Imaging Analysis

Periventricular germinoma with leptomeningeal dissemination: A case report

A 24-year-old Guyanese man presented with a 3-month history of personality changes, abulia, headaches, nausea and vomiting.

On examination, he had diminished verbal output, flat affect and reduced prosody.

Neurologic examination was otherwise unrevealing. Referenced electroencephalography demonstrated mild generalized slowing without epileptiform activity.

Imaging and pathology findings

A contrast-enhanced brain MRI demonstrated extensive confluent periventricular T2 signal abnormality extending into the corpus callosum with numerous small T2-hyperintense cavitary lesions.

Post-contrast images demonstrated nodular enhancement in the corpus callosum and surrounding the lateral ventricles, abutting the ependymal lining of the ventricular system (Figure 1).

 Figure 1. Axial T2-weighted (left) and post-contrast T1-weighted (right) images demonstrate confluent periventricular areas of hyperintense signal abnormality, with associated cavitation and patchy foci of accompanying nodular enhancement on post-contrast images.
Figure 1. Axial T2-weighted (left) and post-contrast T1-weighted (right) images demonstrate confluent periventricular areas of hyperintense signal abnormality, with associated cavitation and patchy foci of accompanying nodular enhancement on post-contrast images.

Source: M. Ghesani, MD reprinted with permission.

There was subtle enhancement along the pial surface of the medulla and possibly the pons, raising the possibility of leptomeningeal spread.

Total spine MRI was notable for diffuse leptomeningeal enhancement along the surface of the spinal cord, including thickening and nodular enhancement of the cauda equina nerve roots with associated mass-like enhancement of the conus and distal thecal sac (Figure 2).

 Figure 2. Sagittal (left) and axial (right) post-contrast T1-weighted images of the lumbar spine demonstrate enhancing, thickened cauda equina nerve roots with associated abnormal mass-like enhancement of the conus and distal thecal sac.
Figure 2. Sagittal (left) and axial (right) post-contrast T1-weighted images of the lumbar spine demonstrate enhancing, thickened cauda equina nerve roots with associated abnormal mass-like enhancement of the conus and distal thecal sac.

Cerebrospinal fluid analysis demonstrated increased white blood cell count (lymphocyte predominance) with cytology negative for malignant cells. Infectious markers — including cerebrospinal fluid encephalitis panel, Epstein-Barr virus, herpes simplex virus, hepatitis C virus, hepatitis B surface antigen, and HIV-1 and HIV-2 antibody screen — as well as cerebrospinal fluid gram stain/cultures all were negative.

Of note, serum alpha-fetoprotein was less than 1.3 ng/mL (reference range, 0-10) and cerebrospinal fluid serum alpha-fetoprotein was less than 1 ng/mL (reference range, 0-1).

Figure 3. Biopsy specimen (Figure 3a) demonstrates hypercellular CNS tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli.
Figure 3. Biopsy specimen (Figure 3a) demonstrates hypercellular CNS tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli. Immunostain (Figure 3b) demonstrates positive staining of tumor cells for OCT-4, CD117 (c-Kit) and placental alkaline phosphatase.

Serum beta-human chorionic gonadotropin was elevated, with a level of 2.5 IU/L (reference range, less than 1 IU/L).

The patient underwent a brain biopsy of the anterior frontal horn lesion.

Pathologic sections demonstrated hypercellular tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli (Figure 3a).

These large atypical cells stained immunopositive for CD117 (c-Kit), OCT-4 and placental alkaline phosphatase (Figure 3b). The results suggested a histopathologic diagnosis of malignant germ cell tumor.

Further diagnostic workup included a scrotal ultrasound and a contrast-enhanced CT of the chest, abdomen and pelvis, all of which were normal. Given the absence of disease outside of the central nervous system, the patient was diagnosed with primary intracranial germ cell tumor.

He began chemotherapy with carboplatin and etoposide. After completing two cycles, he underwent an MRI that showed cavitary change of the periventricular enhancing nodules, and minimal residual enhancement along the ependymal surface.

This suggested a near-complete treatment response.

MRI of the spine showed near resolution of the enhancing lesions in the distal thecal sac of the lumbar spine, with minimal residual enhancement along the ventral dura.

He completed two more cycles of carboplatin and etoposide as an outpatient, and restaging MRI brain and spine was negative for malignancy.

Given the diffuse leptomeningeal spread at presentation and his ultimate complete response to chemotherapy, the decision was made to proceed with craniospinal irradiation to 30.6 Gy with proton therapy.

Case discussion

Intracranial germ cell tumors account for approximately 3% of neoplasms among children in North America aged 0 to 19 years. They account for nearly 15% of intracranial pediatric tumors in Asia.

Ninety percent of patients with pure germinomas who undergo appropriate therapy survive 10 years after diagnosis.

Pure germinomas historically have been treated with craniospinal irradiation.

Whole-ventricular irradiation or whole-brain irradiation with a local boost to the tumor bed have demonstrated similarly high local control rates and gradually are replacing craniospinal irradiation as a localized, reduced-radiation approach given concern about the late toxicities associated with craniospinal irradiation.

