Skip to main content
. 2018 Mar 12;18:275. doi: 10.1186/s12885-018-4178-z

Table 3.

Tumor phenotypes and clinical parameters associated with t-SNE clustering

Feature Cluster Comparison P-value
Breast cancer
 Her2/Neu status Cluster 1–88.2% Other tumors – 7.7% <  0.0001
Hepatocellular carcinoma
 Hepatitis B Cluster 2–38.4% Other tumors – 18.6% <  0.0001
Endometrial carcinoma
 Type II (serous) type Cluster 1–32.4% Other tumors – 5.2% <  0.0001
 Locoregional disease or recurrence Cluster 1–8.6% Other tumors – 2.2% 0.004
Lung cancer
 Adenocarcinoma Cluster 2–80.0% Other tumors – 37.8% <  0.0001
 Squamous cell carcinoma Cluster 4–73.6% Other tumors – 34.2% <  0.0001
Kidney clear cell carcinoma
 Gender Cluster 1–92.9% female Other tumors – 29.0% female
 Longest dimension Cluster 2–1.5 cm Other tumors – 1.8 cm <  0.0001
 Histologic grade 3 or 4 Cluster 3–69.0% Other tumors – 41.5% <  0.0001
 Pathologic stage III or IV Cluster 3–58.5% Other tumors – 25.5% <  0.0001
Thyroid carcinoma
 Gender Cluster 3–35.4% male Other tumors – 11.3% male <  0.0001
 Tall cell subtype Cluster 1–17.2% Other tumors – 5.03% 0.0001
 Follicular subtype Cluster 3–27.7% Other tumors – 10.5% <  0.0001
Acute myeloid leukemia
 Gender Cluster 1–100% female
Cluster 3–81.5% male
Cluster 2–53.6% female <  0.0001
 Complex (> 3 distinct) cytogenetic abnormalities Cluster 3–21.0% Other tumors – 5.5% 0.0025
Testicular cancer
 Seminoma subtype Cluster 1–100% Other tumors – 14.5% <  0.0001
 Embryonal carcinoma subtype Cluster 2–94.6% Other tumors – 20.6% <  0.0001
 Teratoma subtype Cluster 3–84.0% Other tumors – 14.7% <  0.0001
Head and neck cancer
 HPV infection Cluster 1–46.1% Other tumors – 9.8% 0.0001
Skin cutaneous melanoma
 Breslow depth Cluster 1–4.8 cm Other tumors – 7.6 cm 0.0083
Brain cancer
 Glioblastoma Cluster 3–100% Other tumors – 0% <  0.0001
 Astrocytoma low-grade glioma Cluster 2–53.0%
Cluster 4–62.7%
Other tumors – 2.5% <  0.0001
 Non-astrocytoma low-grade glioma Cluster 1–86.7%
Cluster 5–97.0%
Other tumors – 29.1% <  0.0001

Tumor phenotypes and clinical markers were compared between tumor clusters using chi-squared tests, with significance defined as α < 0.01. “Other tumors” are comprised of all tumors from the same cancer cohort not falling into the given cluster. Data were obtained using the UCSC Xenabrowser, under the data heading “Phenotypes.” Cancer cohorts with no tumor clusters significantly associating with any clinical parameter were excluded from this table