Natural Language Processing Algorithm Used for Staging Pulmonary Oncology from Free-Text Radiological Reports: "Including PET-CT and Validation Towards Clinical Use".

Natural language processing (NLP) can be used to process and structure free text, such as (free text) radiological reports. In radiology, it is important that reports are complete and accurate for clinical staging of, for instance, pulmonary oncology. A computed tomography (CT) or positron emission tomography (PET)-CT scan is of great importance in tumor staging, and NLP may be of additional value to the radiological report when used in the staging process as it may be able to extract the T and N stage of the 8th tumor-node-metastasis (TNM) classification system. The purpose of this study is to evaluate a new TN algorithm (TN-PET-CT) by adding a layer of metabolic activity to an already existing rule-based NLP algorithm (TN-CT). This new TN-PET-CT algorithm is capable of staging chest CT examinations as well as PET-CT scans. The study design made it possible to perform a subgroup analysis to test the external validation of the prior TN-CT algorithm. For information extraction and matching, pyContextNLP, SpaCy, and regular expressions were used. Overall TN accuracy score of the TN-PET-CT algorithm was 0.73 and 0.62 in the training and validation set (N = 63, N = 100). The external validation of the TN-CT classifier (N = 65) was 0.72. Overall, it is possible to adjust the TN-CT algorithm into a TN-PET-CT algorithm. However, outcomes highly depend on the accuracy of the report, the used vocabulary, and its context to express, for example, uncertainty. This is true for both the adjusted PET-CT algorithm and for the CT algorithm when applied in another hospital.

Results of a diagnostic imaging audit in a randomised clinical trial in rectal cancer highlight the importance of careful planning and quality control.

BACKGROUND: Magnetic resonance (MR) imaging is the modality used for baseline assessment of locally advanced rectal cancer (LARC) and restaging after neoadjuvant treatment. The overall audited quality of MR imaging in large multicentre trials on rectal cancer is so far not routinely reported. MATERIALS AND METHODS: We collected MR images obtained within the Rectal Cancer And Pre-operative Induction Therapy Followed by Dedicated Operation (RAPIDO) trial and performed an audit of the technical features of image acquisition. The required MR sequences and slice thickness stated in the RAPIDO protocol were used as a reference. RESULTS: Out of 920 participants of the RAPIDO study, MR investigations of 668 and 623 patients in the baseline and restaging setting, respectively, were collected. Of these, 304/668 (45.5%) and 328/623 (52.6%) MR images, respectively, fulfilled the technical quality criteria. The main reason for non-compliance was exceeding slice thickness 238/668, 35.6% in the baseline setting and 162/623, 26.0% in the restaging setting. In 166/668, 24.9% and 168/623, 27.0% MR images in the baseline and restaging setting, respectively, one or more of the required pulse sequences were missing. CONCLUSION: Altogether, 49.0% of the MR images obtained within the RAPIDO trial fulfilled the image acquisition criteria required in the study protocol. High-quality MR imaging should be expected for the appropriate initial treatment and response evaluation of patients with LARC, and efforts should be made to maximise the quality of imaging in clinical trials and in clinical practice. CRITICAL RELEVANCE STATEMENT: This audit highlights the importance of adherence to MR image acquisition criteria for rectal cancer, both in multicentre trials and in daily clinical practice. High-resolution images allow correct staging, treatment stratification and evaluation of response to neoadjuvant treatment. KEY POINTS: - Complying to MR acquisition guidelines in multicentre trials is challenging. - Neglection on MR acquisition criteria leads to poor staging and treatment. - MR acquisition guidelines should be followed in trials and clinical practice. - Researchers should consider mandatory audits prior to study initiation.

Three-month follow-up after severe COVID-19 infection: are chest CT results associated with respiratory outcomes and respiratory recovery in COVID-19 patients?

