Early Estimates of SEER Cancer Incidence, 2014

Denise Riedel Lewis; Huann-Sheng Chen; Myles G Cockburn; Xiao-Cheng Wu; Antoinette M Stroup; Douglas N Midthune; Zhaohui Zou; Martin F Krapcho; Daniel G Miller; Eric J Feuer

doi:10.1002/cncr.30630

. Author manuscript; available in PMC: 2020 Sep 16.

Published in final edited form as: Cancer. 2017 Feb 14;123(13):2524–2534. doi: 10.1002/cncr.30630

Early Estimates of SEER Cancer Incidence, 2014

Denise Riedel Lewis ¹, Huann-Sheng Chen ¹, Myles G Cockburn ², Xiao-Cheng Wu ³, Antoinette M Stroup ⁴, Douglas N Midthune ⁵, Zhaohui Zou ⁶, Martin F Krapcho ⁶, Daniel G Miller ⁶, Eric J Feuer ¹

¹Surveillance Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Maryland

²Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California

³Louisiana Tumor Registry, Louisiana State University, New Orleans, Louisiana

⁴New Jersey State Cancer Registry, Rutgers Cancer Institute of New Jersey, New Brunswick, New Jersey

⁵Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, Bethesda, Maryland

⁶Information Management Systems Incorporated, Calverton, Maryland

^✉

Corresponding author: Denise Riedel Lewis, PhD, MPH, Surveillance Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, 9609 Medical Center Dr, Rm 4E538, MSC 9765, Bethesda, MD 20892; Fax: (240) 276-7908; lewisde@mail.nih.gov

AUTHOR CONTRIBUTIONS

Denise Riedel Lewis: Conceptualization, methodology, validation, formal analysis, investigation, resources, writing–original draft, writing–review and editing, visualization, supervision, and project administration. Huann-Sheng Chen: Conceptualization, methodology, validation, formal analysis, investigation, writing–original draft, writing–review and editing, supervision, and project administration. Myles G. Cockburn: Conceptualization, methodology, writing–review and editing, and visualization. Xiao-Cheng Wu: Writing–review and editing. Antoinette M. Stroup: Conceptualization, investigation, and writing–review and editing. Douglas N. Midthune: Conceptualization, methodology, software, validation, and writing–review and editing. Zhaohui Zou: Methodology, programming, formal analysis, validation, and writing-review and editing. Martin F. Krapcho: Conceptualization, methodology, software, validation, formal analysis, writing–review and editing, visualization, and supervision. Daniel G. Miller: Software, validation, formal analysis, writing–review and editing, and visualization. Eric J. Feuer: Conceptualization, methodology, validation, formal analysis, investigation, resources, writing–review and editing, visualization, and supervision.

PMCID: PMC7493462 NIHMSID: NIHMS1626508 PMID: 28195651

Abstract

BACKGROUND:

Cancer incidence rates and trends for cases diagnosed through 2014 using data reported to the Surveillance, Epidemiology, and End Results (SEER) program in February 2016 and a validation of rates and trends for cases diagnosed through 2013 and submitted in February 2015 using the November 2015 submission are reported. New cancer sites include the pancreas, kidney and renal pelvis, corpus and uterus, and childhood cancer sites for ages birth to 19 years inclusive.

METHODS:

A new reporting delay model is presented for these estimates for more consistent results with the model used for the usual November SEER submissions, adjusting for the large case undercount in the February submission. Joinpoint regression methodology was used to assess trends. Delay-adjusted rates and trends were checked for validity between the February 2016 and November 2016 submissions.

RESULTS:

Validation revealed that the delay model provides similar estimates of eventual counts using either February or November submission data. Trends declined through 2014 for prostate and colon and rectum cancer for males and females, male and female lung cancer, and cervical cancer. Thyroid cancer and liver and intrahepatic bile duct cancer increased. Pancreas (male and female) and corpus and uterus cancer demonstrated a modest increase. Slight increases occurred for male kidney and renal pelvis, and for all childhood cancer sites for ages birth to 19 years.

CONCLUSIONS:

Evaluating early cancer data submissions, adjusted for reporting delay, produces timely and valid incidence rates and trends. The results of the current study support using delay-adjusted February submission data for valid incidence rate and trend estimates over several data cycles.

Keywords: annual percent change, average annual percent change, early estimates of cancer incidence rates, population-based registry data, registry-specific delay-adjustment

INTRODUCTION

This report presents a first look at rates and trends for cases diagnosed through 2014 using data reported by state/regional registries in the Surveillance, Epidemiology, and End Results (SEER) program in February 2016. This marks the third year that the National Cancer Institute has published preliminary estimates of SEER cancer incidence rates and trends using the February data submission. The first 2 reports provided a preliminary indication of changes in cancer incidence trends.^1,2 It is interesting to note that the February data submission is 9 months earlier than the November submission. Using early data, we have shortened the interval between data collection and the reporting of results by 3 to 4 months. As the process becomes more routine, the interval could be shortened by up to 6 months or more.

Typically, there is a 2-year delay between the collection of the data and reporting of cancer incidence trends due to the time necessary to identify cases, abstract the data, and consolidate data from multiple sources. Each April, cancer incidence data are published and public-use cancer surveillance data from state/regional registries are made available via the Cancer Statistics Review and SEER*Stat. These data are based on the previous year’s November data submission, and therefore the most recent diagnosis year (DY) is 22 months after the close of the most recent DY.

Beginning in 2011, SEER registries were required to submit preliminary data each February in advance of the regular November submission. This requirement enables the registries to provide preliminary estimates of cancer incidence rates and trends approximately 6 months to 9 months earlier. In 2015, we published early estimates of SEER cancer incidence rates and trends for the 4 most common cancer sites (ie, lung and bronchus, colon and rectum [CRC], female breast, and prostate) plus skin melanoma.¹ These estimates for cancers diagnosed through 2012 were derived using data submitted 9 months earlier in February 2014 instead of the usual November 2014 submission. In early 2016, we published the rates and trends for cases diagnosed through 2013 based on the February 2015 data submission, and cervical cancer, thyroid cancer, and liver and intrahepatic bile duct (IBD) cancer were added.²

The current report presents cancer incidence trends based on the 2016 February submission and adds pancreas, kidney and renal pelvis, and corpus and uterus for adults, and all sites for childhood cancers between ages birth and 19 years, inclusive. These new sites represent cancers of concern that are important to monitor for early signs of trend changes. Critical to the analysis of February submission data is the necessary adjustment to account for incomplete reporting (ie, delay adjustment). We validated the delay-adjusted rates for 2013 and trends from 2000 through 2013 by comparing the prior February 2015 submission with the more complete November 2015 submission. Finally, we will discuss the analysis using 2016 February submission data.

