Survival trends among adolescent and young adults diagnosed with cancer in the United States: Comparisons to children and older adults

Abstract

PURPOSE:

Although data from 1975–1997 revealed a gap in cancer survival improvement in adolescents and young adults (AYAs; 15–39 years) compared with children and older adults, more recent studies have reported improvements in AYA cancer survival overall. The current analysis provides an update of 5-year relative survival and cancer survival trends among AYAs compared with children and older adults.

METHODS:

We obtained data from the National Cancer Institute Surveillance, Epidemiology and End Results Program for 17 regions to obtain recent (2010–2018) 5-year relative survival estimates by cancer type, stage, sex, and race/ethnicity by age group. In addition, we calculated 5-year relative survival trends during 2000–2014.

RESULTS:

Across 33 common AYA cancers, AYAs and children had high 5-year relative survival (86%) and experienced similar survival improvements over time (average absolute change: AYAs, 0.33%; children 0.36%). Among AYAs, 73% of cancers had improvement in 5-year relative survival since 2000. Despite this overall progress, we identified cancers where survival was worse in AYAs than younger or older patients and cancers that have had either a lack of improvement (osteosarcoma and male breast) or decreases in survival (cervical and female bladder) over time. Further, males had inferior survival to females for all cancers, except Kaposi sarcoma and bladder cancer, and non-Hispanic Blacks/African Americans experienced worse survival than other racial/ethnic groups for many cancers considered in this study.

CONCLUSION:

Future studies should focus on identifying factors impacting survival disparities by age, sex and race/ethnicity. Differences in biology, clinical trial enrollment, delivery of treatment according to clinical guidelines, and supportive and long-term survivorship care may account for the survival disparities we observed and warrant further investigation.

Background

Each year, there are approximately 90,000 new cases of cancer in adolescents and young adults (AYAs; 15–39 years of age at diagnosis) in the United States (US).¹ The gap in survival improvement among AYAs with cancer compared to children and older adults has been the focus of a number of national and international initiatives since the 2006 Report of the AYA Oncology Progress Review Group highlighted this with national Surveillance Epidemiology and End Results (SEER) cancer registry data from 1977–1997.² In 1975–1980, AYAs diagnosed with cancer generally had superior survival than those diagnosed at a younger or older age. However, after that time, this survival advantage diminished relative to children and older adults.^2–4 Subsequent studies have identified subgroups of AYAs with less improvement in survival, including the recognition that the HIV/AIDS epidemic impacted the gap in survival improvement among AYAs with specific cancers (e.g., Kaposi sarcoma, non-Hodgkin lymphoma (NHL)), particularly during 1980s–90s, and outcomes differed by cancer site.^5–8

Using SEER data from 1992–2006, we previously reported improvements in 5-year relative survival for a number of cancers, but also identified cancers where survival improved less in AYAs than in children (acute myeloid leukemia (AML), medulloblastoma) and older adults (thyroid cancer, AML, acute lymphocytic leukemia (ALL), pancreatic, astrocytoma, stomach, kidney, oral cavity and pharynx, Hodgkin lymphoma (HL), ovarian, sarcoma, liver).⁷ A study looking at long-term survival trends (1975–2016) among AYAs for 9 cancers with the highest mortality found that 5-year relative survival has improved for brain and other central nervous system (CNS), colorectal, ovarian, lung, AML and NHL; however, for breast, cervical, and sarcoma, there were no survival improvements.⁹ Data from California (1988–2014) found that, although the overall gap in survival improvement for AYAs with cancer had closed, progress in cancer survival differed not only by cancer site, but also by stage at diagnosis and sociodemographic characteristics.⁸ Specifically, survival differences between early and advanced stage disease widened over time and improvements in invasive brain and other CNS, colorectal, melanoma, stomach, and testicular cancers were only observed among non-Hispanic White patients, with no significant improvements seen among other racial/ethnic groups.⁸

Cancer site-specific analyses are important given that the distribution of cancer varies substantially for AYAs and is unique compared to children and older adult populations.¹ Even among AYAs, the three most common cancers differ by age: thyroid, HL and brain/CNS 15–19 year-olds; thyroid, testicular and melanoma among 20–29 year-olds; and female breast cancer, thyroid and melanoma among 30–39 year-olds.¹ Recent data (2000–2019) suggest that incidence rates of invasive cancers in AYAs are rising, with incidence rates increasing in 84% of cancers in female and 51% of cancers in male AYAs over this time period.¹⁰ While some cancers have excellent survival (e.g., thyroid, testicular, HL), others have substantial room for improvement in AYAs (e.g., leukemia, colorectal, osteosarcoma).^{1, 7, 8} No studies, to our knowledge, have considered the survival gap in AYAs with national data examining trends after 2000 when the use of highly effective HIV/AIDS treatments became more widespread and outcomes for HIV/AIDS-associated cancers improved.^{5, 9, 11} Therefore, using National Cancer Institute (NCI) SEER data we provide an update of recent (2010–2018) 5-year relative survival estimates by cancer type, stage at diagnosis, sex, and race/ethnicity among AYAs compared to children and older adults. In addition, we calculated national survival trends of 5-year cancer survival during 2000–2014. Findings from this update can identify persistent and new inequities in cancer survival, informing priority areas for AYA oncology.