BACKGROUND: CT Severity Score (CT-SS) can be used to assess the extent of severe coronavirus disease 19 (COVID-19) pneumonia. Follow-up CT-SS in patients surviving COVID-19-associated hyperinflammation and its correlation with respiratory parameters remains unknown. This study aims to assess the association between CT-SS and respiratory outcomes, both in hospital and at three months after hospitalization. METHODS: Patients from the COVID-19 High-intensity Immunosuppression in Cytokine storm Syndrome (CHIC) study surviving hospitalization due to COVID-19 associated hyperinflammation were invited for follow-up assessment at three months after hospitalization. Results of CT-SS three months after hospitalization were compared with CT-SS at hospital admission. CT-SS at admission and at 3-months were correlated with respiratory status during hospitalization and with patient reported outcomes as well as pulmonary- and exercise function tests at 3-months after hospitalization. RESULTS: A total of 113 patients were included. Mean CT-SS decreased by 40.4% (SD 27.6) in three months (P < 0.001). CT-SS during hospitalization was higher in patients requiring more oxygen (P < 0.001). CT-SS at 3-months was higher in patients with more dyspnoea (CT-SS 8.31 (3.98) in patients with modified Medical Council Dyspnoea scale (mMRC) 0-2 vs. 11.03 (4.47) in those with mMRC 3-4). CT-SS at 3-months was also higher in patients with a more impaired pulmonary function (7.4 (3.6) in patients with diffusing capacity for carbon monoxide (DLCO) > 80%pred vs. 14.3 (3.2) in those with DLCO < 40%pred, P = 0.002). CONCLUSION: Patients surviving hospitalization for COVID-19-associated hyperinflammation with higher CT-SS have worse respiratory outcome, both in-hospital and at 3-months after hospitalization. Strict monitoring of patients with high CT-SS is therefore warranted.

Systematic screening versus clinical gestalt in the diagnosis of pulmonary embolism in COVID-19 patients in the emergency department.

BACKGROUND: Diagnosing concomitant pulmonary embolism (PE) in COVID-19 patients remains challenging. As such, PE may be overlooked. We compared the diagnostic yield of systematic PE-screening based on the YEARS-algorithm to PE-screening based on clinical gestalt in emergency department (ED) patients with COVID-19. METHODS: We included all ED patients who were admitted because of COVID-19 between March 2020 and February 2021. Patients already receiving anticoagulant treatment were excluded. Up to April 7, 2020, the decision to perform CT-pulmonary angiography (CTPA) was based on physician's clinical gestalt (clinical gestalt cohort). From April 7 onwards, systematic PE-screening was performed by CTPA if D-dimer level was ≥1000 ug/L, or ≥500 ug/L in case of ≥1 YEARS-item (systematic screening cohort). RESULTS: 1095 ED patients with COVID-19 were admitted. After applying exclusion criteria, 289 were included in the clinical gestalt and 574 in the systematic screening cohort. The number of PE diagnoses was significantly higher in the systematic screening cohort compared to the clinical gestalt cohort: 8.2% vs. 1.0% (3/289 vs. 47/574; p<0.001), even after adjustment for differences in patient characteristics (adjusted OR 8.45 (95%CI 2.61-27.42, p<0.001) for PE diagnosis). In multivariate analysis, D-dimer (OR 1.09 per 1000 µg/L increase, 95%CI 1.06-1.13, p<0.001) and CRP >100 mg/L (OR 2.78, 95%CI 1.37-5.66, p = 0.005) were independently associated with PE. CONCLUSION: In ED patients with COVID-19, the number of PE diagnosis was significantly higher in the cohort that underwent systematic PE screening based on the YEARS-algorithm in comparison with the clinical gestalt cohort, with a number needed to test of 7.1 CTPAs to detect one PE.

Pulmonary embolism in hospitalized COVID-19 patients: Short- and long-term clinical outcomes.

