LifePort hypothermic machine perfusion data analysis

Showcase with code

Author

Simon Schwab

Published

March 26, 2025

Abstract

This report demonstrates how to import LifePort hypothermic machine perfusion data for research and data analysis in R.

Objectives

This showcase imports 50 LifePort raw data sets and performs descriptive statistics. This report was created using R, RStudio, Quarto, and the Swisstransplant package (swt). I hope this work will advance research and analysis of perfusion data.

Data import

Code

library(fs)
library(swt)
library(ggplot2)
library(gridExtra)

col = swt::swt_colors()

If these packages are not already installed, you can install them using:

install.packages("foo")

The Swisstransplant package (swt) can be installed from GitHub using the remotes package:

remotes::install_github("Swisstransplant/swt")

Below, I define the path where the data is located in the variable PATH_DATA.

Code

PATH_DATA = file.path(path_home(), "OneDrive - Swisstransplant", 
                      "Data", "Lifeport Rawdata", "2024")
files = list.files(PATH_DATA, pattern = ".txt|.TXT")
files = files[1:50] # only use 50 cases

Raw data files

I read each data file, one by one, from the first to the total number of files length(files), using a for loop. The LifePort data are processed in three steps:

Read the data file with lifeport_read()
Process the data (e.g., signal filtering) with lifeport_process()
Calculate the statistical indicators with lifeport_sumstats()

For more details, refer to the publication at the end.

Reading hundreds of files will inevitably lead to errors, especially if a LifePort data file is corrupt or empty. To handle this, the problematic file must be identified and removed from the data folder. Debugging can be done by adding a print(i) statement in the loop to determine which file (files[i]) caused the issue.

Code

data.device = list()
data.organ = list()
data.timeseries = list()
data.sumstats = list()

for (i in 1:length(files)) {
  
  # print(i) # for debugging
  
  tmp = lifeport_read(file = file.path(PATH_DATA, files[i]), format = "guess")
  tmp = lifeport_process(lpdat = tmp, window_size = 15)
  tmp = lifeport_sumstats(lpdat = tmp, ice_threshold = 2.5)
  
  # add filename as well
  tmp$data.device$Filename = files[i]
  
  data.device[[i]]     = tmp$data.device
  data.organ[[i]]      = tmp$data.organ
  data.timeseries[[i]] = tmp$data
  data.sumstats[[i]]   = tmp$data.sumstats
}

data.device = data.table::rbindlist(data.device)
data.organ = data.table::rbindlist(data.organ)
data.sumstats = data.table::rbindlist(data.sumstats)

The code above is designed to collect all the data and store it temporarily in different lists. Once all the data has been read and processed, the lists can be transformed into data frames.

Data access in R

Device data

The device data contains the serial number, the name of the device, the start, stop, and run time.

Code

n = 5

data.device[1:n, -c("Filename", "Type", "SubType", "DataState", "HasGaps")]

SerialNumber	UnitID	FirmwareVersion	FileID	StartTime	Runtime	StopTime
2326506		768	7	2024-01-02 00:59:26	06:11:10	2024-01-02 07:10:26
1415442	DIEGEM_LOANER	515	10	2024-01-03 14:33:46	03:52:00	2024-01-03 18:25:36
1423194	GDfL	515	97	2024-01-03 14:23:42	02:29:40	2024-01-03 16:53:12
1415442	DIEGEM_LOANER	515	11	2024-01-08 14:23:37	03:16:10	2024-01-08 17:39:37
1423110	SG1	515	48	2024-01-21 14:01:46	03:58:20	2024-01-21 17:59:56

Note

Ideally, the UnitID should be a short name without special characters. I believe the LifePort machines in Geneva have a special character in their UnitID. In such cases, the UnitID may not display correctly, and if it contains a name that could cause serious issues, it will be removed.

Organ data

Organ data includes the kidney side, blood type, and cross-clamp time. This information is entered into the device by healthcare professionals and may be missing if not entered.

Code

data.organ[1:n, c("KidneySide", "BloodType", "CrossClampTime.Date")]

KidneySide	BloodType	CrossClampTime.Date
Right	0	2024-01-01 23:51:00
Left	0	2024-03-01 12:32:00
Right	0	2024-01-03 12:32:00
Right	0	NA
Left	A	NA

Summary statistics data

The summary statistics data are particularly interesting as they include various values such as mean ice temperature, mean discharge, and more. See the publication at the end for more details.

