Use of Electronic Medical Records to Ascertain Depth of SEER-Reported Melanomas of Unknown Tumor Thickness

Cancer surveillance is important for monitoring disease trends and survival, forecasting future directions, and appropriately targeting strategies for prevention and screening. In the United States, the Surveillance, Epidemiology and End Results (SEER) Program is charged with maintaining records of all cancers reported annually in their respective populations, including melanoma, which is the sixth most common cancer in the United States.^1,2

Despite the rigorous methods applied by the SEER tumor registry, there remain limitations on the completeness of data. Approximately 4% of melanomas in the SEER program have no SEER summary stage, which is different from the American Joint Committee on Cancer (AJCC) stage, and approximately 12% of melanomas reported to SEER annually are missing information on tumor thickness.^1,3 An important component of AJCC staging, tumor thickness (aka Breslow thickness) is a key predictor of melanoma survival. These melanomas of unknown thickness represent either incomplete clinical information or inadequate abstracting of existing clinical documentation.⁴

To better understand the composition of melanomas with missing thickness in the SEER data, we examined the electronic medical records (EMR) of health maintenance organization (HMO) melanoma cases. The objectives of this examination were to (1) quantify how often actual tumor depth could be determined from the EMR of originally reported unknown-depth tumors; (2) determine the available abstracted Breslow depths and compare their distribution to that of the reported tumor registry depths; and (3) identify possible areas to target for improving tumor depth reporting by abstractors.

This study was approved by the Henry Ford Health System institutional review board. Data were requested from the Detroit SEER registry for melanoma cases (topography codes C440-C449; morphology code 8720-99) occurring between 1984 and 2008 among the approximately 450  000 HMO plan members annually during the period of interest. Between 1984 and 2008, there were 1156 cutaneous melanomas reported by the HMO and the Detroit SEER. This included 346 melanoma cases with unknown Breslow thickness, defined as EOD-10 (extent of disease length of 10 digits), recorded as unknown (999). The EMR of these 346 cases were reviewed in an attempt to ascertain tumor thickness.

After careful review of the EMR, tumor thickness was ascertained for 249 cases with previously reported unknown depth (72%). Review of EMR failed to provide depth for 97 cases of unknown thickness (28%) (Table 1).

Ascertainment of missing melanoma depth data yielded cases in each of the traditional depth categories used in staging, regardless of behavior code (Table 2). When the distribution was compared with the distribution of all cases in the Detroit SEER and the entire SEER-9 program (Table 3),³ the ascertained HMO cases that were originally reported as unknown depth included more in situ melanomas and fewer invasive melanomas (Figure). Most of these in situ melanomas were recorded in histopathology reports as lentigo maligna (61%).

While demographic characteristics of patients with unstaged disease and missing components of staging have been quantified for select cancers, this study provides information on the distribution of missing cases by tumor prognostic category.^4,5 Review of HMO medical records allowed for tumor depth ascertainment for a large proportion of melanoma cases previously reported as unknown depth by SEER. The distribution of the thickness of these newly ascertained cases demonstrated a disproportionate number of in situ melanomas. The lentigo maligna type of melanoma in situ, was commonly recorded as an unknown depth.

Our study had several limitations. It was conducted at a single institution and so may not be generalizable. The distribution of melanomas with unknown thickness may not reflect the overall distribution of melanoma by thickness. The number of melanomas in situ, as reported by SEER, may be underestimated owing to misclassification, which has implications for the descriptive epidemiology of melanoma.

Future studies are recommended of melanomas of unknown thickness with the goal of identifying areas to improve melanoma reporting, including repeated examination in other settings, such as any of the 6 Cancer Research Network HMO institutions that overlap with SEER registries. Continuing education of registry abstractors may improve SEER registry completeness by focusing on ascertainment of melanoma in situ, including cases recorded as lentigo maligna.

Correspondence: Dr Eide, Department of Dermatology, Henry Ford Hospital, 3031 W Grand Blvd, Ste 800, Detroit, MI 48202 (meide1@hfhs.org).

Accepted for Publication: December 22, 2010.

Author Contributions: All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Eide and Krajenta. Acquisition of data: Eide, Krajenta, and D. Johnson. Analysis and interpretation of data: Eide, Jacobsen, Krajenta, D. Johnson, and C. C. Johnson. Drafting of the manuscript: Eide and D. Johnson. Critical revision of the manuscript for important intellectual content: Jacobsen, Krajenta, and C. C. Johnson. Statistical analysis: Jacobsen. Obtained funding: Eide and C. C. Johnson. Administrative, technical, and material support: Krajenta, D. Johnson, and C. C. Johnson. Study supervision: Eide.

Funding/Support: This study was supported in part by a Dermatology Foundation Career Development Award in Health Care Policy (Dr Eide).