2023 CLB Multisite

DiB badge DOI

This folder contains the pathology report data on lymphatic involvement in 373 patients with head and neck cancer, diagnosed and treated with neck dissection at the Centre Léon Bérard between 2001 and 2018.

The raw.csv data does contain patients which are duplicates of the patient records in the 2021-clb-oropharynx folder, but they have been filtered out in the data.csv file.

Table of Contents

Cohort Characteristics

Below we show some figures that aim to coarsely characterize the patient cohort in this directory.

age distribution
Figure 1: Distribution over age, stratified by sex and smoking status.
T-category distribution subsite distribution
Figure 2: Distribution over age, stratified by sex and smoking status. Figure 3: Distribution over primary tumor subsite.

Curation

Curation and inclusion criteria will be published in a separate Data in Brief article that is currently under review.

Online Interface

We provide a user-friendly and intuitive graphical user interface to view the dataset, which is available at https://lyprox.org/. The GUI has two main functionalities: the patient list and the dashboard. The patient list allows for viewing the characteristics of a patient, corresponding to one row of the csv file, in a visually appealing and intuitive way. The dashboard allows for filtering of the dataset. For example, the user may select all patients with primary tumors extending over the mid-sagittal plane with involvement of ipsilateral level III. The dashboard will then display the number or percentage of patients with metastases in each of the other LNLs.

Data Description

The data is provided as a CSV-table containing one row for each of the 373 patients. The table has a header with three levels that describe the columns. Below we explain each column in the form of a list with three levels. So, for example, list entry 1.i.g refers to a column with the three-level header patient | # | alcohol_abuse and underneath it we report each patient's history of alcohol abuse.

Documentation of Columns

  1. patient: This top-level header contains general patient information.
    1. #: The second level header for the patient columns is only a placeholder.
      1. id: The patient ID.
      2. institution: The institution where the patient was treated.
      3. sex: The biological sex of the patient.
      4. age: The age of the patient at the time of diagnosis.
      5. weight: The weight of the patient at the time of diagnosis.
      6. diagnose_date: The date of surgery because the raw file does not specify a date of diagnosis.
      7. alcohol_abuse: Whether the patient was abusingly drinking alcohol at the time of diagnosis.
      8. nicotine_abuse: Whether the patient was smoking nicotine at the time of diagnosis.
      9. hpv_status: The HPV p16 status of the patient.
      10. neck_dissection: Whether the patient underwent a neck dissection. In this dataset, all patients underwent a neck dissection.
      11. tnm_edition: The edition of the TNM classification used.
      12. n_stage: The pN category of the patient (pathologically assessed).
      13. m_stage: The M category of the patient.
      14. extracapsular: Whether the patient had extracapsular spread. In this dataset, this information is only globally available, not for each individual lymph node level.
  2. tumor: This top-level header contains general tumor information.
    1. 1: The second level header enumerates synchronous tumors.
      1. location: The location of the tumor. This is empty for all patients because we can later infer it from the subsite's ICD-O-3 code.
      2. subsite: The subsite of the tumor, specified by ICD-O-3 code.
      3. central: Whether the tumor is located centrally w.r.t. the mid-sagittal plane.
      4. extension: Whether the tumor extended over the mid-sagittal line.
      5. volume: The volume of the tumor in cm^3.
      6. stage_prefix: The prefix of the T category.
      7. t_stage: The T category of the tumor.
  3. pathology: This top-level header contains information from the pathology that received the LNLs resected during the neck dissection.
    1. info: This second-level header contains general information.
      1. date: The date of the pathology report (same as surgery).
    2. ipsi: This reports the involvement of the ipsilateral LNLs.
      1. III: For example, this column reports the involvement of the ipsilateral LNL III.
    3. contra: This reports the involvement of the contralateral LNLs.
      1. V: This column reports the pathologic involvement of the contralateral LNL V.
  4. diagnostic_consensus: This top-level header contains information about the diagnostic consensus, which we assumed to be negative for each LNL that was not resected during the neck dissection. However, we do not know if it was positive for resected patients. This means, all columns under this top-level header are essentially inferred from looking at missing entries under the pathology columns.
    1. info: This second-level header contains general information.
      1. date: The date of the diagnostic consensus (same as surgery).
    2. ipsi: This reports the diagnostic consensus of the ipsilateral LNLs.
      1. Ib: E.g., this column reports the diagnostic consensus of the ipsilateral LNL Ib.
    3. contra: This reports the diagnostic consensus of the contralateral LNLs.
      1. III: Under this column, we report the diagnostic consensus of the contralateral LNL III.
  5. total_dissected: This top-level header contains information about the total number of dissected and pathologically investigated lymph nodes per LNL.
    1. info: This second-level header contains general information.
      1. date: The date of the neck dissection.
    2. ipsi: This reports the total number of dissected lymph nodes per ipsilateral LNL.
      1. II: For instance, this column reports the total number of dissected lymph nodes in ipsilateral LNL II.
      2. Ib_to_III: This column reports the total number of dissected lymph nodes in ipsilateral LNL Ib to III. This column exists for convenience because we created a figure based on this.
    3. contra: This reports the total number of dissected lymph nodes per contralateral LNL.
      1. VII: While this column reports the total number of dissected lymph nodes in contralateral LNL VII.
      2. Ib_to_III: This column reports the total number of dissected lymph nodes in contralateral LNL Ib to III. This column exists for convenience because we created a figure based on this.
  6. positive_dissected: This top-level header contains information about the number of dissected lymph nodes per LNL that were pathologically found to be positive.
    1. info: This second-level header contains general information.
      1. date: The date of the neck dissection.
    2. ipsi: This reports the number of dissected lymph nodes per ipsilateral LNL that were pathologically found to be positive.
      1. IV: Here, we report the number of metastatic lymph nodes in ipsilateral LNL IV.
      2. Ib_to_III: This column reports the number of metastatic dissected lymph nodes in ipsilateral LNL Ib to III. This column exists for convenience because we created a figure based on this.
    3. contra: This reports the number of dissected lymph nodes per contralateral LNL that were pathologically found to be positive.
      1. Ia: And this column reports the number of metastatic lymph nodes in contralateral LNL Ia.
      2. Ib_to_III: This column reports the number of metastatic dissected lymph nodes in contralateral LNL Ib to III. This column exists for convenience because we created a figure based on this.