Methods

Data source and study population

We obtained data from the NCI’s SEER 17 Research Plus database, 2021 November submission (2000–2019), which covers about 26.5% of the US population based on the 2020 census.¹² Registries that participate in SEER 17 include Alaska Native Tumor Registry, Connecticut, Atlanta, Greater Georgia, Rural Georgia, San Francisco-Oakland, San Jose-Monterey, Greater California, Hawaii, Iowa, Kentucky, Los Angeles, Louisiana, New Mexico, New Jersey, Seattle-Puget Sound, and Utah. We included AYAs diagnosed with a first primary, invasive cancer from 2000–2018 and followed through December 31, 2019. We also included children (0–14 years) and older adults aged ≥40 years for comparisons of survival improvement over time. We used the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) histology and primary site codes, based on the SEER AYA Site Recode 2020 Revision,¹³ to identify 33 common invasive cancers in AYAs (listed in Table 1). Head and neck cancer (HNC) in our study included oral cavity, lip, and pharynx cancers (excludes nasopharynx cancer) and soft tissue sarcoma (STS) included fibromatous neoplasms, liposarcoma, synovial sarcoma, leiomyosarcoma, malignant gastrointestinal stromal tumor, epithelioid sarcoma, desmoplastic small round cell tumor, and other soft tissue sarcomas. Patients diagnosed by death certificate or autopsy only and those with no survival time were excluded from the analyses.

Table 1.

Five-year relative survival in children, adolescents and young adults (AYAs), and older adults diagnosed with cancer, 2010–2018; Surveillance, Epidemiology, End Results (SEER) 17 Research Program

	Children (0–14)			AYAs (15–39)			Older adults (40+)
Cancer site*	N	5-y RS (%)	SE (%)	N	5-y RS (%)	SE (%)	N	5-y RS (%)	SE (%)
All cancer types	15,369	85.5	0.3	158,941	86.3	0.1	1,974,939	64.4	0.0
All cancers except Kaposi sarcoma	15,365	85.5	0.3	157,787	86.5	0.1	1,972,847	64.4	0.0
Acute lymphoblastic leukemia	6,381	91.6	0.4	2,754	63.2	1.0	3,300	32.5	1.0
Acute myeloid leukemia	1,123	67.7	1.5	3,037	62.4	1.0	18,376	21.8	0.4
Chronic myeloid leukemia	182	85.1	2.9	1,853	90.4	0.8	9,861	66.5	0.6
Non-Hodgkin lymphoma	1,616	91.1	0.8	9,328	85.6	0.4	105,197	72.7	0.2
Hodgkin lymphoma	809	98.5	0.5	8,870	95.9	0.2	8,045	78.7	0.6
Myeloma	#			700	83.2	1.7	43,971	55.9	0.3
Oligodendroglioma	#			1,234	85.8	1.1	1,942	66.5	1.3
Glioblastoma	213	17.8	2.8	1,202	22.4	1.5	20,843	4.6	0.2
Ependymoma	429	82.7	2.2	506	91.0	1.5	840	89.7	1.4
Medulloblastoma and other CNS	1,048	66.2	1.6	370	68.9	2.8	121	53.8	5.0
Osteosarcoma	631	68.9	2.1	885	64.0	1.8	574	41.3	2.4
Chondrosarcoma	#			455	90.7	1.6	1,546	75.8	1.4
Ewing sarcoma (bone)	393	76.9	2.4	444	55.3	2.7	#
Ewing sarcoma (soft tissue)	119	84.7	3.6	344	60.8	2.9	196	42.3	4.1
Soft tissue sarcoma**	612	85.8	1.6	4,675	80.2	0.7	25,962	67.8	0.4
Kaposi sarcoma	#			1,154	68.7	1.5	2,092	77.6	1.2
Testicular (germ cell)	133	100.0	0.0	14,549	95.7	0.2	5,734	95.2	0.4
Ovarian (germ cell)	246	95.8	1.5	855	94.5	0.9	147	72.6	4.1
Ovarian (non-germ)	#			2,374	76.8	1.0	34,228	48.3	0.3
Melanoma	296	96.0	1.2	15,787	95.5	0.2	126,906	91.8	0.1
Thyroid@	519	99.6	0.3	27,791	99.8	0.1	67,658	97.9	0.1
Head and neck @&	#			1,709	79.1	1.1	59,143	67.3	0.2
Stomach@	#			1,779	28.1	1.2	40,284	28.9	0.3
Colon@	223	94.1	2.0	6,722	72.8	0.6	164,784	64.8	0.2
Rectal@	#			3,613	70.9	0.9	75,322	67.9	0.2
Liver and intrahepatic bile ducts@	#			907	34.2	1.8	60,967	20.1	0.2
Pancreas@	#			1,216	53.7	1.6	72,558	10.7	0.1
Lung@^	#			2,188	49.2	1.2	293,811	22.8	0.1
Breast	#			24,302	86.1	0.3	419,948	91.1	0.1
Cervical@	#			7,261	80.5	0.5	19,227	62.6	0.4
Uterine corpus@	#			3,827	93.4	0.5	91,688	84.4	0.2
Kidney@	#			4,914	89.6	0.5	90,287	77.1	0.2
Bladder@	#			1,336	89.3	0.9	109,284	77.6	0.2

Abbreviations: CNS, central nervous system; IBD, intrahepatic bile ducts; N, number; RS, relative survival; SE, standard error.

