Scott SimpsonpersName, Fernando Uliana KaypersName, Suhny AbbarapersName, Sanjeev BhallapersName, Jonathan H. ChungpersName, Michael ChungpersName, Travis S. HenrypersName, Jeffrey P. KannepersName, Seth KligermanpersName, Jane P. KopersName, Harold LittpersName (2020date). “Radiological Society of North America Expert Consensus Document on Reporting Chest CT Findings Related to COVID-19.” Radiology: Cardiothoracic Imagingtitle·venue.
Object of study — A REPORTING VOCABULARY WRITTEN IN TWO DAYS BECAUSE THE WORDS DID NOT AGREE - four named categories standing between a chest CT and a diagnosis, with the consequences of each word stated in the same document
Research question — what language radiologists should use when a chest CT shows findings that could be COVID-19term, given that naming the disease sets off a cascade and not naming it may miss it
Contribution — four reporting categories - typicalterm·def, indeterminateterm·def, atypicalterm·def, negative - each with standardized language and a unique coding identifier, plus an explicit pros/cons/limitations account of adopting them
phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | a word in a report acting on the patient before any treatment does — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | |
| claim-type | method — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | |
| method — evidence | — “Imaging appearances in the standardized reporting language are based upon available literature at the time of writing in March 2020, noting the retrospective nature of many reports, including biases related to patient selection in cohort studies, examination timing, and other potential confounders. As radiologists’ experience with COVID-19 increases, our categorization of these findings as typical, indeterminate, or atypical may evolve.” | |
| object of study — evidence | — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | |
| research question — evidence | — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | |
| contribution — evidence | — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | |
| mechanism | THE WORD ITSELF ACTS ON THE PATIENT: `including `COVID-19` in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient`, and `Ordering providers may also feel that having `COVID-19` or `coronavirus` documented in a radiology report constrains their clinical decision making and treatment options`. The shortest path in this corpus from which word to what happens — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | mechanisms[0] |
| mechanism | SO A DELIBERATELY BROADER TERM IS RECOMMENDED: `These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with `VIRAL PNEUMONIA` as a reasonable and inclusive alternative` - a word chosen for its consequences rather than its precision — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | mechanisms[1] |
| mechanism | AND THE COST OF THE BROADER TERM IS NAMED IN THE SAME DOCUMENT: `the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19`. Nothing else in this corpus recommends an imprecise term and states the imprecision in the next paragraph — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | mechanisms[2] |
| mechanism | AND THE VOCABULARY REFUSES TO CARRY A PROBABILITY, ON PURPOSE: `The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on ... the typicality of these features`. Four categories that are statements about THE IMAGE and explicitly not about the patient's probability - and they ship with `unique coding identifiers in brackets that can then be used for future data mining`, which is this project's own move made in March 2020 under pressure — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | mechanisms[3] |
| mechanism | AND EVERY NUMBER AROUND IT IS A RANGE, FOR THE RIFLE REASON: `The reported sensitivities and specificities of CT for COVID-19 vary widely (60%-98% and 25%-53%, respectively) ... likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection` — “The reported sensitivities and specificities of CT for COVID-19 vary widely (60%–98% and 25%–53%, respectively) ( 26 – 29 ), likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection.” | mechanisms[4] |
| extends | the ACR Practice Parameter for Communication of Diagnostic Imaging Findings, which the document repeatedly defers to for how findings are communicated — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | claim_relations[0] |
| contrasts-with | the professional guidance recommending against screening CT for COVID-19, which this document accepts while writing the language for the reports that happen anyway — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | claim_relations[1] |
| sense | `typical` - NOT a probability but a statement about the literature: features `reported in the literature to be frequently and more specifically seen in COVID-19 pneumonia in the current pandemic`. The category is indexed to a moment — “Typical features are those that are reported in the literature to be frequently and more specifically seen in COVID-19 pneumonia in the current pandemic” | senses[0] |
| sense | `indeterminate` - features `REPORTED IN COVID-19 PNEUMONIA BUT NOT SPECIFIC ENOUGH TO ARRIVE AT A relatively confident radiologic diagnosis`. A named slot for `the image cannot decide` — “Indeterminate features are those that have been reported in COVID-19 pneumonia but are not specific enough to arrive at a relatively confident radiologic diagnosis. An example would be diffuse GGO without a clear distribution ( Figs 5 , 6 ). This finding is common in COVID-19 pneumonia but occurs in a wide variety of diseases such as acute hypersensitivity pneumonitis, Pneumocystis infection, and diffuse alveolar hemorrhage, which are difficult to distinguish by imaging alone.” | senses[1] |
| sense | `atypical` - `uncommon or not occurring in COVID-19 pneumonia AND MORE TYPICAL OF OTHER DISEASES`, and the paper prints a positive-RT-PCR case that falls in it: the category is about the image, not the patient — “Atypical features are those that are reported to be uncommon or not occurring in COVID-19 pneumonia and are more typical of other diseases such as lobar or segmental consolidation in the setting of a bacterial pneumonia, cavitation from necrotizing pneumonia, and tree-in-bud opacities with centrilobular nodules, as can occur with a variety of community acquired infections and aspiration” | senses[2] |
| sense | `negative for pneumonia` - `no parenchymal abnormalities that could be attributable to infection. Specifically, GGO and consolidation are absent` - and immediately: `there may be no findings on CT early in COVID-19` — “Negative for pneumonia implies that there are no parenchymal abnormalities that could be attributable to infection. Specifically, GGO and consolidation are absent. Importantly, there may be no findings on CT early in COVID-19.” | senses[3] |
| sense | `viral pneumonia` (as recommended here) - the DELIBERATELY BROADER term, `a reasonable and inclusive alternative` whose breadth is stated as a cost: it `encompasses a range of imaging findings some of which are not typical for COVID-19` — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | senses[4] |
| control | four named categories with fixed language and unique coding identifiers — held — “Typical features are those that are reported in the literature to be frequently and more specifically seen in COVID-19 pneumonia in the current pandemic” — rules out reporting variability between radiologists facing the same image - the stated aim, 'decrease reporting variability, reduce uncertainty in reporting findings' — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | controls[0] |
| control | a likelihood or probability attached to each category — absent-declared-unreported — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” — rules out nothing - deliberately absent: 'The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation' — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | controls[1] |