MATERIALS AND METHODS

Data Source and Methodology for Assessing Trends

Cancer incidence data from 18 registries in the SEER program were used in the current analysis. Registries included in the analysis were San Francisco-Oakland, San Jose-Monterey, Los Angeles, Greater California, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, Atlanta, Alaska Natives, greater Georgia, rural Georgia, Kentucky, Louisiana, and New Jersey. All of these registries have regular data going back at least to the November 2002 submission from DY 2000 onward. Greater Georgia was included for the first time in these analyses, with the early years supplemented with data from the North American Association of Central Cancer Registries (NAACCR) with their permission. The preliminary data files submitted annually each February from 2011 to 2016, plus the usual data submitted in November in the same years, were used to analyze trends for cases diagnosed from 2000 through 2014.

The most recent DY included cases diagnosed up to 2 years before the submission year. For example, cases diagnosed before 2013 were included in the submission for both February and November 2015. The incidence data used in the analysis included DYs 2000 through 2014. These early estimates are based on delay-adjusted rates that were computed to project the eventual case count. The February submission delay model that produces these adjusted rates was modified this year to be more consistent with the one used for the usual November submission, as well as the North American Association of Central Cancer Registries Cancer in North America (CINA) report.³ Previously, the February delay model was estimated for all SEER 18 registries combined, but for this submission the estimation was performed for each registry separately (with rural Georgia and greater Georgia combined). Composite delay-adjusted rates then were calculated for the SEER 18 areas using SEER*Stat.⁴ The standard US population for the same updated US Census population file for the year of diagnosis for a given February submission and its prior November submission were used to calculate age-adjusted rates.

We assessed age-adjusted cancer incidence trends using the Joinpoint Regression Program, and procedures previously published.^1,2 The model allowed up to 2 joinpoints for rates from 2000 to 2014, with 3 maximum segments. The variability of the final segment’s annual percent change (APC) depends on its length, which can be as short as 3 data points. The average annual percent change (AAPC) is a weighted geometric mean of the APC over an interval of fixed length, and has been shown to be more stable than the APC for comparing trends from different series.⁵ Both 5-year and 10-year AAPC trends were reported, representing the most recently available 5-year and 10-year data, respectively. The test statistics follow the default values in the Joinpoint software, in which the Z-test is used to evaluate whether the AAPC (5-year and 10-year) is statistically significant, whereas the Student t test is used for APC trend in the last segment. In the event that an AAPC lies entirely within a single joinpoint segment, the AAPC is equal to the APC for that segment, and the test statistic for the AAPC is modified to be identical to that used for the APC using the Student t test instead of the Z-test. If a test statistic is statistically significant at the .05 level, it is labeled and denoted by ^a (in Tables), ^(in Figures).

Updates to the February Submission Reporting Delay Model

Delay-adjusted rates were produced according to prior methods.^1,2 In the 2016 release of the SEER Cancer Statistics Review,⁶ the National Cancer Institute adopted an updated delay model that included DY 2013. In this new model, an individual registry can have its own delay factor combining cancer site, race, age group (<50 years, 50-64 years, and ≥65 years), and sex if the average number of cases per DY and reporting year for a specific cancer site is at least 50. This threshold ensures that the model produces stable estimates tailored to the data characteristics of the contributing registry. For the February submission delay model with case counts below the threshold, the registry’s delay factor is replaced by the composite factor for all SEER 18 registries combined. In our analysis of the February 2016 submission data, skin melanoma and cancers of the prostate, cervix, thyroid, liver, pancreas, and kidney for Alaska Natives did not meet the criteria. To be consistent with the November submission delay model, we assumed that after 10 years of submissions, the data were complete and that DYs 2000 through 2002 were assigned a delay factor of 1.0.

RESULTS

Model Validation

Because a few new cancer sites were added and with a slightly different, registry-specific delay model, we validated the 2013 delay-adjusted rate and delay-adjusted trends for 2000 to 2013 using the February 2015 submission for comparison with the subsequent November 2015 submission. The observed age-adjusted and delay-adjusted rates for the selected sites, based on the November 2015 and February 2015 submissions for cases diagnosed in 2013, were compared (Table 1). The observed ratios of 2013 incidence rates for the November 2015 versus February 2015 submissions range from 1.02 to 1.16, with an average of 1.06 (Table 1). On average, November 2015 data had approximately 6.3% more cases than the February 2015 data for the selected cancer sites. However, the February undercount was corrected by the delay-adjusted model. The delay-adjusted incidence rate ratios for the February versus November submissions were closer to 1.0 (range, 0.978-1.031), with an average of 0.998, which is only a 0.2% difference between the 2 submissions. The range of November to February 2015 delay-adjusted rate ratios were comparable to those observed using the prior 2014 version of the delay-adjustment model. Overall, the ratios indicated that the delay model provided similar estimates of eventual counts using either February or November submission data.

TABLE 1.

