Data from BCR is estimated to be at least 99% complete for all cancers from 2004 onwards (16); registration is mandatory for all oncological care programs and laboratories for pathological anatomy in Belgium (17). From 2017 onward, UICC TNM8 (18) was used to record tumor stage.

All tumors from the BCR database with incidence 2017-2020, ICD-10 code for HNC (excluding lip; C01-C14;C30-C32) from patients officially residing in Belgium at the time of incidence were included in the study (n=10 582). The following five sub-localizations were considered: oral cavity (C02-C05.0, C06), oropharynx (C01; C05.1-C05.9; C09-C10), hypopharynx (C12-C13), larynx (C32) and other (C07-08;C11;C14;C30-31).

Since surgery with curative intent is performed for fewer than 40% of HNC in Belgium (19), clinical stage (cStage) allows for the broadest comparison of stage at diagnosis. cStage IV tumors were subdivided into those with and without clinically proven distant metastasis (cM1 vs cM0). The few tumor types for which TNM staging was not applicable (n=100; 3.9% of HNC in 2020) were grouped with tumors with unknown cStage.

In 2020 in Belgium, 2491 HNC were diagnosed, which was 9.5% lower than predicted based on the incidence trends 2017-2019 (Figure 1A). Patient and tumor characteristics for all years included in the study can be found in Supplemental Table 1. The decline was significant in the first wave of the pandemic in Mar-Jun and HNC continued to be underdiagnosed in males in Jul-Dec, with no significant rebounds in diagnosis for any group in this period (Figure 1B). The decline in 2020 was much more pronounced in males (n=1799, -11.8%) than in females (n=692, non-significant decline). In the recovery period, Jul-Dec, a significant underdiagnosis remained in males (-13.0%). By age group, the largest declines were seen in the 50-64 (-11.2%) and 65-79 (-11.1%) year age groups. There was a significant underdiagnosis of oral cavity cancer (n=661, -17.6%) but not of other HNC subtypes (oropharynx n= 719, hypopharynx n=249, larynx n=498, other/unspecified n=364; Figures 1C, D).

For all HNC combined, the data give the impression of a decline in cStage I and increase in cStage III tumors particularly in Jul-Dec; however, the incidence by cStage did not differ significantly from the corrected-predicted values (Figures 2A, B). Among both sexes and each of the age groups <80 years, the cStage distributions did not differ significantly from the predicted distributions during the COVID-19 pandemic (Supplemental Figures 2A-E). In the 80+ age group, there was a significant excess of cStage III tumors in 2020, and particularly after the first wave of the pandemic, Jul-Dec (Supplemental Figure 2F) and in males aged 80+ both during (Mar-Jun) and after (Jul-Dec) the first wave (Figures 2C, D).

There was no significant decline in the overall incidence of oropharynx or larynx tumors in 2020 (Figures 1C, D); however, there were significantly more cStage II oropharynx tumors, specifically in Jul-Dec 2020, while there was an excess decline in cStage I tumors (Figure 3A). This trend was not significant in males (Figure 3B) but there was a significant excess cStage II in females in Jul-Dec and overall decline in cStage I (Figure 3C). The excess incidence of cStage II oropharynx tumors was also observed in the 50-64 age group (not shown). In the 65-79 age group, there was a significant decline in incidence of larynx tumors in Mar-Jun (Figure 3D). Overall, in 2020, in Mar-Jun, and specifically in males, there was a significant excess of cStage III larynx tumors (Figures 3E, F).

The largest decline in incidence of all subtypes was in oral cavity cancer (Figure 1C), which was significant in males (-17.6%) and in patients <50 years old (Figure 4A); yet; there was no excess incidence of oral cavity cancer in any cStage overall or when stratified by subgroup (Figure 4B). For tumors of the hypopharynx there were no significant declines in incidence (Figure 4C) or disproportional changes in cStage (Figure 4D). Among tumors with other or unspecified topography, there was a significant decline in incidence in Mar-Jun among patients aged 50-64 (Figure 4E); however, the incidence by cStage did not significantly differ from the predicted distribution (Figure 4F).

To establish whether the COVID-19 pandemic adversely affected the outcome of patients diagnosed with HNC, we compared the evolution of the 1yRS over the period 2017-2019 with 2020. Over the years 2017-2019 there was an increasing trend in 1yRS, which appeared to stagnate or decline in 2020 (Figure 5), particularly for oral cavity tumors (Figure 5A), males, patients aged 50-64 years (Figure 5B) and cStage III (Figure 5C).

To determine whether the 1yRS in 2020 significantly deviated from the trend over the previous years, a linear trend in 1yRS from 2017-2019 was extrapolated to 2020. The 1yRS in 2020 was lower than predicted, though not statistically significantly different (p=0.103, Figure 5D); among the 2 393 patients included in the analysis for 2020, there were 62 more excess deaths from HNC within 1 year of diagnosis for patients diagnosed in 2020 relative to the general population (n=459) than would have been predicted based on the trends 2017-2019 (n=397). Given the apparent decline in 1yRS, even within cStages, we examined whether there was a shift in the distribution in cT or cN category, using the same method applied to stage at diagnosis. The data suggest a generally higher cT category overall and a higher cN category in Jul-Dec (Supplemental Figure 3).

This study was performed using data from a population-based cancer registry covering an entire country, which largely eliminates selection bias that confounds single- and multi-center studies. In 2020, the overlap between registrations of HNC from oncological care programs and laboratories for pathological anatomy (a measure of registration completeness) at BCR was higher than in 2017-2018 and comparable to 2019. We also did not see an increase in non-specific registrations (ICD-10 codes C76.0 or C80) (30).

In this study, we account for temporal trends in cancer incidence and stage distribution. Given the annual increase in cancer incidence (15), taking the average of several previous years as the expected value for 2020 will underestimate the true decline in incidence in 2020. Not only population aging and growth, which could be corrected for by using age-standardized reporting, but also changes in exposure to risk factors and improvements in screening/diagnostic capabilities will impact the incidence and stage at diagnosis of cancer, making it crucial that the complete temporal trends be considered.

COVID-19 infection status may have forced patients to postpone procedures or even have been a direct cause of death. COVID-19 incidence in Belgium was higher in more deprived regions (31) and HNC incidence and mortality is associated with lower socioeconomic position (32, 33). Unfortunately, we lack information on the COVID-19 status of diagnosed or deceased patients (particularly in the first wave of the pandemic, COVID-19 testing was not routinely available). Likewise, patients may have died from COVID-19 before a diagnosis of HNC could be made.

Finally, we are unable to distinguish between HPV-positive and HPV-negative oropharynx tumors in this cohort. HPV-positive oropharyngeal squamous cell carcinomas generally have better overall and progression-free survival than HPV-independent tumors (34). Morphology codes to distinguish HPV-positive and HPV-negative tumors were introduced in 2019 with ICD-O-3.2 (35), and are used in Belgium from 2020. However, it is not possible to compute a predicted value for 2020 since HPV status is not systematically available in the preceding years.

There was substantial underdiagnosis of HNC in Belgium in 2020, considering the recent trends in incidence, particularly for males and oral cavity cancer. Shifts toward higher stage were observed for oropharynx and larynx tumors and for elderly males and there was a reduced 1-year relative survival for HNC patients with diagnosis in 2020. The important impact of the COVID-19 pandemic on HNC diagnosis and outcome should stimulate active planning to ensure continuity of cancer care during future crises affecting the healthcare system. BCR will continue to monitor trends in stage and outcome, as well as treatment patterns, for HNC in the coming years.