| control | a reference standard against which the categories could be scored — failed — “As an example, in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs” — rules out nothing dependable - RT-PCR sensitivity is '42% TO 71%', the collection method halves the yield, and 'Eleven different RT-PCR tests' were in use with varying sensitivity — “Early reports of RT-PCR sensitivity vary considerably, ranging from 42% to 71% ( 26 , 27 ), and an initially negative RT-PCR may take up to 4 days to convert in a patient with COVID-19” | controls[2] |
| control | a category for images carrying both typical and atypical features — absent-declared-unreported — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” — rules out nothing - the case is anticipated and handed to the reader: 'over 20% of patients with COVID-19 may have coexistent infections complicating the categorization ... The radiologist will have to determine' — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” | controls[3] |
| weight | the ACR Practice Parameter for Communication of Diagnostic Imaging Findings, which the document repeatedly defers to for how findings are communicated — 0.8 — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | claim_relations[0] |
| weight | the professional guidance recommending against screening CT for COVID-19, which this document accepts while writing the language for the reports that happen anyway — 0.85 — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | claim_relations[1] |
| audit | COVID-19 —causes→ cascade of events — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says including the TERM 'COVID-19' in a report COULD trigger a cascade; causes is loose. — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | concept_edges[0] |
| note | COVID-19 —causes→ cascade of events — `As a result, including `COVID-19` in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.` The cause is the WORD IN THE REPORT, not the virus - which is why the same document then recommends a different word — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | concept_edges[0] |
| weight | COVID-19 —causes→ cascade of events — 0.95 — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | concept_edges[0] |
| audit | atypical features —causes→ false-negative cases — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says reporting atypical features MAY result in false negatives; hedged. — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | concept_edges[1] |
| note | atypical features —causes→ false-negative cases — `Reporting `ATYPICAL FEATURES` may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.` Weight 0.9 because the paper says `may result`, and the tag should not be more certain than the sentence — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | concept_edges[1] |
| weight | atypical features —causes→ false-negative cases — 0.9 — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | concept_edges[1] |
| audit | COVID-19 pneumonia —is-a→ viral pneumonia — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span offers 'viral pneumonia' as a reasonable alternative LABEL; is-a is loose. — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | concept_edges[2] |
| note | COVID-19 pneumonia —is-a→ viral pneumonia — APPLIED TO MY OWN TAG. The cue is KINDHOOD 'encompasses', in 'the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19' - and its object is 'a range of imaging findings', NOT 'COVID-19'. The relation is real, the passage supports it, and the cue's arguments are one step from the edge's. That is the exact fault I spent today measuring in other tags, so this carries 0.85 and this note rather than 0.95 and silence — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | concept_edges[2] |
| weight | COVID-19 pneumonia —is-a→ viral pneumonia — 0.85 — “These incidental findings do not necessarily need to be reported as COVID-19 pneumonia, with “viral pneumonia” as a reasonable and inclusive alternative.” | concept_edges[2] |
| outcome | control — held — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” — “Typical features are those that are reported in the literature to be frequently and more specifically seen in COVID-19 pneumonia in the current pandemic” | controls[0] |
| outcome | control — absent-declared-unreported — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | controls[1] |
| outcome | control — failed — “Early reports of RT-PCR sensitivity vary considerably, ranging from 42% to 71% ( 26 , 27 ), and an initially negative RT-PCR may take up to 4 days to convert in a patient with COVID-19” — “As an example, in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs” | controls[2] |
| outcome | control — absent-declared-unreported — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” | controls[3] |
| limitation | THE REFERENCE STANDARD IS ITSELF A RANGE: `Early reports of RT-PCR sensitivity vary considerably, RANGING FROM 42% TO 71%, and an initially negative RT-PCR may take up to 4 DAYS TO CONVERT in a patient with COVID-19`. The thing the imaging is scored against is between two-fifths and seven-tenths right — “Early reports of RT-PCR sensitivity vary considerably, ranging from 42% to 71% ( 26 , 27 ), and an initially negative RT-PCR may take up to 4 days to convert in a patient with COVID-19” | limitations[0] |
| limitation | AND THE COLLECTION METHOD HALVES THE ANSWER: `in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs`, while `Eleven different RT-PCR tests were approved for use in China ... with varying levels of sensitivity` — “As an example, in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs” | limitations[1] |
| limitation | AND A NEGATIVE IS A CLAIM ABOUT THE INSTRUMENT AND THE TIMING, NOT THE PATIENT: `A slight majority of patients had a negative CT during the first 2 days after symptom onset ... Therefore, a negative CT should not be used to exclude the possibility of COVID-19, particularly early in the disease` — “A slight majority of patients had a negative CT during the first 2 days after symptom onset with GGO usually developing between day 0 and 4 after symptom onset and peaking at 6–13 days ( 8 , 9 , 22 – 24 ). Therefore, a negative CT should not be used to exclude the possibility of COVID-19, particularly early in the disease.” | limitations[2] |
| limitation | AND THE CATEGORY SYSTEM'S OWN COST IS PRINTED: `Reporting `ATYPICAL FEATURES` may result in false-negative cases, and the risk of missing COVID-19 can have broad implications` — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | limitations[3] |
| limitation | AND THE CATEGORIES MEET CASES THEY CANNOT HOLD: `We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features ... OVER 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations. The radiologist will have to determine whether or not these findings are part of the same process or are unrelated` — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” | limitations[4] |
| limitation | AND THE VOCABULARY DECLARES ITSELF PROVISIONAL: `Imaging appearances ... are based upon available literature at the time of writing in March 2020, noting the retrospective nature of many reports, including biases related to patient selection ... As radiologists' experience with COVID-19 increases, our categorization of these findings as typical, indeterminate, or atypical may evolve` — “Imaging appearances in the standardized reporting language are based upon available literature at the time of writing in March 2020, noting the retrospective nature of many reports, including biases related to patient selection in cohort studies, examination timing, and other potential confounders. As radiologists’ experience with COVID-19 increases, our categorization of these findings as typical, indeterminate, or atypical may evolve.” | limitations[5] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | content_tags.findings[0] |
