Dear Editor,
We read the publication on “Newly developed artificial intelligence algorithm for COVID-19 pneumonia: utility of quantitative CT texture analysis for prediction of favipiravir treatment effect” with a great interest [1]. The goal of this study was to compare the utility of a machine learning (ML)-based algorithm with that of a CT-determined disease severity score and time from disease onset to CT in a prospective, randomized, open-label multicenter trial of favipiravir treatment of COVID-19 patients [1]. Ohno et al. observed that ML-based CT texture analysis is as good as or better than CT disease severity score for forecasting time till CT for favipiravir treatment in COVID-19 patients [1]. We agree that the new intervention might be useful in diagnosis. However, an important concern for applying the algorithm is the concurrent medical problem. In many settings, such as in developing Asia, there is a high incidence of background lung problem [2]. For example, tuberculosis might be a background disease in a patient with COVID-19 [3]. In a recent report from Belarus on the use of further tuberculosis screening for 844 cases with COVID-19 confirmation, 5.6% of the patients had tuberculosis co-infection [4]. Co-infection by bacterial pathogens occurred in 27.7% of COVID-19 patients with pneumonia, with co-infection with several pathogens occurring in 10.6% of the patients, according to a study from France [5]. Also, there is a chance of pre-investigation asymptomatic COVID-19, which is not uncommon [4], and the repeated COVID after a previous asymptomatic COVID-19 is possible [6]. The usefulness of algorithm in those scenarios is doubtful.
Funding
None.
Declarations
Conflict of interest
None.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Ohno Y, Aoyagi K, Arakita K, Doi Y, Kondo M, Banno S, et al. Newly developed artificial intelligence algorithm for COVID-19 pneumonia: utility of quantitative CT texture analysis for prediction of favipiravir treatment effect. Jpn J Radiol. 2022 doi: 10.1007/s11604-022-01270-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sookaromdee P, Wiwanitkit V. COVID-19 and tropical infection: complexity and concurrence. Adv Exp Med Biol. 2021;1318:333–341. doi: 10.1007/978-3-030-63761-3_19. [DOI] [PubMed] [Google Scholar]
- 3.Yasri S, Wiwanitkit V. Tuberculosis and novel Wuhan coronavirus infection: pathological interrelationship. Indian J Tuberc. 2020;67:264. doi: 10.1016/j.ijtb.2020.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Sereda Y, Korotych O, Klimuk D, Zhurkin D, Solodovnikova V, Grzemska M, et al. Tuberculosis co-infection is common in patients requiring hospitalization for COVID-19 in Belarus: mixed-methods study. Int J Environ Res Public Health. 2022;19:4370. doi: 10.3390/ijerph19074370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Feldman C, Anderson R. The role of co-infections and secondary infections in patients with COVID-19. Pneumonia (Nathan) 2021;13:5. doi: 10.1186/s41479-021-00083-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Joob B, Wiwanitkit V. Letter to the editor: coronavirus disease 2019 (COVID-19), infectivity, and the incubation period. J Prev Med Public Health. 2020;53:70. doi: 10.3961/jpmph.20.065. [DOI] [PMC free article] [PubMed] [Google Scholar]