The COVID-19 pandemic, which was declared a national emergency in the United States on March 13, 2020,¹ impacted cancer care in the United States, resulting in delayed cancer screening and diagnosis.^2–4 Past studies using US cancer registry data have described lower cancer incidence rates in 2020 compared to earlier years, coinciding with COVID-related stay-at-home orders during the spring months of 2020.^5–9 Analysis of pathology reports and diagnosis records of cancer—mostly affecting adults during this time period—suggest that decreased cancer incidence was due to decreased diagnosis as opposed to operational delays of cancer reporting.^5,10 However, the impact of the COVID-19 pandemic on US national reporting of childhood, adolescent, and young adult cancer has not been thoroughly studied. Reports of decreased cancer incidence in 2020 compared to earlier years, using population-based cancer registries, have mostly focused on all ages or adults, and focused on cancer types that mostly affect adults (e.g., prostate, colon, lung).^5–7,11 Descriptions of how the COVID-19 pandemic affected cancer diagnoses among children, adolescents, and young adults have been more limited, such as in reports of local areas or lower in population coverage.^12–17

It has also been shown that pediatric cancer care was affected by the COVID-19 pandemic, including delayed treatment and complications due to SARS-CoV-2 infections.^18–21 But it is less clear how pediatric cancer incidence was affected as described by population-based databases. Therefore, we sought to report trends in pediatric and young adult new cancer cases as well as incidence rates—during 2020 and 2021 compared to 2019—using data representing 98% of the US population; the aim was to determine potential disruptions in cancer diagnoses during the initial years of the COVID-19 pandemic.

The US Cancer Statistics (USCS) analytic database²² was used to describe new cases of cancer during 2003 to 2021 using methodology that has been described previously.²³ USCS, which covers all US states and the District of Columbia (DC), encompasses cancer registries that contribute to the National Program of Cancer Registries (NPCR) and to the Surveillance, Epidemiology, and End Results (SEER) program. However, this study included 98% of the US population during this time period. Indiana was excluded because its data did not meet the publication criteria for the entirety of the study period; data from all other states and DC met USCS publication criteria for 2003 to 2021.²⁴ This study included children (ages 0–14 years), adolescents (ages 15–19 years), and young adults (ages 20–39 years) diagnosed with either a malignant or nonmalignant tumor.²⁵ All ages (0–85+ years) were included for comparison and appear only in the supplemental material. Cases were characterized by site and histology according to the International Classification of Diseases for Oncology, third edition.^26,27 Cancer type was grouped according to the International Classification of Childhood Cancer (ICCC)^28,29 for ages 0–19 years (defined here as pediatric ages) at diagnosis and the Adolescent and Young Adult (AYA) Site Recode 2020 Revision for ages 20–39 years.^30,31 Race and ethnicity was categorized as Non-Hispanic American Indian or Alaska Native, Non-Hispanic Asian or Pacific Islander, Non-Hispanic Black (Black), Non-Hispanic White (White), and Hispanic or Latino (Hispanic). Stage was defined by merged summary stage (local, regional, distant, and unknown stage).³² Rural-urban status was characterized by 2013 Beale codes, which categorize counties into metropolitan vs. nonmetropolitan status by population size.³³

Age-adjusted incidence rates were reported per 1 million people for ages 0–19 years and per 100,000 people for ages 20–39 years; all ages were age-adjusted to the 2000 US standard population (19 age groups—Census P25–1130). Incidence rate ratios (IRR) were calculated to compare 2019 to 2020 and 2019 to 2021, with 2019 set as the reference rate; two-sided statistically significant IRRs were different at p < .05. Counts and rates were calculated overall and by sex, age, race and ethnicity, stage, US Census region, metropolitan status, and cancer type. Counts were additionally described by diagnosis month in 2019, 2020, and 2021. Incidence trends during 2003 to 2021 were assessed by calculating annual percent change (APC) to describe a single segment and average APC (AAPC) to describe a fixed interval.²³ Joinpoint analysis during 2003–2021 was performed to provide context for changes observed during 2019–2021. Based upon analysis from the National Cancer Institute, estimation of joinpoint trends excluded the year 2020 due to changes related to the COVID-19 pandemic but included cases diagnosed in 2021.^9,34 Calculations were performed using SEER*Stat, version 8.4.3³⁵ and Joinpoint, version 5.2.0.0.³⁶ This study was a secondary analysis of deidentified data; institutional review board review was not needed.

