TCGA-LIHC
The Cancer Genome Atlas Liver Hepatocellular Carcinoma Collection
Routine clinical CT, MRI and PET of 97 patients with hepatocellular carcinoma from four sites, matched by patient id to the TCGA clinical, pathology and genomic data in the GDC, for radiogenomic research.
Overview
TCGA-LIHC is the radiology part of The Cancer Genome Atlas liver hepatocellular carcinoma project. The Cancer Imaging Archive hosts clinical CT, MRI and PET of TCGA patients, while their clinical, pathology and genomic data sit in the NCI Genomic Data Commons. Both use the same TCGA patient id, so imaging features can be linked to tumor genotype and patient outcome. The first version appeared in 2014, and version 5 followed in 2020.
Composition
Version 5 holds 97 patients with 237 studies, 1,688 DICOM series and 125,397 images (56.38 GB). Counted from the open TCIA series metadata, 75 patients have CT, 40 have MRI and one has a single whole-body PET series. There are 66 men and 31 women. Age at the earliest archived study ranges from 20 to 85 years, with a median of 64. Many patients have several studies: 56 of the 97 have more than one. The collection page also offers a snapshot of the TCGA clinical data from January 2016, plus the TCGA case forms that explain its fields.
Acquisition
Images come from routine care rather than a research protocol, contributed by Mayo Clinic, the University of North Carolina, Alberta Health Services and Lahey Hospital & Medical Center. Scanners are from Siemens, GE, Philips and Toshiba. Protocols differ between sites and studies: CT includes non-contrast and contrast-enhanced abdominal scans, and liver MRI series include T1-weighted gradient-echo before and after gadolinium, T2-weighted fast spin-echo, in- and opposed-phase imaging and, for a few patients, MR elastography. The baseline studies of each case were acquired before surgery. DICOM dates are shifted by a random per-site offset, which keeps intervals between studies intact.
Annotations
The collection itself contains no image labels. Tumor and organ segmentations made later by other groups are published as separate analysis datasets.
Known limitations
- Scanner types, protocols and series naming vary widely, and no harmonized sequence labels are provided.
- Only 97 of the TCGA liver cancer patients have imaging, so the imaged subset may not represent the full TCGA cohort.
- TCIA and TCGA use different date conventions that are not reconciled.
- Two patients have series without a manufacturer value.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 97 subjects.
Sex
- Male 66 68%
- Female 31 32%
Age
mean 61.4± 13.5 · range 20 to 85Modality
subjects
- CT 75 77%
- MRI 40 41%
- PET 1 1%
Scanner vendor
subjects
- Siemens Healthineers 60 62%
- GE HealthCare 55 57%
- Philips 4 4%
- Canon Medical 3 3%
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
Download without an account
DICOM images via the TCIA Data Retriever. Clinical, genomic and histopathology data of the same patients are in the NCI Genomic Data Commons; a clinical snapshot is an open download from the collection page.
Creative Commons Attribution 3.0 Unported
Older version of CC BY. Use, share and adapt the data for any purpose, including commercial use, as long as you credit the creators.
TCIA lists the images under CC BY 3.0. Users must also follow the TCIA Data Usage Policy, which says users must "Not use the requested datasets, either alone or in concert with any other information, to identify or contact individual participants" and must cite the dataset DOI (https://www.cancerimagingarchive.net/data-usage-policies-and-restrictions/). Publications are asked to acknowledge the TCGA Research Network.
What you can do
- Yes
- Yes
- Yes
- Yes
What you can share
- Yes
- Yes
- Yes
What you must do
- Yes
- Share alike No
- No
- No
- Manuscript review No
- Release code No
- Return results No
- Delete after use No
Limits
- No
- Location limits No
Citation
Erickson BJ, Kirk S, Lee Y, Bathe O, Kearns M, Gerdes C, Rieger-Christ K, Lemmerman J. (2016). The Cancer Genome Atlas Liver Hepatocellular Carcinoma Collection (TCGA-LIHC) (Version 5) [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/K9/TCIA.2016.IMMQW8UQ
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total | 97 | tcia-tcga-lihc Data Access table |
| Studies | total | 237 | tcia-tcga-lihc Data Access table |
| Scans | total DICOM series | 1,688 | tcia-tcga-lihc Data Access table |
| Images | total DICOM images (slices) | 125,397 | tcia-tcga-lihc Data Access table |
| Subjects | modality=CT | 75 | tcia-tcga-lihc-series Modality |
| Subjects | modality=MR | 40 | tcia-tcga-lihc-series Modality |
| Subjects | modality=PT | 1 | tcia-tcga-lihc-series Modality |
| Studies | modality=CT distinct StudyInstanceUID with at least one series of this modality | 164 | tcia-tcga-lihc-series Modality |
| Studies | modality=MR distinct StudyInstanceUID with at least one series of this modality | 73 | tcia-tcga-lihc-series Modality |
| Studies | modality=PT distinct StudyInstanceUID with at least one series of this modality | 1 | tcia-tcga-lihc-series Modality |
| Scans | modality=CT DICOM series | 777 | tcia-tcga-lihc-series Modality |
| Scans | modality=MR DICOM series | 910 | tcia-tcga-lihc-series Modality |
| Scans | modality=PT DICOM series | 1 | tcia-tcga-lihc-series Modality |
| Subjects | vendor=siemens subjects with at least one series from this vendor; 2 subjects have series without a Manufacturer value | 60 | tcia-tcga-lihc-series Manufacturer |
| Subjects | vendor=ge subjects with at least one series from this vendor | 55 | tcia-tcga-lihc-series Manufacturer |
| Subjects | vendor=philips subjects with at least one series from this vendor | 4 | tcia-tcga-lihc-series Manufacturer |
| Subjects | vendor=canon Toshiba; subjects with at least one series from this vendor | 3 | tcia-tcga-lihc-series Manufacturer |
| Subjects | sex=male | 66 | tcia-tcga-lihc-patients PatientSex |
| Subjects | sex=female | 31 | tcia-tcga-lihc-patients PatientSex |
| Subjects | age=20-29 age at the earliest archived study | 4 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=30-39 age at the earliest archived study | 3 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=40-49 age at the earliest archived study | 7 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=50-59 age at the earliest archived study | 20 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=60-69 age at the earliest archived study | 37 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=70-79 age at the earliest archived study | 20 | tcia-tcga-lihc-studies PatientAge |
| Subjects | age=80-89 age at the earliest archived study | 6 | tcia-tcga-lihc-studies PatientAge |
| Mean age | total age at the earliest archived study | 61.4 | tcia-tcga-lihc-studies PatientAge |
| Age SD | total age at the earliest archived study | 13.5 | tcia-tcga-lihc-studies PatientAge |
| Median age | total age at the earliest archived study | 64 | tcia-tcga-lihc-studies PatientAge |
| Minimum age | total age at the earliest archived study | 20 | tcia-tcga-lihc-studies PatientAge |
| Maximum age | total age at the earliest archived study | 85 | tcia-tcga-lihc-studies PatientAge |
Sources
The keys used in the table above.
- tcia-tcga-lihc TCIA TCGA-LIHC collection page (version 5, updated 2020/05/29) website
- tcia-tcga-lihc-series TCIA NBIA API getSeries metadata for TCGA-LIHC (open, no account, CC BY 3.0), counted per patient id from the Modality and Manufacturer fields computed
- tcia-tcga-lihc-patients TCIA NBIA API getPatient metadata for TCGA-LIHC (open, no account, CC BY 3.0), PatientSex field computed
- tcia-tcga-lihc-studies TCIA NBIA API getPatientStudy metadata for TCGA-LIHC (open, no account, CC BY 3.0), PatientAge of each patient's earliest study computed