BOB Ped Surg 2023 - Andrew Fleming, AAP - Presentation

The Clinical Reality

Hepatoblastoma is the most common pediatric liver cancer 0:00. Risk stratification follows the AHEAD 1531 trial framework: PRETEXT group — determined by the number of contiguous hepatic sections uninvolved by tumor — combined with metastatic disease status, alpha-fetoprotein at diagnosis, PRETEXT modifiers, patient age, and resectability at diagnosis 0:30 1:00. The fundamental principle is unambiguous: complete surgical resection has been determined to be necessary for patient survival 1:30. Cisplatin-based chemotherapies have improved rates of resectability and subsequently survival 1:45, but chemotherapy alone does not cure this disease.

The question this series addresses is what to do when the primary tumor is resectable but pulmonary metastases remain. The literature on pulmonary metastasectomy for hepatoblastoma has reported varied survival benefits 2:00. The decision point is whether to accept radiographic disease after chemotherapy and primary resection, or to pursue complete clearance through repeated thoracic operations.

The Approach

The St. Jude group analyzed 50 patients treated between 2005 and 2021, of whom 24 were classified as high-risk 4:00. All received chemotherapy 4:15. The institutional philosophy, as Fleming described it: the group takes a fairly aggressive approach to pulmonary metastasectomy for these patients, including bilateral reoperative pulmonary resections if necessary 4:30.

No evidence of disease was defined as a normal AFP and the absence of detectable radiographic disease on interval surveillance imaging 3:00. Forty-one patients were successfully rendered no evidence of disease status 4:15. Fourteen patients underwent a total of 31 pulmonary procedures, with a median of four nodules resected per patient 4:30. Pathology confirmed viable hepatoblastoma in a median of three nodules per patient, demonstrating high concordance between radiographic evidence of disease and pathological confirmation 5:00.

The reasoning was empirical rather than theoretical. If imaging showed disease and resection was technically feasible, the team operated. If disease recurred after initial clearance, they operated again. Three high-risk patients who relapsed after achieving no evidence of disease were successfully salvaged using a combination of chemotherapy and further pulmonary resections 7:30.

The Outcomes

On logistic regression for three-year mortality, only achieving no evidence of disease was associated with a significant decrease in patient mortality 5:30. The overall survival of the entire cohort was quite high 6:00, though event-free survival was lower due to relapse events after achieving no evidence of disease 6:15.

The starkest finding was the survival difference based on disease clearance. Among patients who could not be rendered no evidence of disease, overall survival and event-free survival were significantly lower, with no patients in these cohorts surviving past two years 6:30. This was not a matter of degree — it was binary. Residual disease was incompatible with long-term survival in this series.

Among patients who achieved no evidence of disease, overall survival was similar between high-risk and not-high-risk groups 7:00. Event-free survival remained significantly lower for the high-risk no evidence of disease group due to five high-risk patients who relapsed 7:15, but three of those five were salvaged 7:30. The high-risk designation predicted relapse, but not ultimate survival if the team could clear disease again.

What the Case Changes

The transferable judgment is this: no evidence of disease status is necessary for survival in patients with hepatoblastoma 7:45, and repeated pulmonary metastasectomy and complex local control strategies to obtain no evidence of disease benefit high-risk patients 7:45. The survival curves separate cleanly at the point of complete clearance, not at the point of risk stratification.

This is not a call for futile surgery. The series does not report outcomes for patients with unresectable disease, and the definition of resectability is itself a judgment. But within the bounds of technical feasibility, the data support persistence. A high-risk patient with recurrent pulmonary nodules after initial metastasectomy is not a candidate for palliation — they are a candidate for another operation, because three of five such patients in this series survived after salvage resection 7:30.

The outcome for patients who could not be rendered disease-free — none surviving past two years 6:30 — was not discussed in detail, but that absence is itself informative. The series does not claim that all patients can be cleared, only that clearing them is the only path to cure.

The Clinical Question

A child with hepatoblastoma presents with pulmonary metastases after initial chemotherapy and hepatic resection. Alpha-fetoprotein remains elevated, and surveillance imaging shows multiple bilateral lung nodules. The primary tumor has been resected, but radiographic disease persists. The question is not whether to operate — it is how far to go. Single thoracotomy? Bilateral procedures? Reoperation if nodules recur? The literature on pulmonary metastasectomy for hepatoblastoma reports variable survival benefits 2:00, and the threshold at which aggressive resection becomes futile rather than curative is not well defined.

Fleming's institutional series from St. Jude examined 50 patients treated between 2005 and 2021, of whom 24 met AHEAD 1531 criteria for high-risk disease 4:00. The analysis centered on a single endpoint: achieving no evidence of disease, defined as normal AFP and absence of detectable radiographic disease on surveillance imaging 3:00. This is a binary state — either all measurable tumor is gone, or it is not. The question was whether pursuing that state aggressively, including through repeated pulmonary resections, changed survival in high-risk patients.

What the Team Did

All 50 patients received chemotherapy 4:15. Fourteen patients underwent pulmonary metastasectomy — 31 procedures total, a median of four nodules resected per patient 4:30. The approach was explicitly aggressive: bilateral procedures, reoperative resections when nodules recurred. The pathology confirmed the imaging: a median of three nodules per patient contained viable hepatoblastoma 5:00, meaning the radiographic targets were real disease, not scar or chemotherapy effect.

