Although some studies have confirmed that cancer patients were more susceptible to SARS-CoV-2 than the general population,(Dai et al., 2020;Liang et al., 2020) the clinical features and short-term prognosis of cancer patients infected with SARS-CoV-2 were still unclear. The present study showed that compared to survivors, non-survivors were more male and older, with the vast majority having basic diseases and dyspnea. Moreover, the heart rate of the non-survivors was faster than that of the survivors, and the detection values of non-survivors were higher in most laboratory findings, such as white blood cell counts, neutrophil counts, PCT, CRP, hs-CRP, direct bilirubin, PT, APTT, creatinine, CK-MB, and D-dimer, except for platelet counts and lymphocyte counts. These findings were consistent with previous researches on patients with SARS-CoV-2 infection.(N. Chen et al., 2020;R. Chen et al., 2020;Yang et al., 2020) Although cancer patients infected with SARA-CoV-2 shared common epidemiological characteristics with the general population, they might also have unique clinical features. Thus, this study performed further exploration on cancer patients with SARS-CoV-2 infection to evaluate the potential impact of COVID-19 on cancer patients.
The study of Dai et al. has demonstrated that patients with lung cancer were at the highest risk among patients with solid tumors infected with SARS-CoV-2.(Dai et al., 2020b) They suspected that reduced pulmonary reserve and severe infection were responsible for poor outcomes in these patients. In this study, lung cancer patients had the highest proportion of all cancer patients with SARS-CoV-2 infection. Similarly, lung cancer patients remained the most in 37 non-survivors, which might be associated with the high prevalence of lung cancer. Lung is the leading site of SARS-CoV-2 infection-induced lesions, and it remains unclear whether lung cancer patients have a worse prognosis than other cancer patients during the COVID-19 epidemic. In view of this, we used PSM to balance the differences in clinical characteristics between the lung cancer group and the non-lung cancer group. However, this study did not observe significant difference in short-term outcome between the two groups. Thus, among all cancer patients infected with SARS-CoV-2, we believed that the short-term survival of lung cancer patients was similar to that of other cancer patients.
This study revealed that more patients were treated with antibacterial therapy and glucocorticoid therapy in the non-survivors than in the survivors, suggesting that non-survivors were often associated with bacterial infections, often accompanied by severe symptoms. Higher PCT values and neutrophil counts for non-survivors also appeared to indicate this. PCT, a procalcitonin peptide synthesized from thyroid C cells and released from leukocytes, is a highly specific indicator of bacterial infection and closely related to the prognosis of the disease.(de Jong et al., 2016) In this study, PCT values of non-survivors were significantly higher than that of survivors, and PCT significantly affected the prognosis of patients. This was in line with that of the study by Chen et al..(R. Chen et al., 2020b) Notably, although the participants of this study (cancer patients with SARS-CoV-2 infection) was different from that of Chen et al.(all patients with COVID-19),(R. Chen et al., 2020b) PCT was thought to significantly affect patient outcomes in both studies. In addition, as the first defense barrier against suppurative infection, neutrophils play an essential role in the defense and protection function of the human body and increase significantly in patients with bacterial infections. In this study, the median neutrophil counts of non-survivors was 6.62*10E9, significantly higher than that of the survivors (3.42*10E9). These findings further suggested that patients with bacterial infections were at a higher risk of death and needed to be paid more attention by clinicians.
Of the 223 patients included in this study, 133 patients had dyspnea and up to 75.7% of the 37 non-survivors had dyspnea. If not correctly managed, dyspnea may aggravate hypoxia and lead to acute respiratory failure and other serious complications. In this study, dyspnea was identified to affect the prognosis of patients in univariate Cox analysis and Kaplan-Meier survival analysis. As with dyspnea, sex, heart rate, platelet counts, D-dimer, and AST were also demonstrated to be important factors affecting patient outcomes. Moreover, platelet counts, as well as neutrophil counts, was considered independent prognostic factors in these patients. These patients with low platelets had a higher risk of death, which was similar with the general patients with COVID-19.(Bi et al., 2020;Lippi et al., 2020) The factors mentioned above could significantly affect the prognosis of patients, so clinicians should focus on the changes in these indicators when managing the patients with SARS-CoV-2 infection. We believed that sufficient identification of prognostic factors in such patients would be helpful for clinicians to determine the prognosis of patients in the early stage and timely adjust treatment strategies.
Nomogram, a graphical prediction tool, can use statistical regression to assess the impact of various clinicopathological parameters on the likelihood of events occurring. Compared with the traditional staging system, nomogram has more accurate risk assessment methods, which is helpful to the proper individual treatment of clinical patients.(Balachandran et al., 2015)According to the results of the multivariate Cox analysis, we also constructed a prognostic nomogram that accurately predicted the overall survival of patients at 2, 3, and 5 weeks based on individual characteristics of patients. Importantly, the predictive performance and clinical utility of this model were also well-validated in an independent validation cohort. Furthermore, for the convenience of clinicians using the model constructed in this study, we created a dynamic nomogram on the web. Clinicians could access the website directly through mobile phone or computer anytime and anywhere, and inputting the corresponding information of patients to predict the survival of patients, which would undoubtedly simplify the application process and facilitate clinical use.
Indeed, this study had the following limitations. First, the nature of retrospective research inevitably led to selection bias. Second, despite our efforts to collect some essential clinical information after feedback, it was inevitable to omit some data that may inform our analysis. In addition, although the study used the patients data from 26 clinical centers in Hubei province, Hubei was only one of the epicenter of the outbreak. It was still less representative than the cohort from the whole country or even several countries. Thus, it was still necessary to strengthen the in-depth cooperation among countries as soon as possible and carry out international large-scale or prospective researches.
In conclusion, this study provided evidence that male, dyspnea, elevated PCT, increased heart rate, elevated D-dimers, decreased platelets and so on were risk factors for cancer patients with SARS-CoV-2 infection. Abnormalities of a series of laboratory findings at admission were more common in non-survivors than survivors. Among cancer patients infected with SARS-CoV-2, lung cancer patients had similar short-time survival with other cancer patients. In addition, the nomogram proposed in this study had good predictive performance, which would assist clinicians to predict the prognosis of patients early and perform more reasonable and effective treatment strategies.