Pneumonia and acute pulmonary edema are two important causes of acute respiratory distress and dyspnea encountered in emergency departments and acute medical settings.[1] The underlying processes of pneumonia and pulmonary edema are very different.[2]  Pneumonia may be an infectious inflammatory process, while pulmonary edema often occurs due to increased pulmonary capillary hydrostatic pressure.[3] In both conditions, fever, tachypnea, hypoxemia, cough, crackles, and diffuse pulmonary opacities can be present, further complicating the early differentiation of the two disorders in individuals with non-specific symptoms or concurrent cardiac and respiratory illness.[4] The burden of lower respiratory tract infections (LRTIs) is significant globally.[5] LRTIs were the fifth most common cause of death in the world in 2021 and responsible for some 2.5 million deaths.[6]

Acute pulmonary edema is a cardiovascular disease (CVD) that continues to be a major cause of death in the world, estimated at 17.9 million deaths per year.[7] Pulmonary edema may need to be recognized and treated quickly in patients with acute dyspnea, but the signs and symptoms may mimic those of pneumonia.[8] Traditionally, conventional chest X-ray (CXR) has been applied as the first imaging modality when evaluating suspected pneumonia and pulmonary congestion.[9] However, limitations of CXR have been identified, such as decreased sensitivity in early disease, imaging from the supine position or portable imaging, and the inability to differentiate overlapping airspace and interstitial abnormalities.[10] A meta-analysis of 1827 patients found pooled sensitivity of 73% for CXR and 88% for LUS for the diagnosis of cardiogenic pulmonary edema, and pooled specificity of about 90% for both techniques.[11]

Lung ultrasound has become a rapidly evolving bedside, radiation-free imaging modality capable of recognizing sonographic patterns related to underlying pulmonary pathogenesis.[12] Subpleural consolidation, tissue-like echotexture, dynamic air bronchograms, and focal B-lines can serve as evidence of alveolar involvement in pneumonia.[13] Cardiogenic pulmonary edema, on the other hand, tends to result in bilateral, diffuse, relatively regular B-lines that may also include pleural effusions. Results gathered over the past few years show that high diagnostic accuracy for pneumonia can be obtained with LUS. Pooled sensitivity and specificity of 29 studies with 6702 participants for adult pneumonia were 92% and 94%, respectively.[14]  In more recent years, another updated systematic review limited to studies with CT as a reference standard reported higher sensitivity (90.0%) and specificity (90.8%) for LUS than for CXR (72.6% and 82.0%, respectively).[15] The results indicate that LUS has the potential to identify abnormalities of the lungs which are not easily observed on standard radiography.

Although this is encouraging, the clinically important issue is not just whether LUS or CXR can see an abnormal lung, but whether it can detect pneumonia versus another potentially significant cause of acute respiratory symptoms, pulmonary edema. This distinction is significant because B-lines, interstitial abnormality, and airspace abnormality can be seen in both, and misattribution can result in inappropriate antimicrobial or diuretic therapy. The existing literature has mostly focused on the performance of LUS or CXR individually for pneumonia or pulmonary edema, but not on the relative value of each modality in distinguishing the two frequently co-occurring diagnoses in the same clinical population.

Thus, a direct comparison, based on a predetermined reference standard, might yield clinically relevant evidence of the relative diagnostic performance of these readily available imaging modalities. The aim of the current study was therefore to compare the diagnostic value of lung ultrasound and chest X-ray in discriminating pneumonia from acute pulmonary edema. The goal of the study was to assess and compare the sensitivity, specificity, PPV, NPV, diagnostic accuracy, and AUC of LUS and CXR in distinguishing between these two conditions. The study aimed to determine whether bedside LUS can offer further discriminatory information over routine chest radiography in patients with acute respiratory symptoms to address this diagnostic dilemma.