SEER 18 Cancer Rates for Diagnosis Year 2013 From February and November 2015 Submissions: Validation Check

	2013 Age-Adjusted Rates, 2000 to 2013
	February 2015 Submission		November 2015 Submission		November 2015/ February 2015 Ratios		APC Trends, 2000 to 2013		AAPC Trends, 5-Year (2009 to 2013)/10-Year (2004 to 2013)
Cancer Site	Observed Rate	Delay- Adjusted Rate	Observed Rate	Delay- Adjusted Rate	Observed Rate	Delay- Adjusted Rate	February 2015 Submission	November 2015 Submission	February 2015 Submission	November 2015 Submission
All sites, male	438.1	484.0	469.1	487.0	1.071	1.006	−0.7^a (2000-2009) −2.7^a (2009-2013)	−0.7^a (2000-2009) −2.6^a (2009-2013)	−2.7^a/−1.6^a	−2.6^a/−1.6^a
All sites, female	387.5	420.0	404.6	415.6	1.044	0.989	−0.1 (2000-2013)	−0.1 (2000-2013)	−0.1/−0.1	−0.1/−0.1
Colon and rectum, male	42.1	45.0	43.8	44.9	1.042	0.998	−2.8^a (2000-2013)	−2.8^a (2000-2013)	−2.8^a/−2.8^a	−2.8^a/−2.8^a
Colon and rectum, female	32.1	34.3	33.5	34.2	1.043	0.997	−2.1^a (2000-2008) −3.0^a (2008-2013)	−2.1^a (2000-2008) −3.0^a (2008-2013)	−3.0^a/−2.6^a	−3.0^a/−2.6^a
Lung and bronchus, male	56.5	62.5	60.7	62.6	1.075	1.002	−2.0^a (2000-2009) −3.5^a (2009-2013)	−2.0^a (2000-2009) −3.5^a (2009-2013)	−3.5^a/−2.7^a	−3.5^a/−2.7^a
Lung and bronchus, female	44.0	48.6	46.8	48.3	1.065	0.993	0.0 (2000-2007) −1.8^a (2007-2013)	0.0 (2000-2007) −1.8^a (2007-2013)	−1.8^a/−1.2^a	−1.8^a/−1.2^a
Skin melanoma, male	26.8	29.9	29.3	30.4	1.091	1.017	1.9^a (2000-2013)	1.9^a (2000-2013)	1.9^a/1.9^a	1.9^a/1.9^a
Skin melanoma, female	15.7	17.4	17.1	17.6	1.087	1.011	1.2^a (2000-2013)	1.2^a (2000-2013)	1.2^a/1.2^a	1.2^a/1.2^a
Female breast	122.9	128.6	125.9	128.1	1.025	0.996	−2.4^a (2000-2004) 0.4 (2004-2013)	−2.4^a (2000-2004) 0.4 (2004-2013)	0.4/0.4	0.4/0.4
Prostate	97.7	109.3	106.6	111.0	1.091	1.016	−2.1^a (2000-2009) −13.6 (2009-2013)	−1.8^a (2000-2008) −9.4^a (2008-2013)	−8.0^a/−4.8^a	−7.5^a/−4.4^a
Cervix	6.8	7.2	7.0	7.2	1.039	0.994	−4.3^a (2000-2003) −0.7 (2003-2008) −2.2^a (2008-2013)	−4.3^a (2000-2003) −0.7 (2003-2008) −2.2^a (2008-2013)	−2.2^a/−1.5^a	−2.2^a/−1.6^a
Thyroid, male	7.2	7.6	7.4	7.6	1.028	0.996	6.2^a (2000-2009) 3.3^a (2009-2013)	6.2^a (2000-2009) 3.2^a (2009-2013)	3.3^a /4.9^a	3.2^a/4.8^a
Thyroid, female	21.4	22.6	21.8	22.2	1.017	0.983	6.9^a (2000-2009) 2.9^a (2009-2013)	6.4^a (2000-2006) 7.7^a (2006-2009) 2.3^a (2009-2013)	2.9^a/5.1^a	2.3^a/5.0^a
Liver/IBD, male	11.7	13.9	13.3	14.1	1.140	1.015	4.2^a (2000-2009) 1.9^a (2009-2013)	4.1^a (2000-2009) 2.1^a (2009-2013)	1.9^a/3.1^a	2.1^a/3.2^a
Liver/IBD, female	4.0	4.8	4.7	5.0	1.163	1.031	2.9^a (2000-2013)	3.0^a (2000-2013)	2.9^a/2.9^a	3.0^a/3.0^a
Pancreas, male	13.0	14.8	14.2	14.7	1.091	0.991	1.0^a (2000-2013)	0.9^a (2000-2013)	1.0^a/1.0^a	0.9^a/0.9^a
Pancreas, female	10.1	11.5	11.0	11.4	1.092	0.988	1.0^a (2000-2013)	0.9^a (2000-2013)	1.0^a/1.0^a	0.9^a/0.9^a
Kidney and renal pelvis, male	20.6	22.7	21.6	22.4	1.049	0.990	3.2^a (2000-2008) −0.7 (2008-2011) 2.7 (2011-2013)	3.1^a (2000-2008) 0.3 (2008-2013)	1.0/1.8^a	0.3/1.5^a
Kidney and renal pelvis, female	10.1	11.1	10.5	10.8	1.040	0.978	3.8^a (2000-2007) 0.4 (2007-2013)	3.5^a (2000-2008) −0.4 (2008-2013)	0.4/1.5^a	−0.4/1.3^a
Corpus and uterus NOS	25.2	26.3	25.8	26.1	1.022	0.991	−1.5 (2000-2003) 1.5^a (2003-2013)	−1.4 (2000-2003) 1.4^a (2003-2013)	1.5^a/1.5^a	1.4^a/1.4^a
All sites (ages birth-19 y), male	17.6	19.0	18.2	18.8	1.036	0.992	0.6^a (2000-2013)	0.6^a (2000-2013)	0.6^a/0.6^a	0.6^a/0.6^a
All sites (ages birth-19 y), female	16.3	17.4	16.8	17.2	1.025	0.987	0.9^a (2000-2013)	0.9^a (2000-2013)	0.9^a/0.9^a	0.9^a/0.9^a

Open in a new tab

Abbreviations: AAPC, average annual percent change; APC, annual percent change; IBD, intrahepatic bile duct; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results.

The APC and AAPC trends were statistically significant at P<.05.

