Validation of Testicular Germ Cell Tumor Staging in Nationwide Cancer Registries - Beyond the Abstract

Cancer staging data from the nationwide cancer databases are widely used by researchers to evaluate the efficacy of treatment and propose new strategies for disease management. To make appropriate conclusions, the accuracy of this data is paramount. Tumor registries at individual institutions review patient records to abstract data, which ultimately is incorporated into national cancer databases and accessed by researchers.


Testicular cancer is a rare malignancy occurring in ~1/200 men between 20 and 40 years of age.1 Testis cancer staging is unique in that it incorporates S stage, based on post-orchiectomy serum tumor markers level, to correctly code the composite stage (CS) of disease.

In our study, we conducted a validation analysis of our testicular cancer database and compared urologists staging with the coding done by abstracters at our institutional cancer registry. We found that many inconsistencies exist in the coding of testicular cancer stage by our cancer registry when compared with staging done by urologists. While T, N, M stages had high concordance, S and composite stages had poor concordance rates of 54% of 39%, respectively.

Furthermore, the institutional tumor registrars were unable to abstract some of the data, leading to it being listed as unavailable in national databases; this missing data is likely excluded for research. Our analysis that excludes this missing data demonstrates that CS I is underrepresented, and CS II and III are overrepresented in data used by researchers. Specifically, data used by researchers has 47% fewer CS I, 50% higher CS II, and 69% higher CS III patients than the actual cohort from our institution. For researchers using population databases, the decision to include/exclude patients with missing data appears to introduce a massive selection bias that may lead to errant conclusions.

To understand the source of errant abstraction, we conducted an educational intervention with the registry which led to improved but not high concordance. We gleaned that abstractors have poor instruction for testis cancer staging and have limited patient chart access. We conclude that using national cancer databases to obtain results regarding testis cancer care may lead to inaccurate conclusions. At this time, multi-institutional databases may serve as more credible sources of data though they lack a large sample size. We highlight the need for improved clarification of staging guidelines, an internal institutional audit to evaluate the concordance of testis cancer staging, and evaluation of concordance data for other rare cancer types.

Written by: Rohit R. Badia, BS, Department of Urology, University of Texas Southwestern Medical Center, Dallas, TX & Aditya Bagrodia, MD, Department of Urology, Disease Team Co-Leader Genitourinary Oncology, UC San Diego Health, La Jolla, CA

References:

  1. Kinkade S. Testicular cancer. Am Fam Physician. 1999;59(9):2539-2544, 2549-2550.

Read the Abstract
email news signup