COVID-19 Lung Ultrasound Scores and Lessons from the Pandemic: A Narrative Review

The WHO recently declared that COVID-19 no longer constitutes a public health emergency of international concern; however, lessons learned through the pandemic should not be left behind. Lung ultrasound was largely utilized as a diagnostic tool thanks to its feasibility, easy application, and the possibility to reduce the source of infection for health personnel. Lung ultrasound scores consist of grading systems used to guide diagnosis and medical decisions, owning a good prognostic value. In the emergency context of the pandemic, several lung ultrasound scores emerged either as new scores or as modifications of pre-existing ones. Our aim is to clarify the key aspects of lung ultrasound and lung ultrasound scores to standardize their clinical use in a non-pandemic context. The authors searched on PubMed for articles related to “COVID-19”, “ultrasound”, and “Score” until 5 May 2023; other keywords were “thoracic”, “lung”, “echography”, and “diaphragm”. A narrative summary of the results was made. Lung ultrasound scores are demonstrated to be an important tool for triage, prediction of severity, and aid in medical decisions. Ultimately, the existence of numerous scores leads to a lack of clarity, confusion, and an absence of standardization.


Introduction
Recently, WHO declared that COVID-19 no longer constitutes a public health emergency of international concern [1]. However, pathology remains an ongoing health issue to be dealt with during clinical practice, just like other infectious diseases.
COVID-19 still has a wide range of clinical presentations, from asymptomatic cases to severe cases requiring intensive treatment due to respiratory failure and multi-organ dysfunction [2].
In response to the challenges posed by the COVID-19 pandemic, several protocols and scoring systems have been developed to diagnose and redirect clinical judgment. Lung ultrasound (LUS) has emerged as a suitable alternative to CT scans due to its practicality, low cost, lack of radiation risk, and minimal requirement for health personnel [3]. The peripheral distribution of COVID-19 in the lungs makes ultrasound particularly wellsuited to investigate the disease [4]. Moreover, ultrasound findings have been shown to be compatible with CT findings [5,6]. However, ultrasound provides non-specific signs that can be related to other respiratory diseases commonly encountered in emergency/critical settings [7,8], so interpretation of results must be done within the clinical context. It is also worth noting that the interpretation of ultrasound is operator-dependent and requires specific training [9]. Pleural effusions ( Figure 2): present at LUS as hypo or anechoic regions between parietal and visceral pleura [18]. When scanned with M-Mode, the "sinusoid sign" can be appreciated, which is due to the motion of the floating lung in the pleural effusion fluid [19]. Transudates are usually homogeneous and anechoic, while exudates may appear heterogeneous and loculated [20].   Pleural effusions (Figure 2): present at LUS as hypo or anechoic regions between parietal and visceral pleura [18]. When scanned with M-Mode, the "sinusoid sign" can be appreciated, which is due to the motion of the floating lung in the pleural effusion fluid [19]. Transudates are usually homogeneous and anechoic, while exudates may appear heterogeneous and loculated [20].  Pleural irregularities: The pleural line is seen as a hyperechoic line moving synchronously with breath in a fully aerated lung; in a normal adult it is located 2.5 cm from the skin [21]. Pleural irregularities are characterised as the disappearance of the typical hyperechoic pleural profile [22]. The pleural line may be discontinuous and present an increase in thickness.
Consolidations ( Figure 2): occur when air normally contained in the alveoli is substituted by material of diverse origin. When they are located near the pleura, the acoustic impedance normally seen between pleura and aerated parenchyma is reduced, so that consolidations may be morphologically studied [23]. Consolidations may be translobar which appears as a tissue-like echostructure similar to the one of the liver (hence, it is called hepatization), while non-translobar appears with irregular margins between the consolidation and the fully aerated lung [24]. Consolidations at ultrasound may show air bronchograms that appear as either hyperechoic branching and tubular structures or hyperechoic millimetre long multiple structures [25]. They are the analogues of the bronchograms seen in chest X-rays.

Other Findings
Diaphragmatic findings ( Figure 3): ultrasound study of the diaphragm provides a non-invasive, feasible, and dynamic method to evaluate the movement of the diaphragm together with its characteristics such as thickness [26][27][28]. To evaluate the diaphragm, the probe is positioned below the costal margin either at the midclavicular line or at the anterior-axillary line. The evaluation may be done either at the left or right hemithorax taking advantage of the acoustic window provided by the spleen or liver, respectively. The diaphragm is detected through the two-dimensional mode (2D); then, M-mode is used to evaluate the movement of the structure. Moreover, ultrasound assesses the excursion, the speed of diaphragmatic contraction, the inspiratory time, and the duration of a respiratory cycle, but also the diaphragmatic thickness [29], which is a proxy of diaphragmatic power [30]. Pleural irregularities: The pleural line is seen as a hyperechoic line moving synchro nously with breath in a fully aerated lung; in a normal adult it is located 2.5 cm from th skin [21]. Pleural irregularities are characterised as the disappearance of the typical hy perechoic pleural profile [22]. The pleural line may be discontinuous and present an in crease in thickness.
Consolidations ( Figure 2): occur when air normally contained in the alveoli is subst tuted by material of diverse origin. When they are located near the pleura, the acousti impedance normally seen between pleura and aerated parenchyma is reduced, so tha consolidations may be morphologically studied [23]. Consolidations may be transloba which appears as a tissue-like echostructure similar to the one of the liver (hence, it i called hepatization), while non-translobar appears with irregular margins between th consolidation and the fully aerated lung [24]. Consolidations at ultrasound may show ai bronchograms that appear as either hyperechoic branching and tubular structures or hy perechoic millimetre long multiple structures [25]. They are the analogues of the bron chograms seen in chest X-rays.

Other Findings
Diaphragmatic findings ( Figure 3): ultrasound study of the diaphragm provides non-invasive, feasible, and dynamic method to evaluate the movement of the diaphragm together with its characteristics such as thickness [26][27][28]. To evaluate the diaphragm, th probe is positioned below the costal margin either at the midclavicular line or at the ante rior-axillary line. The evaluation may be done either at the left or right hemithorax takin advantage of the acoustic window provided by the spleen or liver, respectively. The dia phragm is detected through the two-dimensional mode (2D); then, M-mode is used t evaluate the movement of the structure. Moreover, ultrasound assesses the excursion, th speed of diaphragmatic contraction, the inspiratory time, and the duration of a respirator cycle, but also the diaphragmatic thickness [29], which is a proxy of diaphragmatic powe [30]. The standing drawback is that artefacts visualization strongly depends on the ma chine settings (in particular frequency) and the probe utilized and the interpretation o what is seen is strongly operator dependent: actual protocols do not provide a stric  The standing drawback is that artefacts visualization strongly depends on the machine settings (in particular frequency) and the probe utilized and the interpretation of what is seen is strongly operator dependent: actual protocols do not provide a strict quantitative measure to define what is and what is not an artefact and for the evaluation of pleural irregularities and consolidations [31].
For this reason, is of paramount importance the development of standardized practice as it exists for other anatomical regions.

COVID-19 Characteristics at LUS
The ultrasound appearance of COVID-19 is consistent with its pathophysiological basis as an inflammatory interstitial disease that gradually impacts the alveoli and reduces aeration.
Typical findings in COVID-19 pneumonia include a B-line pattern (either focal or diffuse), a thickened or irregular pleural line, and consolidations of various sizes ranging from small subpleural to large translobar [6]. The most common finding is B-lines (shown in Figure 1), which may fuse together to create the characteristic "white lung" appearance. B-lines serve as a densitometer, indicating the progressive loss of air-filled alveoli [3]. Typically, the findings are patchy and have a bilateral distribution, with clusters alternating with spared areas, leading to the definition of a "storm of clusters of B-lines" [31]. Other ultrasound findings in COVID-19 pneumonia include thickening and irregularities of the pleural line, consolidations, and less commonly, pleural effusions [14]. Consolidations are more commonly seen in severe and critical patients, particularly in the posterior fields [32]. Lung parenchyma consolidation indicates complete alveolar de-aeration and corresponds to a more serious stage of the illness.

