Given the several radiological features shared by coronavirus disease 2019 pneumonia and other infective or non-infective diseases with lung involvement, the differential diagnosis is often tricky, and no unequivocal tool exists to help the radiologist in the proper diagnosis. Computed tomography is considered the gold standard in detecting pulmonary illness caused by severe acute respiratory syndrome coronavirus 2.

To conduct a systematic review including the available studies evaluating computed tomography similarities and discrepancies between coronavirus disease 2019 pneumonia and other pulmonary illness, then providing a discussion focus on cancer patients.

Using pertinent keywords, we performed a systematic review using PubMed to select relevant studies published until October 30, 2020.

Of the identified 133 studies, 18 were eligible and included in this review.

Ground-glass opacity and consolidations are the most common computed tomography lesions in coronavirus disease 2019 pneumonia and other respiratory diseases. Only two studies included cancer patients, and the differential diagnosis with early lung cancer and radiation pneumonitis was performed. A single lesion associated with pleural effusion and lymphadenopathies in lung cancer and the onset of the lesions in the radiation field in the case of radiation pneumonitis allowed the differential diagnosis. Nevertheless, the studies were heterogeneous, and the type and prevalence of lesions, distributions, morphology, evolution, and additional signs, together with epidemiological, clinical, and laboratory findings, are crucial to help in the differential diagnosis.

The coronavirus disease 2019 (COVID-19) outbreak began in Wuhan, China, in late 2019 and rapidly spread worldwide at the beginning of 2020, when it was declared a global pandemic by the World Health Organization[1,2]. Many jurisdictions in several states carried out public health interventions to contain the transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)[3]. Europe is now experiencing a second wave of contagion[4].

The virus has a specific tropism for the lower respiratory tract, and it is the cause of mild to severe respiratory infection[5].

Imaging has been widely employed to triage the massive load of acute respiratory referral due to COVID-19 by complementing the nucleic acid testing (i.e., the diagnostic reference standard). Typical manifestations of COVID-19 pneumonia on computed tomography (CT) have become known since the early phase of the pandemic: bilateral peripheral opacities with a lower lung distribution, usually consisting of nodular or mass-like ground-glass opacities (GGO) variably associated with areas of consolidation[6,7]. CT abnormalities may be absent in the earliest disease phases and become more extensive in the peak stage (i.e., around day 9 to 13) before resolve or evolve to a more organized phase, possibly leading to fibrotic-like changes[8,9].

Many of the hallmark CT findings are apparent on chest X-ray, which is prone to miss subtle GGO, even if relatively diffuse in extent. Nevertheless, chest X-ray has shown the potential to predict outcomes in relatively advanced disease stages, assess supervening complications, and monitor the disease course[10]. Lung ultrasound was suggested as a fast and feasible approach for triaging COVID-19 patients by identifying peripheral lung abnormalities such as confluent artifactual signs, small hyperechoic lung regions, thickened pleural lines, and consolidation[11,12]. Though highly sensitive, lung ultrasound is operator-dependent, challenging to perform in obese patients, and has lower specificity than CT[13].

It is worth emphasizing that the imaging appearance of COVID-19 is non-specific, and the performance of different modalities dramatically relies on the disease prevalence. The typical manifestations may mimic or overlap with other infective and non-infectious diseases, including influenza and acute lung injuries from drug reactions and connective tissue diseases[14]. Identifying findings uncommonly seen in COVID-19 pneumonia such as cavitation, tree-in-bud, and pleural effusion may help suggest an alternative diagnosis, which cannot be prescinded from clinical evaluation.

In the present article, we performed a systematic review of the literature focusing on differential radiological diagnosis between COVID-19 pneumonia and other infective and non-infective lung diseases, then discussing possibly helpful clinical elements and finally focusing on this issue in cancer patients.

We conducted this systematic review according to the Preferred Reporting Items guidelines for Systematic Reviews and Meta-Analysis (PRISMA) Statement[15]. The primary aim was to collect, describe, and discuss all the clinical studies evaluating the different CT findings between COVID-19 infection and other infective or non-infective lung diseases.

The current review focuses on the differential diagnosis between COVID-19 disease and other respiratory and non-respiratory disorders.

Since the early phase of the pandemic, radiological imaging has been employed to assess the suspicion of COVID-19 pneumonia in patients selected by clinical triage, demonstrating the potential for a standardized assessment of the degree of pulmonary involvement and prognostication purposes. Moreover, it has been used as a tool capable of complementing the limited sensitivity and time-consuming laboratory testing process for the SARS-CoV-2 infection detection[34-37].