A comparative diagnostic accuracy study was performed at Mayo Hospital, King Edward Medical University (KEMU), Lahore, over a period of six months from 1st January 2026 to 30th June 2026. The sample size for this comparative diagnostic accuracy study was calculated using a 95% confidence level and a 10% margin of error, considering the sensitivity estimates reported in previous diagnostic studies. A previous study documented a lung ultrasound sensitivity of 98.5% for pneumonia and 91.05% for acute pulmonary edema.[16] The sample size for each diagnostic group was estimated to be around 103 participants. Therefore, a total of 206 participants were required, comprising 103 patients with pneumonia and 103 patients with acute pulmonary edema. A non-probability consecutive sampling technique was used. Adult patients aged 18 years or older who presented with acute respiratory symptoms, including dyspnea, cough, fever, tachypnea, hypoxemia, or respiratory distress, and in whom the treating physician clinically suspected either pneumonia or acute pulmonary edema were included. The inclusion criterion was that the patient could have both LUS and CXR performed at the initial diagnostic evaluation. Patients were then divided into the pneumonia group or acute pulmonary edema group based on the predetermined reference diagnosis. Patients under the age of 18 years, those with a previous diagnosis of chronic interstitial lung disease, diffuse pulmonary fibrosis, active pulmonary malignancy, or significant structural lung abnormalities likely to significantly affect interpretation of LUS or CXR were excluded. Patients who had recently undergone major thoracic surgery or trauma, or in whom the imaging examination was not completed satisfactorily, or who declined to participate, were also excluded. Patients who had both confirmed pneumonia and acute pulmonary edema were excluded from the primary comparison since the study was designed to assess the ability of the two modalities to differentiate between these two conditions. Eligible patients were recruited in a consecutive manner after obtaining ethical approval and written informed consent. All demographic and clinical data were collected on a structured data-collection proforma, which included data regarding age, sex, presentation, duration of symptoms, temperature, respiratory rate, oxygen saturation, blood pressure, heart rate, relevant comorbidities, and pertinent laboratory findings. A history of cardiac disease, previous heart failure, hypertension, diabetes mellitus, chronic respiratory disease, fever, productive cough, and other clinical factors relevant to the differential diagnosis were given particular attention. A lung ultrasound and chest X-ray were performed at the time of initial diagnostic evaluation of each participant, preferably within the same clinical evaluation period and before significant changes in the chest X-ray were expected to have occurred due to treatment. Lung ultrasound was carried out using a portable ultrasound machine with an appropriate convex or microconvex probe. The anterior, lateral, and posterior regions of the lungs were systematically examined in both hemithoraces. B-lines, consolidations, air bronchograms, pleural-line abnormalities, pleural effusions, and other sonographic findings were documented. A focal or asymmetric pattern of consolidation with compatible air bronchograms and focal B-lines was considered supportive of pneumonia, whereas bilateral diffuse B-lines with a relatively regular distribution and associated features of pulmonary congestion were considered supportive of pulmonary edema. Chest X-ray was taken as part of the hospital's routine radiological procedure. The radiographic evaluation included evaluation for focal or lobar airspace opacity, consolidation, interstitial infiltrates, bilateral perihilar opacities, vascular congestion, and pleural effusion. To minimize interpretation bias, the LUS and CXR interpretations were recorded independently, and the findings of one modality were not used to determine the interpretation of the other. A reference standard was used to confirm the final diagnosis, which was based on the overall clinical assessment, the available diagnostic evidence, such as clinical presentation, laboratory findings, imaging findings, response to treatment when clinically relevant, and final hospital diagnosis. The reference diagnosis was classified as pneumonia or acute pulmonary edema. The diagnostic findings of LUS and CXR were then compared independently with this reference diagnosis. SPSS version 26 was used to enter and analyses data. The Shapiro–Wilk test was used to determine the normality of continuous variables. Quantitative variables were expressed as mean ± SD and non-normally distributed variables were expressed as median (interquartile range). All categorical variables were presented as frequencies and percentages. Qualitative variables were compared between pneumonia and acute pulmonary edema by the independent-samples t-test and Mann–Whitney U test, while categorical variables were compared by the chi-square test and Fisher's exact test. The results of LUS and CXR were cross-tabulated separately with the reference diagnosis for diagnostic performance analysis. The following sensitivity, specificity, positive predictive value, negative predictive value, overall diagnostic accuracy, positive likelihood ratio, and negative likelihood ratio were calculated with corresponding 95% confidence intervals. Receiver operating characteristic (ROC) curves were plotted for LUS and CXR, and Area under the curve (AUC) was calculated to estimate the overall discriminatory ability of the test. An appropriate paired ROC comparison method was used to compare the AUCs of the two modalities. A two-sided p-value <0.05 was considered statistically significant.