| finding | — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | content_tags.findings[1] |
| finding | — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumonia, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumonia rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.” | content_tags.findings[2] |
| finding | — “The reported sensitivities and specificities of CT for COVID-19 vary widely (60%–98% and 25%–53%, respectively) ( 26 – 29 ), likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection.” | content_tags.findings[3] |
| finding | — “As an example, in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs” | content_tags.findings[4] |
| finding | — “A slight majority of patients had a negative CT during the first 2 days after symptom onset with GGO usually developing between day 0 and 4 after symptom onset and peaking at 6–13 days ( 8 , 9 , 22 – 24 ). Therefore, a negative CT should not be used to exclude the possibility of COVID-19, particularly early in the disease.” | content_tags.findings[5] |
| finding | — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | content_tags.findings[6] |
| finding | — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” | content_tags.findings[7] |
| finding | — “Imaging appearances in the standardized reporting language are based upon available literature at the time of writing in March 2020, noting the retrospective nature of many reports, including biases related to patient selection in cohort studies, examination timing, and other potential confounders. As radiologists’ experience with COVID-19 increases, our categorization of these findings as typical, indeterminate, or atypical may evolve.” | content_tags.findings[8] |
| quantity | 60-98 % sensitivity — the reported range for chest CT in COVID-19, attributed by the authors to the lack of strict diagnostic criteria for imaging — “The reported sensitivities and specificities of CT for COVID-19 vary widely (60%–98% and 25%–53%, respectively) ( 26 – 29 ), likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection.” | content_tags.quantities[0] |
| quantity | 25-53 % specificity — the matching specificity range for the same instrument — “The reported sensitivities and specificities of CT for COVID-19 vary widely (60%–98% and 25%–53%, respectively) ( 26 – 29 ), likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection.” | content_tags.quantities[1] |
| quantity | 42-71 % sensitivity — the reported range for RT-PCR - the reference standard the imaging is scored against — “Early reports of RT-PCR sensitivity vary considerably, ranging from 42% to 71% ( 26 , 27 ), and an initially negative RT-PCR may take up to 4 days to convert in a patient with COVID-19” | content_tags.quantities[2] |
| quantity | 11 RT-PCR tests — the number approved for use in China between 26 January and 12 March 2020, with varying levels of sensitivity — “Eleven different RT-PCR tests were approved for use in China between January 26 and March 12, 2020, with varying levels of sensitivity.” | content_tags.quantities[3] |
| quantity | 20 % — the share of COVID-19 patients who may have coexistent infections, complicating which of the four categories an image belongs to — “We anticipate cases with mixed imaging findings, that is, those that have both typical and atypical imaging features for COVID-19. Recent analysis suggests that over 20% of patients with COVID-19 may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated.” | content_tags.quantities[4] |
| variable | viral pneumonia — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | content_tags.variables[0] |
| definition | typical — “Typical features are those that are reported in the literature to be frequently and more specifically seen in COVID-19 pneumonia in the current pandemic” | content_tags.definitions[0] |
| definition | indeterminate — “Indeterminate features are those that have been reported in COVID-19 pneumonia but are not specific enough to arrive at a relatively confident radiologic diagnosis. An example would be diffuse GGO without a clear distribution ( Figs 5 , 6 ). This finding is common in COVID-19 pneumonia but occurs in a wide variety of diseases such as acute hypersensitivity pneumonitis, Pneumocystis infection, and diffuse alveolar hemorrhage, which are difficult to distinguish by imaging alone.” | content_tags.definitions[1] |
| definition | atypical — “Atypical features are those that are reported to be uncommon or not occurring in COVID-19 pneumonia and are more typical of other diseases such as lobar or segmental consolidation in the setting of a bacterial pneumonia, cavitation from necrotizing pneumonia, and tree-in-bud opacities with centrilobular nodules, as can occur with a variety of community acquired infections and aspiration” | content_tags.definitions[2] |
| definition | negative for pneumonia — “Negative for pneumonia implies that there are no parenchymal abnormalities that could be attributable to infection. Specifically, GGO and consolidation are absent. Importantly, there may be no findings on CT early in COVID-19.” | content_tags.definitions[3] |
| concept | COVID-19 — not defined in this paper's own words; the name whose appearance in a report sets off infection control and anxiety — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | concepts[0] |
| concept | cascade of events — not defined in this paper's own words; what the word triggers - infection control measures and anxiety for provider and patient — “As a result, including “COVID-19” in a radiology report could trigger a cascade of events including infection control measures and anxiety for both the managing provider and the patient.” | concepts[1] |
| concept | viral pneumonia — not defined in this paper's own words; the broader term recommended in its place, whose breadth the paper states as a cost — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | concepts[2] |
| concept | COVID-19 pneumonia — not defined in this paper's own words; the narrower thing the broader term includes without being specific to — “It should be noted that viral pneumonias have a wide variety of imaging manifestations, some of which are atypical or less common in COVID-19 such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumonia encompasses a range of imaging findings some of which are not typical for COVID-19.” | concepts[3] |
| concept | atypical features — not defined in this paper's own words; the reporting category whose use the paper says may produce false negatives — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | concepts[4] |
| concept | false-negative cases — not defined in this paper's own words; the stated cost of the atypical category — “Reporting “atypical features” may result in false-negative cases, and the risk of missing COVID-19 can have broad implications.” | concepts[5] |
| claim type | method — a second reading reaching the same value; recorded: method — “Conclusions We propose four categories for the suggested standardized CT reporting language of COVID-19 based on current literature and expert consensus.” | alternatives.claim_type |
| method | expert-consensus — a second reading DIFFERING from the recorded value; recorded: expert-consensus — “We propose four categories for reporting CT imaging findings potentially attributable to COVID-19, each with suggested standardized language ( Table 1 ).” | alternatives.method |
| phenomenon | standardized language for reporting chest CT findings related to COVID-19 — a second reading DIFFERING from the recorded value; recorded: a word in a report acting on the patient before any treatment does — “We aim to provide guidance to radiologists in reporting CT findings potentially attributable to COVID-19 pneumonia, including standardized language to reduce reporting variability when addressing the possibility of COVID-19.” | alternatives.phenomenon |
Full text — tagged
The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.