APC joinpoint trends for 5-year and 10-year APCs and AAPCs (for the years 2000-2013) (Table 1) for delay-adjusted rates based on the 2 submissions (February and November 2015) were nearly identical for all sites combined, CRC, lung and bronchus, skin melanoma, female breast, cervix, male thyroid, pancreas, and pediatric cancers among children between birth and age 19 years. APC estimates varied slightly for prostate cancer, liver/IBD for males and females, and corpus and uterus. Major changes, such as the decline in prostate cancer incidence, continued based on the most recent estimates, and the joinpoint shifted from 2011 to 2010 in delay models using the February 2016 submission, Table 2. The delay-adjusted rate for the DY 2014 was 97.5 per 100,000, down from 110.5 per 100,000 for the DY 2013 cases. The last joinpoint model segment was found to have a significant decline of 9.7% APC from 2010 to 2014. The 5-year and 10-year AAPC trends were −9.7% and −5.4%, respectively; those statistically significant trends represent trend accelerations observed through DY 2013, which were −8.0% and −4.8%, respectively. Although both data submissions demonstrated a significant increase in trends over time, the number of joinpoint segments changed for female thyroid from 2 segments (February 2015 submission: 2000-2009 APC, 6.9% and 2009-2013 APC, 2.9%) to 3 segments (November 2015 submission: 2000-2006 APC, 6.4%, 2006-2009 APC, 7.7%, and 2009-2013 APC, 2.3%). For male kidney and renal pelvis, the joinpoint model fit changed from 3 segments (2000-2008 APC, 3.2%, 2008-2011 APC, −0.7%, and 2011-2013 APC, 2.7%) to 2 segments (2000-2008 APC, 3.1% and 2008-2013 APC, 0.3%).

TABLE 2.

SEER 18 Cancer Rates for Diagnosis Year 2014 and Trends Through 2014 From the February 2016 Submission

Cancer Site	2014 Age-Adjusted Rates, 2000 to 2014		APC Trends, 2000 to 2014	5-Year AAPC Trends (2010 to 2014)	10-Year AAPC Trends (2005 to 2014)
Cancer Site	Observed Rate	Delay- Adjusted Rate	APC Trends, 2000 to 2014	5-Year AAPC Trends (2010 to 2014)	10-Year AAPC Trends (2005 to 2014)
All sites, male	419.7	463.0	−0.7^a (2000-2009) −2.9^a (2009-2014)	−2.9^a	−1.9^a
All sites, female	381.4	412.0	−0.1 (2000-2014)	−0.1	−0.1
Colon and rectum, male	41.5	44.4	−2.8^a (2000-2014)	−2.8^a	−2.8^a
Colon and rectum, female	31.9	33.9	−2.4^a (2000-2014)	−2.4^a	−2.4^a
Lung and bronchus, male	53.9	59.5	−2.0^a (2000-2009) −3.6^a (2009-2014)	−3.6^a	−2.9^a
Lung and bronchus, female	42.5	46.9	0.0 (2000-2007)	−1.8^a	−1.4^a
Skin melanoma, male	26.2	29.1	1.7^a (2000-2014)	1.7^a	1.7^a
Skin melanoma, female	16.1	17.8	1.2^a (2000-2014)	1.2^a	1.2^a
Female breast	120.4	125.8	−2.3^a (2000-2004) 0.3 (2004-2014)	0.3	0.3
Prostate	87.2	97.5	−1.8^a (2000-2010) −9.7^a (2010-2014)	−9.7^a	−5.4^a
Cervix	7.0	7.4	−3.8^a (2000-2003) −1.3^a (2003-2014)	−1.3^a	−1.3^a
Thyroid, male	7.1	7.5	6.3^a (2000-2009) 2.5^a (2009-2014)	2.5^a	4.2^a
Thyroid, female	21.1	22.1	6.4^a (2000-2006) 8.0^a (2006-2009) 1.9^a (2009-2014)	1.9^a	4.4^a
Liver/IBD, male	11.7	14.0	4.2^a (2000-2009) 1.8^a (2009-2014)	1.8^a	2.8^a
Liver/IBD, female	4.2	5.1	3.1^a (2000-2014)	3.1^a	3.1^a
Pancreas, male	12.9	14.7	0.9^a (2000-2014)	0.9^a	0.9^a
Pancreas, female	9.8	11.2	0.8^a (2000-2014)	0.8^a	0.8^a
Kidney and renal pelvis, male	19.9	21.8	3.1^a (2000-2008) 0.2 (2008-2014)	0.2	1.1^a
Kidney and renal pelvis, female	9.9	10.8	3.5^a (2000-2008) −0.4 (2008-2014)	−0.4	0.9^a
Corpus and uterus NOS	25.7	26.7	−1.4 (2000-2003) 1.4^a (2003-2014)	1.4^a	1.4^a
All sites (ages birth-19 y), male	17.8	19.3	0.6^a (2000-2014)	0.6^a	0.6^a
All sites (ages birth-19 y), female	16.6	17.6	0.9^a (2000-2014)	0.9^a	0.9^a

Open in a new tab

Abbreviations: AAPC, average annual percent change; APC, annual percent change; IBD, intrahepatic bile duct; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results.

The APC and AAPC trends were statistically significant at P<.05.

Female kidney and renal pelvis had 2 segments, but the joinpoint changed from 2007 (February 2015) to 2008 (November 2015). An initial increase in trend for this site was observed in both submissions. The second segment changed from a slight increase to a slight decrease in the rate (−0.4% for 2008-2013, which was nonsignificant). This is consistent with the joinpoint model being sensitive to small changes in rates. To address this issue, we analyzed 5-year and 10-year AAPC trend estimates, providing a better summary for comparison in a longer, more stable time interval. For all sites considered, including those with changes in the numbers of segments (ie, female thyroid, male kidney, and renal pelvis) and segment length (ie, female kidney and renal pelvis), each site had similar 5-year and 10-year AAPC values for both submissions, indicating that the delay-adjusted February submission data provided valid estimates for age-adjusted cancer incidence trends.