A total of 206 patients were included, comprising 103 patients with pneumonia and 103 with acute pulmonary edema. The age of patients with acute pulmonary edema was significantly higher than that of patients with pneumonia (p<0.001), and sex distribution was similar between groups. Median symptom duration was also brief in patients with acute pulmonary edema (2 days vs 4 days, p<0.001). (Table 1).

The clinical presentation was very different for the two groups. Patients with pneumonia had a higher incidence of fever, cough, a productive cough, and higher mean temperature, whereas dyspnea and tachypnea were more common in the patients with acute pulmonary edema. Patients with pulmonary edema also had lower mean oxygen saturation and higher respiratory and heart rates. There were no significant differences between the two groups in either systolic or diastolic blood pressure readings. (Table 2)

Patients with acute pulmonary edema were significantly more likely to have cardiac disease, a history of heart failure, hypertension, diabetes mellitus, or elevated BNP/NT-proBNP, while elevated WBC and CRP were more common in those with pneumonia. The frequency of chronic respiratory disease was comparable between the two groups. (Table 3)

LUS showed different patterns based on the reference diagnosis. Pneumonia had significantly more focal consolidation, air bronchograms, B-lines, and a regular pleural line, as well as focal/asymmetric distribution, while acute pulmonary edema had more B-lines and a regular pleural line. There was also more pleural effusion in the pulmonary edema group. (Table 4)

There were also differences in the findings of the CXR among the diagnostic groups. In pneumonia, focal/lobar airspace opacity was more common, while interstitial infiltrates, bilateral perihilar opacities, vascular congestion, and pleural effusion were more common in acute pulmonary edema. (Table 5)

In comparison to the reference diagnosis, LUS had a sensitivity of 93.2%, a specificity of 94.2%, and an overall diagnostic accuracy of 93.7% for pneumonia cases and 97% for acute pulmonary edema cases. CXR demonstrated lower sensitivity (78.6%), specificity (82.5%), and diagnostic accuracy (80.6%). The positive likelihood ratio and negative likelihood ratio for LUS were also more favorable. (Table 6)

The ROC analysis showed an AUC of 0.937 for LUS and 0.806 for CXR. The AUCs differed by 0.131 (p<0.001), indicating a statistically significant difference in the paired discriminatory performance of the two imaging modalities. (Table 7)

Table 1. Demographic characteristics of study participants

Table 2. Clinical characteristics and vital signs

Table 3. Relevant comorbidities and laboratory findings

Table 4. Lung ultrasound findings according to reference diagnosis

Table 5. Chest X-ray findings according to reference diagnosis

Table 6. 2×2 diagnostic classification and diagnostic performance of LUS and CXR

Table 7. Receiver operating characteristic analysis of LUS and CXR

Figure 1. Receiver operating characteristic (ROC) curves comparing lung ultrasound and chest X-ray for differentiating pneumonia from acute pulmonary edema