Introduction: Pulmonary embolism (PE) is a frequent complication in COVID-19. However, the influence of PE on the prognosis of COVID-19 remains unclear as previous studies were affected by misclassification bias. Therefore, we evaluated a cohort of COVID-19 patients whom all underwent systematic screening for PE (thereby avoiding misclassification) and compared clinical outcomes between patients with and without PE. Materials and methods: We included all COVID-19 patients who were admitted through the ED between April 2020 and February 2021. All patients underwent systematic work-up for PE in the ED using the YEARS-algorithm. The primary outcome was a composite of in-hospital mortality and ICU admission. We also evaluated long-term outcomes including PE occurrence within 90 days after discharge and one-year all-cause mortality. Results: 637 ED patients were included in the analysis. PE was diagnosed in 46 of them (7.2%). The occurrence of the primary outcome did not differ between patients with PE and those without (28.3% vs. 26.9%, p = 0.68). The overall rate of PE diagnosed in-hospital (after an initial negative PE screening in the ED) and in the first 90 days after discharge was 3.9% and 1.2% respectively. One-year all-cause mortality was similar between patients with and without PE (26.1% vs. 24.4%, p = 0.83). Conclusions: In a cohort of COVID-19 patients who underwent systematic PE screening in the ED, we found no differences in mortality rate and ICU admissions between patients with and without PE. This may indicate that proactive PE screening, and thus timely diagnosis and treatment of PE, may limit further clinical deterioration and associated mortality in COVID-19 patients.

Automated Assessment of COVID-19 Reporting and Data System and Chest CT Severity Scores in Patients Suspected of Having COVID-19 Using Artificial Intelligence.

Background The coronavirus disease 2019 (COVID-19) pandemic has spread across the globe with alarming speed, morbidity, and mortality. Immediate triage of patients with chest infections suspected to be caused by COVID-19 using chest CT may be of assistance when results from definitive viral testing are delayed. Purpose To develop and validate an artificial intelligence (AI) system to score the likelihood and extent of pulmonary COVID-19 on chest CT scans using the COVID-19 Reporting and Data System (CO-RADS) and CT severity scoring systems. Materials and Methods The CO-RADS AI system consists of three deep-learning algorithms that automatically segment the five pulmonary lobes, assign a CO-RADS score for the suspicion of COVID-19, and assign a CT severity score for the degree of parenchymal involvement per lobe. This study retrospectively included patients who underwent a nonenhanced chest CT examination because of clinical suspicion of COVID-19 at two medical centers. The system was trained, validated, and tested with data from one of the centers. Data from the second center served as an external test set. Diagnostic performance and agreement with scores assigned by eight independent observers were measured using receiver operating characteristic analysis, linearly weighted κ values, and classification accuracy. Results A total of 105 patients (mean age, 62 years ± 16 [standard deviation]; 61 men) and 262 patients (mean age, 64 years ± 16; 154 men) were evaluated in the internal and external test sets, respectively. The system discriminated between patients with COVID-19 and those without COVID-19, with areas under the receiver operating characteristic curve of 0.95 (95% CI: 0.91, 0.98) and 0.88 (95% CI: 0.84, 0.93), for the internal and external test sets, respectively. Agreement with the eight human observers was moderate to substantial, with mean linearly weighted κ values of 0.60 ± 0.01 for CO-RADS scores and 0.54 ± 0.01 for CT severity scores. Conclusion With high diagnostic performance, the CO-RADS AI system correctly identified patients with COVID-19 using chest CT scans and assigned standardized CO-RADS and CT severity scores that demonstrated good agreement with findings from eight independent observers and generalized well to external data. © RSNA, 2020 Supplemental material is available for this article.

[A boy with a rare pulmonary disease; COVID-19 is clouding our judgment]. / Een jongen met een zeldzame longziekte.

BACKGROUND: In times of coronavirus, a patient with respiratory symptoms whose chest CT scan reveals ground-glass opacities, COVID-19 may seem an obvious diagnosis. CASE DESCRIPTION: At the (currently assumed) peak of the coronavirus crisis, a 12-year-old boy was admitted to the hospital twice for severe respiratory symptoms. A chest CT scan revealed ground-glass opacities.COVID-19 pneumonia was initially thought of. However, it turned out to be a rare interstitial pulmonary disease. CONCLUSION: Due to the increased awareness about COVID-19, tunnel vision is lurking. Even during a health crisis, doctors should remain alert to alternative diagnoses.

Diagnostic performance of chest CT in screening patients with suspected COVID-19 infection in a Western population.