4009 radcardimg Radiology: Cardiothoracic Imaging Radiol Cardiothorac Imaging Radiological Society of North America PMC7233447 7233447 7977964 33778571 10.1148/ryct.2020200152 Radiological Society of North America Expert Consensusterm·mth Document on Reporting Chest CT Findings Related to COVID-19term: Endorsed by the Society of Thoracic Radiology, the American College of Radiology, and RSNA SimpsonpersName Scott DO 1 KaypersName Fernando U MD, PhD 1 AbbarapersName Suhny MD 1 BhallapersName Sanjeev MD 1 ChungpersName Jonathan H MD 1 ChungpersName Michael MD 1 HenrypersName Travis S MD 1 KannepersName Jeffrey P MD 1 KligermanpersName Seth MD 1 Ko JanepersName P MD 1 LittpersName Harold MD, PhD 1 ✉ 1 From the Department of Radiology, Perelman School of Medicine of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104 (S.S., H.L.); Department of Radiology, University of Texas Southwestern Medical Center, Dallas, Tex (F.U.K., S.A.); Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Mo (S.B.); Department of Radiology, University of Chicago Medicine, Chicago, Ill (J.H.C.); Department of Diagnostic, Molecular and Interventional Radiology, Icahn School of Medicine at Mount Sinai, New York, NY (M.C.); Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, Calif (T.S.H.); Department of Radiology, University of Wisconsin School of Medicine and Public Health, Madison, Wis (J.P.K.); Department of Radiology, University of California San Diego, San Diego, Calif (S.K.); and Department of Radiology, New York University Langone Health, New York, NY (J.P.K.). ✉ Address correspondence to H.L. (e-mail: harold.littpersName@uphs.upenn.edu ).
✉ Corresponding author.
25 3 2020 2 2 e200152 e200152 2 6 2020 2020 by the Radiological Society of North America, Inc. Published under a http://creativecommons.org/licenses/by/4.0/ CC BY 4.0 license. Abstract Routine screening CT for the identification of coronavirus disease 19 (COVID-19term) pneumonia is currently not recommended by most radiology societies. However, the number of CT examinations performed in persons under investigation for COVID-19term has increased. We also anticipate that some patients will have incidentally detected findings that could be attributable to COVID-19 pneumoniaterm, requiring radiologists to decide whether or not to mention COVID-19term specifically as a differential diagnostic possibility. We aim to provide guidance to radiologists in reporting CT findings potentially attributable to COVID-19 pneumoniaterm, including standardized language to reduce reporting variability when addressing the possibility of COVID-19term. When typicalterm·def or indeterminateterm·def features of COVID-19 pneumoniaterm are present in endemic areas as an incidental finding, we recommend contacting the referring providers to discuss the likelihood of viral infection. These incidental findings do not necessarily need to be reported as COVID-19 pneumoniaterm. In this setting, using the term viral pneumoniaterm can be a reasonable and inclusive alternative. However, if one opts to use the term COVID-19term in the incidental setting, consider the provided standardized reporting language. In addition, practice patterns may vary, and this document is meant to serve as a guide. Consultation with clinical colleagues at each institution is suggested to establish a consensus reporting approach. The goal of this expert consensusterm·mth is to help radiologists recognize findings of COVID-19 pneumoniaterm and aid their communication with other health care providers, assisting management of patients during this pandemic.
Published under a CC BY 4.0 license.
status released display-pdf yes is-olf no is-manuscript no is-preprint no is-journal-matter no is-scanned no is-retracted no Received 2020 Mar 23; Revision requested 2020 Mar 25; Accepted 2020 Mar 25; Collection date 2020 Apr.
Introduction Coronavirus disease 2019 (COVID-19term) ( 1 ), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) ( 2 ), has become increasingly prevalent worldwide, reaching a pandemic stage in March 2020 ( 3 ). While most radiology professional organizations and societies have recommended against performing screening CT for the identification of COVID-19term ( 4 , 5 ), the number of CT examinations performed in persons under investigation (PUI) for COVID-19term may increase. We also anticipate that patients will have incidental lung findings on CT obtained for unrelated reasons that could be attributable to COVID-19term.