Among children and adolescents ages 0–19 years, 17,593 cases were reported in 2019, 16,622 in 2020, and 16,542 in 2021, with age-adjusted incidence rates of 215.3 per million, 205.0, and 206.8, respectively (Table 1). The IRR comparing 2019 to 2020 was 0.95 (95% confidence interval [CI], 0.93–0.97), indicating a 5% decrease, and comparing 2019 to 2021 was 0.96 (0.94–0.98), indicating a 4% decrease. Among those 0–19 years at diagnosis, rates were significantly lower in 2020 compared to 2019 overall, for both males and females, for ages 10–14 and 15–19 years, for White and Hispanic children and adolescents, and by cancer type for thyroid cancer, melanoma, and nonmalignant brain tumors. By cancer type, only thyroid cancer and nonmalignant central nervous system (CNS) neoplasm rates were significantly lower for both 2020 and 2021 than 2019; both of these tumor types had been previously increasing (Fig. 1). Lymphoma was significantly lower in 2021 than 2019, but this cancer type had a decreasing trend line independent of 2020 (Fig. 1: APC 2015–2021=−2.3).

Among young adults aged 20–39 years, 90,184 cases were reported in 2019, 84,058 in 2020, and 88,871 in 2021, with age-adjusted rates of 107.2 per 100,000, 99.5, and 104.9, respectively (Table 2). The IRR comparing 2019 to 2020 was 0.93 (95% CI, 0.92–0.94), indicating a 7% decrease, and was 0.98 (0.97–0.99) comparing 2019 to 2021, indicating a 2% decrease. Overall, rates were significantly lower in 2020 compared to 2019: for both males and females; for all four 5-year age groups between 20 and 39 years; for all race and ethnicity groups; and by cancer type for leukemia, lymphoma, malignant and nonmalignant CNS neoplasms, sarcomas, melanomas, and several types of carcinoma. For young adults, multiple cancer types had lower rates in both 2020 and 2021 compared to 2019, including lymphoma, malignant and nonmalignant CNS neoplasms, sarcomas, melanoma, thyroid cancer, lung cancer, and carcinoma of the genital sites. Of these cancer types, lymphoma, malignant and nonmalignant CNS neoplasms, and melanoma had a decreasing trend independent of 2020 (Fig. 2).

By diagnosis month, for ages 0–19 years, the lowest counts during 2019 through 2021 occurred in March and April of 2020 (Fig. 3). April 2020 was the lowest count month for both sexes and for childhood and adolescent age groups, although the difference in counts was most pronounced in adolescents.

Similar patterns were seen in young adults aged 20–39 years; the lowest case count months from 2019 through 2021 occurred in March–May of 2020. Cancer case counts were lowest in April 2020 for both sexes and all 4 age groups, although the difference in counts was most pronounced in ages 35–39 years. For all ages 0–39, cancer case counts within each age group were similar in the months June–December across all 3 years. By diagnosis month and by cancer type (Fig. 4), the lowest count month was March 2020 for pediatric patients with leukemia (cases = 260), whereas April 2020 was the lowest count month during 2019–2021 for all other age groups and cancer types.

