a median follow up of 3.6 years (range, 1.7-4.7 years), all patients were still alive. Longitudinal peripheral blood samples were available in 3 of the 10 patients. All 3 patients with longitudinal samples demonstrated expansion and/or persistence of CD30.CAR-T cells following anti-PD-1 therapy (Fig 1B). Patient #9 and #5 had impressive expansion of CD30.CAR-T cell signals shortly after start of anti-PD-1 therapy. Patient #1 had detectable PCR signals in the peripheral blood until at least 1 year after CD30.CAR-T cell infusion; however, sample timepoints were not available to assess for expansion of CD30.CAR-T cell signals shortly after anti-PD-1 therapy. Conclusion: To our knowledge, this is the largest cohort of patients with r/r cHL who have received anti-PD-1 therapies after CD30.CAR-T cell treatment. Clinical efficacy appears strong, with high response rates and durable remissions observed even if most patients had prior exposure to anti-PD-1 therapy. All patients with no prior anti-PD-1 exposure experienced durable remissions. This study is the first to describe CD30.CAR-T expansion following anti-PD-1 therapy. While further analysis is warranted to assess for CD30.CAR-T functionality after anti-PD-1 exposure, our data suggest that anti-PD-1 therapy may rescue CD30.CAR T cells, supporting further investigations in patients with cHL. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Use of YEARS Algorithm in Military Health Care Beneficiaries to Evaluate Use of Computed Tomography Pulmonary Angiography.

Pulmonary embolism (PE) can be difficult to rule out without computed tomography pulmonary angiograms (CTPAs), as presentations vary. Multiple clinical decision rules (CDRs) exist to risk-stratify patients to avoid unnecessary CTPAs. However, the currently used CDRs are complex, and research has shown low compliance with their usage. The YEARS algorithm is less complex and excludes patients from CTPA if PE is not the most likely diagnosis, they do not have a history of hemoptysis, and no clinical signs of deep vein thrombosis, is less complex. However, no studies have evaluated YEARS in the U.S. Military health care beneficiary population. Therefore, this study sought to determine if implementing the YEARS algorithm could decrease the number of CTPAs ordered to rule out PE in low-risk patients. This retrospective, single-center cohort study applied the YEARS algorithm to low-risk military beneficiaries presenting to the emergency department in the calendar year 2020 at a single U.S. Army MTF. The primary outcome was the number of CTPAs indicated by the YEARS algorithm versus the number ordered via standard practice. We used chi-square testing to compare the number of subjects in whom YEARS indicated CTPA (meets criteria/does not meet criteria) versus the actual number of subjects who underwent CTPA (meets criteria/does not meet criteria). The secondary outcomes included applying YEARS similarly to the number of subjects >50 years of age (as opposed to age-adjusted d-dimer), determining the number of pregnant patients who could have avoided CTPA via application of YEARS, and assessing possible cost savings via reduction of CTPA. We included 353 subjects during the study period, 271 of whom underwent CTPA. YEARS would have only indicated 25 of them, P = .018. In patients >50 years of age, 164 underwent CTPA versus nine who met YEARS criteria, P = .014. Among pregnant patients, six underwent CTPA versus one who YEARS would have indicated, P = .130. Application of the YEARS algorithm would have led to a 90.8% reduction in CTPAs ordered with an overall known missed PE rate of 1.1%. Applying the YEARS algorithm in 2020 could have led to 246 fewer CTPAs at a minimum cost savings of $38,762.22 for the MTF based on the coded billing cost of $157.57 per CTPA when local staff radiologists performed image interpretation. This estimate does not consider the additional undisclosed cost of contracted radiologists interpreting after regular duty hours. For our military beneficiaries, our study indicates that the YEARS algorithm would have reduced CTPA utilization in all age ranges and potentially among pregnant patients with a known missed PE rate of 1.1%.

Impact of radiographer immediate reporting of X-rays of the chest from general practice on the lung cancer pathway (radioX): a randomised controlled trial

The National Optimal Lung Cancer Pathway recommends rapid progression from abnormal chest X-rays (CXRs) to CT. The impact of the more rapid reporting on the whole pathway is unknown. The...

Gradual Self-Training via Confidence and Volume Based Domain Adaptation for Multi Dataset Deep Learning-Based Brain Metastases Detection Using Nonlocal Networks on MRI Images.

Research suggests that treatment of multiple brain metastases (BMs) with stereotactic radiosurgery shows improvement when metastases are detected early, providing a case for BM detection capabilities on small lesions. To demonstrate automatic detection of BM on three MRI datasets using a deep learning-based approach. To improve the performance of the network is iteratively co-trained with datasets from different domains. A systematic approach is proposed to prevent catastrophic forgetting during co-training. Retrospective. A total of 156 patients (105 ground truth and 51 pseudo labels) with 1502 BM (BrainMetShare); 121 patients with 722 BM (local); 400 patients with 447 primary gliomas (BrATS). Training/pseudo labels/validation data were distributed 84/51/21 (BrainMetShare). Training/validation data were split: 121/23 (local) and 375/25 (BrATS). A 5 T and 3T/T1 spin-echo postcontrast (T1-gradient echo) (BrainMetShare), 3T/T1 magnetization prepared rapid acquisition gradient echo postcontrast (T1-MPRAGE) (local), 0.5T, 1T, and 1.16 T/T1-weighted-fluid-attenuated inversion recovery (T1-FLAIR) (BrATS). The ground truth was manually segmented by two (BrainMetShare) and four (BrATS) radiologists and manually annotated by one (local) radiologist. Confidence and volume based domain adaptation (CAVEAT) method of co-training the three datasets on a 3D nonlocal convolutional neural network (CNN) architecture was implemented to detect BM. The performance was evaluated using sensitivity and false positive rates per patient (FP/patient) and free receiver operating characteristic (FROC) analysis at seven predefined (1/8, 1/4, 1/2, 1, 2, 4, and 8) FPs per scan. The sensitivity and FP/patient from a held-out set registered 0.811 at 2.952 FP/patient (BrainMetShare), 0.74 at 3.130 (local), and 0.723 at 2.240 (BrATS) using the CAVEAT approach with lesions as small as 1 mm being detected. Improved sensitivities at lower FP can be achieved by co-training datasets via the CAVEAT paradigm to address the problem of data sparsity. 3 TECHNICAL EFFICACY STAGE: 2.

Segmentation of breast tumors using cutting-edge semantic segmentation models

ABSTRACT Early detection of breast cancer is the most important area of mammography research at the moment. It is critical to use computer-aided diagnosis to screen for and prevent breast cancer. In this study, the effectiveness of cutting-edge deep segmentation models for mammography in the detection of breast tumors was investigated. A medical images dataset was compiled and annotated at Lady Reading Hospital, one of the largest teaching hospitals in Pakistan in collaboration with the local health specialists, radiologists, and technologists. A comparison was made between the performance of the segmentation techniques used, and the model that performed the best in detecting tumors and normal breast regions was selected. The evaluation metrics, such as the mean IoU, pixel accuracy, and an in-depth experimental evaluation were used as performance parameters. This investigation determined how well semantic segmentation techniques were performed based on two datasets (cityscapes and mammograms) in this study. The global Dilation 10 semantic segmentation model outperformed the other three semantic segmentation models with a pixel accuracy of 92.98 percent in comparison tests. This paper demonstrates the efficacy of pixel-wise image segmentation techniques and their superiority to other techniques by outperforming other current state-of-the-art automatic image segmentation models.

A Cyber-Security Risk Assessment Methodology for Medical Imaging Devices: the Radiologists' Perspective.

Medical imaging devices (MIDs) are exposed to cyber-security threats. Currently, a comprehensive, efficient methodology dedicated to MID cyber-security risk assessment is lacking. We propose the Threat identification, ontology-based Likelihood, severity Decomposition, and Risk assessment (TLDR) methodology and demonstrate its feasibility and consistency with existing methodologies, while being more efficient, providing details regarding the severity components, and supporting organizational prioritization and customization. Using our methodology, the impact of 23 MIDs attacks (that were previously identified) was decomposed into six severity aspects. Four Radiology Medical Experts (RMEs) were asked to assess these six aspects for each attack. The TLDR methodology's external consistency was demonstrated by calculating paired T-tests between TLDR severity assessments and those of existing methodologies (and between the respective overall risk assessments, using attack likelihood estimates by four healthcare cyber-security experts); the differences were insignificant, implying externally consistent risk assessment. The TLDR methodology's internal consistency was evaluated by calculating the pairwise Spearman rank correlations between the severity assessments of different groups of two to four RMEs and each of their individual group members, showing that the correlations between the severity rankings, using the TLDR methodology, were significant (P < 0.05), demonstrating that the severity rankings were internally consistent for all groups of RMEs. Using existing methodologies, however, the internal correlations were insignificant for groups of less than four RMEs. Furthermore, compared to standard risk assessment techniques, the TLDR methodology is also sensitive to local radiologists' preferences, supports a greater level of flexibility regarding risk prioritization, and produces more transparent risk assessments.