^*Cancer site based on the SEER AYA Site Recode 2020 Revision.

^**Soft tissue sarcoma includes fibromatous neoplasms, liposarcoma, synovial sarcoma, leiomyosarcoma, malignant gastrointestinal stromal tumor, epithelioid sarcoma, desmoplastic small round cell tumor, and other soft tissue sarcomas.

^@All carcinomas.

^&Head and neck cancer includes oral cavity, lip, and pharynx cancers (excludes nasopharynx cancers).

^{^}Includes bronchus and trachea.

^#Cancer sites with less than 100 patients are not presented.

Statistical analysis

We used Surveillance Research Program, NCI SEER*Stat software (www.seer.cancer.gov/seerstat) version 8.4.0.1. to estimate 5-year relative survival (Ederer II Actuarial method) by age group, sex, and race/ethnicity (non-Hispanic (NH) White, NH Black/African American, NH American Indian/Alaska Native, NH Asian/Pacific Islander (PI), Hispanic), overall and for each cancer site from 2010–2018. Relative survival, a net survival measure representing cancer survival in the absence of other causes of death, is defined as the ratio (expressed as a percent) of the proportion of observed survivors in a cohort of cancer patients to the proportion of expected survivors.¹⁴ We also estimated 5-year relative survival for patients diagnosed with distant stage disease, using the SEER Combined Summary Stage (2004+). All leukemias are classified as unstaged in our analyses. We required a minimum of 100 cases of each invasive cancer (Table 1) or 50 cases of each cancer within subgroups defined by distant stage, sex and race/ethnicity to present relative survival estimates in our analyses.

We calculated annual trends in 5-year relative survival and 95% confidence interval (CI) utilizing the NCI’s JPSurv software, a webtool that analyzes survival trends by single year of diagnosis.¹⁵ The joinpoint survival model is an extension of the proportional hazard model for survival where the effect of calendar year at diagnosis is linear on the log hazard of cancer death scale.¹⁶ The survival trend in each joinpoint segment is summarized as the average absolute change (AAC) in survival, which represents the average percentage point difference in cancer survival across a single joinpoint segment for those diagnosed in one calendar year versus the prior year.⁹ The joinpoint model allows for different linear trends between joinpoints (i.e., calendar years where trends in the hazard of cancer death change). We allowed a maximum of two joinpoints in our analyses and also provided the trend measures for the entire period of observation (2000–2014), except for distant disease, where stage information was available from 2004 onwards. For analyses with all cancers combined, we analyzed our cohort with and without Kaposi sarcoma.

Results

For AYAs diagnosed from 2010–2018, the 5-year relative survival for all cancers was 86.3% compared with 85.5% in children and 64.4% in older adults (Table 1). Results were similar when Kaposi sarcoma was excluded. Over half (51.5%) or 17 of the 33 cancers in AYAs had a 5-year relative survival greater than 80%, including nine cancers with >90%. Four cancers (glioblastoma, stomach, liver and lung) had <50% 5-year relative survival. The most common AYA cancers – NHL, HL, testicular, melanoma, thyroid and breast cancers –comprise 64% of all cancers in AYAs and all had a relative survival >85%. Female AYAs generally have better 5-year relative survival than males for all cancers, except Kaposi sarcoma and bladder cancer (Supplementary Table S1). Differences in survival by race/ethnicity was also observed in AYAs, with 5-year relative survival generally higher in NH White and NH Asian/PI patients than NH Black/African American, Hispanic and NH American Indian/Alaskan Native AYAs (Supplementary Table S2).

AYAs had a much lower 5-year relative survival (>20% difference) than children for four cancers, including ALL, Ewing sarcoma (bone), Ewing sarcoma (soft tissue), and colon cancer. Survival was slightly lower for AYAs than children for AML, NHL, HL, osteosarcoma, STS and testicular cancer. On the other hand, survival was superior for AYAs compared to children with chronic myeloid leukemia (CML), glioblastoma, ependymoma and medulloblastoma/CNS. Compared with older adults, AYAs generally had better 5-year relative survival across the 33 cancers, except for Kaposi sarcoma and breast cancer.

Of the cancers considered in this study, 25 had at least 50 AYAs with distant stage disease at diagnosis, excluding leukemias (Table 2). Most cancers (68%) with distant disease at diagnosis had 5-year survival <50%, whereas six cancers had survival <20%: stomach, liver, pancreas, lung, kidney, and bladder. NHL, HL, myeloma, thyroid, testicular and ovarian (germ cell) cancers diagnosed at a distant stage had 5-year relative survival >74%. Children had better survival for the few distant stage cancers that could be assessed (NHL, HL, osteosarcoma, Ewing sarcoma (bone) and STS), but AYAs had better survival than older adults across all distant stage cancers, except for stomach, kidney and bladder cancers.

Table 2.

