Severe Hepatopathy in National Wilms Tumor Studies 3-5: Prevalence, Clinical Features, and Outcomes After Reintroduction of Chemotherapy

Abstract

PURPOSE

The safety of reintroducing chemotherapy in the pediatric renal tumor setting after severe hepatopathy (SH), including sinusoidal obstruction syndrome (SOS), is uncertain. We describe the incidence, severity, outcomes, and impact on subsequent treatment for patients with SH from National Wilms Tumor Study (NWTS) protocols 3-5.

PATIENTS AND METHODS

Archived charts for patients enrolled on NWTS 3-5 who met study inclusion criteria for SH by using established hepatopathy grading scales and clinical criteria were reviewed for demographics, tumor characteristics, radio- and chemotherapy details, SH-related dose modifications, and oncologic outcomes. Genomic analysis for candidate polymorphisms associated with SH was performed in 14 patients.

RESULTS

Seventy-one of 8,862 patients (0.8%) met study inclusion criteria. The median time from therapy initiation to SH was 51 days (range, 2-293 days). Sixty percent received radiotherapy, and 56% had right-sided tumors. Grade 1-4 thrombocytopenia was noted in 70% at initial occurrence of SH (median 22,000/microliter). Among 69 of 71 children with SH occurring before the end of therapy (EOT) and post-SH treatment information available, chemotherapy was delayed posthepatopathy for 65% (69% of these at a reduced dose), continued without delay for 20% (57% of these at reduced dose), and stopped completely for 15% (4 of 10 of whom died of SH). Overall, 42% of patients with dose reductions achieved full dose by EOT. The five-year post-SH event-free survival for patients who continued therapy was 89% (95% CI, 81 to 98), with no significant differences by whether delay or dose reduction occurred. We identified no SH-associated pharmacogenomic polymorphism.

CONCLUSION

The incidence of SH on NWTS 3-5 was low; many had associated severe thrombocytopenia. Careful reintroduction of chemotherapy appeared to be feasible for the majority of patients who developed severe chemotherapy- and/or radiotherapy-induced liver toxicity.

INTRODUCTION

Liver toxicity is a frequent adverse event in the treatment of pediatric cancer, including Wilms tumor (WT), and may be severe. Sinusoidal obstruction syndrome (SOS; historically known as Veno-Occlusive Disease of the liver [VOD]) is a recognized potentially life-threatening complication of radiotherapy and/or chemotherapy with consistent clinicopathologic findings of hepatocyte necrosis and obliteration of small hepatic venules¹ and hepatomegaly, right upper quadrant pain, weight gain, ascites, and variable elevation of liver enzymes and bilirubin.^2,3 SOS has most frequently been reported as a complication of hematopoietic stem-cell transplantation (HSCT), but is well-recognized to occur in treatment of pediatric solid tumors.^2,4-6 Despite this known occurrence of SOS, little information exists on whether it is safe to reintroduce chemotherapy.

CONTEXT

• Key Objective

• What is the clinical impact of reintroduction of chemotherapy for children with renal tumors after an episode of treatment-related severe hepatopathy (SH)?

• Knowledge Generated

• In a cohort of 8,862 children with renal tumors, the incidence of SH was low. Chemotherapy was continued or reintroduced in 85% of patients with SH, 72% of whom had no recurrence of any hepatopathy, and no patient who resumed chemotherapy after SH died of liver toxicity. The majority of children who restarted chemotherapy achieved resumption of full dosing by the end of therapy, and outcomes of patients with delays or dose reductions were not worse than those without modifications.

• Relevance (S. Bhatia)

• Chemotherapy rechallenge for SH patients treated for Wilms tumor appears feasible in the majority, with application in other cancer settings.*

  *Relevance section written by JCO Associate Editor Smita Bhatia, MD, MPH, FASCO.

The incidence of SOS in patients with WT has been reported to be between 1.2% and 8% although precise estimates are limited. Recognized risk factors include younger age, right-sided tumors, and lower body weight.^2,3 SOS has been associated with chemotherapy (such as dactinomycin, doxorubicin, vincristine, and cyclophosphamide) and radiotherapy.^6-8 There are several SOS classification systems, typically describing patients from the HSCT setting.^9,10 Notably, thrombocytopenia and respiratory distress are often observed but are not part of any classification system.

Current Children's Oncology Group (COG) and International Society of Pediatric Oncology (SIOP) renal tumor protocols provide recommendations for subsequent chemotherapy dose modifications after mild or moderate hepatopathy (Appendix Table A1, online only) but not for SOS or severe hepatopathy (SH), which is left to the discretion of the clinician.^2,3,6,11-13 Individual and case series reports suggest that reintroduction in a graduated manner is safe in mild to moderate hepatopathy, but the impact of this strategy in large patient cohorts or in episodes of either SOS or SH has not been reported.^2,6,13 To better understand whether reintroduction of chemotherapy is feasible in severe cases, we reviewed the incidence, clinical characteristics, and laboratory manifestations of SH and/or SOS in all patients with renal tumors treated on National Wilms Tumor Study (NWTS) 3-5 protocols. We describe the effect of chemotherapy dose modifications on patient outcomes and provide considerations for reintroduction of chemotherapy.

PATIENTS AND METHODS

Patients

Written informed consent was obtained from all NWTS patients and/or patients' guardians before treatment, according to national law and regulations in force at the time of enrollment. Ethical approval was obtained from medical ethical committees of all participating US and international centers for the original NWTS trials. Permission to review the NWTS patient charts was obtained through application to the NWTS Data Center and the COG Renal Tumor Committee. Copies of source documents for NWTS study patients were kept in a central repository in Seattle, Washington, until 2016, and have now been scanned into an electronic database overseen by the COG Statistical Data Center. Approval was obtained from Lurie Children's Institutional Review Board to perform WGS on samples where written consent for research use had been given.

Charts from participants enrolled on NWTS 3 (May 1979-September 1986), NWTS 4 (August 1986-August 1995), and NWTS 5 (August 1995-May 2002), which were previously coded with hepatopathy, VOD, or SOS, were examined by five reviewers (E.A.M., T.E.H., C.V.F., S.L.M.G., N.O.). Patients with WT, Clear Cell Sarcoma (CCSK), and Malignant Rhabdoid Tumor of the Kidney (MRTK) were eligible. Therapy and survival data have previously been reported.^14-16 Data on patient and tumor characteristics, symptoms, laboratory abnormalities, severity grade, timing, chemotherapeutic agents, radiotherapy, dose modifications, long-term outcome, and residual toxicity were extracted and later confirmed against the official electronic study databases by the current COG Renal Committee statistician (L.A.R.).

Inclusion Criteria

We reviewed all charts that had been coded as having the occurrence of any hepatopathy by NWTS reviewers. Although reporting of the occurrence of hepatopathy was mandated on all three NWTS studies reviewed, we found that corresponding supportive laboratory data and clinical data were inconsistently recorded. As there are no gold standard criteria for SOS outside the setting of HSCT, we used two different published diagnostic criteria and grading scales for SOS and hepatopathy, World Health Organization (WHO) criteria and Modified Seattle criteria (Table 1).^9,10,17 We allowed inclusion if one of the three following inclusion criteria was met: (1) meeting Modified Seattle Criteria, (2) WHO grade III or IV, or (3) WHO grade II with at least one clinical symptom of SOS (hepatomegaly, ascites, weight gain, pleural effusion). Patients identified in the NWTS database as having hepatopathy, but with only WHO grade I or II transaminase elevation without clinical signs or symptoms of SOS, were excluded (one patient without transaminase levels reported during documented hepatopathy, with weight gain and ascites, was included). Since all patients with SOS were captured within our defined category of SH, we will refer to the studied patients as those with SH.

TABLE 1.

Sinusoidal Obstruction Syndrome and Hepatopathy Grading Scales

Statistical Methods

Clinical Data Analysis

Descriptive statistics on patient and disease characteristics were summarized with rates for categorical variables, and continuous variables were described with medians and ranges. Reasons for holding therapy and complications of restarting chemotherapy were described. Post-SH event-free survival (EFS), defined as the time from first occurrence of SH to the earliest of disease progression, relapse, or death, was reported for the subset of patients who were able to resume therapy (even if delayed or dose-reduced), with patients right-censored at the time of last known disease status. Reported EFS did not include episodes of second SH. Log-rank tests and Cox proportional hazards models were used to compare post-SH EFS by whether subsequent chemotherapy after SH was delayed or dose-reduced and estimate the effects of these strategies. All analyses and graphs were produced using R.¹⁸

Pharmacogenomic Polymorphism Analysis

To identify pharmacogenomic polymorphisms that may predict hepatotoxicity in patients treated for WT, the full PharmGKB database¹⁹ was filtered on August 15, 2021, to identify single-nucleotide and small indel polymorphisms previously associated with hepatotoxicity. This resulted in 170 polymorphisms. In the current study, for patients who had available banked sources of constitutional DNA that met quality controls, materials were submitted to Psomagen²⁰ for Whole Genomic Sequencing. All samples were confirmed to have DNA DIN >7 and sequenced on NovaSeq6000 to 30× coverage (150 bp physical exam). Single-nucleotide variants and small indels were called using the Genome Analysis Toolkit HaplotypeCaller pipeline.²¹ The resulting Genomic Variant Call Format files were filtered to include only the calls for the 170 pharmacogenomic polymorphisms defined above. Patient ethnicity was estimated using the LASER/TRACE method.²²

