Disease Episode Representation Model

Cancer Diagnosis Model is a combination of Base Diagnosis and Diagnostic Modifiers

Base Diagnosis is stored in condition_occurrence table as any other condition in OMOP and are a combination of Histology + Anatomic Site. The concepts in this table come from the ‘Condition’ Domain. The Vocabulary used is a combination of SNOMED and OMOP generated concepts. SNOMED was not granular enough, so we augmented with ICDO3 which is the gold standard for representing cancer diagnosis.

Diagnostic Modifiers are stored in the Measurement table from under ‘Cancer Modifier’ Vocabulary. As any other Measurement record, Cancer Modifiers may be represented by one concept or a pair of concepts. There are different Cancer Modifier types/classes. These include Topography, Histopattern, Staging/Grading, Nodes, Metastasis, Margin, Extension/Invasion and dimension.

There is a way to establish a 1-many relationship between Base Diagnosis and Diagnostic Modifiers using Measurement table.

Cancer Treatments are represented in two levels: (1) Composite (high level) Treatment: representing treatment abstractions such as Regimens, Cycles, Treatment Courses or (2) Individual Treatments: such as Medication or Procedures. High level cancer treatments are covered in Episodes while the low level treatments are recorded in drug_exposure or procedure_occurrence tables with domains being ‘Drug’ and ‘Procedure’, respectively.

Cancer Episode Model Episodes are continuous periods of disease and treatments that have a distinct clinical meaning and are composed of multiple events. For example, ‘Progressive Disease’ Episode is a continuous period of disease when cancer was growing. ‘Treatment Regimen’ is a continuous period in which a treatment regimen was administered. Episodes are integral for conducting cancer research and addressing key questions such as Overall and Progression Free Survival, Time to Progression and Response to Treatment. Episodes can be derived directly from the source data like the Tumor Registry or algorithmically from electronic health records.

Types of Disease Episodes Three types of Episodes are introduced to the OMOP CDM:

1. Overarching Disease Episode: covers the entire disease duration. Overarching Disease Episode is the Parent Episode for the other Disease Episodes, Disease Extent and Disease Dynamic, and is always present.

2. Disease Extent: describes how far the cancer has spread. Episode starts at day of Diagnostic Modifier indicating a new Disease Extent and ends at start of a Treatment Episode or at day of a Diagnostic Modifier indicating different extent. Disease Extent includes:

   • Confined Disease: Organ or site-confined disease (pT1-pT4) without lymph node (regional) metastasis.
   • Regional (Invasive) Disease: Local/regional-confined disease with lymph node (regional) metastasis (any pT and N1 or above). Starts with the first positive Diagnostic Modifier of Concept Class "Invasion/Extension" beyond organ or regional lymph nodes and ends at day of the first positive modifier of Concept Class "Metastasis".
   • Distant (Metastatic) Disease: Evidence of metastasis. Starts with the first positive modifier of Concept Class "Metastasis" and ends at the time of death or the last observation.

3. Disease Dynamic: describes the status of cancer. Episode starts at day of Diagnostic Modifier indicating a new dynamic or Surgery and ends at day of a Diagnostic Modifier indicating different dynamic, or day of treatment. Surgery starts a new "Remission" Episode unless surgery report indicates otherwise.

   • Stable Disease
   • Remission - It is assumed that Surgery triggers a period of Remission. If its unknown whether there is a decrease in or disappearance of signs and symptoms of cancer, the state 'Remission' can be used instead of 'Complete Remission' and 'Partial Remission'.
   • Partial Remission - In 'Partial Remission', some, but not all, signs and symptoms of cancer have disappeared.
   • Complete Remission - In 'Complete Remission', all signs and symptoms of cancer have disappeared, although cancer still may be in the body.
   • Progression

