Introduction
This article serves as a formal correction (erratum) to a previously published study in European Radiology concerning the radiological characteristics of early-stage SARS-CoV-2. The original research focused on identifying lung involvement in asymptomatic and pre-symptomatic patients, a critical task during the early phases of the pandemic for effective triage and disease monitoring.
Problem Statement
The primary objective of this correction is to rectify reported radiodensity values—measured in Hounsfield Units (HU)—for lung infiltrates. In quantitative medical imaging, precise attenuation values are essential for establishing diagnostic benchmarks and training automated detection algorithms. This update ensures the scientific record accurately reflects the density of lung tissue affected by early viral pneumonia.
Quantitative CT Analysis
The study utilized Quantitative Computed Tomography (CT) to measure the radiodensity of lung tissue. The Hounsfield Unit (HU) scale provides a standardized measure of radiodensity, where distilled water is defined as 0 HU and air as -1000 HU. By measuring the average attenuation within specific regions of interest, researchers can distinguish between healthy, air-filled lung tissue and infiltrates, such as ground-glass opacities, where air has been replaced by fluid or cellular debris.
Key Results and Findings
The correction provides definitive quantitative data regarding lung attenuation in early SARS-CoV-2 patients:
- Average Attenuation of Infiltrates: -610 ± 94 HU
- Average Attenuation of Nondiseased Lung: -830 ± 42 HU
These findings reveal a significant "density gap" of approximately 220 HU between healthy tissue and early-stage infiltrates. This measurable difference confirms that physiological changes occur in the lungs of asymptomatic individuals well before the onset of clinical respiratory distress.
Significance
Establishing these corrected values is vital for the development of automated triage tools. By defining a clear reference point for "abnormal" tissue (-610 HU) versus "normal" tissue (-830 HU), clinicians and researchers can better characterize the progression of COVID-19. This correction reflects a collaborative effort involving institutions such as the NIH, Philips, and Nvidia, highlighting the importance of data precision in global health research.
Limitations
As a correction notice, this document does not detail the specific CT hardware or segmentation algorithms used, nor does it account for variability in patient inspiration levels, which can influence HU measurements. These values are specifically representative of early-stage infection and may differ as the disease progresses toward more dense consolidation.
Summary updated