OBJECTIVE: To investigate the diagnostic performance of chest CT in screening patients suspected of Coronavirus disease 2019 (COVID-19) in a Western population. METHODS: Consecutive patients who underwent chest CT because of clinical suspicion of COVID-19 were included. CT scans were prospectively evaluated by frontline general radiologists who were on duty at the time when the CT scan was performed and retrospectively assessed by a chest radiologist in an independent and blinded manner. Real-time reverse transcriptase-polymerase chain reaction was used as reference standard. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. Sensitivity and specificity of the frontline general radiologists were compared to those of the chest radiologist using the McNemar test. RESULTS: 56 patients were included. Sensitivity, specificity, PPV, and NPV for the frontline general radiologists were 89.3% [95% confidence interval (CI): 71.8%, 97.7%], 32.1% (95% CI: 15.9%, 52.4%), 56.8% (95% CI: 41.0%, 71.7%), and 75.0% (95% CI: 42.8%, 94.5%), respectively. Sensitivity, specificity, PPV, and NPV for the chest radiologist were 89.3% (95% CI: 71.8%, 97.7%), 75.0% (95% CI: 55.1%, 89.3%), 78.1% (95% CI: 60.0%, 90.7%), and 87.5% (95% CI: 67.6%, 97.3%), respectively. Sensitivity was not significantly different (p = 1.000), but specificity was significantly higher for the chest radiologist (p = 0.001). CONCLUSION: Chest CT interpreted by frontline general radiologists achieves insufficient screening performance. Although specificity of a chest radiologist appears to be significantly higher, sensitivity did not improve. A negative chest CT result does not exclude COVID-19. ADVANCES IN KNOWLEDGE: Our study shows that chest CT interpreted by frontline general radiologists achieves insufficient diagnostic performance to use it as an independent screening tool for COVID-19. Although specificity of a chest radiologist appears to be significantly higher, sensitivity is still insufficiently high.

CO-RADS: A Categorical CT Assessment Scheme for Patients Suspected of Having COVID-19-Definition and Evaluation.

Background A categorical CT assessment scheme for suspicion of pulmonary involvement of coronavirus disease 2019 (COVID-19 provides a basis for gathering scientific evidence and improved communication with referring physicians. Purpose To introduce the COVID-19 Reporting and Data System (CO-RADS) for use in the standardized assessment of pulmonary involvement of COVID-19 on unenhanced chest CT images and to report its initial interobserver agreement and performance. Materials and Methods The Dutch Radiological Society developed CO-RADS based on other efforts for standardization, such as the Lung Imaging Reporting and Data System or Breast Imaging Reporting and Data System. CO-RADS assesses the suspicion for pulmonary involvement of COVID-19 on a scale from 1 (very low) to 5 (very high). The system is meant to be used in patients with moderate to severe symptoms of COVID-19. The system was evaluated by using 105 chest CT scans of patients admitted to the hospital with clinical suspicion of COVID-19 and in whom reverse transcription-polymerase chain reaction (RT-PCR) was performed (mean, 62 years ± 16 [standard deviation]; 61 men, 53 with positive RT-PCR results). Eight observers used CO-RADS to assess the scans. Fleiss κ value was calculated, and scores of individual observers were compared with the median of the remaining seven observers. The resulting area under the receiver operating characteristics curve (AUC) was compared with results from RT-PCR and clinical diagnosis of COVID-19. Results There was absolute agreement among observers in 573 (68.2%) of 840 observations. Fleiss κ value was 0.47 (95% confidence interval [CI]: 0.45, 0.47), with the highest κ value for CO-RADS categories 1 (0.58, 95% CI: 0.54, 0.62) and 5 (0.68, 95% CI: 0.65, 0.72). The average AUC was 0.91 (95% CI: 0.85, 0.97) for predicting RT-PCR outcome and 0.95 (95% CI: 0.91, 0.99) for clinical diagnosis. The false-negative rate for CO-RADS 1 was nine of 161 cases (5.6%; 95% CI: 1.0%, 10%), and the false-positive rate for CO-RADS category 5 was one of 286 (0.3%; 95% CI: 0%, 1.0%). Conclusion The coronavirus disease 2019 (COVID-19) Reporting and Data System (CO-RADS) is a categorical assessment scheme for pulmonary involvement of COVID-19 at unenhanced chest CT that performs very well in predicting COVID-19 in patients with moderate to severe symptoms and has substantial interobserver agreement, especially for categories 1 and 5. © RSNA, 2020 Online supplemental material is available for this article.

