A Liver Mass in a Young Child: Hepatoblastoma and AFP

Medically reviewed by Dr. Michael Kachur · Frankfurt, Germany·

Bath time is often how these start. A parent's hand moves across a toddler's belly and finds something firm on the right side. The child has been eating a little less. Nothing else looks wrong. Then a blood test comes back with alpha-fetoprotein in the thousands, and nobody has yet explained whether that number is terrifying or simply what a very young liver does. Reading it next to the child's age, the scan and the dates is the kind of case Healz was built for.

Category
Health Insights
Date
Reading time
11 min
Share

This post is about hepatoblastoma, the commonest liver cancer of early childhood, and the test that causes most of the confusion in the first week. Why an AFP result means little without an age beside it. Most abdominal lumps in small children are not this.

A Liver Mass in a Young Child: Hepatoblastoma and AFP

Who gets a liver tumour this young

Hepatoblastoma is the most common type of childhood liver cancer and usually affects children younger than 3 years of age, per the NCI. Its PDQ treatment summary adds that approximately 90% of malignant liver tumours in children aged 4 years and younger are hepatoblastomas. In absolute terms it stays rare.

Most cases have no explanation behind them. A handful of conditions raise the risk enough that children get watched for it: premature birth with a very low birth weight, Beckwith-Wiedemann syndrome, hemihyperplasia, and familial adenomatous polyposis in the family. The scale varies enormously. The PDQ summary puts Beckwith-Wiedemann at a 1,000-fold to 10,000-fold increase and APC carriers at 800-fold, and cites a Japanese series in which infants under 1,000 g at birth carried 15 times the risk of normal-birth-weight children.

Ask Healz.

One chat instead of ten apps. 1M+ rare cases checked. The root cause, not the label. Private.

For the genetic syndromes that carry a 1% or greater risk of hepatoblastoma, Beckwith-Wiedemann syndrome, hemihyperplasia, Simpson-Golabi-Behmel syndrome and trisomy 18, the summary relays an American Association for Cancer Research suggestion: abdominal ultrasound and an AFP blood test every 3 months, from birth or diagnosis through the fourth birthday. PDQ says this will identify 90% to 95% of the hepatoblastomas that develop in these children.

What a parent notices first

A liver tumour in a small child is quiet until it is large. NCI's hepatoblastoma information page says children may not have symptoms until the tumour has grown bigger, and tells parents to check with their child's doctor if they find a lump in the abdomen, swelling or pain there, weight loss for no known reason, loss of appetite, or nausea and vomiting.

One rare presentation looks unrelated to the liver. Some tumours make beta-hCG, and the PDQ summary notes that raised beta-hCG may result in isosexual precocity in boys. Early puberty in a toddler is a hormone finding that points back at an organ nobody was examining.

None of the common complaints is specific. A firm belly in a child who is off their food reads as constipation, as a virus, as a feeding phase.

Why the AFP number means little without an age beside it

Alpha-fetoprotein is a protein the fetal liver makes in quantity, so a newborn's blood is full of it. The PDQ summary states it directly: markedly elevated AFP not caused by the tumour is normal in neonates and steadily falls after birth, the half-life of AFP is 5 to 7 days, and by age 1 year it should be in the reference range, less than 10 ng/mL.

A result in the thousands at two weeks old can be ordinary. At two years old it is not. The value has to be read against an age-specific reference point, not the adult range printed on the form, and in the early months that reference is wide. A 2025 review in the World Journal of Hepatology makes the point from the imaging side: in a suspected infantile hepatic haemangioma, AFP has limited diagnostic value, because AFP tends to be physiologically elevated during the first six months of life. A number read without the age beside it is half a result.

The second thing about AFP inverts the usual instinct about test results: a low value is not the good news. The PDQ summary lists a low AFP level among the statistically significant adverse prognostic factors, and gives it as two bands: under 100 ng/mL, or 100 to 1,000 ng/mL to account for infants with elevated AFP levels. It says absence of an elevated AFP at diagnosis occurs in a small percentage of children and appears to be associated with a very poor prognosis and with the small cell undifferentiated variant. In the Children's Oncology Group risk table it reproduces, any PRETEXT group with AFP under 100 ng/mL falls in the high-risk category.

AFP has a third job once treatment starts, read as a curve rather than a point. The PDQ summary notes that the greatest chemotherapy response, measured as decreases in tumour size and AFP level, occurs after the first two cycles.

What the imaging is for, and what PRETEXT describes

Ultrasound comes first. The 2025 World Journal of Hepatology review states that ultrasonography is typically the first imaging modality for the initial assessment of hepatic lesions, and in a small child it needs no anaesthetic. Our guide to how to read an ultrasound report covers the vocabulary those reports use.

The PDQ summary describes high-quality cross-sectional imaging with three-phase CT or contrast-enhanced MRI. The chest is imaged too, because the lung is the main destination: the summary lists distant metastases as usually lung, and defines the M1 annotation as one noncalcified pulmonary nodule of 5 mm or more in diameter.

PRETEXT is what turns that imaging into a surgical plan. It stands for pretreatment extent of disease. Per the PDQ summary it divides the liver into four sections based on Couinaud's eight-segment anatomy, and the group, I through IV, is set by how many adjoining sections are free of tumour, with group IV meaning all four are involved. The same summary notes that interobserver variability is moderate. It is a planning description, not a verdict.

The other things a liver mass in a toddler can be

Two benign lesions sit squarely in this differential. Infantile hepatic haemangioma is a benign vascular tumour, and the PDQ summary says biopsy is not indicated for infantile haemangioma of the liver when the MRI findings are classic. Mesenchymal hamartoma is a rare benign lesion, usually a large multicystic mass presenting in the first 2 years of life (NCI), and it can raise serum AFP and mimic hepatoblastoma (Boman et al., European Journal of Pediatric Surgery, 2004).

Age shifts the list. The 2025 World Journal of Hepatology review counts a different population and lands in the same place: hepatoblastoma is 91% of primary hepatic malignancies in children presenting before the age of five, and hepatocellular carcinoma is more common than hepatoblastoma in children over six years of age. In adults the rules change again, which our piece on an incidental liver lesion on imaging covers.

Tissue settles most cases. The PDQ summary says a biopsy is always indicated to confirm a paediatric liver tumour, with four named exceptions: classic infantile hepatic haemangioma on MRI, unless the diagnosis is in doubt; focal nodular hyperplasia with classic features on hepatocyte-specific contrast MRI; Children's Oncology Group guidance supporting resection without biopsy for PRETEXT I and selected PRETEXT II tumours; and infantile hepatic choriocarcinoma, where markedly elevated beta-hCG often means chemotherapy without biopsy. The wording of that pathology report names the subtype, and the subtype feeds the risk group.

Why the operation is planned before the first chemotherapy

Cure here is defined by the surgery. The PDQ summary says cure of hepatoblastoma requires gross tumour resection, and that incomplete resection must be avoided, because a rescue transplant after an incompletely resected tumour does worse than a transplant done as the primary operation. A microscopically positive margin is a different matter: the same summary says outcomes after complete resection and after positive microscopic margins are similar once chemotherapy is given.

Chemotherapy before surgery exists to turn a no into a yes. The summary describes preoperative chemotherapy being used for unresectable tumours to enable complete resection, says most stage III and IV hepatoblastomas are rendered resectable after two cycles, and warns that prolonged chemotherapy can lead to unnecessary delays and, rarely, tumour progression.

When too little healthy liver would be left behind, transplantation is the operation, not a partial hepatectomy. That choice is made with the imaging and the PRETEXT group in hand, at a centre that does both operations. Lung metastases do not automatically remove surgery from the plan either: the summary says resection of pulmonary metastases should be considered if the number of metastases is limited, and that any remaining pulmonary metastases should be removed surgically if possible once the primary tumour is resected.

How Healz holds the ultrasound, the AFP and the age in one place

The ultrasound report, the AFP value with the child's age beside it and the imaging that arrives weeks later are three pieces of one case, arriving in three envelopes from three departments. Healz keeps them in one chat, not ten apps and four portals.

Healz is equipped with Frontier AI, so the case is worked whole: the ultrasound wording read against the cross-sectional study that followed it, the AFP value read against the age it was drawn at, the PRETEXT description read for what it says about the operation. A cancer ai reads the pathology report beside the imaging, not as separate documents.

Healz is equipped with root-cause technology, so it does not stop at the first label. It asks what a rushed appointment skips and cross-checks the case against more than a million rare cases, so a benign possibility such as mesenchymal hamartoma stays on the list until the tissue takes it off.

Healz's memory pulls the medical facts out of every upload, dates them and keeps abnormal values up front, so each AFP result sits beside the one before it with its date and the child's age attached, and the decline across cycles reads as a line instead of four letters. That also makes a second opinion quick to ask for, because the case does not have to be reassembled first. You can bring a board-certified doctor into the same chat for a second opinion.

Frequently asked questions

What does a high AFP in a baby or toddler mean?

On its own, not much, until you know the age. AFP is normally very high in newborns and falls with a half-life of 5 to 7 days, reaching the reference range of under 10 ng/mL by age 1, per the NCI's PDQ summary.

What are the first signs of liver cancer in a child?

Usually a lump or swelling in the abdomen. NCI also tells parents to check with their child's doctor about abdominal pain, weight loss for no known reason, loss of appetite, and nausea and vomiting, and says children may not have symptoms until the tumour has grown bigger. Rarely, a tumour making beta-hCG causes signs of early puberty in a boy.

Is a liver mass in a child always cancer?

No. Infantile hepatic haemangioma, a benign vascular tumour, and mesenchymal hamartoma, a rare benign lesion, both present as liver masses in this age group, per a 2025 review in the World Journal of Hepatology. In most cases a biopsy is what settles it.

What does PRETEXT mean on my child's scan report?

It describes how much of the liver the tumour occupies before any treatment. Per the NCI's PDQ summary the liver is divided into four sections and the group, I to IV, reflects how many adjoining sections are free of disease, with IV meaning all four are involved. The same summary notes that preoperative PRETEXT matched postoperative pathology only 51% of the time in SIOPEL-1 data, so it is a plan rather than a final answer.

Bath time is often where these start, and the number that comes back a week later is only readable with the child's age beside it.

Written by Healz Team · Filed under Health Insights

Your health deserves more than guesswork

Get a real investigation, not a symptom list. Root causes found, history remembered, dots connected.

Not a medical provider. Healz.ai is an educational service. Always discuss every decision with your doctor. AI can make mistakes.

© 2026 HealzAI Inc1111B S Governors Ave #92123, Dover, DE 19904, USAMade with for better health