Tempus_QC_Prostate

Requirements

Prostate cancer diagnosis with date
RNA biopsy with date
Pluvicto with date (using the onco_care_plan table) – generic med name is lutetium lu 177 vipivotide tetraxetan
For patients with biopsy post-Pluvicto, please identify patients that have a regimen that begins between Pluvicto and biopsy, we will need counts for this in early January Note: for the purposes of the QC, you can remove all radiotherapy instances; they should not be used to determine if a regimen is flanking a biopsy
Number of patients that were previously delivered (this can be found in docs/Data_Model/legacy_ID_mapping; you can match the ID to the patients delivered in the prior Prostate intermediate files)

library(tidyverse)
library(data.table)
folder <- "/stash/data/clin/external-collaboration/tempus/P-20251203-0001/delivery_20251205/"
drug_name <- "lutetium lu 177 vipivotide tetraxetan"

# Total Patients
onco_patient <- read_csv(paste0(folder, "/data/Clinical/onco_patient.csv"), show_col_types = FALSE)
patients_all <- onco_patient$patient_id %>% unique
print(paste("Total Number of Patients:", length(patients_all)))

## [1] "Total Number of Patients: 211"

# Patients with a diagnosis date (not a year-only date)
onco_diagnosis <- read_csv(paste0(folder, "/data/Clinical/onco_diagnosis.csv"), show_col_types = FALSE)
patients_diagnosis <- onco_diagnosis %>% 
  dplyr::filter(grepl("^\\d{4}-\\d{2}-\\d{2}$", initial_diagnosis_date)) 
rejected_diagnosis <- setdiff(patients_all, patients_diagnosis$patient_id)
print(paste("Patients without a diagnosis full date:", length(rejected_diagnosis)))

## [1] "Patients without a diagnosis full date: 16"

print("List of Patients without a diagnosis full date:")

## [1] "List of Patients without a diagnosis full date:"

df <- onco_diagnosis %>% filter(patient_id %in% rejected_diagnosis) %>% dplyr::select( patient_id, initial_diagnosis_date)
DT::datatable(
  df,
  extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  )
)

# Patient with a RNA biopsy date (not a year-only date)
onco_meta_biospecimen <- read_csv(paste0(folder, "/data/Group_Level_Molecular//onco_meta_biospecimen.csv"), show_col_types = FALSE)
patients_biopsy <- onco_meta_biospecimen %>% 
  dplyr::filter(grepl("RNA", polymer_list)) %>% 
  mutate(RNA_date_check = grepl("^\\d{4}-\\d{2}-\\d{2}$", collection_date))
no_rna_date <- patients_biopsy %>% dplyr::filter(RNA_date_check == FALSE)
rejected_rna_date <- no_rna_date$patient_id
print(paste("Patient without a RNA biopsy full date:", nrow(no_rna_date)))

## [1] "Patient without a RNA biopsy full date: 1"

df <- no_rna_date %>% select(biospecimen_id, patient_id, polymer_list, collection_date)
print("List of Patients without a RNA biopsy full date:")

## [1] "List of Patients without a RNA biopsy full date:"

DT::datatable(  df,
                extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  )
)

# Pluvicto with date
onco_care_plan <- read_csv(paste0(folder, "/data/Clinical/onco_care_plan.csv"), show_col_types = FALSE)

## Warning: One or more parsing issues, call `problems()` on your data frame for details,
## e.g.:
##   dat <- vroom(...)
##   problems(dat)

drug_date <- onco_care_plan %>% 
  dplyr::mutate(
    drug_has_date = case_when(
      grepl(drug_name, agents) & grepl("^\\d{4}-\\d{2}-\\d{2}$", start_date) ~ TRUE,
      grepl(drug_name, agents) & !grepl("^\\d{4}-\\d{2}-\\d{2}$", start_date) ~ FALSE,
      TRUE ~ NA)
  ) %>% filter(!is.na(drug_has_date))
rejected_pluvicto_date <- drug_date %>% dplyr::filter(!drug_has_date) %>% pull(patient_id)
output <- drug_date %>% pull(drug_has_date) %>% table(useNA = "always")
print(paste("Pluvicto without a full date:", output[1]))

## [1] "Pluvicto without a full date: 2"

df <- drug_date %>% filter(drug_has_date == FALSE) %>% dplyr::select(care_plan_id, patient_id, agents, start_date)
print("List of Pluvicto without a full date:")

## [1] "List of Pluvicto without a full date:"

DT::datatable(  df,
                extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  )
)