RESULTS

Composition of the SH Cohort by Study

Of 8,862 patients with renal tumor treated on NWTS 3-5, 146 (1.6%) were coded as having hepatopathy (Fig 1). We identified 71 patients (0.8%) as meeting our inclusion criteria for SH (Table 2), 16 of 2,496 (0.6%) from NWTS 3, 34 of 3,335 (1.0%) from NWTS 4, and 21 of 3,031 (0.7%) from NWTS 5. Patient characteristics are shown in Table 3.

FIG 1.

Flow diagram of patient inclusion. NWTS, National Wilms Tumor Study; SH, severe hepatopathy.

TABLE 2.

Patient Characteristics

TABLE 3.

Clinical Characteristics of Hepatopathy

Treatment Protocol, Onset of SH, and Clinical Symptoms

At the time of data freeze for this study, all 71 patients had either terminated or completed planned therapy. Patients were treated with a variety of published treatment regimens (Table 2),^14,15,23-25 but SH occurred most often (80.3%) after chemotherapy courses of actinomycin (AMD) + vincristine (VCR). Of these, 56.2% occurred after pulse-dose AMD and 43.8% occurred after divided-dose AMD. The median elapsed days from the start of initial chemotherapy until the diagnosis of SH was 51.0 days (range, 2-293 days). Occurrence of clinical symptoms of SH was recorded in 44 patients (Table 3). There was a high incidence of reported thrombocytopenia (67.6%) during episodes of hepatopathy; strikingly, 52.1% of all patients studied (and 84% of those with thrombocytopenia) were classified with Grade 3 or 4 thrombocytopenia. The median of the lowest reported platelet counts during the episodes of hepatopathy was 23 × 10⁹/L (range, 1-92 × 10⁹/L) in those with thrombocytopenia.

Radiation Therapy

Radiation therapy (RT) was administered to 43 of 71 patients (60.6%), 42 of whom had detailed timing, target volume, and dosing data available. Ten patients of the 42 (23.8%) received radiation to the left flank, 12 (28.6%) to the right flank, 19 (45.2%) to the whole abdomen, and 11 (26.2%) to the lung(s), with some patients receiving radiation to more than one site. Of the same 42 patients, 35 (83.3%) started radiation before the onset of SH (for 3 of 35, onset was during radiation (XRT); for 32, onset was after completion of XRT). SH occurred more than 6 weeks after RT initiation for 26 of 35 patients (74.3%). The mean and median times from XRT initiation to SH were 74 days and 52 days (range, 1-287 days).

Chemotherapy Modifications and Patient Status After SH

Patient Inclusion

Of the 71 total patients studied, one patient had received the final planned cycle of chemotherapy before the onset of SH and therefore was not subject to a reduction or delay, and one patient's post-SH treatment was not recorded because of an institution change. They remained alive and disease-free at 17 and 23 years, respectively. The analyses that follow describe post-SH treatment modifications and clinical outcomes of the remaining 69 patients.

Continuation Versus Termination and Chemotherapy Delay

Of the 69 patients with SH occurring during planned chemotherapy whose post-SH treatment histories are known, 10 patients (14.5%) stopped all chemotherapies without resuming and 59 (85.5%) eventually resumed chemotherapy. Of the 10 who stopped chemotherapy, four subsequently died of SH and six were alive and disease-free at last follow-up (range 4 to 26 years post-SH) although one from the latter group was subsequently diagnosed with hepatitis C. The four patients who died of SH were treated according to DD4A (n = 2), K4 (n = 1), or EE4A (n = 1) protocols. DD4A and EE4A protocols contain 45 mcg/kg × 1-day dactinomycin given either once every 3 weeks (EE4A) or 6 weeks (DD4A), compared with some other chemotherapy protocols in the NWTS era, which contained 15 mcg/kg × 5 days dactinomycin given in a row either once every 3 or 6 weeks. Of the 59 patients who resumed chemotherapy post-SH, 45 (76.3%) did so after a delay for symptom recovery, whereas 14 patients (23.7%) continued chemotherapy without a delay. This breakdown, along with further classification of patients who continued chemotherapy by dose reductions and subsequent achievement of full dose by the end of therapy (EOT), is displayed in Figure 2. No patient who resumed chemotherapy after SH died of liver toxicity.

