Establishment and Characterization of a Testicular Yolk Sac Tumor PDX Model

First PDX model for pediatric testicular yolk sac tumor, validated against a 109-patient retrospective series; 91.7% of patients presented at stage I

Journal: Clinical & Translational Oncology | Published: 2026-01-12 | Type: Journal Article | PMID: 41525035 Authors: Cai Jiabin et al. (Children's Hospital of Chongqing Medical University; Nanchang University; Jiangxi Provincial Children's Hospital) Funding/COI: Medical Science and Technology Project of Zhejiang Province. Authors declare no conflicts of interest; funders had no role in study design, data collection, or publication decisions.

Summary

This paper does two things at once: it establishes a patient-derived xenograft (PDX) model of testicular yolk sac tumor (TYST) in immunodeficient mice, and it reports a 16-year retrospective clinical series of 109 pediatric TYST patients from a single Chinese center. The lab model successfully engrafted, retained the original tumor's histological features and biomarker profile, and responded to standard JEB chemotherapy (carboplatin + etoposide + bleomycin) both in live mice and in an ex vivo drug sensitivity assay. Clinically, the series confirms what smaller studies have suggested: most pediatric TYST is caught early and carries an excellent prognosis.

Claims

• PDX model engrafted successfully in both NPG and BALB/c nude mouse strains and remained stable through serial passaging
• Xenograft tumors preserved Schiller-Duval bodies (the histological hallmark of yolk sac tumor) on H&E staining, and maintained AFP and SALL4 immunoreactivity matching the source patient's tumor
• PCR confirmed human origin of engrafted tissue and ruled out murine lymphoma contamination
• JEB chemotherapy induced significant tumor regression in PDX mice without clinically meaningful body weight loss; ex vivo HDST assay independently showed marked sensitivity to JEB
• Clinical series: n=109 patients, median age at diagnosis 18 months (range 2–69 months); 74.3% diagnosed before age 2
• 91.7% of patients staged COG stage I; 5.5% stage II; 2.8% stage IV
• 99.1% had elevated preoperative AFP; 96.3% had AFP >100 ng/ml
• AFP failed to normalize within 2 months postoperatively in 14.8% of patients; this was independently correlated with relapse and worse long-term survival
• Tumor diameter correlated with COG stage, rate of AFP normalization, and survival (R=0.452, R=−0.224, R=−0.357; all P<0.05)
• Most relapses occurred within the first postoperative year

Study Quality

The PDX model component is a proof-of-concept animal study derived from a single patient's tumor — which is standard for initial PDX work but means nothing yet about generalizability across the biological heterogeneity of TYST. The characterization is methodologically appropriate: serial passaging with histological and molecular validation, species-of-origin confirmation, and drug sensitivity testing with both in vivo and ex vivo assays provides reasonable confidence that the model is what the authors say it is. The hydrogel-embedded histoculture drug sensitivity test (HDST) as an ex vivo correlate is a useful methodological addition.

The clinical series is retrospective and single-center, drawing on 16 years of cases from one institution in Chongqing. Despite its limitations, it represents the largest reported Chinese series and one of the largest pediatric TYST series globally — which matters for a disease with fewer than a few hundred published cases worldwide. The statistical analysis is appropriate for the design (Cox regression, chi-square, Pearson correlation). However, the paper blends two distinct study types — a lab model and a clinical outcomes study — without a clear conceptual bridge explaining how the PDX model was validated against the clinical population rather than just developed in parallel.

Red Flags

• PDX model built from a single patient's tumor; engraftment success rates across additional patient samples are not reported
• The full-text clinical sections and the abstract describe what appear to be nearly separate studies; the integration is superficial
• Chemotherapy outcome data in the clinical cohort partially predates 2009, when "no clear guidelines were given" — heterogeneity in treatment regimens across the 16-year period complicates any survival analysis
• Stage IV cohort is only 3 patients — too small for any meaningful prognostic conclusions
• No multi-variate survival model explicitly accounts for treatment era
• Funded by Zhejiang Province but the clinical work is in Chongqing — no explanation of why funding province differs from study institution

Strengths

• Largest pediatric TYST series from China (109 patients); high completeness (88.6% of eligible cases included)
• PDX validation used multiple orthogonal methods: histology, immunohistochemistry, PCR species ID
• Both in vivo and ex vivo drug sensitivity testing strengthens the JEB sensitivity claim
• AFP normalization timeline as a prognostic marker is a clinically actionable finding (even if the paper wisely frames it descriptively)
• Long follow-up period (up to 2021 with regular surveillance protocol)
• No conflicts of interest; government funding only

Verdict

The PDX model is the novel scientific contribution here — it's a technically sound but early-stage tool: one patient's tumor, in mice. That's how all PDX models start, and the characterization is done correctly. The clinical series is the stronger piece on its own terms: 109 patients over 16 years with systematic follow-up is genuinely useful epidemiological data for a disease this rare. The combination of the two in one paper serves the authors' career metrics better than it serves the reader. Worth reading if you work in pediatric oncology or are building a translational research program in germ cell tumors; the AFP normalization finding deserves attention in a prospective study.