Such a practical approach has found application in an unprecedented pandemic scenario, where the prevalence of the disease was extraordinarily high, with the awareness that the imaging findings of COVID-19 pneumonia were non-specific as reflecting the diffuse alveolar damage and organizing pneumonia with features shared by a broad spectrum of disorders[38,39]. Despite the increasing knowledge about radiological imaging’s role in the pandemic, the actual diagnostic performance of different imaging modalities is still unclear, with reported specificities and sensitivities depending on several factors, from the duration of symptoms to the pre-test probability of the disease. Without articulating the relative merit of X-ray or lung ultrasound vs CT, the latter is generally recognized as more sensitive for early parenchymal disease, disease progression, and differential diagnoses, including acute heart failure and pneumonia caused by other pathogens[40]. Remarkably, caution is warranted when analyzing data about the specificity and sensitivity of CT in detecting COVID-19 pneumonia, as some of the most cited studies from the radiology literature seem to suffer from limitations that may lead to overreaching conclusions[16,41-43]. Among the studies included in the present analysis, Bai et al[21] recruited the most extensive study population in which differences between COVID-19 and viral pneumonia were evaluated[21]. Though these authors concluded that radiologists are likely to distinguish COVID-19 from viral pneumonia on chest CT with high specificity (i.e., up to 94%), the lack of training information or specific diagnostic criteria in their study suggests that such results could have been overestimated.

Similarly, other studies included in the present analyses should be interpreted with caution due to limitations such as selection bias or the relatively limited number of patients. Given this awareness, it is noticeable how some CT imaging findings, namely mucoid impactions, centrilobular nodules, lobar consolidation, and significant pleural effusion, have been consistently found to be less frequent in COVID-19 than in other types of pneumonia (Figures 2-4). Thus, they are potentially helpful in everyday practice to complement clinical data in triaging acute respiratory patients[22,23,25,31,44]. Notably, radiologists need to have consciousness of ancillary findings that can be encountered in association with typical pulmonary features of COVID-19 pneumonia, possibly mimicking diseases other than COVID-19 (e.g., centrilobular solid nodules: polyhedral in shape and close to enlarged vessels within ground-glass opacities in COVID-19 pneumonia, while rounded or branching in minor airway diseases)[45].

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Axial computed tomography image of a 36-year-old man shows nodular (arrowheads) and peribronchovascular branching (orange arrows) opacities along with bronchial wall thickening (white arrow), which suggest a diagnosis other than coronavirus disease 2019 pneumonia. The patient was diagnosed with Mycoplasma pneumoniae pneumonia.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Axial computed tomography image shows right lung consolidation (arrow) and unilateral pleural effusion (arrowhead) in a 64-year-old man with bacterial pneumonia.

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Axial computed tomography image in a 50-year-old woman diagnosed with bronchopneumonia shows confluent centrilobular nodules (arrows) and consolidation (arrowheads) mostly located in the lower lobes.

Imaging findings typical of interstitial pneumonia may be found in asymptomatic COVID-19 patients[46]. Interestingly, incidental GGO showing accumulation of fluorine-18 fluorodeoxyglucose at the positron-emission tomography scan has been described in cancer patients with SARS-CoV-2 infection, raising the suspicion of tumor progression in cases of a false-negative RT-PCR result[47-49]. In those situations, an approach that includes a comparison with recent chest CT findings, as well as a close follow-up, would be appropriate.

When dealing with cancer patients, COVID-19 needs to be considered among diseases that may confound staging or treatment response assessment[50]. For obvious reasons, special attention will need to be given to patients being screened or treated for lung cancer. In our experience, different etiopathogenetic factors can coexist, and their respective inflammatory phenomena can be overlapped in the same patient, as shown in Figure 5, collecting the different CT patterns of five cancer patients who underwent differential diagnosis for pneumonitis.

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Axial computed tomography image. A: Axial computed tomography (CT) image of a 45-year-old patient with coronavirus disease 2019 showing left peripheral consolidation with perilobular distribution (arrows) suggesting organizing pneumonia; B: Axial CT image showing bilateral ground-glass opacities distributed in the subpleural regions (arrows) in a renal cancer patient confirmed with coronavirus disease 2019 pneumonia; C: Axial CT image showing multifocal ground-glass opacities in the right lung (orange arrows) and nodular consolidation (black arrow) in a renal cancer patient diagnosed with immune-related pneumonitis after treatment with nivolumab; D: Axial CT image of a patient suffering from immune-related pneumonitis showing multifocal, bandlike consolidation in the left lower lobe (arrows) with peripheral sparing (asterisk) suggesting organizing pneumonia; E and F: The last is a case of immune-related pneumonitis in a patient undergoing ipilimumab plus nivolumab for metastatic soft tissue sarcoma, whose baseline axial CT image (E) shows bilateral solid metastatic nodules; the axial CT image obtained after starting immunotherapy (F) shows new multifocal ground-glass opacities (orange arrows) with interval enlargement and an increasing number of pulmonary nodules (black arrow).