FIG 2.

Posthepatopathy treatment strategy diagram. EOT, end of therapy; SH, severe hepatopathy.

Dosing Among Patients With Chemotherapy Delay

Among the 45 patients rechallenged after a chemotherapy delay, 14 (31.1%) restarted at a full dose, with 3 of 14 experiencing mild serum glutamic-oxaloacetic transaminase (SGOT) elevations. In the same group, 12 patients (85.7%) were alive and disease-free at last follow-up, one patient (2.3%) eventually died of disease, and one (2.3%) remained alive with disease. The other 31 patients with a chemotherapy delay (68.9%) restarted at a reduced dose (drug-specific dose reduction range 20%-70%), with 9 of 31 experiencing restart complications such as mild to moderate SGOT/serum glutamic-pyruvic transaminase elevation. Among the 31 patients who restarted chemotherapy at a reduced dose after delay, 12 (38.7%) achieved a full dose by EOT; nine (75%) were alive and disease-free at last follow-up, two (16.7%) subsequently died of disease, and one (8.3%) died of other causes. The other 19 patients with a reduced dose after a delay (61.3%) maintained drug-specific dose reductions of 20%-50% at EOT; all were alive and disease-free at last follow-up.

Dosing Among Patients Without Chemotherapy Delay

Of the 14 patients without a delay in chemotherapy after the onset of SH, six (42.9%) continued at a full dose, with 2 of 6 re-experiencing elevated liver function tests (LFTs). In the same group, five patients (83.3%) remained alive and disease-free at last follow-up and one patient (16.7%) died of disease. The other eight patients without a post-SH chemotherapy delay (57.1%) continued therapy at a reduced dose (drug-specific dose reduction range, 30%-67%), among whom 3 experienced mild LFT elevation or ascites. Among these eight patients without delay continuing at a reduced dose, whether full dose was reached by EOT was unknown for three patients (all were alive and disease-free at last follow-up). Among the other five patients with available data, three (60.0%) achieved a full dose by EOT (one died because of disease, and two remained alive and disease-free) and two (40.0%) maintained drug-specific dose reductions of 25%-40% at EOT (both remained alive and disease-free).

Post-SH Event-Free Survival by Chemotherapy Continuation, Delay, and Dose Reduction

In the 59 patients who resumed chemotherapy with or without delay or dose reduction after SH, the median follow-up for post-SH EFS was 11 years. The same group had a combined 5-year EFS of 89% (95% CI, 81 to 98; Fig 3A). Among them, post-SH EFS did not differ by whether chemotherapy was delayed (log-rank P = .89; Fig 3B) or whether dose was reduced (log-rank P = .62; Fig 3C).

FIG 3.

Post-SH EFS by treatment factors. (A) Post-SH EFS among patients who resumed chemotherapy. (B) Post-SH EFS by whether resumed chemotherapy was delayed. (C) Post-SH EFS by whether resumed chemotherapy was dose-reduced. EFS, event-free survival; SH, severe hepatopathy.

Fourteen patients from this study had available DNA samples. Binomial tests were used to analyze the frequency of each of the 170 pharmacogenomic polymorphisms (1) in all 14 patients with WT compared with the gnomAD general population and (2) in the nine EUR patients with WT compared with the gnomAD non-Finnish European population in gnomAD. Neither comparison demonstrated polymorphisms with an adjusted P value of < .05.

DISCUSSION

We found a very low incidence (0.8%) of SOS in children treated on NWTS 3-5 protocols, in contrast to 1.2%-8% described in previous reports.^2,3 The delivery of chemotherapy and radiotherapy varied across the three studies, but the incidence we observed was remarkably consistent (Fig 1). Our prevalence rates are likely lower than those previously reported because we accepted only reviewed cases that met stringent SH criteria. As criteria for a diagnosis of SH were not collected prospectively, it is possible that we have underestimated the true incidence. To mitigate this, we identified but excluded cases with milder degrees of hepatopathy (eg, simple rise in transaminase without any other criteria for hepatopathy/SOS) as not representative of the entity of interest of SH, but did include patients with Grade II transaminase elevations and clinical symptoms in our cohort.

