ABSTRACT
For the fight against the COVID-19 pandemic, it is particularly important to map the course of infection, in terms of patients who have currently tested SARS-CoV-2 positive, as accurately as possible. In hospitals, this is even more important because resources have become scarce. Although polymerase chain reaction (PCR) and point of care (POC) antigen testing capacities have been massively expanded, they are often very time-consuming and cost-intensive and, in some cases, lack appropriate performance. To meet these challenges, we propose the COVIDAL classifier for AI-based diagnosis of symptomatic COVID-19 subjects in hospitals based on laboratory parameters. We evaluate the algorithm's performance by unique multicenter data with approximately 4,000 patients and an extraordinary high ratio of SARS-CoV-2-positive patients. We analyze the influence of data preparation, flexibility in optimization targets, as well as the selection of the test set on the COVIDAL outcome. The algorithm is compared with standard AI, PCR, POC antigen testing and manual classifications of seven physicians by a decision theoretic scoring model including performance metrics, turnaround times and cost. Thereby, we define health care settings in which a certain classifier for COVID-19 diagnosis is to be applied. We find sensitivities, specificities, and accuracies of the COVIDAL algorithm of up to 90 percent. Our scoring model suggests using PCR testing for a focus on performance metrics. For turnaround times, POC antigen testing should be used. If balancing performance, turnaround times, and cost is of interest, as, for example, in the emergency department, COVIDAL is superior based on the scoring model. © 2023 Association for Computing Machinery.
ABSTRACT
Introduction: Since the emergence of the novel coronavirus SARS-CoV-2 in December 2019 in Wuhan, cases of the associated disease COVID-19 are seen worldwide. To collect clinical data of the pandemic the international, multicenter Lean European Open Survey on SARS-CoV-2-Infected Patients (LEOSS) registry was established. Here, we present a first description of cancer patients with COVID-19 from LEOSS. Patients and Methods: We retrospectively analyzed a cohort of 283 patients (pts) with cancer and COVID-19 from a total of 1808 pts enrolled between March 6th, 2020, and June 26th, 2020. Baseline characteristics include socio-demographics, comorbidity according to Charlson Comor-bidity Index (CCI), ECOG and outcome of COVID-19. Clinical manifestation of COVID-19 was described in four phases: uncomplicated (asymptomatic/mild symptoms), complicated (need for oxygen supplementation), critical (need for life supporting therapy) and recovery (clinical improvement/discharge). Result(s): Median observational period was 11 (range 0-48) days, median inpatients stay 12.5 (range 0-72) days. Most patients were aged 66 years or older (75.5%), 112 (39.5%) pts were female. Median CCI was 4 (0-15), 46/119 (16.5%) pts had an ECOG >2. Solid tumors were seen in 61%, lymphoma and leukemia in 14.5% and 10.5% respectively. One hundred and seven pts (38%) had an active malignant disease and 76 (27%) had received anti-cancer treatment within the last 3 months. In 181 (64%) pts COVID-19 remained in the uncomplicated phase whereas 93 (33%) pts developed a complicated or critical phase. Sixty-three (22.5%) pts required intensive care, 35 out of 63 needed mechanical ventilation. A total of 79 (28%) pts died, 67 (23.5%) from COVID-19. Median survival was 33 days and worse compared to non-cancer pts (non-cancer pts: med. survival not reached, p-value < 0.001). Conclusion(s): As expected, cancer patients hospitalized for COVID-19 frequently have severe disease and an adverse outcome. To confrm these results, age-and comorbidity adjusted analysis are needed. An update of the analysis will be presented at the DGHO Annual Meeting.
ABSTRACT
Background: Active cancer has been identified as an independent risk factor for severity and mortality in COVID-19. However, direct comparisons of SARS-CoV-2 infected patients (pts) with active and non-active cancers remain scarce. Method(s): We retrospectively analyzed a cohort of pts with cancer with confirmed SARS-CoV-2 infection, enrolled 03/16/2020 - 07/31/2021. Data on demographics, cancer and laboratory findings were collected. Descriptive and subsequent regression analysis was performed. Endpoints were progression to severe COVID-19 and infection-associated mortality. Result(s): In total, 987 pts with cancer (510 active vs 477 non-active) were included in our analysis. Majority was male and > 55 years, with a higher number of elderly pts with non-active cancer. CCI was 4.75 vs 3.85 in pts with active and non-active cancer (p<0.001). Localized solid tumors were reported in 38 vs 79% (p<0.001), metastasized in 37.5 vs 5.5% (p<0.001) and hematological diseases in 37.5 vs 19.5% (p<0.001) pts with active and non-active cancer, respectively. At virus detection, majority of pts showed mild to moderate symptoms, while deterioration to severe COVID-19 was slightly more common in pts with active cancer (19% vs 16%;p=0.284). COVID-19 related mortality was significantly higher in pts with active cancer (24% vs 17.5%, p<0.001). In line, severe cytopenia and an increase of inflammatory markers were common findings in pts with active cancer at baseline, particularly in those who developed severe infection or died. Multivariate analysis revealed that ferritin (14.24 [2.1-96], p=0.006) and CRP (2.85 [1.02-8.02], p=0.046) were associated with severe COVID-19 and infection-related mortality. In pts with non-active cancer, association was seen for ferritin only (4.1 [1.51-11.17], p=0.006). Conclusion(s): Comparing pts with active and non-active cancer, mortality rate was significantly higher in pts with active cancer. Also inflammatory markers were significantly increased assuming higher levels of inflammation may play a role in adverse outcome of COVID-19 in pts with active cancer.
ABSTRACT
INTRODUCTION: Since the early SARS-CoV-2 pandemic, cancer patients have been assumed to be at higher risk for severe COVID-19. Here, we present an analysis of cancer patients from the LEOSS (Lean European Open Survey on SARS-CoV-2 Infected Patients) registry to determine whether cancer patients are at higher risk. PATIENTS AND METHODS: We retrospectively analyzed a cohort of 435 cancer patients and 2636 non-cancer patients with confirmed SARS-CoV-2 infection, enrolled between March 16 and August 31, 2020. Data on socio-demographics, comorbidities, cancer-related features and infection course were collected. Age-, sex- and comorbidity-adjusted analysis was performed. Primary endpoint was COVID-19-related mortality. RESULTS: In total, 435 cancer patients were included in our analysis. Commonest age category was 76-85 years (36.5%), and 40.5% were female. Solid tumors were seen in 59% and lymphoma and leukemia in 17.5% and 11% of patients. Of these, 54% had an active malignancy, and 22% had recently received anti-cancer treatments. At detection of SARS-CoV-2, the majority (62.5%) presented with mild symptoms. Progression to severe COVID-19 was seen in 55% and ICU admission in 27.5%. COVID-19-related mortality rate was 22.5%. Male sex, advanced age, and active malignancy were associated with higher death rates. Comparing cancer and non-cancer patients, age distribution and comorbidity differed significantly, as did mortality (14% vs 22.5%, p value < 0.001). After adjustments for other risk factors, mortality was comparable. CONCLUSION: Comparing cancer and non-cancer patients, outcome of COVID-19 was comparable after adjusting for age, sex, and comorbidity. However, our results emphasize that cancer patients as a group are at higher risk due to advanced age and pre-existing conditions.