# - The regimen containing the Pluvicto occurs either directly before or directly 
# after the biopsy (it should be within X days, with no other regimens occurring between Pluvicto and biopsy). 
# All regimens can be grouped together using care_plan_id
# Extract drug-only regimens
drug_plans <- onco_care_plan %>% dplyr::filter(grepl(drug_name, agents) & !grepl(".*radiotherapy.*", therapy_class, ignore.case = T)) %>% 
  dplyr::select(care_plan_id, patient_id, regimen_id, agents, start_date_indexed, end_date_indexed)
# Extract non-edrug regimens
non_drug_plans <- onco_care_plan %>% dplyr::filter(!grepl(drug_name, agents) & 
                                                  !is.na(regimen_id)  & !grepl(".*radiotherapy.*", therapy_class, ignore.case = T)) %>% 
  dplyr::select(care_plan_id, patient_id, regimen_id, agents, start_date_indexed, end_date_indexed)
# Extract relevant biopsis
biopsis <- onco_meta_biospecimen %>%   
  dplyr::filter(grepl("RNA", polymer_list)) %>%  
  dplyr::filter(grepl("^\\d{4}-\\d{2}-\\d{2}$", collection_date_indexed)) %>%
  dplyr::select(meta_biospecimen_id, patient_id, collection_date_indexed) 
# Merge everything together
joined <- full_join(drug_plans, non_drug_plans, by=c( "patient_id"), relationship = "many-to-many") %>%
  full_join(biopsis, by=c( "patient_id"), relationship = "many-to-many") %>% 
  full_join(onco_diagnosis %>% dplyr::select(patient_id, initial_diagnosis_date_indexed), by=c( "patient_id"), relationship = "many-to-many") 
  
# Set some inclusions flags
joined <- joined %>% 
  # First replace NA date with today date
  mutate(
         start_date_indexed.x = if_else(is.na(start_date_indexed.x), initial_diagnosis_date_indexed, start_date_indexed.x),
         end_date_indexed.x = if_else(is.na(end_date_indexed.x), today(), end_date_indexed.x),
         start_date_indexed.y = if_else(is.na(start_date_indexed.y), initial_diagnosis_date_indexed, start_date_indexed.y),
         end_date_indexed.y = if_else(is.na(end_date_indexed.y), today(), end_date_indexed.y),
         ) 
# For patients with biopsy post-Pluvicto, 
# identify patients that have a regimen that begins between Pluvicto and biopsy
patients_post_drug <- joined %>% 
  dplyr::select(patient_id, meta_biospecimen_id, regimen_id.x, regimen_id.y, agents.x, agents.y, 
                start_date_indexed.x, end_date_indexed.x, start_date_indexed.y, end_date_indexed.y, 
                collection_date_indexed) %>% 
  dplyr::filter(collection_date_indexed > start_date_indexed.x & 
                    (between(start_date_indexed.y, start_date_indexed.x, collection_date_indexed) | 
                       between(start_date_indexed.y, end_date_indexed.x, collection_date_indexed))
  )
df <- patients_post_drug %>% dplyr::select(patient_id, regimen_id.x, regimen_id.y, agents.x, agents.y, start_date_indexed.x, end_date_indexed.x, start_date_indexed.y, end_date_indexed.y, collection_date_indexed)
print(paste("Patients with biopsy post-Pluvicto that have a regimen that begins between Pluvicto and biopsy:", length(df$patient_id %>% unique)))

## [1] "Patients with biopsy post-Pluvicto that have a regimen that begins between Pluvicto and biopsy: 24"

rejected_regimen_between <- df$patient_id %>% unique
print(paste("List of biopsies post-Pluvicto that have a regimen that begins between Pluvicto and the biopsy:"))

## [1] "List of biopsies post-Pluvicto that have a regimen that begins between Pluvicto and the biopsy:"

DT::datatable(  df,
                extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  )
)

# Number of patients that were previously delivered
legacy_mapping <- read_csv(paste0(folder, "docs/Data_Model/legacy_ID_mapping_20251204.csv"), show_col_types = FALSE)
previous_deliveries <- c(
  "/stash/results/dev/teams/trans_epi/prostate/tempus/intermediate_files/delivery_2024_01_16/internal/",
  "/stash/results/dev/teams/trans_epi/prostate/tempus/intermediate_files/delivery_2025_08_21/internal/"
)
previous_patients <- unique(unlist(lapply(previous_deliveries, 
                                   function(f) read_csv(paste0(f, "demographic.csv"), show_col_types = F)$pt_id)))
previously_delivered <- legacy_mapping %>% dplyr::filter(id_value %in% onco_patient$patient_id & legacy_value %in% previous_patients) 

print(paste("Number of patients previously delivered:", nrow(previously_delivered)))

## [1] "Number of patients previously delivered: 3"

print("Patients previously delivered:")

## [1] "Patients previously delivered:"

df <- previously_delivered
DT::datatable(  df,
                extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  ))

rejected_patients <- c(rejected_diagnosis, rejected_rna_date, rejected_pluvicto_date, rejected_regimen_between) %>% unique
print(paste("Total Rejected Patients:", rejected_patients %>% length))

## [1] "Total Rejected Patients: 39"

df <- rejected_patients %>% as.data.frame()
DT::datatable(  df,
                extensions = "Buttons",
  options = list(
    pageLength = nrow(df),
    scrollX = FALSE,
    scrollY = FALSE,
    paging = FALSE,
    dom = "Bfrtip",
    buttons = c("csv")
  )
)

Tempus_QC_Prostate

2026-01-05

Requirements