By stage at diagnosis, for ages 0–19 and 20–39 years, rates decreased for localized and regional stage, but rates were stable for distant stage when comparing 2019 to 2020. For ages 0–19 years, by month, the percentage of cases with distant stage was highest in April and June of 2020 compared to any other month during 2019 to 2021 (Supplemental Figure S1). For ages 20–39 years, the percentage of cases with distant stage was highest in April 2020 compared to any other month in 2019 to 2021. By US Census region, for ages 0–19 years, cases decreased in 2020 and 2021 compared to 2019 in the Midwest and South regions. For ages 20–39 years, cases decreased in 2020 for all regions and only for the Northeast and West in 2021. For pediatric ages, the IRR for 2020 showed a significant decrease in metropolitan but not nonmetropolitan counties. For ages 20–39 years, the IRR for 2020 showed a significant decrease in both metropolitan and nonmetropolitan counties, but it was lower in metropolitan counties.

For all ages combined (0–85+ years), rates were 10% lower in 2020 than 2019 and 4% lower in 2021 than 2019 (Supplemental Table S1). The percent decrease between 2019 and 2020 was higher in every 10-year age group ≥40 years (range 9%–10%) than in every age group <40 years (range 2%–7%). Decreases in 2020 were more pronounced for ages ≥40 years compared to ages <19 and 20–39 years (Supplemental Figure S2).

Our large population-based study demonstrated that reported cancer incidence in the US decreased by 5% and 7% between 2019 and 2020 for patients aged 0–19 and 20–39 years, respectively, which was a smaller rate decrease than all ages combined (10%) (Supplemental Table S1). Significant decreases between 2019 and 2020 were seen only for certain cancer types and were not seen for ages 0–4 or 5–9 years. For ages 0–39 years by month, new diagnoses were lowest during March or April 2020 but then returned to prepandemic levels during the second half of 2020.

Previous studies reported that for all ages combined (0–85+ years), cancer incidence rates decreased by 9% to 11% from 2019 to 2020, and decreases in 2020 were more pronounced for older ages compared to pediatric and young adult ages.^9,37,38 Notably for this study, rates did not significantly decrease comparing 2019 to 2020 for children diagnosed at ages 0–4 or 5–9 years, which is consistent with previous reports.^16,37 Less pronounced decreases in pediatric ages could be related to differences in health care utilization, access to care, and cancer type³⁹—especially because cancer screening is not relevant to most pediatric cancer types.

By cancer type in ages 0–19 years, significant decreases in 2020 were seen only for thyroid cancer, malignant melanomas, and nonmalignant CNS neoplasms. These three cancer types were also among the types with the largest 2020 decreases among young adults ages 20–39 years. Thyroid and melanoma cancers are often diagnosed in outpatient settings in adolescents and adults, and decreases might be related to decreased routine health care utilization during 2020—including reduced imaging examinations, and discovery of thyroid cancer on imaging orders for reasons unrelated to cancer.^5,6,9 Similarly, decreases in nonmalignant CNS neoplasms could be related to decreased incidental diagnoses secondary to decreased access to care or imaging.³⁹ The most common cancer types diagnosed in pediatric and young adults—such as leukemia, lymphoma, and malignant brain cancer, which are not subject to population-wide screening recommendations (e.g., for cervical, prostate, or lung cancer)^40,41—might be less affected by decreases in routine primary care visits compared to cancers associated with population-level screening recommendations.

Cancer incidence rates in 2021 were significantly lower than in 2019. However, rates were decreasing independently of 2020 rates for pediatric lymphoma, malignant CNS neoplasms, retinoblastoma, and germ cell tumors, and for ages 20–39 years, for lymphoma, malignant CNS neoplasms, and melanoma. For some pediatric cancer types that are associated with imaging-based or incidental diagnoses, such as thyroid cancer and nonmalignant CNS neoplasms, decreased rates in both 2020 and 2021 compared to 2019 might reflect continued impact of the COVID-19 pandemic and decreased utilization of health care services.^39,42 These findings are consistent with data of mostly adult ages showing decreased cancer rates in 2020 and 2021 (with 2020 decreases being most pronounced) compared to 2019, and the decreased rates vary by cancer type.⁴³

