Clinical context: Endometrial cancer (EC) ranks sixth among female cancers globally, with 417,367 new cases and 97,000 deaths in 2020. Mortality continues to rise, and identifying biomarkers that can stratify high-risk patients is essential for enabling individualized treatment and improving outcomes.
The PLK family: Polo-like kinases (PLKs) are serine/threonine kinases (PLK1-5) involved in centrosome biology and cell cycle regulation. PLK4 is unique among the family, possessing three polo boxes and consisting of 970 amino acids with a molecular mass of 109 kDa. It serves as a pivotal regulator of tumor growth and metastasis.
PLK4 in gynecologic malignancy: Prior work showed PLK4 is overexpressed in epithelial ovarian cancer and associated with shortened survival. Bioinformatic analysis suggested PLK4 may be a key oncogene in EC through cancer stem cell regulation. However, clinical evidence directly measuring PLK4 protein in EC patients was lacking before this study.
Patient cohort: This retrospective study enrolled 142 EC patients who underwent surgical resection between April 2017 and December 2022. Inclusion required pathology-confirmed EC diagnosis, surgical resection, available tumor tissue for IHC, complete clinicopathological data, and at least one follow-up data point.
IHC assay protocol: Tumor tissue samples were stained using an anti-PLK4 rabbit polyclonal antibody (1:100 dilution, overnight incubation at 4 degrees C), followed by a secondary antibody. Color staining was terminated when a positive result appeared with clear background visualization under microscope.
Scoring system: The IHC score ranged from 0 to 12, calculated by multiplying staining intensity (0-3) by staining density (1-4). Patients were classified into four groups: score = 0, score 1-3, score 4-6, and score 7-12. Multiple cutoff values (greater than 0, greater than 3, and greater than 6) were used for survival comparisons.
Statistical methods: Associations were assessed using chi-square, Fisher's exact, or Kruskal-Wallis tests. Survival was evaluated using Kaplan-Meier estimators with log-rank comparisons. Univariable and multivariable Cox regression analyses identified independent risk factors for DFS and OS.
Patient demographics: The mean patient age was 60.1 +/- 9.0 years; 79.6% were postmenopausal. The majority had endometrioid carcinoma G1/G2 (67.6%), while 10.6% had endometrioid G3, 13.4% serous, and 8.5% clear cell subtypes. FIGO stage I was most common (62.0%).
PLK4 expression distribution: The mean PLK4 IHC score was 3.8 +/- 3.3. Among patients, 26.1% had a score of 0, 24.6% scored 1-3, 27.5% scored 4-6, and 21.8% scored 7-12. Overall, 73.9% of patients had some measurable PLK4 expression (IHC score greater than 0).
Association with tumor features: Higher PLK4 expression was significantly associated with lymphovascular invasion (P = 0.008) and higher FIGO stage (P = 0.005). No statistically significant differences were found for age, menopausal status, diabetes, hypertension, histological subtype, myometrial invasion, or tumor markers (CA125, CA19-9, CEA).
Disease-free survival (DFS): No DFS difference was found between patients with PLK4 IHC score greater than 0 vs. 0 (P = 0.154). However, DFS was significantly shorter in patients with scores greater than 3 vs. 3 or less (P = 0.009) and even more markedly worse for scores greater than 6 vs. 6 or less (P less than 0.001).
Overall survival (OS): Similarly, no difference was found at the greater than 0 threshold (P = 0.322). OS was significantly shorter with PLK4 scores greater than 3 (P = 0.011) and greater than 6 (P = 0.006), demonstrating a dose-response relationship between PLK4 expression level and mortality risk.
Independent predictors of DFS (multivariate): PLK4 IHC score greater than 6 remained independently associated with shortened DFS (HR: 3.156, P = 0.008) after adjustment for histological subtype (endometrioid G3 HR: 7.617; serous HR: 5.393; clear cell HR: 6.339) and FIGO stage (HR: 1.983).
Independent predictors of OS (multivariate): PLK4 IHC score greater than 3 independently predicted shortened OS (HR: 3.918, P = 0.026), alongside hypertension (HR: 3.108), endometrioid G3 (HR: 19.661), clear cell subtype (HR: 30.569), and high FIGO stage (HR: 2.413).
Invasion and metastasis mechanisms: PLK4 may promote EC invasion through Arp2/3-mediated actin cytoskeletal rearrangement, enabling tumor cell movement. Additionally, PLK4 may induce epithelial-mesenchymal transition (EMT) via the Wnt/beta-catenin signaling pathway, enhancing metastatic potential and explaining the association with lymphovascular invasion.
Cross-cancer evidence: The negative association between PLK4 and survival aligns with findings in other cancers: high PLK4 transcripts predict poor relapse-free survival in breast cancer; high PLK4 expression associates with unfavorable survival in glioma; and high PLK4 is an independent predictor of poor OS in colorectal cancer.
Limitations: The study was retrospective and single-center with a relatively small sample size (142 patients), limiting statistical power and introducing selection bias. Only patients who underwent surgical resection were included, meaning findings do not apply to non-surgical cases. The optimal PLK4 cutoff value and its relationship to Lynch syndrome screening require further investigation.
Prognostic tool potential: PLK4 IHC scoring offers a practical, tissue-based method for identifying EC patients at elevated risk of recurrence and death following surgical resection. The use of multiple cutoffs (greater than 3 for OS, greater than 6 for DFS) provides clinical flexibility depending on the endpoint being predicted.
Integration with AI and radiomics: The authors note that radiomic analysis of radiological images represents a major frontier in EC investigation. Combining PLK4 protein expression with AI-driven radiomic approaches may enhance preoperative risk stratification, diagnosis, treatment planning, and prognosis assessment in EC patients.
Molecular subtyping synergy: EC molecular subtyping (POLEmut, MMRd, copy-number low, copy-number high) reveals distinct prognostic groups. Integrating PLK4 expression within this molecular classification framework is a promising direction for future investigation to further personalize EC management.