Several recent publications have described CT imaging features of COVID-19term, the evolution of these features over time, and the performance of radiologists in distinguishing COVID-19term from other viral infections ( 6 – 10 ). These studies have shown that COVID-19term often produces a CT pattern resembling organizing pneumonia, notably peripheral ground-glass opacities (GGOs) and nodular or masslike GGO that are often bilateral and multilobar ( 11 ). However, additional imaging findings have also been reported including linear, curvilinear, or perilobular opacities, consolidation, and diffuse GGO, which can mimic several disease processes including other infections, inhalational exposures, and drug toxicities ( 12 – 15 ).
COVID-19 pneumoniaterm has a high mortality rate in some populations, including the elderly and those with diabetes, hypertension, and other comorbidities ( 16 – 18 ), and is spreading rapidly and sustainably in the community ( 19 ). As a result, including “COVID-19term” in a radiology report could trigger a cascade of eventsterm including infection control measures and anxiety for both the managing provider and the patient. This potentially can complicate interpretations, as CT imaging features can overlap significantly with other causes of acute lung injury and organizing pneumonia ( 20 ). Standardized COVID-19term reporting language will improve communication with referring providers and has the potential to enhance efficiency and aid in management of patients during this pandemic.
This document aims to provide guidance to radiologists reporting CT findings potentially attributable to COVID-19 pneumoniaterm in both PUI and when discovered incidentally. The potential role of CT in COVID-19term; parameters for structured reporting; and the pros, cons, and limitations of adopting this strategy are discussed. In addition, practice patterns may vary by institution, and this document is meant to serve as a guide. If a radiologist, in accordance with one’s respective institutional procedures, chooses to mention COVID-19term specifically in CT reports, this standard framework can be adopted accordingly. Consultation with clinical colleagues at each institution is suggested to establish an agreed upon approach, which may evolve over time and be dependent upon the prevalence of the disease in the local population and other factors.
Chest CT in COVID-19term Infection CT Imaging Features Several papers have found that COVID-19term typically presents with GGO with or without consolidation in a peripheral, posterior, and diffuse or lower lung zone distribution ( 6 – 11 ). GGO has also been frequently reported to have round morphology or a “crazy paving” pattern ( 6 , 8 ). However, a significant portion of cases have opacities without a clear or specific distribution ( 8 ). A predominant perihilar pattern was not reported ( 8 ). Bronchial wall thickening, mucoid impactions, and nodules (“tree-in-bud” and centrilobular) seen commonly in infections, are not typically observed ( 8 ). Lymphadenopathy and pleural effusion have been rarely reported ( 6 , 21 ).
The frequency of imaging findings also depends on when infected patients are imaged. A slight majority of patients had a negative CT during the first 2 days after symptom onset with GGO usually developing between day 0 and 4 after symptom onset and peaking at 6–13 days ( 8 , 9 , 22 – 24 ). Therefore, a negative CT should not be used to exclude the possibility of COVID-19term, particularly early in the disease. Later in the course of the disease, the frequency of consolidation increases as does the likelihood of seeing a reverse halo or atoll sign, typically absent near the time of symptom onset ( 8 ). Available evidence regarding these CT findings is limited, and new patterns of pulmonary involvement may eventually be reported ( 25 ).
Diagnostic CT Performance and Screening Chest CT findings can precede positivity on reverse-transcription polymerase chain reaction testing (RT-PCR). Early reports of RT-PCR sensitivity vary considerably, ranging from 42% to 71% ( 26 , 27 ), and an initially negative RT-PCR may take up to 4 days to convert in a patient with COVID-19term ( 26 ). The reported sensitivities and specificities of CT for COVID-19term vary widely (60%–98% and 25%–53%, respectively) ( 26 – 29 ), likely due to the retrospective nature of the currently published studies, including lack of strict diagnostic criteria for imaging and procedural differences for confirming infection. The positive and negative predictive value of chest CT for COVID-19term are estimated at 92% and 42%, respectively, in a population with high pretest probability for the disease (eg, 85% prevalence by RT-PCR) ( 27 ). The relatively low negative predictive value suggests that CT may not be valuable as a screening test for COVID-19term at least in earlier stages of the disease.
Literature comparing individual CT features of COVID-19term or radiologists’ performance in correctly choosing COVID-19term as a first-choice diagnosis on imaging is limited. In one study, six of seven radiologists demonstrated 93%–100% specificity in correctly distinguishing CT features of COVID-19term from other viral infections ( 10 ). A peripheral distribution of GGO was found to correctly distinguish COVID-19term from other viral causes 63%–80% of the time. However, the authors did not include high numbers of influenza-A or any noninfectious causes such as drug reaction, which could degrade radiologists’ performance.
Viral Testing—Implications for CT In reviewing CT publications on COVID-19term, it is important to consider the accuracy of the laboratory viral testing used. This applies both to the collection method and the laboratory testing method ( 30 ), as many articles published on chest CT do not specify the sample collection or RT-PCR method used ( 31 ). With respect to collection method, bronchoalveolar lavage fluid (BALF) testing is the most sensitive, but not for general use given the invasive nature of fluid collection, and because it is an aerosol-generating procedure that could place health care workers at greater risk. Sputum and nasopharyngeal swab collection are considered equivalent in sensitivity, while throat swab testing is less sensitive. As viral pneumoniasterm typically do not result in production of purulent sputum, nasopharyngeal swab is the preferred method for sample collection ( 31 ). As an example, in a recently published series of 1070 patients, the majority of samples collected were throat swabs, and throat swabs detected only approximately half of the positive cases that were detected by nasal swabs ( 32 ).
Rapid antigen tests are fast but have poor sensitivity. While RT-PCR is the most accurate, not all tests are equivalent. Eleven different RT-PCR tests were approved for use in China between January 26 and March 12, 2020, with varying levels of sensitivity. In a report of CT findings in 1014 patients ( 26 ), with 59% having a positive RT-PCR and 88% having a positive chest CT, the method of swab collection was not described. Two different RT-PCR tests were used, one of which does not appear on the list of approved tests, and the other approved for use in nasal, throat and sputum collection. The sensitivity of tests approved for use in the United States is high, with emergency use authorizations available on the website of the U.S. Food and Drug Administration ( 33 ).
Structured Reporting Rationale and use.— The goal of structured reporting in the setting of COVID-19 pneumoniaterm is to help radiologists recognize the findings seen, decrease reporting variability, reduce uncertainty in reporting findings potentially attributable to this infection, and enhance the referring provider’s understanding of those radiologic findings, thereby allowing better integration into clinical decision making. While we do not currently recommend the use of CT screening for COVID-19 pneumoniaterm, we suggest using a standardized language when specifically asked to address whether or not findings of COVID-19 pneumoniaterm may be present on CT and propose language that could be placed in the impression of the report.
How to report incidentally discovered features potentially attributable to COVID-19 pneumoniaterm is more complex. When typicalterm·def features of COVID-19 pneumoniaterm are present in an endemic area as an incidental finding, we recommend direct communication with the referring provider to discuss the likelihood of viral infection and to try to reach consensus. As always, radiologists should follow the ACR Practice Parameter for Communication of Diagnostic Imaging Findings ( 34 ). These incidental findings do not necessarily need to be reported as COVID-19 pneumoniaterm, with “viral pneumoniaterm” as a reasonable and inclusive alternative. However, if consensus is reached, and COVID-19term is mentioned as a potential diagnosis in the radiology report, we suggest using the provided standardized reporting language. Additionally, staff at the site performing the examination should be notified to initiate standard operating procedures (SOP) for potential exposure.
It should be noted that viral pneumoniasterm have a wide variety of imaging manifestations, some of which are atypicalterm·def or less common in COVID-19term such as tree-in-bud opacities and other small nodules, bronchial wall thickening, and bronchial mucus plugs ( 12 ). Thus, the term viral pneumoniaterm encompasses a range of imaging findings some of which are not typicalterm·def for COVID-19term. It is also important to describe other lung abnormalities that may be associated with increased morbidity in the setting of COVID-19term, such as emphysema and diffuse parenchymal lung disease.
Categories.— We propose four categories for reporting CT imaging findings potentially attributable to COVID-19term, each with suggested standardized language ( Table 1 ). The reporting language does not offer an exact likelihood for COVID-19 pneumoniaterm, which depends on several factors including prevalence in a community, exposure, risk factors, and clinical presentation. Rather, the reporting language focuses on CT findings reported in the literature and the typicality of these features in COVID-19 pneumoniaterm rather than other diseases. Included in the reporting language are unique coding identifiers in brackets that can then be used for future data mining.
Table 1: Proposed Reporting Language for CT Findings Related to COVID-19term, including Rationale, CT Findings, and Suggested Reporting Language for each Category Typicalterm·def features are those that are reported in the literature to be frequently and more specifically seen in COVID-19 pneumoniaterm in the current pandemic ( 10 , 11 ). ( Figs 1 – 4 ). The principal differential diagnosis includes some viral pneumoniasterm, especially influenza, and acute lung injury patterns, particularly organizing pneumonia, either secondary, such as from drug toxicity and connective tissue disease, or idiopathic.
Figure 1: Typicalterm·def CT imaging features for COVID-19term. Unenhanced thin-section axial images of the lungs in a 52-year-old man with a positive RT-PCR ( A–D ) show bilateral, multifocal rounded (asterisks) and peripheral GGO (arrows) with superimposed interlobular septal thickening and visible intralobular lines (“crazy-paving”). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Typicalterm·def CT imaging features for COVID-19term. Unenhanced thin-section axial images of the lungs in a 52-year-old man with a positive RT-PCR (A–D) show bilateral, multifocal rounded (asterisks) and peripheral GGO (arrows) with superimposed interlobular septal thickening and visible intralobular lines (“crazy-paving”). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 4: Typicalterm·def CT imaging features for COVID-19term and other diseases with similar findings. Posterior, peripheral, and rounded GGO and consolidation in axial images of four patients; COVID-19term ( A, B ), organizing pneumonia secondary to dermatomyositis ( C ), and influenza A pneumonia ( D ). Organizing pneumonia and influenza pneumonia can be indistinguishable from COVID-19term by CT. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Typicalterm·def CT imaging features for COVID-19term and other diseases with similar findings. Posterior, peripheral, and rounded GGO and consolidation in axial images of four patients; COVID-19term (A, B), organizing pneumonia secondary to dermatomyositis (C), and influenza A pneumonia (D). Organizing pneumonia and influenza pneumonia can be indistinguishable from COVID-19term by CT. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 2: Typicalterm·def CT imaging features for COVID-19term. Unenhanced thin-section axial ( A–C ) and coronal multiplanar reformatted images ( D ) of the lungs in a 77-year-old man with a positive RT-PCR show bilateral, multifocal rounded and peripheral GGO. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Typicalterm·def CT imaging features for COVID-19term. Unenhanced thin-section axial (A–C) and coronal multiplanar reformatted images (D) of the lungs in a 77-year-old man with a positive RT-PCR show bilateral, multifocal rounded and peripheral GGO. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 3: Typicalterm·def CT imaging features for COVID-19term. Unenhanced axial ( A–C ) and sagittal multiplanar reformatted ( D ) images of the lungs in a 29-year-old man with a positive RT-PCR show multiple bilateral, rounded consolidations with surrounding GGO. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Typicalterm·def CT imaging features for COVID-19term. Unenhanced axial (A–C) and sagittal multiplanar reformatted (D) images of the lungs in a 29-year-old man with a positive RT-PCR show multiple bilateral, rounded consolidations with surrounding GGO. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Indeterminateterm·def features are those that have been reported in COVID-19 pneumoniaterm but are not specific enough to arrive at a relatively confident radiologic diagnosis. An example would be diffuse GGO without a clear distribution ( Figs 5 , 6 ). This finding is common in COVID-19 pneumoniaterm but occurs in a wide variety of diseases such as acute hypersensitivity pneumonitis, Pneumocystis infection, and diffuse alveolar hemorrhage, which are difficult to distinguish by imaging alone.
Figure 5: Indeterminateterm·def CT imaging features for COVID-19term. Unenhanced axial images in two patients showing patchy GGO with nonrounded morphology and no specific distribution, in a case of COVID-19 pneumoniaterm ( A, B ) and acute lung injury from presumed drug toxicity ( C, D ). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Indeterminateterm·def CT imaging features for COVID-19term. Unenhanced axial images in two patients showing patchy GGO with nonrounded morphology and no specific distribution, in a case of COVID-19 pneumoniaterm (A, B) and acute lung injury from presumed drug toxicity (C, D). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 6: Indeterminateterm·def CT imaging features for COVID-19term. Widespread GGO with nonrounded morphology and no specific distribution in unenhanced axial images from two different patients secondary to acute lung injury from presumed drug toxicity ( A ) and Pneumocystis pneumonia ( B ). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Indeterminateterm·def CT imaging features for COVID-19term. Widespread GGO with nonrounded morphology and no specific distribution in unenhanced axial images from two different patients secondary to acute lung injury from presumed drug toxicity (A) and Pneumocystis pneumonia (B). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Atypical featuresterm are those that are reported to be uncommon or not occurring in COVID-19 pneumoniaterm and are more typicalterm·def of other diseases such as lobar or segmental consolidation in the setting of a bacterial pneumonia, cavitation from necrotizing pneumonia, and tree-in-bud opacities with centrilobular nodules, as can occur with a variety of community acquired infections and aspiration ( Figs 7 – 9 ).
Figure 7: Atypicalterm·def CT imaging features for COVID-19term. Contrast-enhanced axial CT image ( A ) and frontal chest radiograph ( B ) showing segmental consolidation without significant GGO. Although this patient tested positive for COVID-19term, the imaging features are not typicalterm·def and could represent pneumonia related to COVID-19term or a secondary infectious process. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Atypicalterm·def CT imaging features for COVID-19term. Contrast-enhanced axial CT image (A) and frontal chest radiograph (B) showing segmental consolidation without significant GGO. Although this patient tested positive for COVID-19term, the imaging features are not typicalterm·def and could represent pneumonia related to COVID-19term or a secondary infectious process. Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 9: Atypicalterm·def CT imaging features for COVID-19term. Axial CT images from two different patients showing tree-in-bud opacities and centrilobular nodules, caused by respiratory syncytial virus ( A ) and active tuberculosis ( B ). A small cavity (arrow) is also present in ( B ) Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Atypicalterm·def CT imaging features for COVID-19term. Axial CT images from two different patients showing tree-in-bud opacities and centrilobular nodules, caused by respiratory syncytial virus (A) and active tuberculosis (B). A small cavity (arrow) is also present in (B) Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Figure 8: Atypicalterm·def CT imaging features for COVID-19term. Axial images of the lungs of two patients showing cavitation (arrow) in Klebsiella pneumonia ( A ) and tree-and-bud opacities (circle) and a cavity (arrow) in nontuberculous mycobacterial infection ( B ). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention.
Atypicalterm·def CT imaging features for COVID-19term. Axial images of the lungs of two patients showing cavitation (arrow) in Klebsiella pneumonia (A) and tree-and-bud opacities (circle) and a cavity (arrow) in nontuberculous mycobacterial infection (B). Routine screening CT for diagnosis or exclusion of COVID-19term is currently not recommended by most professional organizations or the U.S. Centers for Disease Control and Prevention. Negative for pneumoniaterm·def implies that there are no parenchymal abnormalities that could be attributable to infection. Specifically, GGO and consolidation are absent. Importantly, there may be no findings on CT early in COVID-19term. Conversely, CT has been reported to be more sensitive than RT-PCR earlier in the course of the disease ( 29 ), although this result may change with local RT-PCR test characteristics.
Pros, Cons, and Limitations of Standardized Reporting There are compelling arguments both for and against the use of standardized reporting language in describing CT findings potentially attributable to COVID-19term ( Table 2 ). This project largely stemmed from the expectation that despite most current professional guidelines recommending against the routine use of screening CT for COVID-19term, CT may be requested for potential assistance in diagnosis and management, particularly if RT-PCR is not readily available.
Table 2: Pros and Cons of Standardized Reporting for Chest CT Findings Related to COVID-19term Pros.— Without expert consensusterm·mth, radiologists may be left with uncertainty as to how to convey the presence, absence, or likelihood of COVID-19term when confronted with this as a specific indication or as an incidental finding. Standardized reporting can provide guidance and confidence to radiologists as well as increased clarity to providers through reduced reporting variability. Providing unique identifiers for each category facilitates mining data for future educational, research, and quality improvement. In addition, standardized radiology reports combined with clinical assessment may enable future care pathways to determine which patients may preferentially undergo RT-PCR should testing capacity be exceeded. Initial RT-PCR testing may also be negative, and typicalterm·def imaging findings may encourage repeat testing.