Besides cancer progression, COVID-19 pneumonia has been investigated as a mimicker of early lung cancer, both potentially displaying as single or multifocal GGOs[29]. Unsurprisingly, in their retrospective study, Zhang et al[29] found a single location and nodular morphology as significantly more frequent in lung cancer than in COVID-19[29]. Although the authors are alert about the consequences of an inappropriate surgical approach in these patients, it is reasonable to assume that evaluating the temporal evolution of CT findings, symptoms, and molecular test results would allow avoiding such a diagnostic and therapeutic pitfall in most cases.

Radiotherapy and oncologic treatment, such as target therapy and immunotherapy, may induce lung toxicity, mimicking COVID-19 illness. Zeng et al[28] recruited suspected COVID-19 patients diagnosed with cancer and treated with radiation to explore the differential diagnosis between COVID-19 pneumonia and radiation pneumonitis. The location, extent, and distribution of the lung CT abnormalities were considered useful to differentiate these two entities, with acute radiation-induced pneumonitis usually displaying GGOs or consolidation in the irradiated lung in contrast to the predominantly peripheral, subpleural opacities described in COVID-19 pneumonia. Otherwise, the differential diagnosis between COVID-19 and immuno-related pneumonia is likely to be less straightforward. The issue is highly relevant, especially considering the expanding immune checkpoint inhibitor indications and their potential to induce unique pulmonary toxicities[51,52]. Pneumonitis is a rare but potentially severe side effect of immune checkpoint inhibitors, involving 2.7% of the patients treated with anti-programmed cell death 1 and anti-programmed death-ligand 1 monotherapy and 6.6% of the patients receiving the combination of anti-programmed cell death 1 and anti-cytotoxic T-lymphocyte antigen 4[53]. Several clinical and radiological presentations have been described. Dyspnea and cough are the most frequent symptoms, while fever occurs in 12% of cases[54]. These clinical manifestations could be further confounding for the differential diagnosis. From a radiology perspective, COVID-19 and immune-related pneumonia have a range of imaging manifestations that can substantially overlap, particularly in cases of organizing pneumonia pattern (i.e., the most common pattern seen across all tumor treatments and regimens) and when leading to diffuse GGO and consolidation as a result of diffuse alveolar damage[51,52]. The varying sensitivity of molecular confirmation of SARS-CoV-2 infection and the low specificity of both entities’ clinical manifestation renders even more complicated the correct diagnosis, mostly requiring a multidisciplinary discussion before deciding on patient management[55]. However, the simultaneous presence of other immune-related adverse events, such as diarrhea, skin toxicities, and thyroid alterations, with or without a high level of inflammatory factors (i.e., interleukin-6, C-reactive protein) involved in the cytokine storm (the latter also shared with COVID-19), could lead to the hypothesis of pneumonitis most likely due to immunotherapy.

In addition to immune checkpoint inhibitors, other anticancer drugs such as tyrosine kinase inhibitors (i.e., gefitinib, erlotinib, crizotinib, osimertinib, panitumumab, cetuximab, and others), mTOR inhibitors (everolimus, temsirolimus), and chemotherapy (topotecan, bleomycin, gemcitabine, and others) can induce an interstitial pneumonitis[56]. To date, neither radiological nor clinical features can help the physician in the differential diagnosis. The rapid development of cardiovascular complications, such as acute pericarditis, left ventricular dysfunction, acute myocardial injury, embolic complications due to coagulopathy, such as disseminated intravascular coagulation, venous thromboembolism, or massive pulmonary embolism, with or without detection of the virus, can address the diagnosis of COVID-19[57].

Although lymphopenia and a high level of D-dimer and C-reactive protein are often identified in COVID-19 patients, these laboratory findings are not unique and are inadequate to address the proper diagnosis, especially in cancer patients.

Our systematic review has several limitations, including the mostly retrospective nature and the heterogeneity of the included studies.

The patient’s global view of epidemiological, clinical, radiological, and laboratory elements could help the physician overcome the diagnostic difficulties in the COVID-19 era.