Data elements to use for analyses • prostateredcap

This data dictionary describes the full set of derived variables to use for analyses after running load_prostate_redcap() and check_prostate_redcap(recommended_only = TRUE). Additional variables may be available, especially if disabling recommended_only.

For definitions and underlying data collection, see the data dictionary of the REDCap database.

Patient-level data

Available in the datasets$pts dataset.

At diagnosis

Variable	Description	Levels	Reasons for missing values
`ptid`	Patient ID. Will be sequential integer number in deidentified data.	–	(No missing values)
`age_dx`	Age at prostate cancer diagnosis (in years)	Continuous	Date of birth or diagnosis unavailable (unusual¹)
`race4`	Self-reported race, 4 categories	Asian; Black or African American; White; Other (the latter category to avoid identifiability in uncommon categories)	Not reported
`race3`	Self-reported race, 3 categories	Asian; White; Black	Another category or not reported
`smoking`	Self-reported smoking status around first contact	Current; Former; Never	Not reported
`bx_gl34`	Gleason score at biopsy (diagnosis), grade-grouped	<7; 3+4; 4+3; 8; 9-10	Not available or Gleason score sum 7 with unknown major/minor pattern
`psa_dx`	Prostate-specific antigen at cancer diagnosis (ng/ml)	Continuous	Unavailable
`psa_dxcat`	Prostate-specific antigen at cancer diagnosis (ng/ml)	<4; 4-10; 10-20; >20	Unavailable
`lnpsa_dx`	Prostate-specific antigen at cancer diagnosis (ng/ml), log_e-transformed	Continuous	Unavailable
`stage`	Clinical TNM stage at diagnosis	N0/NX M0; N1 M0; M1	Metastasis (M) stage component unavailable (unusual¹)
`clin_tstage`	Clinical T (tumor) stage at diagnosis	T1/T2; T3; T4	Not available (many are M1)
`clin_nstage`	Clinical N (nodal) stage at diagnosis	N1 M0; N1 M0	Not available (many are M1)
`mstage`	M (metastasis) stage at diagnosis	M0; M1	Unavailable (unusual¹)

¹ A dataset processed through check_prostate_redcap() excludes records with missing values in the key characteristics, age at diagnosis and M stage. Missingness will only occur if QC filters in check_prostate_redcap() were manually disabled.

Primary treatment

Variable	Description	Levels	Reasons for missing values
`rxprim`	Primary treatment	Many levels due to treatment combinations; also “No Primary Therapy” and “Other”	Unknown
`rxprim_rp`	Primary treatment included radical prostatectomy	`TRUE`; `FALSE`	None (`FALSE` if no report about prostatectomy)
`rxprim_adt`	Primary treatment included androgen deprivation therapy	`TRUE`; `FALSE`	None (`FALSE` if no report about primary ADT)
`rxprim_chemo`	Primary treatment included chemotherapy	`TRUE`; `FALSE`	None (`FALSE` if no report about primary chemotherapy)
`rxprim_xrt`	Primary treatment included radiation therapy	`TRUE`; `FALSE`	None (`FALSE` if no report about XRT)
`rxprim_other`	Primary treatment included free-text	`TRUE`; `FALSE`	None (`FALSE` if no free-text)
`rxprim_oth`	Primary treatment, other	Freetext	Predefined treatment combination used, or primary treatment unknown
`rp_gl34`	Gleason score at radical prostatectomy, grade-grouped	<7; 3+4; 4+3; 8; 9-10	Not available, no radical prostatectomy, or Gleason score sum 7 with unknown major/minor pattern
`path_t`	pT (tumor) stage at radical prostatectomy	2; 2a; 2b; 2c;; 3; 3a; 3b; 4	Unknown or no radical prostatectomy
`path_n`	pN (nodal) stage at radical prostatectomy	0; 1	Unknown or no radical prostatectomy

Sample-level data

Available in the datasets$smp dataset.