Lung Ultrasound Scores for Predicting Severity, Treatment Response, and Outcomes
The scores presented in Table 1 demonstrated to be useful in predicting the severity of disease, the need for oxygen support, NIMV, and evaluating whether the support chosen has been efficient. The statistical data and advantages of each score are presented in Table 2.
Quantitative LUS score (q-LUS) and coalescent LUS score (c-LUS) were extensively used in the pre-COVID-19 era to assess lung aeration [10,11]. In the setting of COVID-19 pneumonia, c-LUS has been found to strongly correlate with Chest-CT [32] and is useful for assessing the benefits of recruiting maneuvers and changes in ventilation settings [33]. It can also predict outcomes in asymptomatic frail patients [34]. Falgarone et al. proposed an index with 89% sensitivity and 100% specificity in predicting an abnormal CT exam: a value of 0.32 was set as a threshold for the need for oxygen support [35]. Conversely, Soldati et al. proposed a lung ultrasound protocol specific to COVID-19 [36] with a high negative predictive value [37]. The score showed to be associated with patient worsening in medium to low-intensity care units (AUC 0.82), with the need for high-flow oxygen support, ICU admission, and death [38]. A higher total score is associated with pleural effusions, a lower P/F ratio, and higher lactate dehydrogenase (LDH) [38].
When combined with coagulation parameters and compressive ultrasound (CUS), the Soldati protocol proved to be useful in evaluating the length of hospitalization and the need for O2 therapy [39]. Notably, the Soldati score demonstrated being a good predictor of fatality (AUC 0.878, sensitivity 87.5%, specificity 81.7%), and helpful in discerning whether patients would benefit from HFNC or MV [40].
The use of lung ultrasound scores to predict the appropriateness and efficacy of NIV support has generated considerable interest. The Integrated LUS score (I-LUS) [41] has been shown to be effective in distinguishing patients who would benefit from oxygen therapy alone versus those who require NIV. Patients redirected to intensive care units had higher I-LUS scores compared to those redirected to low-intensity units. Similarly, Casella et al. proposed a score to predict the need for Continuous Positive Airway Pressure ventilation, which was predictive of death and transfer to the ICU [42].
A simplified LUS score was proposed for the early assessment of the lung as a predictor of NIV support failure (defined as the death of the patient or the need for endotracheal intubation) in the first 24 h [43]. The authors found that a score >11 is associated with worsening clinical outcome and admission to ICU in 72 h, while a score >12 and ≥5 areas involved are indicative of NIV failure (sensitivity of 88%, specificity of 93%).
Interestingly, using the same score it was found a significant association between simplified LUS and hypercoagulability state largely known to be a cornerstone of COVID-19 pathophysiology [44] to have a robust correlation with disease severity [45].
Dargent et al. proposed a score in which a rating of 27 is related to extubation failure and prompt need for NIV support [46]. Conversely, a reduction of the modified LUS score is associated with successful extubation [47].
Several of these scores have been correlated with other laboratory or clinical parameters to increase their utility in clinical settings [39,48]. Notably, Boero et al. developed the "COVID-19 Worsening Score" (COWS) [49]. A LUSS >15, in combination with four other variables (P/F ratio, dyspnea at admission, number of comorbidities, and days from symptom onset), is useful in estimating a patient's risk of disease progression and can discriminate those at high or low risk of worsening, with an overall accuracy of 80% and a negative predictive value of 93% [49]. Coalescent Lung score c-LUS [10,11] Score 0: presence of A-lines, maximum 2 B-lines Score 1: ≥3 well-spaced B-lines Score 2: coalescent B-lines Score 3: tissue-like pattern ening clinical outcome and admission to ICU in 72 h, while a score >12 and ≥5 areas involved are indicative of NIV failure (sensitivity of 88%, specificity of 93%).
Interestingly, using the same score it was found a significant association between simplified LUS and hypercoagulability state largely known to be a cornerstone of COVID-19 pathophysiology [44] to have a robust correlation with disease severity [45].
Dargent et al. proposed a score in which a rating of 27 is related to extubation failure and prompt need for NIV support [46]. Conversely, a reduction of the modified LUS score is associated with successful extubation [47]. Coalescent Lung score c-LUS [10,11] Score 0: presence of A-lines, maximum 2 B-lines Score 1: ≥3 well-spaced B-lines Score 2: coalescent B-lines Score 3: tissue-like pattern
Interestingly, using the same score it was found a significant association between simplified LUS and hypercoagulability state largely known to be a cornerstone of COVID-19 pathophysiology [44] to have a robust correlation with disease severity [45].
Dargent et al. proposed a score in which a rating of 27 is related to extubation failure and prompt need for NIV support [46]. Conversely, a reduction of the modified LUS score is associated with successful extubation [47]. Coalescent Lung score c-LUS [10,11] Score 0: presence of A-lines, maximum 2 B-lines Score 1: ≥3 well-spaced B-lines Score 2: coalescent B-lines Score 3: tissue-like pattern

AREAS
Falgarone index [35] 1: normal pleural image with A-lines 2: B lines 3: Multiple B-lines (ground glass) 4: consolidations 5: neo organisation known as hepatisation 6: pleurisy (not found among the patients) Index: a score from 1 to 5 is given to each area, the scores obtained are added up together. The resulting number is divided to the maximal score possible considering only the evaluated areas.

