Why New Rules Are Necessary

Over the course of time, the existing rules, which had been in place for over 25 years, were found to be no longer optimally effective. SEER Quality Improvement staff had identified problems with the existing rules through a number of sources. Various quality improvement studies, including field re-abstracting audits, casefinding audits, and centralized web-based reliability studies performed over the past 5 years were reviewed. These quality improvement studies identified growing problems in consistency and coding accuracy in determining the most representative ICD-O histology code, particularly for complex or mixed histologic types, and in making correct decisions regarding single versus multiple primary cancers using the existing rules. There were numerous problems, for example:

• The existing rules were not written with computers in mind. SEER staff tried to flowchart the current rules & the first step in developing computer programs for determining multiple primaries and assigning histology codes & and found that it was impossible. This indicated that the logical order of the rules needed to be reviewed and revised.
• The rules for multiple primaries were set up as general rules only, but were full of site-specific exceptions. They did not address such issues as ICD-O site groups.
• Modifications, exceptions, and add-ons in the histology coding rules were made over the years in an attempt to cope with the increasing specificity of pathologic descriptive terminology and additional ICD-O histology codes, changing names within ICD-O histology groups, and mixed and combination histology codes.
• As a result, the rules were difficult to apply and confusing to navigate. Most registrars found it difficult to follow the logic.
• Central registries and facility-based registries began to develop idiosyncratic rules to meet the needs of their own registries and the caseload that they saw. Rules were becoming less and less consistent from agency to agency.
• Consequently, it was difficult to train a new registrar on determining multiple primaries and coding complex histologies.

When the International Classification of Diseases for Oncology, third edition (ICD-O-3) was implemented in 2001 in North America, over 400 new terms were added to the morphology codes list. These included several nonspecific terms, such as T-cell lymphoma, NOS; several combination morphology codes, such as combined small cell-large cell carcinoma; and many newly-identified histologic entities. Some of the new histologic entities had to be placed where codes were available, and the result was not always neatly in the same numeric range of codes. For example, non-small cell carcinoma (8046) has the same first three digits as small cell carcinoma.

ICD-O-3 does not define the terms it includes; therefore, registrars did not necessarily understand when some of the histologic entities should be used or when certain combination codes should be used. The coded terms were defined differently by different registrars. Furthermore, ICD-O rules do not define the circumstances when combination codes should be used to describe multiple tumors.

In addition, over the past 25 years, there have also been vast changes in clinical medicine.

• More sophisticated imaging techniques such as CT scanning and magnetic resonance imaging (MRI) have permitted diagnostic needle biopsies rather than open resection of tumors.
• Contemporary understanding of the cancer disease process has determined that some cell types are more relevant to the patient's prognosis than others.
• Advanced pathology and cytology laboratory techniques have given the pathologist much more insight into the composition of a tumor, including cytogenetics, tumor markers, special stains, and electron microscopy. This increased the number of cases in which multiple histology descriptors are applied to a single case.
• The resulting expanded descriptions in pathology reports may or may not have meaning for the registrar (and the clinician). Registrars in particular could not discern between descriptive characteristics, diagnostic histology terms, and other non-standard use of nomenclature by pathologists in the process of selecting the appropriate histology code.
• Also, our understanding of how humans process information has improved. Printed statements, such as standardized rules, are no longer 'one size fits all.'

Since the implementation of ICD-O-3 in 2001, both the SEER Inquiry System (SINQ) and the Commission on Cancer (CoC) Inquiry and Response (I&R) system have been inundated with requests for clarification of histology coding rules and multiple primary rules. Determination of single versus multiple primary decisions and histology coding rules have led requests for additional training and have been the subject of informal inquiries by researchers using cancer registry data.

In summary, the problems with the existing rules include:

• Difficulties determining whether there is one primary or more
• Too many descriptors and ambiguous terms in the pathology report
• Multiple choices for histology codes
• No hierarchy of rules when there are choices

Note: You will need your ICD-O-3 manual and pre-2007 FORDS or SEER Program Coding and Staging Manual for this exercise.

The Histology Project Committee set goals before starting to change the rules. The main goals were:

• Coding consistency. By clarifying the multiple primary and histology coding rules and prioritizing the sequence of the rules, including providing explanations and definitions of terms used in the rules themselves, coding reproducibility will be improved. Reproducibility means that, given the same case, registrars would consistently record the same histology.
• No change in incidence rates. The incidence rates are the measure of the cancer risk in a population. The incidence rates can be used to look at trends over time; in other words we can look at increases or decreases in risk for certain types of cancer, special populations, or age groups.
  • Example: If the rules were changed to say that only one breast cancer was collected per patient for their lifetime, the trends would show that there was a sharp decline in the risk of a woman having breast cancer.
• Improved data quality overall. If registrars are consistent in determining multiple primaries and coding histologies, the overall result will be a marked improvement in the data collected.