Variable	Description	Levels	Reasons for missing values
`ptid`	Patient ID. Will be sequential integer number in deidentified data. Matches `ptid` in the `pts` dataset.	–	(No missing values)
`dmpid`	Sample ID	–	(No missing values)
`hist_smp`	Histology of the sample	Adenocarcinoma / poorly differentiated carcinoma; Adenocarcinoma/poorly differentiated carcinoma with ductal or intraductal features; Adenocarcinoma/poorly differentiated carcinoma with Neuroendocrine features; Other (Text box); Pure Small Cell / Neuroendocrine Carcinoma	Unknown
`dzextent_smp`	Disease extent at sample (biopsy)	Localized; Regional nodes; Metastatic hormone-sensitive; Non-metastatic castration-resistant; Metastatic castration-resistant; Metastatic, variant histology	Unavailable (unusual, see footnote above)
`dzextent_seq`	Disease extent at sequencing	Localized; Regional nodes; Metastatic hormone-sensitive; Non-metastatic castration-resistant; Metastatic castration-resistant; Metastatic, variant histology	Unavailable (unusual, see footnote above)
`ext_pros`	Disease extent at sampling includes prostate	`TRUE`; `FALSE`	None (`FALSE` if no known prostate disease)
`ext_lndis`	Disease extent at sampling includes distant lymph nodes	`TRUE`; `FALSE`	None (`FALSE` if no known distant lymph node)
`ext_bone`	Disease extent at sampling includes bone	`TRUE`; `FALSE`	None (`FALSE` if no known bone disease)
`ext_vis`	Disease extent at sampling includes visceral disease (liver, lung, other soft tissue)	`TRUE`; `FALSE`	None (`FALSE` if no known visceral disease)
`ext_liver`	Disease extent at sampling includes liver	`TRUE`; `FALSE`	None (`FALSE` if no known liver disease)
`ext_lung`	Disease extent at sampling includes lung	`TRUE`; `FALSE`	None (`FALSE` if no known lung disease)
`ext_other`	Disease extent at sampling includes other soft tissue	`TRUE`; `FALSE`	None (`FALSE` if no known other soft tissue)
`bonevol`	Volume of bone disease at sample (biopsy)	High-Volume Bone Metastases; Low-Volume Bone Metastases	Unknown or none bone metastases
`cntadt`	Sample on continuous androgen deprivation therapy	Yes; No	Unknown
`tissue`	Sample tissue	Bone; Liver; Lung; Lymph Node; Other soft tissue; Prostate	None
`smp_pros`	Sample tissue is prostate	Prostate; Non-prostate	None
`smp_tissue`	Sample tissue (collapse)	Prostate; Lymph node; Bone; Visceral; Other soft tissue	None
`primmet_smp`	Disease extent at sample: primary or regional nodes (“Primary”) vs. all others (“Metastatic”)	Primary; Metastatic	Unknown
`age_smp`	Age at sample (biopsy, in years)	Continuous	Unavailable (unusual, see footnote above)
`age_seq`	Age at sequencing (in years)	Continuous	Unavailable (unusual, see footnote above)
`dx_smp_mos`	Time from diagnosis to sample (biopsy, in months)	Continuous	Unavailable (unusual, see footnote above)
`adt_smp_mos`	Time from ADT initiation to sample (biopsy, in months)	Continuous	Unavailable
`dx_seq_mos`	Time from diagnosis to sequencing (in months)	Continuous	Unavailable (unusual, see footnote above)
`dzvol`	Disease volume at sample (biopsy): composite of disease extent (high if visceral) and bone volume (high if >=4 bone metastases)	Low-volume disease; high-volume disease	Non-metastatic disease at sample
`denovom_smp`	De-novo metastatic disease: metastases since diagnosis (M1) vs. M0 at diagnosis and metastases detected after diagnosis but before sample	Metastatic recurrence; de-novo metastatic	Non-metastatic disease at sample
`denovom_seq`	De-novo metastatic disease: metastases since diagnosis (M1) vs. M0 at diagnosis and metastases detected after diagnosis but before sequencing	Metastatic recurrence; de-novo metastatic	Non-metastatic disease at sequencing

Outcomes

Setup:

Indicator (event) variables are in datasets$pts, because events can only occur once per person.
Time variables are in datasets$smp, because time intervals between start of follow-up and event depend on when start of follow-up is set and thus differ for different samples from the same person. For analyses, select samples and then merge patient data by ptid.

Time variables will be missing:

If the patient is not at risk of the event at the start of follow up because the event already occurred, e.g., metastasis when starting follow-up in metastatic castration-resistant disease.
If the patient is not at risk because the event cannot occur, e.g., a variant histology indicates castration resistance and the patient will by definition not be followed for this outcome.
If the event time is unknown.

Outcome: Transition	Population of Interest	Event variable (Descriptive)	Event variable (modeling)	Time variable (from sequencing)	Time variables (late-entry models)¹
Metastasis: Sequencing to metastasis	Disease extent “Localized” or “Regional nodes” at sequencing	`is_met`: No; Yes	`met_event`: 0; 1	`seq_met_mos`: Months from sequencing to metastasis or last clinic visit	`dx_met_mos`: Months from diagnosis to metastasis or last clinic visit
Metastasis-free survival: Sequencing to metastasis or death from any cause (composite endpoint)	Disease extent “Localized” or “Regional nodes” at sequencing	`is_mfs`: Event-free; Metastasis/death	`mfs_event`: 0; 1	`seq_mfs_mos`: Months from sequencing to metastasis, death, or last clinic visit (if neither)	`dx_mfs_mos`: Months from diagnosis to metastasis, death, or last clinic visit (if neither)
CRPC: Sequencing to castration resistance	Disease extent “Localized”, “Regional nodes”, or “Metastatic hormone-sensitive” at sequencing	`is_crpc`: No; Yes; N/A-Pure Other Histology at Diagnosis; N/A-Pure small cell/neuroendocrine at diagnosis	`crpc_event`: 0; 1; missing (if variant histology)	`seq_crpc_mos`: Months from sequencing to castration resistance or last clinic visit	`dx_crpc_mos`: Months from diagnosis to castration resistance or last clinic visit
Overall survival: Sequencing to death	All	`is_dead`: Alive; Dead	`death_event`: 0; 1	`seq_os_mos`: Months from sequencing to death or last contact	`dx_os_mos`: Months from diagnosis to death or last contact; `adt_os_mos`: Months from ADT initiation to death or last contact; `met_os_mos`: Months from metastasis to death or last contact

¹ In late entry models, participants should enter risk sets at the time of sequencing using appropriate variables, i.e., dx_seq_mos (diagnosis to sequencing), adt_seq_mos (ADT initiation to sequencing), or met_seq_mos (metastasis to sequencing).