Forty-one of the 50 patients were rendered no evidence of disease 4:15. Nine could not be cleared. The survival difference between these groups was absolute. On logistic regression for three-year mortality, only achieving no evidence of disease was associated with a significant decrease in mortality 5:30. Patients who could not be rendered disease-free did not survive past two years 6:30. The overall survival of the entire cohort was high 6:00, but event-free survival was lower due to relapses after initial clearance 6:15.

The High-Risk Subgroup

Among high-risk patients who achieved no evidence of disease, overall survival matched that of lower-risk patients who were similarly cleared 7:00. Event-free survival was lower in the high-risk group — five patients relapsed 7:15 — but three of those five were salvaged with further chemotherapy and additional pulmonary resections 7:30. The implication is that high-risk designation does not preclude cure if complete resection can be achieved and sustained, even if that requires multiple operations.

The series does not report how many nodules were too many, or which anatomic distributions were unresectable, or what imaging findings predicted futility. It reports only that 41 of 50 patients could be cleared and that clearing them was necessary for survival. The nine who could not be cleared all died within two years 6:30.

What the Case Changes

Complete surgical resection is necessary for survival in hepatoblastoma 1:30, and cisplatin-based chemotherapy has improved resectability 1:45, but this series argues that "complete resection" must include the lungs if that is where residual disease sits. The survival curves separate cleanly at the no-evidence-of-disease threshold 6:30. The conclusion is that repeated pulmonary metastasectomy and complex local control strategies to obtain no evidence of disease benefit high-risk patients 7:45, and that no evidence of disease status is necessary for survival 7:45.

The transferable judgment is this: if imaging shows resectable pulmonary disease and AFP remains elevated, the question is not whether metastasectomy is justified — it is whether complete clearance is achievable. If it is, the operation is indicated regardless of how many nodules are present or how many procedures it takes. If it is not, the prognosis is measured in months. The decision point is resectability, not risk stratification. High-risk patients who can be cleared survive at the same rate as lower-risk patients who can be cleared 7:00. Patients who cannot be cleared do not survive, regardless of risk group 6:30.

The Clinical Problem

Hepatoblastoma is the most common pediatric liver cancer 0:00. Complete surgical resection has been determined to be necessary for patient survival 1:30, and cisplatin-based chemotherapies have improved rates of resectability and subsequently survival for these patients 1:45. But achieving complete disease clearance often requires more than resecting the primary tumor. Some patients harbor pulmonary metastases that must be addressed before they can be considered disease-free. The question is whether aggressive pursuit of that endpoint — including repeated thoracic operations — changes outcomes, particularly in high-risk patients.

The reported survival benefits of pulmonary metastasectomy for hepatoblastoma have varied in the literature 2:00. This uncertainty creates a decision point: when radiographic disease persists in the lungs after chemotherapy and liver resection, should the team pursue repeated pulmonary resections, or accept that some disease will remain?

The Institutional Approach

Fleming's group at St. Jude takes an aggressive stance. Among 50 patients treated between 2005 and 2021, all received chemotherapy and 41 were successfully rendered no evidence of disease status 4:15. No evidence of disease is defined as a normal AFP and the absence of detectable radiographic disease on interval surveillance imaging 3:00.

Fourteen patients underwent a total of 31 pulmonary procedures, with a median of four nodules resected per patient 4:30. These were not single operations. The approach included bilateral and reoperative pulmonary resections when necessary. The radiographic targets proved accurate: a median of three nodules per patient contained pathologically confirmed viable hepatoblastoma, confirming a high degree of concordance between radiographic evidence of disease and pathological confirmation of disease 5:00.

Twenty-four patients in the cohort were determined to be high risk 4:00 according to AHEAD 1531 trial criteria, which incorporate PRETEXT group, presence or absence of metastatic disease, alpha fetoprotein at diagnosis, PRETEXT modifiers, patient age, and resectability at diagnosis 1:00. PRETEXT group itself is determined by the number of contiguous hepatic sections uninvolved by tumor 0:30. High-risk designation matters because these patients face worse prognoses and the question of whether aggressive local control changes that trajectory is not settled.

What Happened

On logistic regression for three-year mortality, only achieving no evidence of disease was associated with a significant decrease in patient mortality 5:30. The overall survival of the entire cohort is quite high 6:00, but event-free survival is lower than overall survival due to relapse events after achieving no evidence of disease 6:15.

The divergence becomes stark when patients are stratified by whether the team succeeded in rendering them disease-free. The overall survival and event-free survival of the group in whom they were unable to render them no evidence of disease is significantly lower, with no patients in these cohorts surviving past two years 6:30.

Among patients who did achieve no evidence of disease, the overall survival is similar between high-risk and not high-risk patients 7:00. But the event-free survival remains significantly lower for the high-risk no evidence of disease group due to five high-risk patients who relapsed 7:15. Three of those five high-risk patients who relapsed were successfully salvaged using a combination of chemotherapy and further pulmonary resections 7:30.

What the Case Changes

No evidence of disease status is necessary for survival in patients with hepatoblastoma 7:45. That statement is stronger than saying it correlates with survival — the data suggest it may be causal. Repeated pulmonary metastasectomy and complex local control strategies to obtain no evidence of disease benefit high-risk patients 7:45.

The transferable judgment is this: when imaging shows resectable pulmonary disease after chemotherapy and liver resection, the threshold for operating should be low, even if it means bilateral procedures or reoperations. The high concordance between radiographic and pathologic disease means you are resecting real tumor, not chasing artifacts. And in high-risk patients who relapse after achieving disease-free status, the same aggressive approach to salvage — more chemotherapy, more pulmonary resections — can succeed. The patients who did not survive past two years were those in whom no evidence of disease could not be achieved. That is the outcome to avoid.