Five-year relative survival in children, adolescents and young adults (AYAs), and older adults diagnosed with distant cancer stage, 2010–2018; Surveillance, Epidemiology, End Results (SEER) 17 Research Program

	Children (0–14)			AYAs (15–39)			Older adults (40+)
Cancer site*	N	5-y RS (%)	SE (%)	N	5-y RS (%)	SE (%)	N	5-y RS (%)	SE (%)
All cancer sites	1,940	75.6	1.0	21,095	54.2	0.4	482,314	22.8	0.1
Non-Hodgkin lymphoma	860	88.3	1.2	3,911	76.6	0.7	54,043	65.6	0.3
Hodgkin lymphoma	328	97.7	1.0	3,293	92.8	0.5	3,822	71.1	0.9
Myeloma	#			559	79.6	2.0	42,059	55.1	0.3
Oligodendroglioma	#			#			#
Glioblastoma	#			#			330	2.7	1.0
Ependymoma	#			#			#
Medulloblastoma and other CNS	178	47.5	4.0	#			#
Osteosarcoma	154	38.1	4.5	180	24.7	3.7	141	7.9	2.9
Chondrosarcoma	#			#			199	29.6	3.9
Ewing sarcoma (bone)	129	58.8	4.9	183	37.5	4.1	#
Ewing sarcoma (soft tissue)	#			104	28.6	5.0	74	13.8	4.5
Soft tissue sarcoma**	68	33.8	7.2	608	23.4	2.2	4,303	20.8	0.8
Kaposi sarcoma	#			90	50.7	6.0	54	30.6	8.5
Testicular (germ cell)	#			1,779	75.5	1.1	644	71.2	1.9
Ovarian (germ cell)	#			137	81.8	3.6	50	44.9	7.5
Ovarian (non-germ cell)	#			742	48.7	2.1	20,869	30.7	0.4
Melanoma	#			410	36.3	2.6	5,754	27.9	0.7
Thyroid@	#			436	90.1	1.6	2,864	49.4	1.1
Head and neck@&	#			204	55.8	3.7	9,554	39.2	0.6
Stomach@	#			955	4.0	0.8	15,217	4.3	0.2
Colon@	#			1,585	23.6	1.3	37,493	14.2	0.2
Rectal@	#			873	21.8	1.8	14,400	16.5	0.4
Liver and intrahepatic bile ducts@	#			249	5.4	2.4	11,261	2.2	0.2
Pancreas@	#			476	18.1	2.0	38,655	2.9	0.1
Lung@^	#			1,182	19.3	1.4	165,832	6.6	0.1
Breast@	#			1,915	41.5	1.4	23,855	29.1	0.4
Cervical@	#			526	20.2	2.0	3,513	17.3	0.8
Uterine Corpus@	#			185	34.7	4.2	7,070	19.8	0.6
Kidney@	#			329	10.6	2.0	13,737	13.6	0.4
Bladder@	#			56	2.6	2.5	5,245	7.0	0.5

Abbreviations: CNS, central nervous system; y, year; N, number; RS, relative survival; SE, standard error.

^*Cancer site based on the SEER AYA Site Recode 2020 Revision. All cancer sites exclude leukemias because staging does not apply for these malignancies.

^**Soft tissue sarcoma includes fibromatous neoplasms, liposarcoma, synovial sarcoma, leiomyosarcoma, malignant gastrointestinal stromal tumor, epithelioid sarcoma, desmoplastic small round cell tumor, and other soft tissue sarcomas.

^@All carcinomas.

^&Head and neck cancer includes oral cavity, lip, and pharynx cancers (excludes nasopharynx cancers).

^{^}Includes bronchus and trachea.

^#Cancer sites with less than 50 patients are not presented.

Among AYAs, 25 (72.7%) cancers had significant improvements in 5-year relative survival over time, with 7 cancers (ALL, AML, myeloma, Ewing sarcoma (soft tissue), liver, pancreatic, and lung cancer) having an AAC in survival of >1.0% (Table 3; Figure 1; Supplementary Figure S1). Seventeen other cancers had survival improvements at <1.0% since 2000: CML, NHL, HL, oligodendroglioma, glioblastoma, medulloblastoma/CNS, Ewing sarcoma (bone), Kaposi sarcoma, melanoma, ovarian (non-germ cell), thyroid, HNC, stomach, colon, rectal, breast, and kidney cancers. For CML and Kaposi sarcoma, survival improved considerably (>2.0%) from 2000–2007 then plateaued thereafter. Improvements in HNC survival were from 2000–2003 (3.5%), while improvements in colon cancer survival were more substantial from 2012–2014 (1.9%). Survival decreased in AYAs for cervical and bladder cancers and showed no improvement for ependymoma, osteosarcoma, chondrosarcoma, STS, testicular, ovarian (germ cell) and uterine corpus cancers (selected types shown in Figure 2).

Table 3.