Preliminary Report of Cancer Cases Diagnosed Through 2014

Table 2 describes the observed 2014 age-adjusted and delay-adjusted 2014 incidence rates and delay-adjusted incidence trends from 2000 through 2014. The data are based on the February 2016 submission, which is an early look at 2014 age-adjusted rates for cancer sites included. The full set of observed and delay-adjusted rates is presented in Supporting Information Table 1. Using the February 2016 submission, large declines were evident for all cancer sites for males (delay-adjusted rates of 463.0 for 2014 [Table 2] and 485.1 for 2013 per 100,000 males [see Supporting Information Table 1]) and for prostate cancer (delay-adjusted rates of 97.5 for 2014 [Table 2] and 110.5 for 2013 per 100,000 males [see Supporting Information Table 1]). These more substantial declines are supported by the joinpoint analyses: the APC for all male cancers accelerated from −2.6% for 2009 through 2013 (Table 1) to −2.9% for 2009 through 2014 (Table 2) and the APC for prostate cancer accelerated from −9.4% for 2008 through 2013 to −9.7% for 2010 through 2014. Major changes, such as the decline in prostate cancer incidence, continued based on the most recent estimates, and the joinpoint shifted from 2011 to 2010 in delay models using the February 2016 submission (Table 2). The delay-adjusted rate for DY 2014 was 97.5 per 100,000, which is down from 110.5 per 100,000 for the DY 2013 cases (see Supporting Information Table 1). The last joinpoint model segment was found to have a significant decline of 9.7% in the APC from 2010 to 2014 (Fig. 1). The 5-year and 10-year AAPC trends were −9.7% and −5.4%, respectively, in the February 2016 data (Table 2) and were statistically significant trends representing a trend acceleration observed through DY 2013, which were −8.1% and −4.8%, respectively.²

The trends shown in Table 2 can be seen in Figure 1, which shows the fitted joinpoint model for both the observed and delay-adjusted age-adjusted incidence rates for selected sites. The sites shown in Figure 1 were selected because they represent sites that have had noteworthy rate changes recently: female CRC, female melanoma of the skin, female breast, male prostate, female cervix, and female thyroid. Although incidence trends for female CRC cancers continued to decline, there was a change in trajectory from 1 joinpoint through 2013 (−2.1% [2000-2008] and −3.0% [2008-2013]) (Table 1) to no joinpoints when 2014 data were added (−2.4% [2000-2014]) (Table 2) (Fig. 1), with the decline of 2.4% noted to not be as rapid as last year’s estimate of −3.0%.² The rise in female melanoma of the skin (APC, 1.17) was now significant as opposed to last year when it was flat, the slight increase in trend for female breast remained the same as last time with the same joinpoint in 2004, and estimates remained stable from 2004 to 2014 (APC, 0.3%). The joinpoint for prostate cancer shifted back 1 year to 2010 from last year and the decline was still very pronounced (APC, −9.74) and now was statistically significant. Cervix cancer lost a joinpoint from last year, with the last joinpoint segment demonstrating a steady decline (APC, −1.35) from 2003 to 2014 (which is statistically significant). This represents a slight flattening of the rate from the previous APC of −2.2% through DY 2013 using the November 2015 submission (Table 1). The smaller downward trend for the later submission for cervix cancer also was reflected in the 5-year and 10-year AAPC, changing from −2.2% and 1.6%, respectively, through DY 2013 (Table 1) to −1.3% and −1.3% (Table 2), respectively, after including the DY 2014 estimates. Female thyroid cancer rates continued to rise with the same joinpoints in 2006 and 2009, although the rise in the last segment diminished to an APC of 1.92 from an APC of 2.6. The most recent trend segments also decreased for male thyroid from 3.2% to 2.5% through 2013 and 2014 (Table 2).

Skin melanoma among males continued to increase, but with the preliminary 2014 data, the APC was found to have decreased from 1.9% (2000-2013) (Table 1) to 1.7% (2000-2014) (Table 2). Comparing the November 2015 submission and February 2016 submissions, APC trends were nearly identical for liver/IBD (male and female), pancreas (male and female), kidney and renal pelvis (male and female), corpus and uterus, and all sites for children aged birth to 19 years.

Figure 2 shows trends for the sites that were added for the February 2016 submission. For each site, joinpoints or segments remained the same between the November 2015 submission and the February 2016 submission. Exceptions include a slowing of the female pancreas positive trend from 0.9 (2000-2013) (Table 1) to 0.8 (2000-2014) (Table 2) and for male kidney and renal pelvis from 0.3 (2008-2013) (Table 1) to 0.2 (2008-2014) (Table 2).

Long-Term Trends (10-Year AAPC)

The long-term trends (10-year AAPC results) shown in Tables 1 and 2 revealed continuing declines for all cancer sites and 3 of 4 major cancers: male CRC, lung and bronchus, and prostate. Lung and bronchus and prostate cancer appeared to have a steeper decline in trends in the new data for 2014 than previous estimates. Although the 10-year AAPC for breast cancer remained positive, the magnitude was slightly diminished based on the 2014 estimate. Female CRC and cervical cancer had slightly smaller decreasing trends. Male skin melanoma, female breast (previously mentioned), thyroid, male liver cancer, and kidney and renal pelvis continued to have increasing 10-year trends, but the trends were slightly smaller in the most recent data for 2014. Trends for female skin melanoma, male pancreas, corpus and uterus, and all sites aged birth to 19 years were similar and increasing in the 2 estimates. Female pancreas continued to increase, but less so in the more recent data. Female liver/IBD trends were more positive, demonstrating a continuing increase.

DISCUSSION

Early estimates for a variety of cancers, derived using improved delay-adjustment models, are of interest to the cancer control community.⁷ For 3 years, we have reported on early estimates of cancer incidence rates and trends using SEER data. The delay-adjusted rates were validated through comparison with the usual subsequent November submission data. With additional new sites, our current report contains estimates for 13 cancer sites.

With changes to the February submission delay-adjustment model for consistency with the model used for the November submission, the model validation demonstrated consistency of the estimates, further substantiating that valid estimates of cancer incidence are obtainable from more incomplete data in February submissions, in comparison with the November submissions.

The sites that were added in last year’s report (cervix, thyroid, liver/IBD, and pediatric cancers)² were validated when comparing the February and November 2015 submissions for DY 2013 (Table 1) and continued to demonstrate good agreement. Trends for female thyroid cancer continued to show 3 segments using the February 2016 submission, a change that was evident in last year’s report.²

The direction of trends for early estimates from the November 2015 submission was consistent with the February 2016 submission, with a decline in rates noted for all sites. Incidence reduction continued for male skin melanoma, female breast, thyroid, male liver cancer, and kidney and renal pelvis cancers. Despite increasing trends, the trend magnitude for these sites increased more slowly than previously noted. Joinpoints did not shift for cancers of the lung and bronchus, female breast, male and female thyroid, and male liver/IBD, or among the sites introduced in this report (pancreas, kidney and renal pelvis, corpus and uterus, and all sites for those aged birth to 19 years).

