Trends in pediatric thyroid cancer incidence in the United States, 1998-2013.

BACKGROUND: Pediatric differentiated thyroid cancer (DTC) rates have increased over time in the United States and worldwide. Improvements in imaging for the diagnosis of DTC have been hypothesized as a potential driver of these increases. This study stratifies temporal trends in pediatric DTC by stage and tumor size to assess whether rates of large, late-stage cancers, which are likely to be clinically meaningful, are increasing over time. METHODS: Age-standardized incidence rates (ASRs) of DTC and annual percent changes (APCs) in primary DTC rates were estimated for 0- to 19-year-olds with data from 39 US cancer registries during 1998-2013. RESULTS: During 1998-2013, 7296 cases of DTC were diagnosed (6652 papillary cases and 644 follicular cases). APCs of pediatric DTCs significantly increased by 4.43%/y [95% CI, 3.74%/y-5.13%/y], primarily because of increases in papillary histologies. Increasing trends were observed for children aged 10 to 19 years for both sexes and for non-Hispanic whites, non-Hispanic blacks, and Hispanics. Rates increased significantly over the time period for all tumor stages (APClocalized , +4.06%/y [95% CI, 2.84%/y-5.29%/y]; APCregional , +5.68%/y [95% CI, 4.64%/y-6.73%/y]; APCdistant , +8.55%/y [95% CI, 5.03%/y-12.19%/y]) and across tumor sizes (APC<1 cm , +9.46%/y [95% CI, 6.13%/y-12.90%/y]; APC1-2 cm , +6.92%/y [95% CI, 4.31%/y-9.60%/y]; APC>2 cm , +4.69%/y [95% CI, 2.75%/y-6.67%/y]). CONCLUSIONS: Significantly increasing rates of DTC over time among 10- to 19-year-olds in the United States are unlikely to be entirely explained by increases in medical surveillance during childhood because rates of large and late-stage DTC are increasing over time. Future studies should examine environmental and other factors that may be contributing to rising DTC rates.

Validation of prostate-specific antigen laboratory values recorded in Surveillance, Epidemiology, and End Results registries.

BACKGROUND: Researchers have used prostate-specific antigen (PSA) values collected by central cancer registries to evaluate tumors for potential aggressive clinical disease. An independent study collecting PSA values suggested a high error rate (18%) related to implied decimal points. To evaluate the error rate in the Surveillance, Epidemiology, and End Results (SEER) program, a comprehensive review of PSA values recorded across all SEER registries was performed. METHODS: Consolidated PSA values for eligible prostate cancer cases in SEER registries were reviewed and compared with text documentation from abstracted records. Four types of classification errors were identified: implied decimal point errors, abstraction or coding implementation errors, nonsignificant errors, and changes related to "unknown" values. RESULTS: A total of 50,277 prostate cancer cases diagnosed in 2012 were reviewed. Approximately 94.15% of cases did not have meaningful changes (85.85% correct, 5.58% with a nonsignificant change of <1 ng/mL, and 2.80% with no clinical change). Approximately 5.70% of cases had meaningful changes (1.93% due to implied decimal point errors, 1.54% due to abstract or coding errors, and 2.23% due to errors related to unknown categories). Only 419 of the original 50,277 cases (0.83%) resulted in a change in disease stage due to a corrected PSA value. CONCLUSIONS: The implied decimal error rate was only 1.93% of all cases in the current validation study, with a meaningful error rate of 5.81%. The reasons for the lower error rate in SEER are likely due to ongoing and rigorous quality control and visual editing processes by the central registries. The SEER program currently is reviewing and correcting PSA values back to 2004 and will re-release these data in the public use research file. Cancer 2017;123:697-703. © 2016 American Cancer Society.

Melanoma Tumor Depth Quality Audit: A Nonmatch Analysis.

Background: The National Cancer Institute's Surveillance Research Program (SRP) received reports from cancer registries in the Surveillance, Epidemiology, and End Results (SEER) Program concerning the coding of melanoma tumor depth. To address these concerns, SRP developed an algorithm to identify melanoma depth measurement values and conducted a nonmatch analysis. Methods: A nonmatch analysis was conducted on 1,117 cases diagnosed between 2010 and 2017. With the help of Information Management Services, a natural language processing algorithm was developed to identify melanoma tumor depth values along with a gold standard for comparison. A randomly sampled data set was created to compare the algorithm-generated and gold standard values to the originally reported values; these were analyzed using SAS software version 9.4. Analyses were conducted to determine the distribution of nonmatches by demographics and estimate the distribution of nonmatches by the derived T variable according to the 7th edition of the American Joint Committee on Cancer (AJCC)'s AJCC Cancer Staging Manual. Results: Of the 1,117 cases, 849 cases (76%) were a match between the originally reported values and the gold standard. The majority of cases were found to be in male patients (60%) and non-Hispanic White patients (93%). When comparing derived AJCC-7 T based on the originally reported value to the gold standard, 16% of the original derived AJCC-7 T values were incorrect, with most of the nonmatches resulting in incorrectly coding a case as TX instead of T1. Conclusion: In total, 24% of cases were found to have a discrepancy in the originally recorded values. Decimal errors made up 3% of all cases in this nonmatch analysis. This algorithm may prove to be an essential tool in optimizing registry resources by flagging inconsistencies via automated text review to be adjudicated by registrars, improving their quality of data as needed.

Machine Learning Methods to Identify Missed Cases of Bladder Cancer in Population-Based Registries.

PURPOSE: Population-based cancer incidence rates of bladder cancer may be underestimated. Accurate estimates are needed for understanding the burden of bladder cancer in the United States. We developed and evaluated the feasibility of a machine learning-based classifier to identify bladder cancer cases missed by cancer registries, and estimated the rate of bladder cancer cases potentially missed. METHODS: Data were from population-based cohort of 37,940 bladder cancer cases 65 years of age and older in the SEER cancer registries linked with Medicare claims (2007-2013). Cases with other urologic cancers, abdominal cancers, and unrelated cancers were included as control groups. A cohort of cancer-free controls was also selected using the Medicare 5% random sample. We used five supervised machine learning methods: classification and regression trees, random forest, logic regression, support vector machines, and logistic regression, for predicting bladder cancer. RESULTS: Registry linkages yielded 37,940 bladder cancer cases and 766,303 cancer-free controls. Using health insurance claims, classification and regression trees distinguished bladder cancer cases from noncancer controls with very high accuracy (95%). Bacille Calmette-Guerin, cystectomy, and mitomycin were the most important predictors for identifying bladder cancer. From 2007 to 2013, we estimated that up to 3,300 bladder cancer cases in the United States may have been missed by the SEER registries. This would result in an average of 3.5% increase in the reported incidence rate. CONCLUSION: SEER cancer registries may potentially miss bladder cancer cases during routine reporting. These missed cases can be identified leveraging Medicare claims and data analytics, leading to more accurate estimates of bladder cancer incidence.

Estimating Chemotherapy Use Among Patients With a Prior Primary Cancer Diagnosis Using SEER-Medicare Data.

Cancer treatment studies commonly exclude patients with prior primary cancers due to difficulties in ascertaining for which site treatment is intended. Surveillance, Epidemiology, and End Results-Medicare patients 65 years and older diagnosed with an index colon or rectal cancer (CRC) or female breast cancer (BC) between 2004 and 2013 were included. Chemotherapy, defined as "any chemotherapy" and more restrictively as "chemotherapy with confirmatory diagnoses," was ascertained based on claims data within 6 months of index cancer diagnosis by prior cancer history. Any chemotherapy use was slightly lower among patients with a prior cancer (CRC: no prior = 17.4%, prior = 16.1%; BC: no prior = 12.9%, prior = 12.0%). With confirmatory diagnoses required, estimates were lower, especially among patients with a prior cancer (CRC: no prior = 16.8%, prior = 13.6%; BC: no prior = 12.6%, prior = 11.0%). These findings suggest that patients with prior cancers can be included in studies of chemotherapy use; requiring confirmatory diagnoses can increase treatment assignment confidence.

Using the SEER-Medicare Data to Assess Incident Chronic Myeloid Leukemia and Bladder Cancer Cases Missed by Cancer Registries.

The growing use of oral systemic therapies and transition of some cancer treatments to the outpatient setting makes capturing all cancer case patients more difficult. We aim to develop algorithms to identify potentially missed incident case patients and estimate impact on incidence rates. We reviewed claims from SEER-Medicare 5% noncancer control patient sample to identify potentially missed chronic myeloid leukemia (CML) and bladder case patients based on diagnosis codes, cancer-related treatments, and oncology consultations. Observed rates of definite missed CML and definite and probable missed bladder case patients were calculated and the impact of missed case patients of these two cancers on SEER 65+ incidence rates were estimated. From 2008 to 2015, the algorithm estimated 781 definite CML case patients missed, increasing the number by 10.7%. From 2007 to 2015, the algorithm estimated 4629 definite and 5772 probable bladder case patients missed, increasing the number by 3.8% to 8.1%. Our algorithms offer potential methods for identifying missed case patients and validating the completeness of cancer registries.

Utilization of the Cancer Medications Enquiry Database (CanMED)-National Drug Codes (NDC): Assessment of Systemic Breast Cancer Treatment Patterns.

Cancer Medications Enquiry Database (CanMED) is comprised of two interactive, nomenclature-specific databases within the Observational Research in Oncology Toolbox: CanMED-Healthcare Common Procedure Coding System (HCPCS) and CanMED-National Drug Code (NDC), described through this study. CanMED includes medications with a) a US Food and Drug Administration-approved cancer treatment or treatment-related symptom management indication, b) inclusion in treatment guidelines, or c) an orphan drug designation. To demonstrate the joint utility of CanMED, medication codes associated with female breast cancer treatment were identified and utilization patterns were assessed within Surveillance Epidemiology and End Results-Medicare (SEER) data. CanMED-NDC (11_2018 v.1.2.4) includes 6860 NDC codes: chemotherapy (1870), immunotherapy (164), hormone therapy (3074), and ancillary therapy (1752). Treatment patterns among stage I-IIIA (20 701) and stage IIIB-IV (2381) breast cancer patients were accordant with guideline-recommended treatment by stage and molecular subtype. CanMED facilitates identification of medications from observational data (eg, claims and electronic health records), promoting more standardized and efficient treatment-related cancer research.

Development and Utility of the Observational Research in Oncology Toolbox: Cancer Medications Enquiry Database-Healthcare Common Procedure Coding System (HCPCS).

PURPOSE: Health-care claims are of increasing utility as a rich, real-world data resource for conducting treatment-related cancer research. However, multiple dynamic coding nomenclatures exist, leading to study variability. To promote increased standardization and reproducibility, the National Cancer Institute (NCI) developed the Cancer Medications Enquiry Database (CanMED)-Healthcare Common Procedure Coding System (HCPCS) within the Observational Research in Oncology Toolbox. METHODS: The CanMED-HCPCS includes codes for oncology medications that a) have a US Food and Drug Administration-approved indication for cancer treatment or treatment-related symptom management; b) are present in National Comprehensive Cancer Network guidelines; or c) carry an orphan drug designation for treatment or management of cancer. Included medications and their HCPCS codes were primarily identified based on Center for Medicare and Medicaid Services annual HCPCS Indices (2012-2018). To demonstrate the utility of the CanMED-HCPCS, use of systemic treatment for stage II-IV colorectal cancer patients included in the Surveillance, Epidemiology, and End Results-Medicare data (2007-2013) was assessed. RESULTS: The CanMED-HCPCS (v2018) includes 332 HCPCS codes for cancer-related medications: chemotherapy (156), immunotherapy (74), hormonal therapy (54), and ancillary therapy (48). Observed treatment trends within the NCI Surveillance, Epidemiology, and End Results-Medicare data were as expected; utilization of each treatment type increased with stage, and immunotherapy was largely confined to use among stage IV patients. CONCLUSION: The CanMED-HCPCS provides a comprehensive resource that can be used by the research community to facilitate systematic identification of medications within claims or electronic health data using the HCPCS nomenclature and greater reproducibility of cancer surveillance and health services research.

Trends in Pediatric Central Nervous System Tumor Incidence in the United States, 1998-2013.

BACKGROUND: Brain and other central nervous system (CNS) cancers are the leading cause of U.S. pediatric cancer mortality. Incidence trends can provide etiologic insight. We report trends in incidence rates of pediatric malignant CNS cancers and pilocytic astrocytoma (nonmalignant but historically registered) in the United States. METHODS: Age-standardized incidence rates and annual percent changes (APC) in rates during 1998 to 2013 were calculated for children aged 0 to 19, stratified by subtype, age, sex, and for gliomas, histology and location. We estimated the absolute change in number of cases diagnosed U.S.-wide during 2013 compared with the expected number of cases had 1998 rates remained stable. RESULTS: Rates of all pediatric malignant CNS cancer combined (n = 18,612) did not change [APC: 0.16; 95% confidence interval (CI): -0.21-0.53]. There were statistically significant changes in several subtypes; however, glioma incidence (n = 10,664) increased by 0.77% per year (95% CI: 0.29-1.26), embryonal cancer rates (n = 5,423) decreased by 0.88% per year (95% CI: -1.33 to -0.43), and pilocytic astrocytoma rates (n = 6,858) increased by 0.89% per year (95% CI: 0.21-1.58). Of the 1,171 malignant tumors and 450 pilocytic astrocytomas diagnosed in U.S. children in 2013, we estimated 120 excess gliomas, 94 excess pilocytic astrocytomas, and 72 fewer embryonal CNS tumors than would be expected had 1998 rates remained stable. CONCLUSIONS: The gradual changes in incidence we observed for specific types of pediatric CNS cancers are likely due to a combination of changes in classification and diagnosis and true changes in CNS cancer. IMPACT: Continued surveillance of pediatric CNS tumors should remain a priority, given their significant contribution to pediatric cancer-related deaths.

Association of Chemotherapy for Solid Tumors With Development of Therapy-Related Myelodysplastic Syndrome or Acute Myeloid Leukemia in the Modern Era.

Importance: Therapy-related myelodysplastic syndrome or acute myeloid leukemia (tMDS/AML) is a rare, usually fatal complication of chemotherapy, including certain alkylating agents, topoisomerase II inhibitors, and platinum compounds. With the introduction of new chemotherapeutic agents, expanded indications for established agents, and increased neoadjuvant and adjuvant chemotherapy, tMDS/AML risks in the modern age are poorly understood. Objectives: To quantify tMDS/AML risk after chemotherapy for solid cancer among United States adults since 2000 and correlate tMDS/AML risk patterns with chemotherapy treatment practices. Design, Setting, and Participants: A population-based cohort study was conducted using cancer registries from the Surveillance, Epidemiology, and End Results (SEER) Program and Medicare claims. Risk analyses included 1619 tMDS/AML cases among 700 612 adults (age, 20-84 years) who were diagnosed with first primary solid cancer during 2000 to 2013 (followed up through 2014), received initial chemotherapy, and survived 1 year or longer, as reported to SEER. Descriptive analyses were conducted of SEER records linked with Medicare claims for chemotherapy in 165 820 older adults (age, 66-84 years) receiving initial chemotherapy for a first primary solid cancer in 2000-2013. Data analysis was conducted from October 2017 to April 2018. Exposures: Receipt of initial chemotherapy for solid cancer. Main Outcomes and Measures: Second primary tMDS/AML. Results: Based on 1619 tMDS/AML cases in the SEER database (mean [SD] age, 64.3 [12.2] years; 1148 [70.9%] female), tMDS/AML risks were statistically significantly elevated after chemotherapy for 22 of 23 solid cancers (all except colon). Relative risks ranged from 1.5 to greater than 10 and excess absolute risks from 1.4 to greater than 15 cases per 10 000 person-years compared with the general population. Overall survival following tMDS/AML diagnosis was poor (1270 of 1619 patients [78.4%] died; median overall survival, 7 months). For patients treated with chemotherapy at the present time, approximately three-quarters of tMDS/AML cases expected to occur within the next 5 years will be attributable to chemotherapy. In the SEER-Medicare database, use of known leukemogenic agents, particularly platinum compounds, in initial chemotherapy increased substantially since 2000, most notably for gastrointestinal tract cancers (esophagus, stomach, colon, and rectum; 10% in 2000-2001 to 81% during 2012-2013). Conclusions and Relevance: Large-scale, United States population-based data demonstrate excess tMDS/AML risks following chemotherapy for nearly all solid tumor types, consistent with expanded use of known leukemogenic agents in the 21st century. Continued efforts to reduce treatment-related adverse events, particularly for solid cancer patients with favorable prognosis, are needed.

Risk Factors for Melanoma Among Survivors of Non-Hodgkin Lymphoma.

PURPOSE: Previous studies have reported that survivors of non-Hodgkin lymphoma (NHL) have an increased risk of developing cutaneous melanoma; however, risks associated with specific treatments and immune-related risk factors have not been quantified. PATIENTS AND METHODS: We evaluated second melanoma risk among 44,870 1-year survivors of first primary NHL diagnosed at age 66 to 83 years from 1992 to 2009 and included in the Surveillance, Epidemiology, and End Results-Medicare database. Information on NHL treatments, autoimmune diseases, and infections was derived from Medicare claims. RESULTS: A total of 202 second melanoma cases occurred among survivors of NHL, including 91 after chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) and 111 after other NHL subtypes (cumulative incidence by age 85 years: CLL/SLL, 1.37%; other NHL subtypes, 0.78%). Melanoma risk after CLL/SLL was significantly increased among patients who received infused fludarabine-containing chemotherapy with or without rituximab (n=18: hazard ratio [HR], 1.92; 95% CI, 1.09 to 3.40; n=10: HR, 2.92; 95% CI, 1.42 to 6.01, respectively). Significantly elevated risks also were associated with T-cell activating autoimmune diseases diagnosed before CLL/SLL (n=36: HR, 2.27; 95% CI, 1.34 to 3.84) or after CLL/SLL (n=49: HR, 2.92; 95% CI, 1.66 to 5.12). In contrast, among patients with other NHL subtypes, melanoma risk was not associated with specific treatments or with T-cell/B-cell immune conditions. Generally, infections were not associated with melanoma risk, except for urinary tract infections (CLL/SLL), localized scleroderma, pneumonia, and gastrohepatic infections (other NHLs). CONCLUSION: Our findings suggest immune perturbation may contribute to the development of melanoma after CLL/SLL. Increased vigilance is warranted among survivors of NHL to maximize opportunities for early detection of melanoma.