Like monthly patterns seen for cancer types mostly affecting adults,^5,6 for ages 0–39 years, decreases during the beginning of the COVID-19 pandemic in early 2020 generally were not present in the second half of 2020. Notably, in June 2020 for ages 0–19 years, lymphoma counts were higher than in any other month in 2019, 2020, or 2021. In adults, decreases in new diagnoses of cancer in early 2020 were likely related to decreases in screening, decreases in surgical procedures, or other disruptions in care.^3,5,44,45 Past studies have described reduced pediatric and young adult emergency department^46–49 and primary care or routine visits^49–51 during the spring months of 2020, which rebounded somewhat in later months of 2020. Emergency room visits were lower in 2020 than 2019 for all ages, but the percent decrease was larger for children compared to adolescents^46,47 and children compared to adults.⁴⁸ A delay in medical care could explain the cancer incidence pattern observed in the spring of 2020 for pediatric and young adult ages, and this pattern may have impacted solid tumors or tumors with an outpatient diagnosis more than other cancer types.^14,39 Some delays in diagnosis might have been related to symptom overlap of fatigue, distress, and disturbed sleep attributed to the COVID-19 pandemic.³⁹ Decreases in malignant and nonmalignant CNS tumors were similar to patterns described for ages 0–39 years by Cioffi et al., whereby the largest decreases in 2020 were seen in April 2020, and decreases in nonmalignant CNS tumors were greater than among people with malignant brain cancer.³⁹

Previous data looking at all ages showed decreases in 2020 for all racial and ethnic groups, which is consistent with findings here for ages 20–39 years.⁶ However, in this study we did not observe decreases in pediatric ages for Black, American Indian/Alaska Native, and Asian and Pacific Islander groups. By stage, the rate of patients diagnosed at distant stage did not decrease from 2019 to 2020 for both pediatric and young adult ages; however, the rate of localized and distant stage decreased from 2019 to 2020. This is consistent with previous data of adult cancers showing larger decreases in localized stage compared to distant stage between 2019 and 2020.^11,43,52 It is possible that for pediatric cancer, some localized stage cancers, especially those diagnosed incidentally, might have not been detected or had delayed diagnosis, while distant stage cases were not widely missed during 2020, potentially due to greater likelihood of symptoms.³⁹ Differences by US Census region or metropolitan status might be related to differing impacts of the pandemic on cancer reporting, access to care, or utilization of route health services.⁵³

Our study must be considered in the light of certain limitations. First, the COVID-19 pandemic disrupted cancer reporting in both hospitals and central cancer registries;⁵⁴ it also may have impacted reported data as early as the 2018 diagnosis year and may have continued through the 2021 diagnosis year. It is possible that disruptions resulted in some underreporting of cases throughout this study’s reporting period. It is not clear how potential underreporting might have affected cancers among those aged 0–39 years differently from cancers among older individuals. Some cancer types are reported late, especially those diagnosed at outpatient clinics, and might not be included in this database version. However, including 2021 data in this study allowed for additional time to add late-reported cases diagnosed during 2020. Second, our evaluation of stage excluded leukemia (because almost all cases were coded as distant) and germ cell tumors (>45% of cases were coded as unknown or blank stage). Third, merged summary stage captured in the USCS database does not match clinical staging for many cancer types. Finally, not all nonmalignant cases are collected by central cancer registries, and nonmalignant CNS cases were included in this study only beginning in 2005.⁵⁵

This study provides context for how the COVID-19 pandemic may have impacted pediatric and young adult cancer diagnoses. The COVID-19 pandemic did not seem to have as large of an impact on cancer incidence in pediatric patients compared to adult patients; the impact was greater for cancers diagnosed in adolescents and young adults and in certain cancer types, such as thyroid and nonmalignant brain tumors. Analysis of incidence data over subsequent years will help determine if patterns observed here during 2020 and 2021 resulted from decreased diagnosis as opposed to decreased reporting. Nevertheless, diagnosis patterns observed in this study could be helpful to researchers in interpreting cancer mortality and survival data during this time period, evaluating other markers of cancer care (e.g., intensive care unit admissions), assessing future incidence trends, and informing responses to future public health crises that affect pediatric and young adult populations.^43,56,57