Prostate cancer incidence decreased dramatically based on the DY 2013 data, as observed in the last early estimates (February 2015 submission). Female CRC and cervical cancer each demonstrated a smaller rate of decline in the more recent data. Evaluations of CRC and cervical cancer screening use and, in the case of cervical cancer, uptake of the human papillomavirus vaccine, are needed to promote prevention efforts. Cancer sites to be watched in the future include female skin melanoma, female liver/IBD, male pancreas, corpus and uterus, and all sites aged birth to 19 years, all of which had similar but increasing trends after the February 2016 data were added.

The pronounced decline in prostate cancer rates described in the results of the current study is consistent with an analysis of SEER and National Health Interview Survey data⁸ that found the percentage of men reporting prostate-specific antigen screening within the past 12 months to have decreased from 2005 to 2013 among men aged ≥ 50 years. The US Preventive Services Task Force recommended against prostate-specific antigen screening in 2012.⁹ Currently, the US Preventive Services Task Force has published the “Final Research Plan,” which will guide a systematic review of the evidence by researchers adhering to several key questions.¹⁰

The accelerated decline noted in female CRC was reported in the November 2015 submission for DY 2013. However, for female CRC cases diagnosed in 2014, the accelerated decline appeared to be attenuated. The final joinpoint segment APC from 2000 through 2014 was 2.4%, changing from an APC of −3.0% (2008-2013) in the 2015 November submission.

Among the new sites introduced herein, joinpoints did not shift for pancreas, kidney and renal pelvis, corpus and uterus, and all sites for those aged birth to 19 years. Pancreatic cancer was found to have steady, increasing trends. An analysis of population-based discharge diagnosis data from the Nationwide Inpatient Sample database found a nearly 25% increase in the numbers of cases diagnosed with pancreatic cancer between 1997 and 2012.¹¹

An analysis of international renal cell carcinoma (RCC)¹² revealed increasing trends among men; the majority of trend increases among women were approximately one-half as great as those among men over recent 10-year periods. Although our early estimates continue to demonstrate a marked increase in kidney cancer (including RCC) among men, rates for women continue in a slight, nonsignificant decrease. Znaor et al¹² suggested that changes in the distribution of RCC risk factors (smoking, obesity, hypertension, and diet) could be a factor, and found that incidence rates are higher in developed countries. It is not possible to connect the changing distribution of risk factors influencing rates and trends in our early estimates; however, the difference in rates and trends of kidney cancer by sex could suggest that use of modern imaging techniques, which is higher in Western populations,^13,14 is more prevalent among males. This possibility needs to be investigated further to determine whether improved diagnosis and treatment protocols impact rates in light of known risk factors.¹⁵ Recent research regarding uterine or endometrial cancer^16-18 has highlighted the importance of adjusting for hysterectomy rates when examining uterine cancer trends. Adjustment for hysterectomy diminishes differences in rates between black and white women, and differences in rates that are explained by risk factors, including obesity. Although our trend analysis of corpus and uterine cancer did not provide an adjustment for hysterectomy, the adjustment of these rates is warranted in future trend evaluations. Corrections for hysterectomy using state-level uterine corpus cancer incidence data and hysterectomy prevalence data from the Behavioral Risk Factor Surveillance System have been discussed, along with methods for correcting cervical cancer mortality rates for hysterectomy.^18-20

A previous analysis of childhood cancer diagnosed among individuals aged birth to 14 years between 1975 and 1995²¹ reported modest incidence rate increases for selected cancer sites in this population (brain and central nervous system, leukemia, and infant neuroblastoma) followed by declining rates. These findings were believed to be due to diagnostic improvements or reporting changes. A more recent analysis discussed the increase in rates of childhood cancer overall and head and neck malignancies in a pediatric population from 1973 to 2010.²² An evaluation of leukemia and lymphoma among international populations among children and adolescents aged birth to 19 years detected mostly stable rates for these cancer sites, with the exception of increases in Hodgkin lymphoma in selected European and Asian populations.²³ Although our rates included all cancer sites among those diagnosed in the group aged birth to 19 years, it is worth noting that our finding represents a steady increase. Further investigation of cancer site trends in this grouping is warranted.

The new sites presented in this report (pancreas, kidney and renal pelvis, and corpus and uterus) do not typically have screening tests available. However, much of the literature cites the increasing use of imaging as possibly contributing to the increased detection of some these cancers.^11,24 It becomes more important to carefully assess early estimates of these cancers to monitor population changes, perhaps with an eye toward the use of diagnostic technology.

The February delay model performs well, especially given that the delay factors necessary to inflate the observed rates to delay-adjusted rates are much larger for the February submission compared with the November submission. The February submission delay model can successfully project incidence rates that are similar to those based on the November submission, which are substantially more complete.

There appears to be continued good agreement between the delay-adjusted rates and trends from the February and November 2014 submissions and the February and November 2015 submissions. Additional years of February submissions contribute to more stable delay-adjustment estimates, and their interpretation becomes clearer. As our processes for assembling the data and estimating the February submission delay model become more routine, we should be able to present these estimates more rapidly.

Supplementary Material

NIHMS1626508-supplement-SM.xlsx^{(36.5KB, xlsx)}

Acknowledgments

We thank Susan Scott, MPH, for assistance with editing of the article. We also thank the Surveillance, Epidemiology, and End Results registries for submitting February data submissions.

FUNDING SUPPORT

No specific funding was disclosed.

Footnotes

This article has been contributed to by US Government employees and their work is in the public domain in the USA.

Additional supporting information may be found in the online version of this article

CONFLICT OF INTEREST DISCLOSURES

Zhaohui Zou’s contribution to the current study was funded under a contract between Mr. Zou’s employer and the National Cancer Institute. The two other IMS co-authors on this work, MF Krapcho and DG Miller, were also funded by the same contract between IMS and the NCI.