Table 3.1

Name	episode_concept_id	Description	Rule
Episode of care	32533	Overarching disease episode	Starts from the initial diagnosis and ends at the time of death or the time of most recent follow-up. We always need to have the parent episode.
Distant (Metastatic) Disease	32944	Disease Extent	Any of the following applies (1) Modifier of Concept Class "Metastasis" OR (2) If any cM+'TNM Finding by AJCC/UICC 8th edition' OR any pM+TNM Finding by AJCC/UICC 8th edition' of Concept Class "AJCC Category" OR (3) for Summary Stage (Concept ID=35918652;Class='NAACCR Variable'; Domain='Measurement'; Vocabulary='NAACCR') there exists a 'Has Answer' relationship to 'Distant site(s) involved|- Distant site(s)....' (Concept ID=35919435;Class='NAACCR Value'; Domain='Meas Value'; Vocabulary='NAACCR') (See description in table 3.2 below)
Regional (Invasive) Disease	32943	Disease Extent	For Summary Stage (Concept ID=35918652;Class='NAACCR Variable'; Domain='Measurement'; Vocabulary='NAACCR') there exists a 'Has Answer' relationship to 'Regional by direct extension only....' (Concept ID=35919900;Class='NAACCR Value'; Domain='Meas Value'; Vocabulary='NAACCR') OR 'Regional lymph node(s) involved...' (Concept ID=35919773;Class='NAACCR Value'; Domain='Meas Value'; Vocabulary='NAACCR') OR 'Regional by BOTH direct extension AND regional lymph node...' (Concept ID =35919198;Class='NAACCR Value'; Domain='Meas Value'; Vocabulary='NAACCR') OR 'Regional (NOS)..' (Concept ID is not in Athena)(See description in table 3.2 below)
Confined Disease	32942	Disease Extent	For Summary Stage (Concept ID=35918652;Class='NAACCR Variable'; Domain='Measurement'; Vocabulary='NAACCR') there exists a 'Has Answer' relationship to 'Localized only|- Involvement of multiple nodal chains in the SAME lymph node....' (Concept ID=35919274;Class='NAACCR Value'; Domain='Meas Value'; Vocabulary='NAACCR')(See description in table 3.2 below)
Remission	32945	Disease Dynamic	Start with either diagnostic modifier or surgery. They always end with next treatment episode or new evidence.
Partial Remission	32949	Disease Dynamic	Start with evidence. In this case evidence is either diagnostic modifier or surgery. They always end with next treatment episode or new evidence.
Complete Remission	32945	Disease Dynamic	Start with evidence. In this case evidence is either diagnostic modifier or surgery. They always end with next treatment episode or new evidence.
Stable Disease	32948	Disease Dynamic	Start with evidence. In this case evidence is either diagnostic modifier or surgery. They always end with next treatment episode or new evidence.
Progression	32949	Disease Dynamic	Start with evidence. In this case evidence is either diagnostic modifier or surgery. They always end with next treatment episode or new evidence.

Table 3.2: The below table describes the Summary Stage codes. Applies to every site and/or histology combination, including lymphomas and leukemias:

Code	Description
0	In sit: Intraepithelial, noninvasive, or non-infiltrating cancer
1	Localized only: confined to the organ of origin without extension beyond the primary organ
2	Regional by direct extension only: direct extension to adjacent organs or structures
3	Regional lymph nodes involved only: direct extension to adjacent organs or structures by lymph node involvement
4	Regional by BOTH direct extension AND lymph node involvement: direct extension to adjacent organs or structures or by lymph node involvement
5	Regional, NOS (Not Otherwise Specified)
7	Distant site(s)/node(s) involved: spread to parts of the body remote from the primary tumor
9	Unknown/Death certificate only case

EPISODE Table

For a Disease Episode, to describe the remission of Bladder Cancer, the episode_concept_id field will contain the Concept ID for ‘Remission’. The episode_object_concept_id contains the Concept ID for Bladder Cancer. Together they form a sentence of the type ‘Remission of Bladder Cancer’.

For a treatment plan of a 12 month course of intervescial installation of epirubicin, the episode_concept_id will contain the Concept_ID for ‘Treatment Regimen’ and the episode_object_concept_id describes the value of the Treatment Regime which in this case will be the Concept ID of Epirubicin Monotheraphy. Together they mean ‘Treatment Regimen of Epirubicin Monotheraphy’.

To nest Treatment Episodes, we will be using the episode_parent_id field. In this case of the Treatment Episodes we’re representing Treatment Regimens and Treatment Cycles. The episode_parent_id field of the Treatment Cycle records will have the same value as the episode_id field of the Treatment Regimen record. So in this case the episode records that represent the Treatment Cycle will be considered childern of the episode record that represents the Treatment Regimen record. The episode_concept_id of the Treatment Cycle records will be the Concept_ID of ‘Treatemtn Cycle’ which is 32532. While the episode_object_concept_id of the Treatment Cycle records will be the same as the episode_object_concept_id of the Treatment Regime record which in this case is ‘Epirubicin Monotheraphy’.

To nest Disease Episodes, we will be using the episode_parent_id field. The parent episode is the overarching disease episode is a parent of the 3 extent of disease episodes -> confined disease, invasive disease and metastatic disease. The episode_parent_id of the disease extent episodes will be the same as the episode_concept_id of the overarching disease episode record.

Episodes are composed of two concepts (1) episode concept -> Defines the kind of epi Disease progression or treatment regimen (2) episode object -> describing nature of the episode these are concepts from Condition, Procedure and Drug domain. Here are some examples: