Stage shift and relative survival for head and neck cancer during the 2020 COVID-19 pandemic: a population-based study of temporal trends

Abstract

Objective: During the first wave of the COVID-19 pandemic in 2020, non-essential health services were suspended in Belgium, and the public was ordered to socially isolate. Underdiagnosis of cancer during this period was reported worldwide. Certain risk factors for head and neck cancer (HNC) overlap with those for COVID-19 incidence and mortality, making underdiagnosis and subsequent stage shift of this potentially rapidly progressing cancer a major concern. We aimed to analyze incidence, clinical stage at presentation, and survival of patients diagnosed with HNC in 2020 in Belgium, considering recent temporal trends.

Methods: Using population-based data from the Belgian Cancer Registry (BCR), we extrapolated 2017-2019 trends in incidence, clinical stage, and 1-year relative survival (1yRS) of HNC to create an expected value for 2020 and compared this to the observed value.

Results: There were 9.5% fewer HNCs diagnosed in 2020, compared to the predicted incidence. Underdiagnosis was larger for males (-11.8%), patients aged 50-64 (-11.2%) and 65-79 (-11.1%), and for oral cavity cancer (-17.6%). Shifts to more advanced stages were observed in larynx and oropharynx tumors and for (male) patients aged 80+. A 2.4 percentage point decline in 1yRS was observed, relative to the increasing trends in 1yRS (2017-2019).

Conclusion: The COVID-19 pandemic led to underdiagnosis of HNC, resulting in shifts to more advanced stage at presentation in certain subgroups. A stage shift can be expected for the 9.5% of tumors not yet diagnosed at the end of 2020. HNC patients diagnosed in 2020 suffered higher than expected mortality.

Keywords: COVID-19; delayed diagnosis; head and neck neoplasm; incidence; mortality; routinely collected health data.

Figure 1

Observed versus predicted incidence of head and neck cancer in Belgium in 2020, by sex, age group and tumor localization. (A) Predicted versus observed incidence by sex and age group in 2020. (B) Difference between predicted and observed incidence by sex and age group in 2020 and by time period. (C) Predicted versus observed incidence by tumor localization. (D) Difference between predicted and observed incidence by tumor localization in 2020 and by time period. Predicted values are extrapolated from the incidence trends 2017-2019. Difference between predicted and observed incidence is significant if the 95% confidence interval (error bar) of the difference does not include zero. NS, non-significant.

Figure 2

Observed versus predicted incidence by clinical stage of head and neck cancer, corrected for total decline in incidence (Corr-Pred), by sex, age group and tumor localization, in Belgium in 2020. (A) Predicted and observed clinical stage distribution for all patients in 2020. (B) Difference between observed and corrected-predicted incidence per clinical stage for all patients in 2020 and by time period. (C) Predicted and observed clinical stage distribution for males aged 80+ in 2020. (D) Difference between observed and corrected-predicted incidence per clinical stage for males aged 80+ in 2020 and by time period. Difference between predicted-corrected and observed incidence is significant if the 95% confidence interval (error bar) of the difference does not include zero.

Figure 3

Difference between observed and predicted incidence by clinical stage of oropharynx and larynx tumors, corrected for total decline in incidence, in Belgium in 2020. Difference between observed and corrected-predicted incidence per clinical stage and by time period for (A) all oropharynx tumors, (B) oropharynx tumors in males, and (C) oropharynx tumors in females. (D) Difference between observed and predicted incidence of larynx tumors by sex and age group in 2020 and by time period. Difference between observed and corrected-predicted incidence per clinical stage in 2020 and by time period for (E) all larynx tumors, and (F) larynx tumors in males. Difference between predicted-corrected and observed incidence is significant if the 95% confidence interval (error bar) of the difference does not include zero.

Figure 4

Difference between observed and predicted incidence in 2020 and by clinical stage corrected for total decline in incidence, for oral cavity, hypopharynx and other/unspecified head and neck cancers, in Belgium in 2020. Difference between observed and predicted incidence by sex and age group in 2020 and by time period for (A) oral cavity, (C) hypopharynx, and (E) other/unspecified head and neck tumors. Difference between observed and corrected-predicted incidence per clinical stage in 2020 and by time period for (B) oral cavity, (D) hypopharynx and (F) other/unspecified head and neck tumors Difference between predicted-corrected and observed incidence is significant if the 95% confidence interval (error bar) of the difference does not include zero.

Figure 5

1-year relative survival for head and neck cancer in Belgium in 2020. One-year relative survival for head and neck cancer in 2017-2020 by (A) tumor localization, (B) sex and age group, and (C) clinical stage at diagnosis. (D) The trend in 1-year relative survival 2017-2019 was extrapolated to 2020 (predicted, blue line) and compared to the observed (green bars) 1-year Relative Survival. Error bars indicate 95% confidence intervals.