Global registry, collective intelligence, big data and machine learning should be the new methodological approach in cancer care after COVID-19 pandemic

Reviewing cancer concepts and guidelines

James Fleck: Anticancerweb 06(03), 2020

Global COVID-19 cases are still growing. On April 6th, 2020 1:55 PM, data provided by the Center for Systems Science and Engineering at Johns Hopkins University pointed to 1 309 439 cases and 72 638 deaths. Currently, the case fatality rate (CFR) is 5.5% in the general population. In cancer patients, a more severe clinical presentation combined with a higher COVID-19 CFR has been first described in China and the preliminary data published in Lancet Oncology. COVID-19-related serious events were observed in 39% of cancer patients versus the 8% reported in patients without cancer (P = 0.0003). A retrospective study describing the clinical characteristics of 28 infected cancer patients from three designated hospitals in Wuhan, China, was recently accepted for publication by Annals of Oncology. Severe events were observed in 53.6% of COVID19-infected cancer patients, leading to a CFR of 28.6%, which is more than 5-fold the currently CFR of the general population. The COVID19-infected cancer patient’s mean age was 65 years-old and the majority were male (60.7%). Most patients had stage as I to III tumors and were supposed to be treated with curative intent. The small number of patients limited the subset analysis. But, interestingly, a multivariate analysis for risk of severe events pointed out to anti-tumor treatment within 14-days (P = 0.037) and the patchy consolidation on the admitting CT (P = 0.010) as predictive factors for developing severe events. However, COVID-19-infected cancer patients corresponded only to 2.2%, which could limit the conclusions.

Nowadays, the best methodological approaching is systematic review and metanalysis. A single arm metanalysis, published by the Journal of Medical Virology, including 50 466 SARS-CoV2-infected patients, reported 18.1% severe events, 14.8% incidence of acute respiratory distress syndrome (ARDS) and a CFR of 4.3%. A second metanalysis published at the International Journal of Infectious Diseases, including 46 248 COVID-19-infected patients, showed comorbidities (hypertension, diabetes, cardiovascular and previous respiratory disease) as risk factors for severe events, however conclusion was limited by a significant level of heterogeneity (I2 index ranging from 39.9 to 87.5%). The Latin American Network of Coronavirus Disease (LANCOVID-19) had pulled data from 19 articles for qualitative and quantitative analysis plus 39 case reports for systematic review. The methodology was a random-effects metanalysis and the paper in-press has been accepted on March 11th, 2020 by Travel Medicine and Infectious Disease. A third of the patients presented comorbidities. In patients requiring hospitalization, 20.3% were admitted in intensive care unit, 32.8% presented with ARDS, leading to a CFR of 13.9%. None of the publications, described in detail the COVID19-infected cancer patient’s outcome. 

There is an urgent need to mobilize efforts of Cancer Societies and Comprehensive Cancer Centers around the world toward a Global COVID-19-infected cancer patient’s registry. The American Society for Clinical Oncology Survey at COVID-19 (ASCO Registry) is a well-received initiative, which I hope will move towards its globalization. The initiative will generate collective intelligence, creating a big data, which could be further analyzed using machine-learning and deep-learning resources. Physician has always been trained to provide the best recommendation to the patient. Eventually, delay could be responsible for mixed feelings and ethical dilemmas. The gap might be solved with a well-designed medicine and technology interaction. In health care, man and machine are not competitive or mutually exclusive. They need to cooperate by exploiting complementary skills. Nowadays, IoT are mapping market behavior and anticipating individual trends. The strategy could be useful in health care, especially in times when old concepts should be reviewed.  

 IoT = Internet-of-things

 

References:

1.     Center for Systems Science and Engineering at Johns Hopkins University 

(link: https://www.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6)

2.     Wenhua Liang, Weijie Guan, Ruchong Chen, et al: Cancer patients in SARS-CoV-2 infection: a nationwide analysis in China, Lancet Oncol, Feb 14th, 2020

3.     L.Zhang ,F.Zhu ,L.Xie ,et al: Clinical characteristics of COVID-19-infected cancer patients: A retrospective case study in three hospitals within Wuhan, China, Ann. Oncol 2020 Mar 26 EPub Ahead of Print

4.     Pengfei Sun, Shuyan Qie, Zongjian Liu, et al:  Clinical characteristics of hospitalized patients with SARS CoV‐2 infection: A single arm meta‐analysis, Journal of Medical Virology, Feb 28th, 2020

5.     Yang J, Zheng Y, Gou X, Pu K, Chen Z, Guo Q, Ji R, Wang H, Wang Y, Zhou Y, Prevalence of comorbidities in the novel Wuhan coronavirus (COVID-19) infection: a systematic review and meta-analysis, International Journal of Infectious Diseases, DOI: https://doi.org/10.1016/j.ijid.2020.03.017

6.     Alfonso J. Rodriguez-Morales, Jaime A. Cardona-Ospina, Estefanía Gutiérrez-Ocampo, et al: Clinical, laboratory and imaging features of COVID-19: A systematic review and meta-analysis, Travel Medicine and Infectious Disease, Received 29 February 2020; Received in revised form 11 March 2020; Accepted 11 March 2020 DOI: https://doi.org/10.1016/j.tmaid.2020.101623