1762. Spectrum and Predictors of Long-Term Radiological Abnormalities in COVID-19 Survivors
Authors * Denotes Presenting Author
  1. Aarushi Gupta *; Sunnybrook Health Sciences Center
  2. Abhenil Mittal; Princess Margaret Cancer Center
  3. Reut Anconina; Sunnybrook Health Sciences Center
The spectrum of radiological abnormalities and their predictors on computed tomography (CT) in COVID-19 survivors are not well defined, especially outside of Asian populations. However, data from Caucasian populations is limited.

Materials and Methods:
Patients with confirmed COVID-19 who had baseline and follow-up imaging at-least three months after recovery were identified retrospectively. Relevant clinical and laboratory data were collected, and baseline and follow-up CT scans were evaluated for pattern (normal vs abnormal and fibrotic vs nonfibrotic) and distribution of abnormalities. CT severity scores (CTSS) at both time points were calculated. Chi square test and logistic regression analysis were used to identify predictors for abnormal and fibrotic follow-up imaging. Optimal cutoff for independent variables was defined using ROC analysis.

A total of 55 patients with a mean age of 62.7 years (28–87 years) were included; 31 (56.4%) were women. Of those, 15 patients (27.2%) had mild, 21 (38.2%) had moderate, and 19 (34.5%) had severe COVID-19 with mean baseline CTSS of 15.4 (range 3–25). Follow-up imaging was done at a median interval of 6 months (range 3–12 months). A total of 45 patients (81.8%) had abnormal imaging and 22 patients (40%) had evidence of fibrosis with a mean CTSS of 8.44 (0–25). Multiple factors were associated with persistently abnormal CT scans including age (p = 0.001), COVID-19 severity (p = 0.003), higher baseline CTSS (p = 0.001), prolonged hospital stay (p = 0.028), admission to ICU (p = 0.015), oxygen requirement at baseline (p < 0.001), and persistent symptoms at follow-up (p < 0.001). Severe COVID-19 (p = 0.006), longer duration of hospital stay (p = 0.009), ICU admission (p = 0.011), invasive ventilation (p = 0.005), and CTSS at follow-up (p < 0.001) were predictive for fibrosis. Among these, age (OR 1.125, 0.99–1.268, p = 0.05, AUC = 0.751 [95% CI 0.56–0.94], cutoff = 54 years) and baseline CTSS (OR 1.466, 95% CI 1.022–2.102, p = 0.038, AUC = 0.919 [95% CI 0.824–1] cutoff = 11) were independent predictors for abnormal follow-up imaging and CTSS at follow-up for fibrosis (OR 1.926, 95% CI 1.217–3.048, p = 0.005, AUC = 0.89 [95% CI 0.8–0.98] cutoff = 7.5).

In this series of COVID-19 survivors, a high rate of persistent radiological abnormalities and fibrotic opacities at follow-up was observed. Age and baseline CTSS were independent predictors of abnormal imaging whereas CTSS at follow-up was predictor of fibrosis. If validated in larger prospective cohorts, these could be potential markers to select patients for more intensive monitoring.