The present study demonstrated that lung ultrasound (LUS) had a higher diagnostic performance than chest X-ray (CXR) for differentiating pneumonia from acute pulmonary edema. LUS showed a sensitivity of 93.2%, specificity of 94.2%, and a diagnostic accuracy of 93.7%, while CXR showed a sensitivity of 78.6%, specificity of 82.5%, and diagnostic accuracy of 80.6%. The higher AUC (p<0.001) reinforced the superior discriminatory performance of LUS. The results have clinical implications because pneumonia and acute pulmonary edema can have overlapping clinical findings, for example, dyspnea, hypoxemia, tachypnea, and diffuse pulmonary abnormalities, and distinguishing them on clinical examination alone can be difficult. The findings are consistent with Ghauri et al. (2021), who specifically compared LUS and CXR in patients presenting with acute dyspnea and suspected acute pulmonary edema. The sensitivity and specificity for LUS were 91.05% and 91.18%, respectively, compared with 60.16% and 66.67% for CXR. The close agreement between their LUS estimates and the present sensitivity of 93.2% and specificity of 94.2% supports the usefulness of LUS for identifying pulmonary edema-related changes.[17] Likewise, Dhawan & Singh (2022) compared bedside LUS with CXR in 85 critically ill adults with suspected pneumonia, against the reference standard of CT. LUS was able to achieve a higher sensitivity of 88.1% than that of CXR at 67.8%, with the reported ROC AUCs of 0.94 and 0.66, respectively. The direction of the difference between LUS and CXR was similar, but not the magnitude of the difference, as the population and reference standard differ from this study.[18] A meta-analysis of eight studies, which included 2787 patients with suspected acute decompensated heart failure (ADHF) in 2022, revealed that the sensitivity and specificity of LUS were also higher than those of CXR in the detection of cardiogenic pulmonary edema, with a sensitivity of 91.8% compared to 76.5% and specificity of 92.3% compared to 87.0%, respectively. The present results are similar to these estimates, with the exception of the greater sensitivity of LUS than CXR. The consistency is important because the same technique of B-line distribution was also utilized to differentiate pulmonary congestion from focal infectious abnormalities in the current study.[19] The sonographic features noted in the present study also support the diagnostic distinction between the two conditions noted above. Among patients with pneumonia, focal consolidation, air bronchograms, focal B-lines, pleural-line abnormalities, and asymmetric distribution were significantly more common in these patients than were bilateral diffuse B-lines and a regular pleural line in acute pulmonary edema patients. This is in keeping with the 2023 European Association of Cardiovascular Imaging consensus statement, which defined multiple, diffuse, bilateral B-lines as the typical LUS pattern of acute heart failure with pulmonary edema, and highlighted the need to differentiate this pattern from focal or patchy B-lines.[20] The present results also corroborate the findings of Desai et al. (2024), who conducted a systematic review and meta-analysis of 29 studies with 6,702 adults diagnosed with suspected pneumonia. Pooled sensitivity and specificity for the LUS were 92% and 94%, respectively; the positive likelihood ratio was 16, the negative likelihood ratio was 0.08, and AUC was 0.9712. The present LUS sensitivity of 93.2%, specificity of 94.2%, LR+ of 16.02, LR− of 0.07, and AUC of 0.937 are therefore remarkably similar, despite differences in study populations and reference standards.[14] A prospective study conducted in a TB-endemic area published in 2025 also reported better sensitivity of LUS compared to CXR for pneumonia, 96.0% vs. 82.8%, respectively, but generally low specificity for both. In addition, CT was used as a secondary reference standard, which was also highly sensitive. The slightly higher sensitivity found in that study than in the current 93.2% may have to do with disease spectrum differences and with methodology differences in the reference standards. Importantly, the authors noted that LUS could miss lesions that did not reach the pleural surface, illustrating an important limitation when interpreting LUS findings.[21] A recent systematic review and meta-analysis of the literature (2025) that directly compared pulmonary ultrasound with CXR for adult community-acquired pneumonia found that ultrasound had a sensitivity of 90.0% and a specificity of 90.8%, whereas CXR had a sensitivity of 72.6% and a specificity of 82.0%. The corresponding LR− values were 0.12 for ultrasound and 0.36 for CXR. These results are similar to the present study, which reported a sensitivity of 93.2% and a specificity of 94.2%, with a considerably lower LR− of 0.07 compared to the specificity of the CXR of 0.26. This is a comparable pattern across studies, which confirms that LUS has a greater potential to decrease diagnostic uncertainty when pneumonia is suspected.[22] This is corroborated by Orso et al. (2025), who conducted a Bayesian network meta-analysis of nine prospective studies with 746 critically ill adults. The pooled sensitivity of LUS was 0.93 and CXR was 0.65, and specificity was 0.83 and 0.81, respectively. The HSROC AUC was also higher for LUS (0.88) than CXR (0.76). Although the absolute AUCs were lower than those observed in the present study, the relative difference between the modalities was similar. The authors also noted moderate to substantial heterogeneity and operator dependence and blinding concerns that should be taken into account when interpreting diagnostic accuracy estimates.[23] A previous systematic review and meta-analysis of LUS in acute heart failure (2025) also noted a pooled sensitivity of 92%, specificity of 90%, and accuracy of 96.0%. These values are very similar to the current LUS performance differential for pulmonary edema/pneumonia. There was also significant variation between the diagnostic studies in this study, suggesting that LUS may differ depending on patient population, scan protocol, operator experience, and reference diagnosis. This is probably why the estimates presented here are not identical to those found in single studies.[24] Another 2025 study on the value of POCUS in acute heart failure yielded similar results, with additional features of lung POCUS (B-line pattern and pleural effusion) increasing the sensitivity to 80% and specificity to 82%, while combining cardiac and lung POCUS improved specificity to 96%. These slightly lower figures than those in the current study may be due to a wider definition of acute heart failure and the use of patients who have been difficult to diagnose for dyspnoea. However, the results emphasize the importance of B-lines and pleural signs in clinical bedside assessment, not only based on one but on several sonographic signs.[25] The differences in clinical features of the two groups also have implications for interpretation. Patients with acute pulmonary edema had a significantly older age, and higher frequencies of cardiac disease, previous history of heart failure, hypertension and BNP/NT-proBNP. Those with pneumonia had higher frequencies of fever, productive cough, leukocytosis, and elevated CRP. The results confirm the importance of considering imaging in conjunction with the clinical and laboratory setting. Dyspnea and Hypoxemia overlap, for example, between the two conditions, as an example, demonstrating the continued relevance of a multimodal diagnostic approach. This present study possesses a few methodological advantages. LUS and CXR were conducted in the same first diagnostic examination; each interpretation was made independently, and diagnostic performance was assessed against a predetermined reference diagnosis. However, incorporation or verification-related bias can occur with the composite reference standard since imaging results contribute to the clinical diagnosis. In addition, operator dependence of LUS and differing scanning experience can be important in considering reproducibility. In addition, excluding patients with simultaneous pneumonia and pulmonary edema reduces direct applicability to patients with mixed pathology, which can be the most difficult. LIMITATIONS There were a number of limitations to this study. It took place in one tertiary-care hospital, and results may not be generalizable to other health care facilities. Verification or incorporation bias may have occurred if the imaging findings were related to the final diagnosis. The diagnostic performance of lung ultrasound is operator-dependent and may vary based on the experience of the individual examiner. Simultaneous pneumonia and acute pulmonary edema were excluded, yet clinical practice may present mixed pathology that poses a greater challenge for diagnosis. Furthermore, the study excluded the evaluation of interobserver agreement among LUS operators, as well as the assessment of the influence of various ultrasound protocols and operators' skills on diagnostic accuracy.

Lung ultrasound demonstrated higher sensitivity, specificity, diagnostic accuracy, and discriminatory performance than chest X-ray for differentiating pneumonia from acute pulmonary edema. The diagnostic value of LUS was further affirmed by the significantly greater AUC and paired diagnostic performance. Its characteristic patterns of focal consolidation and asymmetric abnormalities in pneumonia and bilateral diffuse B-lines in pulmonary edema provided useful differentiation between the two conditions. LUS may therefore serve as an important imaging modality in the diagnostic assessment of patients presenting with acute respiratory symptoms.