REFERENCES

1.Lewis DR, Chen HS, Midthune DN, Cronin KA, Krapcho MF, Feuer EJ. Early estimates of SEER cancer incidence for 2012: approaches, opportunities, and cautions for obtaining preliminary estimates of cancer incidence. Cancer. 2015;121:2053–2062. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Lewis DR, Chen HS, Cockburn M, et al. Preliminary estimates of SEER cancer incidence for 2013. Cancer. 2016;122:1579–1587. [DOI] [PubMed] [Google Scholar]
3.North American Central Cancer Registries. Cancer Incidence in North America (CINA) publication (2009-2013). www.naaccr.org/dataandpublications/cinapubs.aspx. Accessed January 11, 2017.
4.Surveillance, Epidemiology, and End Results Program. SEER*Stat software version 8.3.2. seer.cancer.gov/seerstat. Accessed January 31, 2017.
5.Clegg LX, Hankey BF, Tiwari R, Feuer EJ, Edwards BK. Estimating average annual per cent change in trend analysis. Stat Med. 2009;28:3670–3682. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Howlader N, Noone AM, Krapcho M, et al. , eds. SEER Cancer Statistics Review, 1975-2013. Bethesda, MD: National Cancer Institute; 2016. [Google Scholar]
7.North American Central Cancer Registries. Delay adjustment. www.naaccr.org/research/delayadjustment.aspx. Accessed January 12, 2017.
8.Jemal A, Fedewa S, Ma J, et al. Prostate cancer incidence and PSA testing patterns in relation to USPSTF screening recommendations. JAMA. 2015;314:2054–2061. [DOI] [PubMed] [Google Scholar]
9.US Preventive Services Task Force. Prostate cancer: screening. https://www.uspreventiveservicestaskforce.org/tools/consumerinfo/index/information-for-consumers. Accessed November 25, 2016.
10.US Preventive Services Task Force. Final research plan: final research plan for prostate cancer: screening. https://www.uspreventiveservices-taskforce.org/Page/Document/final-research-plan/prostate-cancer-screening1. Accessed January 26, 2017.
11.Wadhwa V, Patwardhan S, Garg SK, Lopez R, Sanaka MR. Inpatient burden of pancreatic cancer in the United States: an analysis of national trends in the United States from 1997 to 2012. Pancreas. 2016;45:e41–e42. [DOI] [PubMed] [Google Scholar]
12.Znaor A, Lortet-Tieulent J, Laversanne M, Jemal A, Bray F. International variations and trends in renal cell carcinoma incidence and mortality. Eur Urol. 2015;67:519–530. [DOI] [PubMed] [Google Scholar]
13.Hollingsworth JM, Miller DC, Daignault S, Hollenbeck BK. Rising incidence of small renal masses: a need to reassess treatment effect. J Natl Cancer Inst. 2006;98:1331–1334. [DOI] [PubMed] [Google Scholar]
14.Sun M, Thuret R, Abdollah F, et al. Age-adjusted incidence, mortality, and survival rates of stage-specific renal cell carcinoma in North America: a trend analysis. Eur Urol. 2011;59:135–141. [DOI] [PubMed] [Google Scholar]
15.Adibi M, Karam JA, Wood CG. Reporting geographic and temporal trends in renal cell carcinoma: why is this important? Eur Urol. 2015;67:531–532. [DOI] [PubMed] [Google Scholar]
16.Temkin SM, Minasian L, Noone AM. The end of the hysterectomy epidemic and endometrial cancer incidence: what are the unintended consequences of declining hysterectomy rates? Front Oncol. 2016;6:89. doi 10.3389/fonc.2016.00089 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Jamison PM, Noone AM, Ries LA, Lee NC, Edwards BK. Trends in endometrial cancer incidence by race and histology with a correction for the prevalence of hysterectomy, SEER 1992 to 2008. Cancer Epidemiol Biomarkers Prev. 2012;22:233–241. [DOI] [PubMed] [Google Scholar]
18.Siegel R, Devesa SS, Cokkinides V, Ma J, Jemal A. State-level uterine corpus cancer incidence rates corrected for hysterectomy prevalence, 2004 to 2008. Cancer Epidemiol Biomarkers Prev. 2013;22:25–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Stang A, Hawk H, Knowlton R, Gershman ST, Kuss O. Hysterectomy-corrected incidence rates of cervical and uterine cancers in Massachusetts, 1995 to 2010. Ann Epidemiol. 2014;24:849–854. [DOI] [PubMed] [Google Scholar]
20.Beavis AL, Gravitt PE, Rositch AF. Hysterectomy-corrected cervical cancer mortality rates reveal a larger racial disparity in the United States [published online ahead of print January 23, 2017]. Cancer. doi 10.1002/cncr.30507. [DOI] [PubMed] [Google Scholar]
21.Linet MS, Ries LA, Smith MA, Tarone RE, Devesa SS. Cancer surveillance series: recent trends in childhood cancer incidence and mortality in the United States. J Natl Cancer Inst. 1999;91:1051–1058. [DOI] [PubMed] [Google Scholar]
22.Schwartz I, Hughes C, Brigger MT. Pediatric head and neck malignancies: incidence and trends, 1973-2010. Otolaryngol Head Neck Surg. 2015;152:1127–1132. [DOI] [PubMed] [Google Scholar]
23.Linet MS, Brown LM, Mbulaiteye SM, et al. International long-term trends and recent patterns in the incidence of leukemias and lymphomas among children and adolescents ages 0-19 years. Int J Cancer. 2016;138:1862–1874. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Hock LM, Lynch J, Balaji KC. Increasing incidence of all stages of kidney cancer in the last 2 decades in the United States: an analysis of Surveillance, Epidemiology, and End Results program data. J Urol. 2002;167:57–60. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS1626508-supplement-SM.xlsx^{(36.5KB, xlsx)}

[R1] 1.Lewis DR, Chen HS, Midthune DN, Cronin KA, Krapcho MF, Feuer EJ. Early estimates of SEER cancer incidence for 2012: approaches, opportunities, and cautions for obtaining preliminary estimates of cancer incidence. Cancer. 2015;121:2053–2062. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Lewis DR, Chen HS, Cockburn M, et al. Preliminary estimates of SEER cancer incidence for 2013. Cancer. 2016;122:1579–1587. [DOI] [PubMed] [Google Scholar]

[R3] 3.North American Central Cancer Registries. Cancer Incidence in North America (CINA) publication (2009-2013). www.naaccr.org/dataandpublications/cinapubs.aspx. Accessed January 11, 2017.