Average annual change in 5-year relative survival among children, adolescents, and young adults (AYAs), and older adults diagnosed with cancer; Surveillance, Epidemiology, End Results 17 Research Program

	Children (0–14)			AYAs (15–39)			Older adults (40+)
Cancer site*	Interval	AAC (%)	SE (%)	Interval	AAC (%)	SE (%)	Interval	AAC (%)	SE (%)
All cancers	2000–2014	0.33 &	0.04	2000–2007	0.55 &	0.05	2000–2009	0.58 &	0.04
				2007–2014	0.24 &	0.03	2009–2012	0.34 &	0.11
				2000–2014	0.36 &	0.02	2012–2018	0.63 &	0.07
							2000–2014	0.56 &	0.02
Acute lymphoblastic leukemia	2000–2014	0.31 &	0.07	2000–2014	1.29 &	0.14	2000–2014	1.24 &	0.09
Acute myeloid leukemia	2010–2014	0.51 &	0.18	2000–2014	1.10 &	0.12	2000–2014	0.70 &	0.04
Chronic myeloid leukemia	2000–2014	1.15 &	0.46	2000–2007	2.23 &	0.51	2000–2011	1.97 &	0.14
				2007–2014	−0.06	0.20	2011–2014	−0.12	0.22
				2000–2014	0.83 &	0.17	2000–2014	1.16 &	0.08
Non-Hodgkin lymphoma	2000–2014	0.48 &	0.11	2000–2014	0.80 &	0.05	2000–2003	1.65 &	0.28
							2003–2014	0.33 &	0.04
							2000–2014	0.55 &	0.04
Hodgkin lymphoma	2000–2014	0.34 &	0.09	2000–2014	0.26 &	0.04	2000–2014	0.47 &	0.08
Myeloma	#			2000–2014	1.42 &	0.20	2000–2014	1.46 &	0.04
Oligodendroglioma	#			2000–2014	0.49 &	0.19	2000–2014	0.76 &	0.15
Glioblastoma	2000–2014	−0.07	0.32	2000–2014	0.50 &	0.18	2000–2006	0.55 &	0.03
							2006–2014	0.15 &	0.04
							2000–2014	0.28 &	0.02
Ependymoma	2000–2014	1.05 &	0.28	2000–2014	0.08	0.19	2000–2014	0.50 &	0.18
Medulloblastoma and other CNS	2000–2014	0.38	0.20	2000–2014	0.82 &	0.33	2000–2014	1.26	0.67
Osteosarcoma	2000–2014	−0.09	0.28	2000–2014	−0.12	0.21	2000–2014	−0.03	0.30
Chondrosarcoma	#			2000–2014	0.31	0.23	2000–2014	−0.19	0.16
Ewing sarcoma (bone)	2000–2014	0.72 &	0.31	2000–2014	0.63 &	0.29	2000–2014	−1.45 &	0.61
Ewing sarcoma (soft tissue)	#			2000–2014	1.17 &	0.39	2000–2005	−6.16 &	0.43
							2005–2014	1.58 &	0.58
							2000–2014	−0.47	0.37
Soft tissue sarcoma**	2000–2014	0.15	0.19	2000–2014	−0.01	0.08	2000–2014	0.21 &	0.05
Kaposi sarcoma	#			2000–2007	2.05 &	0.56	2000–2014	0.68 &	0.15
				2007–2014	−0.31	0.36
				2000–2014	0.61 &	0.20
Testicular (germ cell)	#			2000–2014	−0.03	0.03	2000–2014	−0.04	0.04
Ovarian (germ cell)	#			2000–2014	−0.10	0.11	2000–2014	0.49	0.53
Ovarian (non-germ cell)	#			2000–2014	0.39 &	0.12	2000–2014	0.56 &	0.04
Melanoma	#			2000–2014	0.10 &	0.02	2000–2014	0.25 &	0.02
Thyroid@	#			2000–2014	0.03 &	0.01	2000–2011	0.25 &	0.03
							2011–2014	−0.08 &	0.03
							2000–2014	0.12 &	0.02
Head and neck@$	#			2000–2003	3.77 &	1.22		0.77 &	0.03
				2003–2014	−0.18	0.14
				2000–2014	0.48 &	0.17
Stomach@	#			2000–2014	0.37 &	0.15	2000–2014	0.62 &	0.04
Colon@	#			2000–2012	0.37 &	0.15	2000–2008	0.51 &	0.09
				2012–2014	1.93 &	0.30	2008–2011	−0.58 &	0.22
				2000–2014	0.89 &	0.09	2011–2014	0.19	0.12
							2000–2014	0.20 &	0.04
Rectal@	#			2000–2014	0.32 &	0.11	2000–2014	0.22 &	0.03
Liver and intrahepatic bile ducts@	#			2000–2014	1.33 &	0.21	2000–2006	1.18 &	0.13
							2006–2014	0.69 &	0.07
							2000–2014	0.85 &	0.04
Pancreas@	#			2000–2014	2.27 &	0.20	2000–2014	0.58 &	0.02
Lung@^	#			2000–2014	1.47 &	0.14	2000–2013	0.53 &	0.02
							2013–2014	1.67 &	0.09
							2000–2014	0.84 &	0.02
Breast@	#			2000–2006	0.73 &	0.11	2000–2014	0.12 &	0.01
				2006–2014	−0.04	0.05
				2000–2014	0.22 &	0.03
Cervical@	#			2000–2014	−0.22 &	0.07	2000–2014	−0.16 &	0.05
Uterine corpus@	#			2000–2014	0.07	0.06	2000–2014	0.04	0.03
Kidney@	#			2000–2014	0.55 &	0.08	2000–2008	1.42 &	0.09
							2008–2011	−0.13	0.19
							2011–2014	0.74 &	0.09
							2000–2014	0.90 &	0.04
Bladder@	#			2000–2014	−0.30 &	0.12	2000–2014	−0.02	0.02

Abbreviations: AAC, average absolute change; SE, standard error; CNS, central nervous system.

