Machine Learning Classification of Body Part, Imaging Axis, and Intravenous Contrast Enhancement on CT Imaging.

Purpose: The development and evaluation of machine learning models that automatically identify the body part(s) imaged, axis of imaging, and the presence of intravenous contrast material of a CT series of images. Methods: This retrospective study included 6955 series from 1198 studies (501 female, 697 males, mean age 56.5 years) obtained between January 2010 and September 2021. Each series was annotated by a trained board-certified radiologist with labels consisting of 16 body parts, 3 imaging axes, and whether an intravenous contrast agent was used. The studies were randomly assigned to the training, validation and testing sets with a proportion of 70%, 20% and 10%, respectively, to develop a 3D deep neural network for each classification task. External validation was conducted with a total of 35,272 series from 7 publicly available datasets. The classification accuracy for each series was independently assessed for each task to evaluate model performance. Results: The accuracies for identifying the body parts, imaging axes, and the presence of intravenous contrast were 96.0% (95% CI: 94.6%, 97.2%), 99.2% (95% CI: 98.5%, 99.7%), and 97.5% (95% CI: 96.4%, 98.5%) respectively. The generalizability of the models was demonstrated through external validation with accuracies of 89.7 - 97.8%, 98.6 - 100%, and 87.8 - 98.6% for the same tasks. Conclusions: The developed models demonstrated high performance on both internal and external testing in identifying key aspects of a CT series.

Radiomic Features Based on MRI Predict Progression-Free Survival in Pediatric Diffuse Midline Glioma/Diffuse Intrinsic Pontine Glioma.

Purpose: Biopsy-based assessment of H3 K27 M status helps in predicting survival, but biopsy is usually limited to unusual presentations and clinical trials. We aimed to evaluate whether radiomics can serve as prognostic marker to stratify diffuse intrinsic pontine glioma (DIPG) subsets. Methods: In this retrospective study, diagnostic brain MRIs of children with DIPG were analyzed. Radiomic features were extracted from tumor segmentations and data were split into training/testing sets (80:20). A conditional survival forest model was applied to predict progression-free survival (PFS) using training data. The trained model was validated on the test data, and concordances were calculated for PFS. Experiments were repeated 100 times using randomized versions of the respective percentage of the training/test data. Results: A total of 89 patients were identified (48 females, 53.9%). Median age at time of diagnosis was 6.64 years (range: 1-16.9 years) and median PFS was 8 months (range: 1-84 months). Molecular data were available for 26 patients (29.2%) (1 wild type, 3 K27M-H3.1, 22 K27M-H3.3). Radiomic features of FLAIR and nonenhanced T1-weighted sequences were predictive of PFS. The best FLAIR radiomics model yielded a concordance of .87 [95% CI: .86-.88] at 4 months PFS. The best T1-weighted radiomics model yielded a concordance of .82 [95% CI: .8-.84] at 4 months PFS. The best combined FLAIR + T1-weighted radiomics model yielded a concordance of .74 [95% CI: .71-.77] at 3 months PFS. The predominant predictive radiomic feature matrix was gray-level size-zone. Conclusion: MRI-based radiomics may predict progression-free survival in pediatric diffuse midline glioma/diffuse intrinsic pontine glioma.

Post-mortem CT with macroscopic and microscopic correlation in a case of sudden death due to systemic sarcoidosis.

We report a case of sudden death due to systemic sarcoidosis in a fifty-four year old male who was reportedly healthy. A computerized tomography (CT) scan was performed before the autopsy. It showed cardiomegaly with hilar and abdominal lymphadenopathy. The autopsy showed pale yellow plaque deposition on the heart surface which was infiltrating the myocardium. Histological examination of the heart, lungs, liver, and spleen showed extensive sarcoid granulomata which helped in establishing the cause of death.

Effect of multifocality and multicentricity on outcome in early stage breast cancer: a systematic review and meta-analysis.

Women with multifocal or multicentric breast tumors (multifocality henceforth) have been reported to have greater probability of nodal metastasis and relapse and worse survival than women with unifocal tumors. However, these associations have been inconsistent and multifocality is not taken into account by staging guidelines and prognostic models. A systematic review of electronic databases identified publications exploring the association between multifocality and overall survival (OS), disease-free survival (DFS), disease-specific survival (DSS), and loco-regional relapse (LRR). The hazard ratios (HRs) for OS and DFS for multifocal compared to unifocal tumors were extracted from multivariable analyses and included in a meta-analysis. For studies not reporting multivariable analyses, odds ratios (OR) were estimated from Kaplan-Meier curves for all endpoints at 5 and 10 years. Twenty-two studies comprising 67,557 women were included. Multifocality was reported in 9.5 % of patients. Classical prognostic factors were well balanced between unifocal and multifocal populations. In multivariable analyses, multifocality was associated with significantly worse OS (HR 1.65; P = 0.02), and a non-significant association with worse DFS (HR 1.96; P = 0.07). In univariable analyses, multifocality was associated with worse OS, DFS, DSS, and LRR at 5 years (OR 1.39, P = 0.02; OR 1.52, P = 0.02; OR 1.56, P = 0.03; and OR 3.23, P = 0.02, respectively). Similar estimates were observed at 10 years, but statistical significance was only reached for DSS and LRR. Mutifocality appears to be associated with a worse prognosis, however, substantial inter-study heterogeneity limits the precise determination of increased risk. Further validation of the independent prognostic impact of multifocality is warranted.

