according to a pooled analysis of cooperative group trials.
The findings suggest a role for the routine use of tumor markers for surveillance in CNS-NGGCT patients, said Adriana Fonseca, MD, a pediatric neuro-oncology fellow at the Hospital for Sick Children in Toronto.
She presented these findings as part of the American Society of Clinical Oncology virtual scientific program.
This pooled analysis represents the largest prospective cohort to date of relapsed intracranial germ cell tumors, Dr. Fonseca said. The analysis included 483 patients enrolled in five prospective CNS-NGGCT trials between 1989 and 2016. There were 106 patients who relapsed after the end of therapy; the relapsed patients had a median age of 13 years (range, 1-30 years) at diagnosis and 82% were male.
Tumor marker utility
There were 86 patients with tumor marker assessments at diagnosis, and 83 had tumor marker elevations in serum, cerebrospinal fluid (CSF), or both.
The three patients without tumor marker elevations at diagnosis had mixed GCT, choriocarcinoma, and yolk sac tumor, which are usually associated with tumor marker elevation, so this will be investigated further, Dr. Fonseca said.
The sensitivity of tumor markers at diagnosis was 94% for serum, 83% for CSF, and 97% for either serum or CSF.
The median time to relapse was 15.5 months. Relapses were local in 45 patients (44%), distant in 32 (33%), and combined in 22 (21%). Three intracranial relapses were located outside of the radiation field and were classified as distant.
Only two patients presented with isolated tumor marker elevations as the sole evidence of relapse, and the elevations usually preceded the presence of macroscopic disease, Dr. Fonseca said.
At the time of relapse, 88% of patients (n = 73) had tumor marker elevations. The sensitivity of tumor markers was 82% in serum, 85% in CSF, and 88% in either.
To better understand if tumor markers can be used for surveillance, the researchers analyzed data from patients who had either serum or CSF tumor marker levels available at both diagnosis and relapse.
Of the 74 evaluable patients who had elevated tumor markers at diagnosis, 68 had elevated tumor markers at relapse as well. This means 92% of relapsed patients were detectable by tumor markers, Dr. Fonseca said.
“Only six patients had tumor marker–negative relapses, and interestingly, one patient who was tumor marker negative at diagnosis relapsed with tumor markers positive,” she added.
Rationale and next steps
CNS-NGGCTs are rare and heterogeneous tumors that respond best when treated using multimodal approaches, including surgical resection, chemotherapy, and radiation, according to Dr. Fonseca. The 5-year event-free and overall survival rates range from 72%-84% and 82%-93% respectively.
“GCTs are unique as they express tumor markers, such as AFP and beta-HCG, which we know are sensitive and specific and used for diagnostic and monitoring purposes,” Dr. Fonseca said.
Current surveillance strategies use a combination of brain and spine MRI and serum tumor markers with declining frequency over time.
“CSF tumor markers are not performed during follow-up, and they are usually obtained only at the time of relapse,” Dr. Fonseca said. “But what is the best surveillance strategy? We have to remember that some of our patients require sedation to undergo MRI, and recurrent sedations in children have been recently associated with potential detrimental neurocognitive effects.”
Similarly, the administration of gadolinium used for MRI has been associated with an increased risk of renal fibrosis and negative neurological outcomes.
“Additionally, nonspecific areas of enhancement are commonly encountered and can lead to unnecessary further investigations,” Dr. Fonseca said, adding that this can contribute to patients’ and parents’ anxiety and to increased overall health care costs and resource utilization.
Recent Children’s Oncology Group data showed that 98% of patients with extracranial germ cell tumors who relapsed were detectable by tumor markers alone, and this led to a change in surveillance guidelines for those patients. This raised the question as to whether a similar approach could be used in CNS-NCCGTs, Dr. Fonseca explained.
“We hypothesized that tumor markers alone may be sufficient for relapse detection in children and adolescents treated for CNS-NGGCT, and hence, the frequency and associated risk with serial MRIs could be safely avoided,” she said.
Though this study was limited by missing data in some cases, the inclusion of trials from different eras, and the use of different detection techniques across trials, the findings confirm the sensitivity of tumor markers in this setting.
“Tumor markers represent a valuable surveillance strategy with the potential to reduce MRI frequency in these patients,” Dr. Fonseca said. “Additionally, the higher proportion of tumor marker–negative relapses, compared to extracranial GCTs, suggests a different biological behavior. Further studies to investigate the biology of the primary versus relapsed samples in GCTs are currently needed.”
Dr. Fonseca and colleagues are “currently undertaking some correlative outcomes analyses to try to understand if the elevation or nonelevation to tumor markers is correlated with survival. We also would like to elucidate the optimal MRI frequency required for surveillance,” she said.
Dr. Fonseca reported having no disclosures, and the researchers disclosed no funding for the study.
SOURCE: Fonseca A et al. ASCO 2020, Abstract 2503.