Code

n = 10

data.sumstats[1:n, ]

perfusion.dur	perfusion.dur.str	systolicPressure.md	diastolicPressure.mean	flowRate.mean	organResistance.mean	organResistance.sd	organResistance.x1	organResistance.y1	organResistance.x2	organResistance.y2	organResistance.delta	organResistance.slope	iceContainerTemperature.mean	iceContainerTemperature.sd	iceContainerTemperature.minAbove	iceContainerTemperature.minAbove.str	infuseTemperature.mean	infuseTemperature.sd	infuseTemperature.start	infuseTemperature.minAbove	infuseTemperature.minAbove.str
368.83333	06:08:50	28.33333	19.67842	151.66601	0.1499605	0.0106282	8.00000	0.3328040	96.42424	0.2153304	-0.1174736	-0.0080094	2.354513	0.0728145	0.00000	00:00:00	6.778806	1.2528984	10.236667	7.1666667	00:07:10
35.66667	00:35:40	24.66667	21.93645	123.50081	0.1453492	0.0491878	NA	NA	NA	NA	NA	NA	1.873707	0.1100868	0.00000	00:00:00	6.424103	0.1477794	6.413333	0.0000000	00:00:00
142.16667	02:22:10	29.46667	21.78759	49.16569	0.5064565	0.0560590	8.00000	0.7881980	35.18270	0.6431914	-0.1450066	-0.0322236	1.313920	0.0794070	0.00000	00:00:00	6.253837	0.4731581	6.423333	0.0000000	00:00:00
193.83333	03:13:50	29.46667	20.69923	143.20896	0.1635259	0.0125971	8.00000	0.2880239	66.88864	0.1913324	-0.0966915	-0.0100026	2.342481	0.1482780	17.33333	00:17:20	6.552988	0.1423970	6.783333	0.0000000	00:00:00
189.33333	03:09:20	29.46667	20.68374	126.49927	0.1829073	0.0188346	8.00000	0.3763685	37.31903	0.2488760	-0.1274925	-0.0263778	1.908881	0.1506162	0.00000	00:00:00	5.136852	0.6166316	5.486667	0.0000000	00:00:00
475.33333	07:55:20	29.53333	16.02805	78.10237	0.2776885	0.0218842	60.00002	0.5077676	105.81427	0.3577929	-0.1499747	-0.0199968	1.995274	0.3623524	31.83333	00:31:50	4.955487	0.8323820	7.333333	0.1666667	00:00:10
26.66667	00:26:40	13.40000	10.02441	NA	NA	NA	NA	NA	NA	NA	NA	NA	1.657855	0.1108550	0.00000	00:00:00	8.026923	0.3899790	8.250000	0.0000000	00:00:00
336.33333	05:36:20	29.60000	16.03300	74.53828	0.2960289	0.0178520	8.00000	0.9659789	26.47657	0.5348632	-0.4311158	-0.1437060	1.656053	0.1851524	0.00000	00:00:00	6.060091	0.4616435	8.016667	0.0000000	00:00:00
103.16667	01:43:10	22.60000	19.86160	165.24365	0.1234410	0.0288686	32.66376	0.1663610	237.97514	0.1105713	-0.0557897	-0.0015989	2.139178	0.1180563	0.00000	00:00:00	7.204429	0.7044574	8.476667	0.0000000	00:00:00
262.00000	04:22:00	29.46667	19.01612	49.97994	0.4795192	0.0423279	8.00000	1.1968849	55.91632	0.7775040	-0.4193810	-0.0535380	1.519420	0.0863631	0.00000	00:00:00	6.450427	0.8935817	9.143333	0.0000000	00:00:00

Time series data

The complete time series data is also available. Below is an example for the first kidney, the first 10 samples, along with the filtered (smoothed) time series. Due to filtering, the first few and the last few values are missing.

Code

data.timeseries[[1]][1:10, c("FlowRate", "FlowRate.flt")]

FlowRate	FlowRate.flt
14	NA
10	NA
30	NA
52	NA
63	NA
67	NA
71	NA
75	61.86667
75	66.53333
77	71.40000

Descriptive statistics

Here are some examples of how to visualize the data and perform some statistics.

Plotting the data

Code

k = 4 # which time series to show
t = 1:1000 # show first 100 seconds

p1 = ggplot(data.sumstats, aes(x = iceContainerTemperature.mean)) + 
  geom_histogram(bins = 10, alpha = 1, fill = col$blue.alt) +
  xlab("Mean ice temperatue") +
  labs(tag = "A") + swt_style()

p2 = ggplot(data.sumstats, aes(x = flowRate.mean, y = organResistance.mean)) + 
  geom_point(size = 2, alpha = 0.5, col = col$strongred.akzent) +
  scale_color_manual(values = col$blue.swt) +
  xlab("Mean flow rate") + ylab("Mean organ resistance") +
  labs(tag = "B") + swt_style()

p3 = ggplot(data.timeseries[[k]][t,], aes(x = time.clock, y = FlowRate)) + 
  geom_line(size = 1, alpha = 1, col = col$turkis.tpx) +
  ylim(c(0, 200)) +
  xlab("Clock Time") + ylab("Flow rate (raw)") + 
  labs(tag = "C") + swt_style()

p4 = ggplot(data.timeseries[[k]][t,], aes(x = time.zero, y = FlowRate.flt)) + 
  geom_line(size = 1, alpha = 1, col = col$purple.alt) +
  ylim(c(0, 200)) +
  xlab("Time (starting from 0)") + ylab("Flow rate (filtered)") +
  labs(tag = "D") + swt_style()

grid.arrange(p1, p2, p3, p4, nrow = 2, ncol = 2)

Mean (IQR) of the perfusion duration

Flow rate was converted to the unit hours.

Code

data.frame(median = median_iqr(data.sumstats$perfusion.dur/60))

median
5.8 (from 3.4 to 8.4)

Mean (IQR) of flow rate

To clarify, each kidney has a mean flow rate, and I calculate the median across all the mean flow rates.

Code

data.frame(median = median_iqr(data.sumstats$flowRate.mean))

median
104.5 (from 68.4 to 133.0)

References

Please be sure to cite the following work in your research when utilizing EXAM or the Swisstranplant swt package for machine perfusion data analysis.

Schwab S, Steck H, Binet I, Elmer A, Ender W, Franscini N, Haidar F, Kuhn C, Sidler D, Storni F, Krügel N, Immer F. EXAM: Ex-vivo allograft monitoring dashboard for the analysis of hypothermic machine perfusion data in deceased-donor kidney transplantation. PLOS Digit Health. 2024;3(12):e0000691. doi:10.1371/journal.pdig.0000691

Schwab S. EXAM: Ex Vivo Allograft Monitoring Dashboard. Github; 2025. Accessed February 13, 2025. https://github.com/Swisstransplant/EXAM

Schwab S. swt: Swisstransplant R Package. Github; 2024. Accessed February 13, 2025. https://github.com/Swisstransplant/swt

Contact

Please feel free to contact me at simon.schwab@swisstransplant.org with any inquiries regarding the analysis of perfusion data; I’m happy to assist.

Computing information

Code

sessionInfo()

R version 4.4.3 (2025-02-28 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 10 x64 (build 19045)

Matrix products: default


locale:
[1] LC_COLLATE=English_Switzerland.utf8  LC_CTYPE=English_Switzerland.utf8   
[3] LC_MONETARY=English_Switzerland.utf8 LC_NUMERIC=C                        
[5] LC_TIME=English_Switzerland.utf8    

time zone: Europe/Zurich
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] gridExtra_2.3 ggplot2_3.5.1 swt_0.3       fs_1.6.5     

loaded via a namespace (and not attached):
 [1] gtable_0.3.6      jsonlite_1.9.1    dplyr_1.1.4       compiler_4.4.3   
 [5] tidyselect_1.2.1  splines_4.4.3     scales_1.3.0      yaml_2.3.10      
 [9] fastmap_1.2.0     lattice_0.22-6    R6_2.6.1          labeling_0.4.3   
[13] generics_0.1.3    knitr_1.50        htmlwidgets_1.6.4 MASS_7.3-65      
[17] tibble_3.2.1      munsell_0.5.1     lubridate_1.9.4   pillar_1.10.1    
[21] rlang_1.1.5       xfun_0.51         segmented_2.1-4   timechange_0.3.0 
[25] cli_3.6.4         withr_3.0.2       magrittr_2.0.3    digest_0.6.37    
[29] grid_4.4.3        hms_1.1.3         lifecycle_1.0.4   nlme_3.1-167     
[33] vctrs_0.6.5       evaluate_1.0.3    glue_1.8.0        data.table_1.17.0
[37] farver_2.1.2      colorspace_2.1-1  rmarkdown_2.29    tools_4.4.3      
[41] pkgconfig_2.0.3   htmltools_0.5.8.1