AREAS
Falgarone index [35] 1: normal pleural image with A-lines 2: B lines 3: Multiple B-lines (ground glass) 4: consolidations 5: neo organisation known as hepatisation 6: pleurisy (not found among the patients) Index: a score from 1 to 5 is given to each area, the scores obtained are added up together. The resulting number is divided to the maximal score possible considering only the evaluated areas.

AREAS
Falgarone index [35] 1: normal pleural image with A-lines 2: B lines 3: Multiple B-lines (ground glass) 4: consolidations 5: neo organisation known as hepatisation 6: pleurisy (not found among the patients) Index: a score from 1 to 5 is given to each area, the scores obtained are added up together. The resulting number is divided to the maximal score possible considering only the evaluated areas. Simplified LUS score [43] 1: a small loss of aeration characterised by more than three B-lines or presence of multiple sub-pleuric consolidations separated by normal pleura; 2: a moderate loss of aeration consisting of multiple and coalescent B-lines and/or multiple sub-pleuric consolidations 1 × 2 cm or smaller and separated by thickened or irregular pleura; 3: a severe loss of aeration described as parenchymal consolidation or subpleuric consolidations greater than 1 × 2 cm.

Diagnostics 2023, 13, x FOR PEER REVIEW 8 of 16
Simplified LUS score [43] 1: a small loss of aeration characterised by more than three B-lines or presence of multiple sub-pleuric consolidations separated by normal pleura; 2: a moderate loss of aeration consisting of multiple and coalescent B-lines and/or multiple sub-pleuric consolidations 1 × 2 cm or smaller and separated by thickened or irregular pleura; 3: a severe loss of aeration described as parenchymal consolidation or subpleuric consolidations greater than 1 × 2 cm.

AREAS
Integrated ultrasound score (I-LUS) [ Simplified LUS score [43] 1: a small loss of aeration characterised by more than three B-lines or presence of multiple sub-pleuric consolidations separated by normal pleura; 2: a moderate loss of aeration consisting of multiple and coalescent B-lines and/or multiple sub-pleuric consolidations 1 × 2 cm or smaller and separated by thickened or irregular pleura; 3: a severe loss of aeration described as parenchymal consolidation or subpleuric consolidations greater than 1 × 2 cm.

AREAS
Integrated ultrasound score (I-LUS) [ Several of these scores have been correlated with other laboratory or clinical parameters to increase their utility in clinical settings. [39,48] Notably, Boero et al. developed the "COVID-19 Worsening Score" (COWS) [49]. A LUSS >15, in combination with four other variables (P/F ratio, dyspnea at admission, number of comorbidities, and days from symptom onset), is useful in estimating a patient's risk of disease progression and can discriminate those at high or low risk of worsening, with an overall accuracy of 80% and a negative predictive value of 93% [49].  High sensitivity for asymptomatic patients Sensitivity and specificity scores are represented. In case of missing data, the correlation between the topic and scores has been provided by supplying either the value of p or the AUC (area under the curve) value.

Discussion
The massive engorgement of emergency departments during the pandemic has prompted the usage of lung ultrasound as a point-of-care methodology to dispatch patients in wards with different levels of intensities [50,51], highlighting the importance of lung ultrasound for triage. It can predict the severity of COVID-19, guide treatment decisions, and detect the need for respiratory support [14,43,[52][53][54].
The implementation of scores to stratify COVID-19 has obvious advantages from an economic standpoint, avoiding unnecessary hospitalization and redirecting intensive support. The goal of applying scoring systems is to translate an ensemble of ultrasonographic qualitative features and patterns into numbers [55]. Efforts have been made towards this goal during and even before the COVID-19 pandemic, resulting in a myriad of scores that often bear little differences between each other. However, the establishment of new scores and revision of pre-existing scoring systems have led to a lack of clarity. We speculate that disarrayed information may pose a difficulty for those who are first approaching the technique. Moreover, in our opinion, this may increase the lack of standardization and inter-operator variability which is already a main limitation of the methodology.
An issue that has raised a great deal of interest is the number of areas that should be scanned. In COVID-19, where the lung findings are preferentially located in the posterior regions [56], an accurate evaluation of these areas may be advantageous. However, a common field should be defined to address underestimation related to the limitation of inspected areas [57][58][59].
Various studies have been conducted with the aim of finding the appropriate number of areas to be scanned in COVID-19-affected patients. From their results, a scanning protocol including 12 zones is comparable to the commonly used protocol of 14 zones [56,60,61]. When considering a 10-area protocol [60] it seems feasible only if it considers posterior and basal areas, according to the already stressed knowledge of the preferential posterior distribution of COVID-19-related lung damage. However, it is not recommended to scan fewer than 10 areas as this would lead to underestimation of the damage. [61]. Some protocols with less than 10 areas were proposed [43], which bear the advantage of reducing the spread of infection during examinations.
On the other hand, even more extensive protocols have been proposed. It has been suggested that an 18-zone protocol would be more accurate if performed with the patient in a lateral decubitus position [62]. The authors suggest that this protocol would allow for a more extensive evaluation of the thorax, a simultaneous evaluation of anterior lateral and posterior regions, and the lateral decubitus position helps to reduce the gravity-related confounding effect on lung aeration [62].
It is important to emphasize the evaluation of posterior regions in COVID-19 patients. Soldati et al. recommend the observation of three posterior regions in their protocol, but it may not always be possible for the patient to maintain a sitting position [36]. In such cases, the authors suggest evaluating postero-basal regions instead.
However, Casella et al. demonstrated that a score obtained using only anterior and lateral areas is still a reliable predictive tool, significantly associated with respiratory failure progression [42]. This may be useful when the patient's position is limited, and posterior areas are difficult to assess.
Another important issue is the orientation of the probe. Scans are traditionally performed longitudinally, allowing for easy identification of the pleural line, but the visualization of the parenchyma may be limited by the size of the intercostal space.
The reliability of a lung ultrasound (LUS) score can be limited when it is based on the number of artefacts seen, which is why a transverse approach is preferred to evaluate lung aeration [12]. To address this issue, a quantitative LUS score was developed [12], which focuses on the percentage of pleura involved rather than the presence of B-Lines, as was used in the previously developed coalescent LUS score. By considering the amount of pleura involved, nonhomogeneous diseases such as ARDS, ventilator-associated pneumonia (VAP), and lung contusion can be better evaluated for the severity of loss of aeration.
The intercostal approach is recommended in the Soldati score [36] and the modified LUS score [47], where not only whether the pleura is involved but also its appearance (i.e., indented or broken in the Soldati score, irregular or blurred in the modified LUS score) is considered. In particular, the modified LUS score evaluates the parenchyma and the pleura separately, which may provide an advantage in prognostic value. Serial m-LUS score evaluations were found to be more sensitive than LUS score evaluations for predicting the need for prolonged mechanical ventilation. It is important to note that, unlike imaging of other organs, lung parenchymal ultrasound does not rely on the direct visualization of anatomical landmarks, and its use is based on the interpretation of artefacts. Therefore, the visualization of artefacts is highly operator dependent, and features obtained from non-optimized settings could over or under-estimate the severity of illness. This issue is amplified in an emergency setting, such as during a pandemic. Consequently, clear definitions for features should be decided, as qualitative descriptions are highly susceptible to personal interpretation and may lead to confusion. The same considerations should be adopted when areas are graded.
Regarding lung ultrasound scoring, the Soldati score [36] proposes a grading system that avoids counting B lines, as their enumeration may not be a reliable parameter due to differences in probe and imaging settings [36,63]. Instead, the presence of lung consolidations (also known as a tissue-like pattern) is often used to indicate more severe stages of the disease. However, the lack of sizing for consolidations can lead to an overestimation of loss of aeration, particularly in non-homogeneous pathologies such as ARDS [64,65]. Some authors have proposed cut-offs and rankings for consolidations in their scores [42,43,46], but a clear and standardized definition is still lacking.
As COVID-19 continues to be encountered in the clinical setting alongside other pathologies, clear and standardized protocols for diagnosis and clinical decision making should be established. Efforts towards these goals have been made [66], but a consensusbased clinical tool that is easily implemented should be considered mandatory.

Limitations
A limitation of the study is that the scores mentioned pursue different outcomes, so results are not easily comparable to each other.

Conclusions
During the pandemic, lung ultrasounds gained considerable popularity. Many authors worked to reduce the main limit of lung ultrasound as operator dependence and lack of specificity. Scores were proposed, but inevitably this created confusion and a lack of standardization. Additional research and a consensus seem mandatory for a standardized approach to lung ultrasound for COVID-19 as the pandemic has finally ended.
Author Contributions: All authors made substantial contributions to all of the following: conception and design of the review, drafting the article, revising it critically for important intellectual content, and final approval before submission. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.