module mapping

Map the raw.csv data from the 2023-clb-multisite cohort to the data.csv file.

This module defines how the command lyscripts data lyproxify (see here for the documentation of the lyscripts module) should handle the raw.csv data that was extracted at the Centre Léon Bérard in order to transform it into a LyProX-compatible data.csv file.

The most important definitions in here are the list EXCLUDE and the dictionary COLUMN_MAP that defines how to construct the new columns based on the raw.csv data. They are described in more detail below:

global EXCLUDE

List of tuples specifying which function to run for which columns to find out if patients/rows should be excluded in the lyproxified data.csv.

The first element of each tuple is the flattened multi-index column name, the second element is the function to run on the column to determine if a patient/row should be excluded:

python EXCLUDE = [ (column_name, check_function), ]

Essentially, a row is excluded, if for that row check_function(raw_data[column_name]) evaluates to True.

More information can be found in the documentation of the lyproxify function.

global COLUMN_MAP

This is the actual mapping dictionary that describes how to transform the raw.csv table into the data.csv table that can be fed into and understood by LyProX.

See here for details on how this dictionary is used by the lyproxify script.

It contains a tree-like structure that is human-readable and mimics the tree of multi-level headers in the final data.csv file. For every column in the final data.csv file, the dictionary describes from which columns in the raw.csv file the data should be extracted and what function should be applied to it.

It also contains a __doc__ key for every sub-dictionary that describes what the respective column is about. This is used to generate the documentation for the README.md file of this data.

Global Variables

  • TNM_COLS
  • IB_TO_III_DISSECTED
  • EXCLUDE
  • COLUMN_MAP

function smpl_date

python smpl_date(entry: str) → str

Parse date from string.

function smpl_diagnose

python smpl_diagnose(entry: str | int, *_args, **_kwargs) → bool

Parse the diagnosis.

function robust

python robust(func: Callable) → Optional[Any]

Wrapper that makes any type-conversion function 'robust' by simply returning None whenever any exception is thrown.

function get_subsite

python get_subsite(entry: str, *_args, **_kwargs) → str | None

Get human-readable subsite from ICD-10 code.

function parse_pathology

python parse_pathology(entry, *_args, **_kwargs) → bool | None

Transform number of positive nodes to True, False or None.

function set_diagnostic_consensus

python set_diagnostic_consensus(entry, *_args, **_kwargs)

Return False, meaning 'healthy', when no entry about a resected LNL is available or when the pathology report says it was healhty. This is a hack to tackle the issue described here:

https://github.com/rmnldwg/lyprox/issues/92

function extract_hpv

python extract_hpv(value: int | None, *_args, **_kwargs) → bool | None

Translate the HPV value to a boolean.

function strip_letters

python strip_letters(entry: str, *_args, **_kwargs) → int

Remove letters following a number.

function clean_cat

python clean_cat(cat: str) → int

Extract T or N category as integer from the respective string. I.e., turn 'pN2+' into 2.

function get_tnm_info

python get_tnm_info(ct7, cn7, pt7, pn7, ct8, cn8, pt8, pn8) → tuple[int, int, int, str]

Determine the TNM edition used based on which versions are available for T and/or N category.

function get_t_category

python get_t_category(*args, **_kwargs) → int

Extract the T-category.

function get_n_category

python get_n_category(*args, **_kwargs) → int

Extract the N-category.

function get_tnm_version

python get_tnm_version(*args, **_kwargs) → int

Extract the TNM version.

function get_tnm_prefix

python get_tnm_prefix(*args, **_kwargs) → str

Extract the TNM prefix.

function check_excluded

python check_excluded(column: pandas.core.series.Series) → Index

Check if a patient/row is excluded based on the content of a column.

For the 2022 CLB multisite dataset this is the case when the first column with the three-level header ("Bauwens", "Database", "0_lvl_2") is not empty or does not contain the character 'n'.

function sum_columns

python sum_columns(*columns, **_kwargs) → int

Sum the values of multiple columns.