Development and multicenter validation of chest X-ray radiography interpretations based on natural language processing

BackgroundArtificial intelligence can assist in interpreting chest X-ray radiography (CXR) data, but large datasets require efficient image annotation. The purpose of this study is to extract CXR labels from diagnostic reports based on natural language processing, train convolutional neural networks (CNNs), and evaluate the classification performance of CNN using CXR data from multiple centersMethodsWe collected the CXR images and corresponding radiology reports of 74,082 subjects as the training dataset. The linguistic entities and relationships from unstructured radiology reports were extracted by the bidirectional encoder representations from transformers (BERT) model, and a knowledge graph was constructed to represent the association between image labels of abnormal signs and the report text of CXR. Then, a 25-label classification system were built to train and test the CNN models with weakly supervised labeling.ResultsIn three external test cohorts of 5,996 symptomatic patients, 2,130 screening examinees, and 1,804 community clinic patients, the mean AUC of identifying 25 abnormal signs by CNN reaches 0.866 ± 0.110, 0.891 ± 0.147, and 0.796 ± 0.157, respectively. In symptomatic patients, CNN shows no significant difference with local radiologists in identifying 21 signs (p > 0.05), but is poorer for 4 signs (p < 0.05). In screening examinees, CNN shows no significant difference for 17 signs (p > 0.05), but is poorer at classifying nodules (p = 0.013). In community clinic patients, CNN shows no significant difference for 12 signs (p > 0.05), but performs better for 6 signs (p < 0.001).ConclusionWe construct and validate an effective CXR interpretation system based on natural language processing.

Deep learning assistance for tuberculosis diagnosis with chest radiography in low-resource settings.

Tuberculosis (TB) is a major health issue with high mortality rates worldwide. Recently, tremendous researches of artificial intelligence (AI) have been conducted targeting at TB to reduce the diagnostic burden. However, most researches are conducted in the developed urban areas. The feasibility of applying AI in low-resource settings remains unexplored. In this study, we apply an automated detection (AI) system to screen a large population in an underdeveloped area and evaluate feasibility and contribution of applying AI to help local radiologists detect and diagnose TB using chest X-ray (CXR) images. First, we divide image data into one training dataset including 2627 TB-positive cases and 7375 TB-negative cases and one testing dataset containing 276 TB-positive cases and 619 TB-negative cases, respectively. Next, in building AI system, the experiment includes image labeling and preprocessing, model training and testing. A segmentation model named TB-UNet is also built to detect diseased regions, which uses ResNeXt as the encoder of U-Net. We use AI-generated confidence score to predict the likelihood of each testing case being TB-positive. Then, we conduct two experiments to compare results between the AI system and radiologists with and without AI assistance. Study results show that AI system yields TB detection accuracy of 85%, which is much higher than detection accuracy of radiologists (62%) without AI assistance. In addition, with AI assistance, the TB diagnostic sensitivity of local radiologists is improved by 11.8%. Therefore, this study demonstrates that AI has great potential to help detection, prevention, and control of TB in low-resource settings, particularly in areas with more scant doctors and higher rates of the infected population.