MRI predicts increased eligibility for sphincter preservation after CRT in low rectal cancer.

Chemoradiation increases the eligibility for sphincter preservation in low rectal cancer, as assessed by MRI. INTRODUCTION: We evaluated whether MRI can predict sphincter preservation after chemoradiation (CRT), and whether the feasibility of sphincter preservation increases after CRT, when compared with MRI before neoadjuvant treatment. METHODS: 85 patients with low rectal tumour (≤5 cm from anorectal junction (ARJ)) were included. Radiologist and a surgeon measured the tumour distance to ARJ, and assigned confidence level scores (CLS) for the feasibility of sphincter preserving surgery on MRI. Reference standard was the type of surgery, sphincter preserving vs. non-preserving. RESULTS: Tumour distance from the ARJ increased after CRT by 9 mm (p < 0.001). Eligibility for sphincter preservation increased by 21% for the radiologist and 25% for the surgeon, based on CLS. Cut-off for distance to the ARJ after CRT was 28 mm, aiming for optimal specificity. Diagnostic performance after CRT based on CLS yielded an AUC of 0.81 [95%CI 0.70-0.91] for the radiologist and 0.82 [95%CI 0.72-0.92] for the surgeon (p = 0.78). AUCs for tumour distance to the ARJ were 0.85 [95%CI 0.77-0.94] and 0.84 [95%CI 0.75-0.94], respectively (p = 0.84). Interobserver agreement for CLS was moderate before CRT (Κ 0.51; 95%CI 0.36-0.66) and after (K 0.54; 95%CI 0.39-0.69). Measurement of tumour distance to ARJ showed good agreement before (ICC 0.76; 95%CI 0.65-0.84) and after CRT (ICC 0.77; 95%CI 0.66-0.84). CONCLUSION: MRI can be a valuable adjunct in the decision making for sphincter preservation after CRT, with distance from the tumour to the ARJ as an accurate and reliable factor. CRT increases the tumour distance to the ARJ, leading to an estimated increase of sphincter preserving surgery in up to 21-25% of patients.

Radiological Society of North America Chest CT Classification System for Reporting COVID-19 Pneumonia: Interobserver Variability and Correlation with Reverse-Transcription Polymerase Chain Reaction.

PURPOSE: To evaluate the Radiological Society of North America (RSNA) chest CT classification system for reporting coronavirus disease 2019 (COVID-19) pneumonia. MATERIALS AND METHODS: Chest CT scans of consecutive patients suspected of having COVID-19 were retrospectively and independently evaluated by two chest radiologists and a 5th-year radiology resident using the RSNA chest CT classification system for reporting COVID-19 pneumonia. Interobserver agreement was evaluated by calculating weighted κ coefficients. The proportion of patients with real-time reverse-transcription polymerase chain reaction (RT-PCR)-confirmed COVID-19 in each of the four chest CT categories (typical, indeterminate, atypical, and negative features for COVID-19) was calculated. RESULTS: In total, 96 patients (61 men; median age, 70 years [range, 29-94]) were included, of whom 45 had RT-PCR-confirmed COVID-19. The number of patients assigned to chest CT categories typical, indeterminate, atypical, and negative by the three readers ranged from 18 to 29, 26 to 43, 19 to 31, and 5 to 8, respectively. The κ coefficient among the chest radiologists was 0.663 (95% confidence interval [CI]: 0.565, 0.761). κ coefficients among the chest radiologists and the 5th-year radiology resident were 0.570 (95% CI: 0.443, 0.696) and 0.564 (95% CI: 0.451, 0.678), respectively. The proportion of patients with RT-PCR-confirmed COVID-19 in the chest CT categories typical, indeterminate, atypical, and negative for the three readers ranged from 76.9% to 96.6%, 51.2% to 64.1%, 2.8% to 5.3%, and 20% to 25%, respectively. CONCLUSION: The RSNA chest CT classification system for reporting COVID-19 pneumonia has moderate-to-substantial interobserver agreement. However, the proportion of RT-PCR-confirmed COVID-19 cases in the categories atypical appearance and negative for pneumonia is nonnegligible.Supplemental material is available for this article.© RSNA, 2020.