Cons.— The true sensitivity and specificity of CT for COVID-19term remains relatively unknown. One study showed that radiologists identified COVID-19term versus other viral pneumoniasterm correctly 60%–83% of the time based on typicalterm·def CT imaging features ( 10 ). However, the results of this study must be evaluated cautiously as all of the COVID-19term cases came from one country (China) and most of the control cases from a single institution in another country (United States). Additionally, this moderate level of distinction may be reduced in clinical practice as the control cases included a low proportion of influenza-A, which is the major viral pneumoniaterm that must be differentiated from COVID-19term during the winter and spring months across the northern hemisphere. Reporting “atypical featuresterm” may result in false-negative casesterm, and the risk of missing COVID-19term can have broad implications. Ordering providers may also feel that having “COVID-19term” or “coronavirus” documented in a radiology report constrains their clinical decision making and treatment options. This concern is less relevant in PUIs, as clinical suspicion already exists. However, difficulties may arise in patients with findings suggestive of COVID-19term that are incidentally detected. Direct communication with the referring provider about the likelihood of COVID-19term is recommended to avoid surprising providers and patients. We again emphasize that as an incidental finding, particularly with indeterminateterm·def or atypical featuresterm, “viral pneumoniaterm” may be preferable to “COVID-19term” or “coronavirus.”
Limitations.— We anticipate cases with mixed imaging findings, that is, those that have both typicalterm·def and atypicalterm·def imaging features for COVID-19term. Recent analysis suggests that over 20% of patients with COVID-19term may have coexistent infections complicating the categorization of imaging observations ( 35 ). The radiologist will have to determine whether or not these findings are part of the same process or are unrelated. For example, a hospitalized patient undergoing chest CT for fever could have lower lobe tree-in-bud opacities as well as peripheral GGO, which could reflect aspiration superimposed on viral pneumoniaterm. It is also possible that atypical featuresterm such as lobar consolidation may reflect a secondary bacterial pneumonia even in patients who test positive for COVID-19term. Available evidence is still limited concerning the appearance of COVID-19term in the presence of secondary disease processes such as coexistent infections and aspiration. In scenarios such as these, discussion with the treating team would be prudent.
Imaging appearances in the standardized reporting language are based upon available literature at the time of writing in March 2020, noting the retrospective nature of many reports, including biases related to patient selection in cohort studies, examination timing, and other potential confounders. As radiologists’ experience with COVID-19term increases, our categorization of these findings as typicalterm·def, indeterminateterm·def, or atypicalterm·def may evolve.
Conclusions We propose four categories for the suggested standardized CT reporting language of COVID-19term based on current literature and expert consensusterm·mth. We acknowledge that for patients with unexpected findings that could be attributed to COVID-19term, the matter is complex and that “viral pneumoniaterm” is a reasonable alternative. As always, radiologists should follow the ACR Practice Parameters for Communication of Diagnostic Imaging Findings. If COVID-19term is a potential incidental diagnosis, staff at the site performing the examination should be notified to initiate SOP for potential exposure. We also acknowledge that practice patterns vary, and this document is intended to provide guidance. If a radiologist chooses to mention COVID-19term in CT reports, this is a standard framework that can be adopted. Consensus between local imaging and clinical providers is essential to establish an agreed-upon approach.
At this time, CT screening for the detection of COVID-19term is not recommended by most radiologic societies. However, we anticipate that the use of CT in clinical management as well as incidental findings potentially attributable to COVID-19term will evolve. We believe it important to provide radiologists and referring providers guidance and confidence in reporting these findings and a more consistent framework to improve clarity. Clear and frequent communication among health care providers, including radiologists, is imperative to improving patient care during this pandemic.
Acknowledgments Acknowledgments We gratefully acknowledge colleagues who contributed to the development of these guidelines. University of Pennsylvania: Ana Kolansky, MD, Maya Galperin-Aizenberg, MD, Eduardo Barbosa, MD. University of Texas Southwestern Medical Center: Neil M. Rofsky, MD, Cecelia Brewington, MD, Travis Browning, MD, Eric Zeikus, MD. Washington University: Constantine Raptis, MD, Andrew Bierhals, MD. University of Chicago: Steven M. Montner, MD. Mount Sinai: Corey Eber, MD, Bradley N. Delman, MD, Adam Bernheim, MD, Adam Jacobi, MD. University of California San Francisco: Brett Elicker, MD, Michael Hope, MD, Kimberly Kallianos, MD, Karen Ordovas, MD, Brian Haas, MD, Thienkhai Vu, MD. University of Wisconsin School of Medicine and Public Health: M. Daniela Martin, MD. NYU Langone Health: William Moore, MD. University of Michigan: Ella Kazerooni, MD.
Footnotes * S.S. and F.U.K. contributed equally to this work.
Disclosures of Conflicts of Interest: S.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: employed by University of Pennsylvania. Other relationships: disclosed no relevant relationships. F.U.K. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: employed by University of Texas Southwestern Medical Center. Other relationships: disclosed no relevant relationships. S.A. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution receives grants from NIH; author receives royalties from Elsevier for authoring of textbooks; Editor of RCTI (RSNA). Other relationships: disclosed no relevant relationships. S.B. disclosed no relevant relationships. J.H.C. disclosed no relevant relationships. M.C. disclosed no relevant relationships. T.S.H. disclosed no relevant relationships. J.P. KannepersName Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: author is consultant for Parexel Informatics (clinical trial support, not related to topic at hand). Other relationships: disclosed no relevant relationships. S.K. disclosed no relevant relationships. J.P. Ko Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution receives money for consulting for Siemens Healthineers (research collaboration); author’s spouse is employed by Allovir (biotechnology company developing cell therapies for treatment of viral diseases). Other relationships: disclosed no relevant relationships. H.L. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution receives grant from Siemens Healthineers; author receives travel/accommodations/meeting expenses from GE Healthcare and Edwards LifeSciences. Other relationships: editorial board member, RCTI .