Combination brentuximab vedotin and bendamustine for pediatric patients with relapsed/refractory Hodgkin lymphoma

AbstractIn patients with relapsed/refractory (R/R) Hodgkin lymphoma (HL), achieving a complete metabolic response (CMR) after salvage therapy is associated with superior outcomes, and optimal treatments must be identified. The combination of brentuximab vedotin and bendamustine (BVB), although highly active in adult patients, has not been extensively evaluated in pediatric patients with R/R HL. We performed a multicenter, retrospective review of pediatric patients <21 years of age with R/R HL treated with BVB from January 2016 through July 2019. Response was assessed by local radiologists according to Lugano classification criteria. Twenty-nine patients (17 relapsed, 12 refractory) with a median age of 16 years (range, 10-20) were treated with BVB and received a median of 3 cycles of therapy (range, 2-7). Patients received an infusion of 1.8 mg/kg of BV on day 1 with bendamustine 90 mg/m2 on days 1 and 2 of 3-week cycles. Nineteen patients (66%) achieved a CMR (95% CI, 46-82). An objective response was observed in 23 patients (objective response rate, 79%; 95% CI, 60-92). The most common grade 3 and 4 toxicities were hematologic, and 3 patients (10%) experienced grade 3 infusion reactions. Seventeen of 18 patients underwent successful mobilization and collection of stem cells. Sixteen patients (13 autologous, 3 allogeneic) received a consolidative transplant after BVB. The 3-year post-BVB event-free and overall survival were 65% (95% CI, 46-85) and 89% (95% CI, 74-100), respectively. For pediatric patients with R/R HL, BVB was well tolerated and compared favorably with currently accepted salvage regimens.

Comparison of Multiple Deep Models on Semantic Segmentation for Breast Tumor Detection

The early diagnosis of breast tumor detection is the most significant research issue in mammography. Computer-aided diagnosis (CAD) is one of the highly essential methods to prevent breast cancer. This research work explored the effectiveness of deep-based pixel-wise segmentation models for low energy X-rays (mammographic imagery) to detect tumors in the breast region. For this purpose, various semantic segmentation models were incorporated into the experimental procedure. All the models were analyzed using the medical images dataset, which was gathered and annotated from one of the largest teaching hospitals in the Khyber Pakhtunkhwa province, known as Lady reading hospital. It is coordinated in cooperation with local health specialists, radiologists, and technologists. The comparative analysis of the incorporated segmentation techniques' performance was observed, selecting the most appropriate model for detecting tumors and normal breast regions. The experimental evaluation of the proposed models performs efficient detection of tumor and non-tumor areas in breast mammograms using traditional evaluation metrics such as mean IoU and Pixel accuracy. The performance of the semantic segmentation techniques was evaluated on two datasets (Cityscapes and mammogram). Dilation 10 (global) performed the best among the four semantic segmentation models by achieving a higher pixel accuracy of 93.69%. It reflects the effectiveness of the pixel-wise segmentation techniques by outperforming other state-of-the-art automatic image segmentation models.

Abstract P856: Contributors to Internal Primary Stroke Center Transfer Delays: Factors Beyond Stroke Coordinators’ Control

Background: Stroke coordinators often are challenged to transfer patients to higher levels of care, facing internal system, transport, and comprehensive stroke center (CSC) factors that may slow the transfer process. We aimed to understand the internal challenges faced by stroke coordinators at primary stroke centers/acute stroke ready hospitals (PSC/ASRH) in the transfer process. Methods: Six national focus groups with PSC/ASRH staff experienced with stroke transfers were conducted by web conference. Interviews were conducted using pre-scripted open-ended questions; information was recorded and data were transcribed for theme identification. Results: Participants were from Northeast, Mid-Atlantic, Southeast, Midwest, and West Coast USA. Internal PSC/ASRH problems were tied exclusively to physicians: Emergency department (ED) physicians’ lack knowledge of how/when to order advanced imaging; delays initiating telemedicine guidance were common; and, negative attitudes towards stroke emergencies among “tenured” ED physicians were identified. All participants noted that they were compliant with door-to-noncontrast CT and alteplase treatment time metrics, however, radiologists commonly used the full 2-hours allotted for CTA and CTP reads, and many telemedicine neurologists refused to review/interpret advanced imaging choosing to rely on the local general radiologists’ formal interpretation. Conclusions: Major factors contributing to internal PSC/ASRH transfer delays are exclusively physician-based. Physician education, standardized algorithms for advanced imaging selection, more-timely regulatory metrics for advanced imaging interpretation, and improved telemedicine physician expertise and guidance would improve transfer timeliness in these centers.