Registration of molecular markers in population-based cancer registries: Hopes and hurdles Free

In 2018 Brain Molecular Markers (BMM) were incorporated as a site-specific data item in the Uniform Data Standards (UDS) of the North American Association of Central Cancer Registries (NAACCR).¹ The NAACCR is setting data standards used by cancer registries in the US and Canada. The development of the BMM was partially driven by a lack of unique International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) morphology codes for tumor types with a “histomolecular” definition in the revised fourth edition of the World Health Organization Classification of Tumors of the Central Nervous System (WHO-CNS4R).² Incorporating BMM into registration practices would allow for more accurate reporting of WHO-CNS4R entities.

Using the National Cancer Database as a reference, the recent publication by Neff et al³ provides insight in the reporting completeness of and testing patterns for the BMM for adult-type diffuse gliomas. The authors report a completeness of 81% for patients aged 40 years and older. However, testing/reporting completeness varied by extent of resection and hospital characteristics. As testing/reporting completeness serves as the basis to draw valid conclusions on BMM, the authors provided a first insight into the translational value of these markers for epidemiological and clinical research purposes.

Accurate and detailed registration is key for monitoring the burden of cancer and to inform health decision makers. Although the NAACCR is setting standards for registration practices in North America, variability in registration and data quality remain important issues when analyzing registry data in a broader international context. A good example of the wide variability was recently presented by Girardi et al⁴ who analyzed registry data for brain tumors on a global scale for the period 2000–2014.

Inherently, cancer registries are lagging behind on changes in tumor classification. This is reflected by the ICD-O-3 coding scheme used in the most recent WHO-CNS5 (2021) classification where codes have not yet been updated and different tumors are still grouped under the same ICD-O morphology code (eg, ICD-O-M9385/3 for all pediatric-type diffuse high-grade gliomas). Obviously, this makes it impossible to discriminate the tumor types within this group based on the ICD-O codes.⁵

Including molecular markers like the BMM in cancer registries is potentially a valuable effort to improve data accuracy. In addition, including molecular information in cancer registries can facilitate health care evaluation (eg, for quality assurance purposes). Moreover, molecular markers have the potential to provide insight into prognostic factors (eg, isocitrate dehydrogenase [IDH] mutation status) and predictive factors (eg, MGMT methylation status), and can facilitate research efforts related to cancer risk and therapy response monitoring and evaluation.⁶ However, population-based registries are intrinsically less suitable for novel molecular marker discovery and their role in population-based registries will therefore most likely be limited to markers that are already established as important in the field.

Also, an important complexity when registering the molecular markers themselves is that the underlying diagnostic techniques, which have an important impact on the quality of the results obtained, are not accounted for. For example, the diagnosis of 1p/19q-codeleted status as diagnosed by fluorescence in situ hybridization (FISH) can produce false positive results in cases with an incomplete loss of 1p and 19q, but such nuances are lost when only 1p/19q is registered. On the other hand, according to the WHO-CNS5 one does not necessarily need to perform molecular testing for a diagnosis of Glioblastoma, IDH-wildtype in patients over 55 years of age of which the CNS tumor has a phenotype fully compatible with that diagnosis and immunohistochemistry for the IDH1 R132H mutant protein is negative. In other words: Lack of information on molecular tests for such patients in cancer registries does not necessarily mean that, according to WHO-CNS5 standards, the work-up of the case has been suboptimal. Ideally, in such a situation the diagnosis in the cancer registry would be Glioblastoma, IDH-wildtype, indicating that the IDH-status in that case has been assessed by other means. These examples stress the need for detailed guidelines and variables accounting for the underlying diagnostic process when registering molecular markers.

It remains challenging for registries to keep up with the rapidly advancing field of molecular pathology and the clinical implications that are the result thereof. In this context it is also important to note that cancer registries are historically not equipped to register molecular characteristics. In addition, for most countries implementing molecular markers will be strenuous due to costs and time restrictions as many cancer registries are under-resourced.

Overall, the bigger issue seems to be that registries currently do not have a tumor coding system in place that is able to keep up with the latest developments.

To address this issue a more flexible coding system, the 11th revision of the International Classification of Diseases and Related Health Problems (ICD-11), has been suggested.⁷ ICD-11 is updated on a yearly basis, making it easier to more closely follow the latest developments in tumor classification. Additional advantages are the possibility to register more detailed information on tumor location, which may make the registry data more useful for clinicians, and the fact that in ICD-11 the behavior code (5th digit in ICD-O morphology code) has been abandoned for CNS tumors. As classifying tumors based on their behavior can result in biased estimates of incidence and survival, this latter change can be expected to facilitate more accurate comparison between countries.⁸ However, in line with molecular markers implementing ICD-11 comes with comparable challenges such as additional costs and time investments.

In conclusion, although registration for a selected group of molecular markers in population-based cancer registries have the potential to improve data quality and facilitate health care evaluation processes, their role should not be overestimated as registration comes with several limitations and complexities. Meanwhile, cancer registries inherently lag behind, and flexibility in tumor coding practices is key for keeping up with the nowadays rapidly evolving changes in classification of CNS tumors. Therefore, if population-based cancer registries want to increase their clinical relevance, novel coding strategies need to be considered.

Funding

Not applicable.

Conflict of Interest:

The authors declare that they have no conflict of interest.

Authorship

Study design: RH, PW; data collection and analyses: NA; data interpretation and writing of the manuscript: RH, PW; revision of the manuscript, and final approval of the manuscript: RH, PW. The text is the sole product of the authors and no third party had input or gave support to its writing.

References