GPT-4 Vision on Medical Image Classification -- A Case Study on COVID-19 Dataset (preprint)

This technical report delves into the application of GPT-4 Vision (GPT-4V) in the nuanced realm of COVID-19 image classification, leveraging the transformative potential of in-context learning to enhance diagnostic processes.

Detection of serum KL-6 and SARS-CoV-2 antibody in patients with coronavirus disease 2019 and the diagnostic value in severe disease (preprint)

Background Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a significant threat to human health, but its clinical manifestations vary greatly among individuals. Early detection and treatment are important for severely ill patients to improve their prognosis and reduce the risk of death.Methods In the present study, serum markers were detected and analyzed in moderately ill and severely ill patients.Results The results found that there were statistically significant differences in age, serum Krebs von den Lungen-6 (KL-6) and Immunoglobulin A (IgA) levels between severely ill patients and moderately ill patients (P < 0.05). The cut-off of using KL-6 alone for the diagnosis of severely ill patients was 298.91 U/mL, with an AUC of 0.737, a sensitivity of 100%, and a specificity of 43%. When the diagnosis was performed using KL-6 in combination with Interleukin-6 (IL-6), an indicator of infection, the AUC was 0.776, with a sensitivity and specificity of 82% and 69%, respectively. When the three above were used in combination for diagnosis, the AUC was 0.785, and the sensitivity and specificity were 100% and 59%, respectively. After rehabilitation, the serum levels of KL-6, C-reactive protein (CRP), as well as antibodies, IgA, IgM and IgG, were significantly lower than those in the early stage of hospitalization.Conclusion In the present study, KL-6 and IgA were found to have some diagnostic efficacy for severely ill patients with COVID-19, but larger cohort studies are still needed for further confirmation, which in turn improves the diagnostic and therapeutic efficiency of severely ill patients.

Analysis and validation of a highly sensitive one-step nested quantitative real-time polymerase chain reaction assay for specific detection of severe acute respiratory syndrome coronavirus 2 (preprint)

Background Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, is posing a serious threat to global public health. Reverse transcriptase real-time quantitative polymerase chain reaction (qRT-PCR) is widely used as the gold standard for clinical detection of SARS-CoV-2. Due to technical limitations, the reported positive rates of qRT-PCR assay of throat swab samples vary from 30–60%. Therefore, the evaluation of alternative strategies to overcome the limitations of qRT-PCR is required. A previous study reported that one-step nested (OSN)-qRT-PCR revealed better suitability for detecting SARS-CoV-2. However, information on the analytical performance of OSN-qRT-PCR is insufficient.Method: In this study, we aimed to analyze OSN-qRT-PCR by comparing it with droplet digital PCR (ddPCR) and qRT-PCR by using a dilution series of SARS-CoV-2 pseudoviral RNA and a quality assessment panel. The clinical performance of OSN-qRT-PCR was also validated and compared with ddPCR and qRT-PCR using specimens from COVID-19 patients.Result The LoD (copies/ml) of qRT-PCR, ddPCR, and OSN-qRT-PCR were 520.1 (95% CI): 363.23–1145.69) for ORF1ab and 528.1 (95% CI: 347.7–1248.7) for N, 401.8 (95% CI: 284.8–938.3) for ORF1ab and 336.8 (95% CI: 244.6–792.5) for N, and 194.74 (95% CI: 139.7–430.9) for ORF1ab and 189.1 (95% CI: 130.9–433.9) for N, respectively. Of the 34 clinical samples from COVID-19 patients, the positive rates of OSN-qRT-PCR, ddPCR, and qRT-PCR were 82.35% (28/34), 67.65% (23/34), and 58.82% (20/34), respectively.Conclusion In conclusion, the highly sensitive and specific OSN-qRT-PCR assay is superior to ddPCR and qRT-PCR assays, showing great potential as a technique for detection of SARS-CoV-2 in patients with low viral loads.

Analysis and validation of a highly sensitive one-step nested quantitative real-time polymerase chain reaction assay for specific detection of severe acute respiratory syndrome coronavirus 2 (preprint)

Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, is posing a serious threat to global public health. Reverse transcriptase real-time quantitative polymerase chain reaction (qRT-PCR) is widely used as the gold standard for clinical detection of SARS-CoV-2. Due to technical limitations, the reported positive rates of qRT-PCR assay of throat swab samples vary from 30%-60%. Therefore, the evaluation of alternative strategies to overcome the limitations of qRT-PCR is required. A previous study reported that one-step nested (OSN)-qRT-PCR revealed better suitability for detecting SARS-CoV-2. However, information on the analytical performance of OSN-qRT-PCR is insufficient. In this study, we aimed to analyze OSN-qRT-PCR by comparing it with droplet digital PCR (ddPCR) and qRT-PCR by using a dilution series of SARS-CoV-2 pseudoviral RNA and a quality assessment panel. The clinical performance of OSN-qRT-PCR was also validated and compared with ddPCR and qRT-PCR using specimens from COVID-19 patients. The LoD (copies/ml) of qRT-PCR, ddPCR, and OSN-qRT-PCR were 520.1 (95% CI): 363.23-1145.69) for ORF1ab and 528.1 (95% CI: 347.7-1248.7) for N, 401.8 (95% CI: 284.8-938.3) for ORF1ab and 336.8 (95% CI: 244.6-792.5) for N, and 194.74 (95% CI: 139.7-430.9) for ORF1ab and 189.1 (95% CI: 130.9-433.9) for N, respectively. Of the 34 clinical samples from COVID-19 patients, the positive rates of OSN-qRT-PCR, ddPCR, and qRT-PCR were 82.35% (28/34), 67.65% (23/34), and 58.82% (20/34), respectively. In conclusion, the highly sensitive and specific OSN-qRT-PCR assay is superior to ddPCR and qRT-PCR assays, showing great potential as a technique for detection of SARS-CoV-2 in patients with low viral loads.