Our findings demonstrate that the continuation/reintroduction of chemotherapy was safe and feasible in the majority of patients after an episode of SH. Although a minority of rechallenged patients (28%) experienced recurrence of mild LFTS elevations or symptoms, none of the subsequent hepatopathy cases were fatal, and the outcome of patients with delay or dose reduction was not worse than those without these modifications. We acknowledge that the outcomes observed may not be representative of all patients with SH, because of overall low incidence, patient heterogeneity, and selection of the first episode of SH as the start of outcome calculations (rather than diagnosis or treatment initiation). There were no obvious baseline characteristics that could predict those likely to develop some degree of hepatopathy on reintroduction of chemotherapy, nor which patients could be started on full versus reduced dose. Our results parallel SIOP reports and other small case series, wherein reintroduction of chemotherapy, generally at reduced dose, appears to be feasible.^2,3,6,11-13 In these previous reports, a significant proportion of patients never received further chemotherapy although the selection criteria for this strategy are not described.

A striking finding was the common occurrence of very severe thrombocytopenia during episodes of SH. Currently used classification schemas for SOS were designed to be used in the context of myeloablative therapy, and thus, thrombocytopenia is not part of their criteria.^9,11,17 Grade 3 or 4 thrombocytopenia was identified in the majority of our patients with reported thrombocytopenia. This likely underestimates the true incidence of severe thrombocytopenia, as reporting of platelet count in the setting of SH was not required on NWTS 3-5. Given our observation, SH/SOS should be considered early in the differential diagnosis of ill patients with renal tumor with low platelets, as is supported by European Society for Blood and Marrow Transplantation (ESBMT) pediatric SOS recognition guidelines and others.²⁶

To define SH and SOS for our study, we chose two published scales (WHO and Modified Seattle). We recognize that other classifications (Seattle, Baltimore, McDonald)^10,27 that describe SOS exist, but we chose these as the most commonly used and relevant. We believe that the small differences noted in the criteria (primarily focused on the degree of hyperbilirubinemia) are very unlikely to change our conclusions. Our expansion to include patients with WHO grade II criteria with at least one clinical symptom of SOS (hepatomegaly, ascites, weight gain, pleural effusion) is justified by a past observation that up to 30% of children do not have significant hyperbilirubinemia despite other features of classical SOS.²⁸ Patients in our study were unlikely to be exposed to defibrotide as the first clinical report of efficacy in VOD was published in 1998.²⁹ Nevertheless, early recognition using the proposed ESBMT classification system is important given the efficacy of defibrotide for this complication.^28,29

It is possible that we have underestimated the true incidence of SH as this was not collected prospectively on NWTS. However, we believe that it is very unlikely that the exclusion of patients with milder hepatopathy will change the conclusions about safety of reintroduction of chemotherapy. A further limitation may be the general applicability of our findings as chemotherapy and radiotherapy schedules for renal tumors in North America and Europe have evolved over time, nor may the findings be fully translatable to nonrenal tumor patients. However, many of the regimens and drugs remain in common use.

We were unable to demonstrate a specific pharmacogenomic polymorphism that differed in frequency from population norms although the number of available samples was small. The strength of these observations is the large data set of almost 9,000 patients at risk for hepatopathy, the use of prospectively collected data on NWTS trials, and the use of standardized definitions of SOS.

In conclusion, chemotherapy rechallenge for patients with SH treated on NWTS protocols appears to be feasible in the majority of patients and may be possible in other cancer settings. A conservative reintroduction strategy that is likely to be successful is to wait for resolution of severe liver toxicity and initiate the putative responsible chemotherapy at 50% total dosing over additional days. Specific guidelines for recognition and management of SH including recording of platelet count and reintroduction options should be included in future prospective renal tumor therapeutic protocols.

APPENDIX

TABLE A1.

Therapy Modification for Hepatopathy According to COG and SIOP Recommendations

AUTHOR CONTRIBUTIONS

Conception and design: Natanja Oosterom, Marry M. van den Heuvel-Eibrink, Paul E. Grundy, Najat C. Daw, Jeffrey S. Dome, Conrad V. Fernandez, Elizabeth A. Mullen

Provision of study materials or patients: Elizabeth J. Perlman, Paul E. Grundy

Collection and assembly of data: Natanja Oosterom, Saskia L.M. Gooskens, Lindsay A. Renfro, Marry M. van den Heuvel-Eibrink, Thomas E Hamilton, Paul E. Grundy, Conrad V. Fernandez, Elizabeth A. Mullen

Data analysis and interpretation: Natanja Oosterom, Saskia L.M. Gooskens, Lindsay A. Renfro, Elizabeth J. Perlman, Marry M. van den Heuvel-Eibrink, Daniel M. Green, James I. Geller, Jeffrey S. Dome, Conrad V. Fernandez, Elizabeth A. Mullen

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Severe Hepatopathy in National Wilms Tumor Studies 3-5: Prevalence, Clinical Features, and Outcomes After Reintroduction of Chemotherapy

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).