[R4] 4.Surveillance, Epidemiology, and End Results Program. SEER*Stat software version 8.3.2. seer.cancer.gov/seerstat. Accessed January 31, 2017.

[R5] 5.Clegg LX, Hankey BF, Tiwari R, Feuer EJ, Edwards BK. Estimating average annual per cent change in trend analysis. Stat Med. 2009;28:3670–3682. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Howlader N, Noone AM, Krapcho M, et al. , eds. SEER Cancer Statistics Review, 1975-2013. Bethesda, MD: National Cancer Institute; 2016. [Google Scholar]

[R7] 7.North American Central Cancer Registries. Delay adjustment. www.naaccr.org/research/delayadjustment.aspx. Accessed January 12, 2017.

[R8] 8.Jemal A, Fedewa S, Ma J, et al. Prostate cancer incidence and PSA testing patterns in relation to USPSTF screening recommendations. JAMA. 2015;314:2054–2061. [DOI] [PubMed] [Google Scholar]

[R9] 9.US Preventive Services Task Force. Prostate cancer: screening. https://www.uspreventiveservicestaskforce.org/tools/consumerinfo/index/information-for-consumers. Accessed November 25, 2016.

[R10] 10.US Preventive Services Task Force. Final research plan: final research plan for prostate cancer: screening. https://www.uspreventiveservices-taskforce.org/Page/Document/final-research-plan/prostate-cancer-screening1. Accessed January 26, 2017.

[R11] 11.Wadhwa V, Patwardhan S, Garg SK, Lopez R, Sanaka MR. Inpatient burden of pancreatic cancer in the United States: an analysis of national trends in the United States from 1997 to 2012. Pancreas. 2016;45:e41–e42. [DOI] [PubMed] [Google Scholar]

[R12] 12.Znaor A, Lortet-Tieulent J, Laversanne M, Jemal A, Bray F. International variations and trends in renal cell carcinoma incidence and mortality. Eur Urol. 2015;67:519–530. [DOI] [PubMed] [Google Scholar]

[R13] 13.Hollingsworth JM, Miller DC, Daignault S, Hollenbeck BK. Rising incidence of small renal masses: a need to reassess treatment effect. J Natl Cancer Inst. 2006;98:1331–1334. [DOI] [PubMed] [Google Scholar]

[R14] 14.Sun M, Thuret R, Abdollah F, et al. Age-adjusted incidence, mortality, and survival rates of stage-specific renal cell carcinoma in North America: a trend analysis. Eur Urol. 2011;59:135–141. [DOI] [PubMed] [Google Scholar]

[R15] 15.Adibi M, Karam JA, Wood CG. Reporting geographic and temporal trends in renal cell carcinoma: why is this important? Eur Urol. 2015;67:531–532. [DOI] [PubMed] [Google Scholar]

[R16] 16.Temkin SM, Minasian L, Noone AM. The end of the hysterectomy epidemic and endometrial cancer incidence: what are the unintended consequences of declining hysterectomy rates? Front Oncol. 2016;6:89. doi 10.3389/fonc.2016.00089 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Jamison PM, Noone AM, Ries LA, Lee NC, Edwards BK. Trends in endometrial cancer incidence by race and histology with a correction for the prevalence of hysterectomy, SEER 1992 to 2008. Cancer Epidemiol Biomarkers Prev. 2012;22:233–241. [DOI] [PubMed] [Google Scholar]

[R18] 18.Siegel R, Devesa SS, Cokkinides V, Ma J, Jemal A. State-level uterine corpus cancer incidence rates corrected for hysterectomy prevalence, 2004 to 2008. Cancer Epidemiol Biomarkers Prev. 2013;22:25–31. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Stang A, Hawk H, Knowlton R, Gershman ST, Kuss O. Hysterectomy-corrected incidence rates of cervical and uterine cancers in Massachusetts, 1995 to 2010. Ann Epidemiol. 2014;24:849–854. [DOI] [PubMed] [Google Scholar]

[R20] 20.Beavis AL, Gravitt PE, Rositch AF. Hysterectomy-corrected cervical cancer mortality rates reveal a larger racial disparity in the United States [published online ahead of print January 23, 2017]. Cancer. doi 10.1002/cncr.30507. [DOI] [PubMed] [Google Scholar]

[R21] 21.Linet MS, Ries LA, Smith MA, Tarone RE, Devesa SS. Cancer surveillance series: recent trends in childhood cancer incidence and mortality in the United States. J Natl Cancer Inst. 1999;91:1051–1058. [DOI] [PubMed] [Google Scholar]

[R22] 22.Schwartz I, Hughes C, Brigger MT. Pediatric head and neck malignancies: incidence and trends, 1973-2010. Otolaryngol Head Neck Surg. 2015;152:1127–1132. [DOI] [PubMed] [Google Scholar]

[R23] 23.Linet MS, Brown LM, Mbulaiteye SM, et al. International long-term trends and recent patterns in the incidence of leukemias and lymphomas among children and adolescents ages 0-19 years. Int J Cancer. 2016;138:1862–1874. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Hock LM, Lynch J, Balaji KC. Increasing incidence of all stages of kidney cancer in the last 2 decades in the United States: an analysis of Surveillance, Epidemiology, and End Results program data. J Urol. 2002;167:57–60. [PubMed] [Google Scholar]

PERMALINK

Early Estimates of SEER Cancer Incidence, 2014

Denise Riedel Lewis, PhD, MPH

Huann-Sheng Chen, PhD

Myles G Cockburn, PhD

Xiao-Cheng Wu, MD

Antoinette M Stroup, PhD

Douglas N Midthune, MS

Zhaohui Zou, MS

Martin F Krapcho, BS

Daniel G Miller, BA

Eric J Feuer, PhD

Abstract

BACKGROUND:

METHODS:

RESULTS:

CONCLUSIONS:

INTRODUCTION

MATERIALS AND METHODS

Data Source and Methodology for Assessing Trends

Updates to the February Submission Reporting Delay Model

RESULTS

Model Validation

TABLE 1.

TABLE 2.

Preliminary Report of Cancer Cases Diagnosed Through 2014

Figure 1.

Figure 2.

Long-Term Trends (10-Year AAPC)

DISCUSSION

Supplementary Material

Acknowledgments

Footnotes

REFERENCES

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases