Background: Cancer registries are the mainstay for Canadian population-based cancer statistics and research. Data is collected in provincial and territorial registries including the Nova Scotia Cancer Registry (NSCR). The goal of this study was to determine the accuracy of the NSCR data regarding the diagnosis, primary site, histological subtype, and stage of germ cell tumors (GCT) when compared to the individual pathology reports and staging investigations from the chart. Methods: This analysis included all NSCR patients diagnosed with GCT from 2006-2015. From the NSCR, the date and method of diagnosis, primary site, histology, and stage were recorded. This data was also extracted from each patient’s chart record. Any discrepancies between the two sources were reviewed and reasons behind the discrepancies were recorded. Results: 239 unique patients were identified in the NSCR during the specified time period. Ten were excluded as no chart records were available to confirm accuracy. 229 patients make up the study cohort. Using NSCR data 57.6% had seminoma, 34.5% nonseminoma (NSGCT), and 7.9% other. Discrepancies in pathology were noted in 29 patients (12.7%) for a number of reasons including no appropriate coding option in the NSCR, multiple tumors, biopsy only specimens with misinterpretation of tumor marker elevation, and true coding error. Using NSCR staging data (available in 185 cases) 71.9% had stage I, 12.4% stage II, 11.9% stage III, and 3.8% other. Discrepancies in stage were noted in 33 patients (17.9%) with NSCR data downstaging 10 patients (5.4%) and upstaging 19 patients (10.3%) predominantly due to miscoding patients as stage IS. The site of the primary GCT was discrepant in 12 patients (5.2%) due primarily to difficulty coding post chemotherapy orchiectomy specimens and burnt out primary testicular lesions. The date of diagnosis was accurate within 1 week for all patients except one which differed by several months. Conclusions: Canadian cancer registry data is population based and used for many purposes including policy decisions and research. The NSCR higher level data such as date of diagnosis and overall pathological diagnosis appears relatively accurate. However, there are inaccuracies in more detailed information like histological subtype and stage. This study raises awareness of these gaps. It also highlights key areas for improvement in educating registry personnel who interpret and enter data about the uniqueness of GCT pathology, staging and interpretation of tumor markers.