The hallmark of the L phenotype is the vasoplegia, as confirmed by the rapid change in density and distribution of CT findings from the supine to the prone position. The benefit of a prone position in awake, nonintubated, spontaneously breathing Covid-19 patients has been emerging as potential tool to improve oxygenation and to prevent the access to ICU.

No evidence of radiological modifications related to Aim of our study is to evaluate CT changes in terms of extension, distribution and prevalence of findings, in the supine compared with the prone position.

Patients were considered eligible if were with a confirmed diagnosis of Covid-19 infection, as documented by RT-PCR nasal swab, or, confirmed by a subsequent BAL, aged between 18 and 90 years old, able to maintain both the supine and prone position on the CT scan table. Patients who were pregnant, scarcely collaborative were excluded.

Once adopted the supine position on the CT scan table, the patient received the oximeter. At the end of the supine scan, arterial oxygenation was recorded. Soon after the supine scan the patient adopted prone position with both arms extended anteriorly. CT scan was performed with a low dosage protocol, finally the second arterial oxygenation was recorded at the end of the prone scan. Patients who had oxygen requirement below 3 L/min temporarily suspended the supply; whilst those with an oxygen supply more than 3 L/min maintained it.

Three experienced radiologists analyzed CT images on a PACS workstation, in three stages. First to score the total severity score, secondly to analyze the supine scan and finally the prone CT images. The images were viewed in the lung window settings (width 1000-1500 HU; level 700 to -550 HU).

The TSS was calculated dividing each lung into three zones. The upper zone included parenchyma above the carina; the middle zone from carina and inferiori pulmonary vein and, finally the lower zone included the parenchyma below the inferior pulmonary vein. Each lung zone was further subdivided into anterior and posterior, so that the resulting zones were 12. Each zone was scored as follows: 0 with no changes; 1 with changes extent <25%; 2 with changes between 26 and 50%; 3, 51-75% and finally 4, >75%. To obtain the final score, the sum of each zone score was calculated, with a maximum value of 48.

After this first step, for each zone, the percentage of extension of patterns was calculated visually and recorded. Patterns included: pure ground glass; crazy paving; part-solid ground glass; perilobular pattern, and consolidation. This semiquantitative evaluation of this five patterns, was then evaluated for the prone scanning. The extension of the five patterns were averaged multiplying 1 for ground glass, 2 for crazy paving; 3 for part solid ground glass; 4 for perilobular pattern and 5 for consolidation. Each zone showed a final averaged score. Summarization of the anterior zones and posterior allows to calculate the anterior-posterior ratio both in supine and prone. Moreover, summarizing the scores in supine and prone, the prone-supine ratio has been calculated. A target vein, representing a "outside" vessel enlargement" has ben identified in dorsal segment of right upper lobe or in the posterobasal segment of right lower lobe, measured in supine and prone, and calculated prone-supine ratio.

In a final step, CT scan has been reviewed in consensus with pulmonologists in order to define qualitative pattern.

Inclusion Criteria:

• Confirmed diagnosis of Covid-19 infection, as documented by RT-PCR nasal swab, or, confirmed by a subsequent BAL
• Age between 18 and 90 years old
• Ability to maintain both the supine and prone position on the CT scan table.

Exclusion Criteria:

• Patients who were pregnant, scarcely collaborative

Patients were considered eligible if were with a confirmed diagnosis of Covid-19 infection, as documented by RT-PCR nasal swab, or, confirmed by a subsequent BAL, aged between 18 and 90 years old, able to maintain both the supine and prone position on the CT scan table. Patients who were pregnant, scarcely collaborative were excluded.

CT analysis was conducted by two experienced thoracic radiologists and one pulmonologist.

Extension of each CT finding and presence of vessel enlargement has been scored.