^*Cancer site based on the SEER AYA Site Recode 2020 Revision.

^**Soft tissue sarcoma includes fibromatous neoplasms, liposarcoma, synovial sarcoma, leiomyosarcoma, malignant gastrointestinal stromal tumor, epithelioid sarcoma, desmoplastic small round cell tumor, and other soft tissue sarcomas.

^@All carcinomas.

^$Head and neck cancer includes oral cavity, lip, and pharynx cancers (excludes nasopharynx cancer).

^{^}Includes bronchus and trachea.

^&AAC results in bold indicate significant increased or decreased trends (SEER JoinPoint Survival Analysis tool; maximum of 2 joinpoints allowed).

^#Could not be estimated dues to small number of patients.

Figure 1.

Average absolute change (AAC) in 5-year relative survival for selected cancer types showing significant improvement in adolescents and young adults: A. acute lymphoblastic leukemia, B. acute myeloid leukemia, C, liver, D. colorectal. Asterisk indicates a significant trend at P< 0.05 level.

Figure 2.

Average absolute change (AAC) in 5-year relative survival for selected cancer types showing no or little improvement in adolescents and young adults: A. female breast, B, cervical, C. female bladder, D. osteosarcoma. Asterisk indicates a significant trend at P< 0.05 level and NS indicates a non-significant trend.

When survival is considered for all cancers, AYAs had an AAC in survival similar to children from 2000–2014 (AYAs, 0.36%; children, 0.33%). However, survival improved more from 2000–2008 (0.55%) for AYAs than from 2008–2014 (0.24%) (Table 3). Older adults had a higher AAC (0.56%) in survival than AYAs and children. When considering specific cancers, survival improved less in AYAs than children for ependymoma. In addition, survival improved less in AYAs than in older adults for CML, HL, oligodendroglioma, STS, ovarian (non-germ cell), melanoma, HNC, stomach and kidney cancers.

Among AYAs with distant disease, 12 (48.0%) cancers had significant improvements in 5-year relative survival over time, with 7 cancers (NHL, myeloma, Ewing sarcoma (soft tissue), melanoma, lung, breast and uterine corpus cancers) having an AAC in survival of >1.0% (Supplemental Table S3). HL, stomach, colon, rectal and pancreatic cancers with distant disease had survival improvements at <1.0% since 2000.

Survival trends in AYA were similar for men and women for a number of cancers (Supplemental Table S1). However, men experienced 5-year relative improvements at <1.0% for oligodendroglioma, chondrosarcoma, Ewing sarcoma (bone) and rectal cancer, with no corresponding improvements seen in women. Conversely, women, but not men, experienced improvements at <1.0% for glioblastoma, HNC, stomach and breast cancers. Of note, survival for bladder cancer decreased by 1% in women (Figure 2) with no change over time for men.

When we considered trends by race/ethnicity for common cancers, we observed improvements in survival across all racial/ethnic groups for leukemia (ALL, AML, CML combined), NHL, lung and breast cancers (Supplemental Table S4). For melanoma, improvements in survival were observed among NH White and NH Asian/PI AYAs, with suggestions that survival is decreasing for NH Black/African American and NH American Indian/Alaskan Native AYAs. Survival after colorectal cancer appeared to improve in all racial/ethnic groups, with significant improvements observed for NH White, NH Asian/PI and Hispanic AYAs. For cervical cancer, we observed decreases in survival, which were most pronounced for NH White and Hispanic women.

Discussion

Using a large, population-based database in the US, we found that AYAs and children had high 5-year relative survival and experienced similar survival improvements over time. Both children and AYAs had somewhat less improvement in survival than older adults, a finding that may relate to the much lower 5-year relative survival and larger room for improvement in older adults for these cancers. Among AYAs in our study, 73% of cancers had improvement in 5-year relative survival since 2000, with fewer differences in survival improvement by age than previously reported.⁷ We also observed survival improvements in many cancers that were diagnosed at a distant stage of disease, which is promising. Despite this progress, we found two cancers in women, cervical and bladder, where 5-year relative survival decreased over time. In addition, there were still a number of cancers where AYAs experienced worse survival than younger and older patients, and where survival differed by sex and race/ethnicity among AYAs. Along with recent studies identifying increasing incidence for most AYA cancers,¹⁰ including cervical, stomach, kidney, breast, ALL and colon cancers, our study highlights priority areas to address survival gaps in AYA oncology (Figure 3).

Figure 3.

Gaps in 5-year relative survival in adolescents and young adults. Asterisk indicates cancers with significantly increasing cancer incidence (P<0.05) during 2000–2019. ¹⁰

Our findings are consistent with other recent reports^{8, 17} noting the improvement in survival among AYAs with cancer and the closure of the survival gap compared to younger and older adults. This progress is likely related to both effective treatments for HIV/AIDS-related cancers as well as national initiatives to improve outcomes in AYAs.^{2, 5, 9, 11, 17–19} These initiatives included efforts to address low clinical trial enrollment among AYAs through increased collaboration between pediatric and medical oncology cooperative groups across the NCI Clinical Trials Network to create clinical trials inclusive of the entire AYA age range that improve treatment and understanding of potentially unique biological characteristics of cancers impacting AYAs.^{17, 18, 20–22} Efforts have also recognized the importance of health insurance to improve access to care and the need for AYAs to receive treatment in specialized cancer centers for certain cancers.^23–31 In particular, the Affordable Care Act has increased insurance coverage for AYAs, who were historically the most highly uninsured population in the US.^{20, 32–34} Health insurance is not only important for early-stage diagnosis and cancer treatment, but also for long-term cancer survivorship care.^{23, 24, 35–37} Despite this overall progress, studies have also shown that improvements are not uniform across all AYAs, but rather vary by cancer type, stage at diagnosis and sociodemographic factors, including race/ethnicity, sex, and socioeconomic status.^{8, 17, 38}

When considering specific cancers, AYAs with ALL, AML, myeloma, Ewing sarcoma (soft tissue), liver, pancreatic and lung cancers had substantial improvements in survival (AAC in survival of >1.0%). Our prior study from 1992–2006 had previously identified AYAs with AML, ALL and liver cancer as having less improvement in survival than children (AML) or older adults (AML, ALL and liver),⁷ differences we no longer observe. Substantial improvements in survival among AYAs diagnosed with distant disease also were found over time, specifically for NHL, myeloma, Ewing sarcoma (soft tissue), melanoma, lung, breast and uterine corpus cancers, which is encouraging. Our findings for improvement in distant stage lung and breast cancers are consistent with a prior study that focused on selected AYA cancers with high mortality.⁹

Survival improved less in AYAs than children for ependymoma and older adults for CML, HL, oligodendroglioma, STS, ovarian (non-germ cell) cancer, melanoma and HNC, which may relate to AYAs having a higher 5-year relative survival for these cancers. However, survival also improved less in AYAs than older adults for stomach cancer, where survival is similarly poor (~28%) in both age groups, and kidney cancer, where survival is worse for AYAs with distant stage disease than older adults. Similar to our previous study (1992–2006), survival for both stomach and kidney cancers were found to improve less in AYAs than older adults.⁷ Studies using SEER data have identified increasing stomach cancer incidence rates in AYAs^{10, 39} and higher stomach cancer-specific mortality among AYAs compared to older adults.³⁹ The worse outcomes among AYAs with stomach cancer may be partially attributed to the higher rates of advanced stage at diagnosis, more aggressive disease progression, and higher relapse rates in AYAs than older adults.⁴⁰ Further, our findings for kidney cancer are concerning given that this is the most rapidly increasing cancer across all AYA age groups, particularly for high-grade, aggressive tumors.¹

AYAs had lower survival than younger or older patients for a number of cancers. Using 2010–2018 data, the 5-year relative survival for children with ALL, Ewing sarcoma (bone and soft tissue) and colon cancers were significantly better than that of AYAs diagnosed with the same disease. These survival differences have been noted previously and may relate, in part, to the unique biological/genomic characteristics of AYA cancers, not completing planned treatment, as well as reduced access to appropriate cancer treatment, supportive care for treatment-related toxicities and long-term survivorship care.^{17, 18, 23, 26, 41, 42} In addition, children had superior 5-year relative survival than AYAs for distant stage NHL, HL, osteosarcoma, Ewing sarcoma (bone) and STS. While there have been improvements in survival for most cancers, there was no improvement in survival observed for AYAs with osteosarcoma over time. This finding may relate to the lack of changes in osteosarcoma therapy over the past two decades.^{43, 44} Hopefully, the new Children’s Oncology Group clinical trials for osteosarcoma will lead to improved outcomes in the future.

Compared to older adults, 5-year relative survival after breast cancer (86%, AYAs; 91%, older adults) and Kaposi sarcoma (69%, AYAs; 78%, older adults) was lower in AYAs. Kaposi sarcoma in AYAs is more likely to be HIV/AIDS-associated than in older adults and carries an inferior prognosis.^{45, 46} The worse breast cancer outcomes in AYAs likely relate to breast cancer being more aggressive in this age group (larger, higher grade and more likely triple-negative disease).^{18, 47–49} In addition, AYAs experience a much higher rate of disease recurrence than older adults.⁴⁶ National Comprehensive Cancer Network guidelines suggest that AYAs with breast cancer, especially those with locally advanced and metastatic disease, be referred to tertiary/specialized cancer centers where clinical trials are offered, so they can be managed according to tumor stage and tumor biology rather than based on age at diagnosis.^{49, 50} Survival has been reported to be improved at specialized cancer centers for many cancer types,^{23, 26–31} providing further evidence that such centers have the potential to enhance outcomes in this population.

Female AYAs generally had better 5-year relative survival than males across cancers, consistent with prior reports in AYAs and across age groups.⁵ While survival improved for both females and males for many cancers, there were some notable differences. Despite the rarity of breast cancer in men, survival is worse than in females and there has been no improvement in 5-year relative survival since 2000. Consistent with our findings, a US national study showed that male patients were more frequently diagnosed with late-stage and estrogen receptor-positive breast cancer and less likely to receive radiotherapy and adjuvant endocrine therapy, experiencing worse survival than females across all stages.⁵¹ Further, consistent with previous reports,⁹ survival improvements in AYA women with all stages of breast cancer plateaued after 2006. However, substantial improvements in survival for distant stage breast cancer were observed. In contrast, survival for bladder cancer was lower in female AYAs, as found previously,⁵² and decreased by 1% per year with no change in survival observed over time for males. In population-based data, excess bladder cancer mortality in females compared to males in Sweden was restricted to muscle-invasive tumors, whereas worse bladder cancer survival in females compared to males in Japan was likely due to more advanced stage disease at diagnosis.^{53, 54} More studies are needed to better understand lack of improvement of this disease in AYAs.

While prior studies have not observed improvements in cervical cancer survival,^{7, 9} we observed that survival has decreased over time across most racial/ethnic groups, with decreases most pronounced for NH White and Hispanic women. While it is unclear why survival is decreasing, it may relate to screening delays or more aggressive histologic subtypes in AYAs, as approximately one-third of AYAs are diagnosed with regional or distant stage disease.¹ Whereas early-stage cervical cancer can be cured with appropriate surgery, chemoradiation or a combination of treatment modalities, advanced-stage disease is frequently incurable, with high relapse rate and treatment resistance.^{55, 56} Implementation of the new FIGO 2018 staging system is expected to improve the management of cervical cancer in women by distinguishing groups with different prognosis, specifically stage IB1 and IB2.^{57, 58} Although targeted therapy and immunotherapy may help to treat AYAs with recurrent or metastatic cervical cancer, efforts should be made to increase access to screening and early detection.

We identified racial/ethnic differences in 5-year relative survival, with generally higher survival among NH White and NH Asian/PI AYAs compared to NH Black/African American, Hispanic and NH American Indian/Alaskan Native AYAs. Our findings are consistent with numerous studies identifying racial/ethnic disparities in cancer survival among AYAs.^{8, 38, 59–64} Importantly, we observed improvements in survival across all racial/ethnic groups for leukemia, NHL, lung and breast cancers. However, for melanoma, improvements in survival were only observed among NH White AYAs, as found previously.⁸ Particularly concerning is the suggestion that survival decreased among NH Black/African Americans and NH American Indian/Alaskan Native AYAs with melanoma over time, consistent with a recent study suggesting that racial/ethnic melanoma survival disparities are widening.⁶⁵ While a prior study in California only observed improvements in colorectal cancer survival among NH White patients,⁸ we observed improvements across all racial/ethnic groups, with significant improvements for NH White, NH Asian/PI and Hispanic AYAs. Over the last 20 years, advances in surgery techniques (e.g., minimally invasive laparoscopic or robotic surgery), new targeted therapies (e.g., bevacizumab for metastatic colon cancer), a multidisciplinary approach, and refinement of treatment protocols, may have all contributed to the colorectal cancer survival improvement we observed in AYAs. ^66–70

This study is subject to some limitations. Because at least 5-years of follow-up are necessary to calculate relative survival trends, our analyses are limited to patients diagnosed through 2014; thus, more recent changes in survival are not presented. In addition, due to small numbers of AYA patients, survival estimates for some types of cancers diagnosed at a distant stage are not presented. For the same reason, some cancer sites were combined or not presented in analyses of survival trends by race/ethnicity. Although socioeconomic status has been found to be associated with survival among AYAs,^{17, 38} we lacked this information to consider these associations in our study. A major strength of this study is the inclusion of a large number of patients from population-based cancer registries, comprising approximately 27% of the US population, who received their care across all types of institutions, increasing the generalizability of our findings. Further, the use of NCI JPSurv analysis tool allowed us to investigate the magnitude of the change in 5-year survival over time and identify when changes in survival occurred.

Conclusion

The findings from our study suggest that there have been improvements in survival for most cancer types in AYAs, with survival improving to the same extent in AYAs as in children. Despite this overall progress, we identified cancers where survival was worse in AYAs than younger or older patients, and cancers that have had either a lack of improvement (osteosarcoma and male breast) or decreases in survival (cervical and female bladder) over time. Further, males had inferior survival to females for all cancers except Kaposi sarcoma and bladder cancer, and NH Blacks/African Americans experienced worse survival than other racial/ethnic groups for many of the cancers studied. Future studies should prioritize the identification of factors impacting these survival differences, especially in populations where there has been no improvement or decreases in survival (Figure 3). The substantial cost of cancer in AYAs, including lost productivity, loss of well-being and loss of life, underscore the need for improvements in treatment and access to care for these patients.¹⁹ Differences in biology, clinical trial enrollment, delivery of treatment according to clinical guidelines, supportive care and long-term survivorship care may account for the survival disparities we observed and warrant further investigation.

Supplementary Material

PV Appendix Figure 1

Supplemental Figure 1. Average absolute change (AAC) in 5-year relative survival for selected cancer types showing significant improvement in adolescents and young adults: A. Ewing sarcoma. B. myeloma, C. pancreas, D. lung. Asterisk indicates a significant trend at P< 0.05 level.