Augmentation of the RSNA Pulmonary Embolism CT Dataset with Bounding Box Annotations and Anatomic Localization of Pulmonary Emboli.

Supplemental material is available for this article. Keywords: CT, Pulmonary Arteries, Embolism/Thrombosis, Feature Detection © RSNA, 2023.

Local tumor progression patterns after radiofrequency ablation of colorectal cancer liver metastases.

PURPOSE: We aimed to evaluate patterns of local tumor progression (LTP) after radiofrequency ablation (RF ablation) of colorectal cancer liver metastases (CRCLM) and to highlight the percentage of LTP not attributable to lesion size or RF ablation procedure-related factors (heat sink or insufficient ablation margin). METHODS: CRCLM treated by RF ablation at a single tertiary care center from 2004-2012, with a minimum of six months of postprocedure follow-up, were included in this retrospective study. LTP morphology was classified as focal nodular (<90° of ablation margin), circumferential (>270°), or crescentic (90°-270°). Initial metastasis size, minimum ablation margin size, morphology of LTP, presence of a heat sink, and time to progression were recorded independently by two radiologists. RESULTS: Thirty-two of 127 RF ablation treated metastases (25%) with a mean size of 23 mm (standard deviation 12 mm) exhibited LTP. Fifteen of 32 LTPs (47%) were classified as focal nodular, with seven having no procedure-related factor to explain recurrence. Ten of 32 LTPs (31%) were circumferential, with four having no procedure-related factor to explain recurrence. Seven of 32 LTPs (22%) were crescentic, with two having no procedure-related factor to explain recurrence. Of the 13 lesions without any obvious procedure-related reason for LTP, six (46%) were <3 cm in size. CONCLUSION: Although LTP in RF ablation treated CRCLM can often be explained by procedure-related factors or size of the lesion, in this study up to six (5%) of the CRCLM we treated showed LTP without any reasonable cause.

Honorary and ghost authorship in reports of randomised clinical trials in oncology.

BACKGROUND: The International Committee of Medical Journal Editors (ICMJE) has developed guidelines for responsible and accountable authorship. Few studies have assessed the frequency and nature of ghost and honorary authorship in publications of oncology trials. MATERIALS AND METHODS: Reports of randomised clinical trials evaluating systemic cancer therapy published from July 2010 to December 2012 in six high-impact journals were identified systematically. Ghost authorship was determined to be present in any scenario where investigators or statisticians listed in the protocol were not included as authors and not acknowledged in the report of the trial. The list of contributions for authors of published articles was recorded, and we defined an article as having an honorary author if any author did not meet all three criteria established by ICMJE in 1985. RESULTS: Two hundred publications were identified. For 61 articles, protocols with listed investigators were available, and 40 (66%) of these articles met our definition of ghost authorship. Medical writers were involved in 89 articles (45%), and assistance was acknowledged only in sponsored trials. Contributions of each author were provided in 195 articles, and 63 (33%) articles met our definition for honorary authorship. Funding source was not a predictor for either honorary or ghost authorship. Journal impact factor was positively associated with honorary authorship (odds ratio = 1.03; 95% confidence interval = 1.004-1.065; P = 0.03), but not with ghost authorship. CONCLUSION: Ghost and honorary authorship are prevalent in articles describing trials for systemic therapy of cancer. Guidelines should be enforced to improve transparency and accountability.

Bias in reporting of randomised clinical trials in oncology.

BACKGROUND: Bias in reporting efficacy and toxicity in clinical trials may impact treatment decisions. Here, we report quality of reporting of efficacy and of toxicity in articles describing randomised controlled trials (RCTs) of cancer therapy and the association between biased reporting and study results, funding and financial relationships of the authors with the sponsor. MATERIALS AND METHODS: We reviewed articles published from July 2010 to December 2012 in six high-impact journals reporting RCTs of systemic treatment for cancer. Bias in reporting of the primary end-point and toxicity were assessed. Associations between biased reporting and study results, funding source and financial ties of the author with the funding source were evaluated using logistic regression. RESULTS: Two hundred articles were identified. Among 107 RCTs where there was no statistically significant difference in the primary end-point between the two arms, 50 (47%) reports used biased reporting in the abstract of the paper to imply benefit of the experimental treatment. Toxicity was not reported in the abstract in 18.5% of the studies and this was associated with a positive primary end-point. Source of funding and financial ties were not associated with biased reporting. CONCLUSIONS: Bias in reporting of efficacy outcomes is common for studies with a negative primary end-point and can lead to off-label misuse of experimental therapies, if they are approved for other indications. Toxicity is under-reported, especially for studies with a positive primary end-point, leading to a biased view of the safety of new treatments.