Additionally, preradiation chemotherapy has successfully allowed a reduction in radiation dose to decrease the risk for long-term toxicity.

An intriguing aspect of intracranial germ cell tumors is their predilection for midline structures surrounding the third ventricle; specifically, the pineal, neurohypophyseal and suprasellar regions of the brain.

Germinomas with malignant characteristics — as was the case with our patient — will infiltrate into the subependymal lining of the ventricular system, spreading in a periventricular pattern.

A modern challenge is identifying germinomas that contain malignant nongerminomatous elements (eg, mixed malignant germ cell tumors) that may benefit from more intensive treatment aimed at reducing the risk for leptomeningeal dissemination.

Primary tumor location is an important consideration.

Germinomas centered in the basal ganglia and thalamus account for only 5% to 10% of all intracranial germ cell tumors and are thought to have an increased risk for tumor recurrence.

Tumors in these locations traditionally have been treated with craniospinal irradiation. Off-midline structures are less commonly involved, with only nine reported cases of primary frontal lobe germinomas in the literature.

It is unclear whether bifocal disease represents true synchronous lesions vs. early locoregional metastatic spread. Regardless, craniospinal irradiation has been associated with excellent disease control in such cases.

Serum and cerebrospinal fluid beta-human chorionic gonadotropin and alpha-fetoprotein, now commonly used as surrogate diagnostic tools in lieu of pathological confirmation of disease, are additional prognostic considerations.

Classically, pure germinomas are associated with normal serum and cerebrospinal fluid alpha-fetoprotein and low levels of beta-human chorionic gonadotropin.

Germinomas that secrete beta-human chorionic gonadotropin and/or alpha-fetoprotein may contain nongerminomatous elements, even among patients with histological confirmation of a pure germinoma.

An emerging challenge is setting uniform diagnostic criteria based on biomarkers alone, given highly variable levels in cerebrospinal fluid and serum of patients with pathologically confirmed pure germinomas.

Contemporary European and North American trials have used beta-human chorionic gonadotropin cutoff values of 50 IU/L to 100 IU/L for diagnosis and initiation of chemotherapy, yet a research group in Japan found no overall difference in 5-year survival or recurrence rates among a pediatric population when serum and cerebrospinal fluid beta-human chorionic gonadotropin were less than 200 IU/L.

Another study recommended new diagnostic beta-human chorionic gonadotropin cutoff points of more than 8.2 IU/L in the cerebrospinal fluid and more than 2.5 IU/L in the serum confirmed by pathology and cerebrospinal fluid cytology, of which our patient exceeded.

With the emerging role of serum biomarkers in the diagnosis of intracranial germ cell tumors, it is possible to identify mixed malignant germinomas without the need for surgical biopsy. Accordingly, all patients suspected to have an intracranial germ cell tumor should have serum and cerebrospinal fluid tumor marker assays performed.

Tumor location — specifically involvement of off-midline structures and the presence of bifocal disease — may be useful prognostic indicators.

Evidence-based and uniform clinical diagnostic criteria using serum and cerebrospinal fluid beta-human chorionic gonadotropin and alpha-fetoprotein are needed before neuroimaging and tumor assays alone can guide treatment decisions.

References:

Allen JC, et al. Cancer. 1994;74(3):940-944.

Cho J, et al. Radiother Oncol. 2009;doi:10.1016/j.radonc.2008.10.012.

Hu M, et al. Eur J Med Res. 2016;doi:10.1186/s40001-016-0204-2.

Khatua S, et al. Pediatr Blood Cancer. 2010;doi:10.1002/pbc.22468.

Kim JY and Park J. J Korean Neurosurg Soc. 2015;doi:10.3340/jkns.2015.57.5.315.

Luo Z, et al. World Neurosurg. 2017;doi:10.1016/j.wneu.2016.12.004.

Matsutani M, et al. J Neurooncol. 2001;54(3):311-316.

Phi JH, et al. J Neurosurg Pediatr. 2013;doi:10.3171/2012.10.PEDS11487.

Weksberg DC, et al. Int J Radiat Oncol Biol Phys. 2012;doi:10.1016/j.ijrobp.2011.04.033.

For more information:

Munir Ghesani, MD, FACNM, is assistant professor of radiology and director of PET/CT fellowship at NYU Langone Medical Center in New York. He also is a HemOnc Today Editorial Board Member. He can be reached at munir.ghesani@nyumc.org.

Benjamin T. Cooper, MD, is an attending radiation oncologist at NYU Langone Medical Center.

Ana M. Franceschi, MD, is a neuroradiology fellow at NYU Langone Medical Center.

Jonathan Howard, MD, is assistant professor of neurology and psychiatry at NYU Langone Medical Center, and associate director of the neurology service at Bellevue Hospital Center.

Juhi M. Purswani is a medical student at New York University School of Medicine.

Nikhil A. Sahasrabudhe, MD, is a neurosurgeon at Loma Linda University Health.

David Zagzag, MD, is professor of pathology and neurosurgery, as well as chief of the division of neuropathology, at NYU Langone Medical Center.

Disclosures: The authors report no relevant financial disclosures.

A 24-year-old Guyanese man presented with a 3-month history of personality changes, abulia, headaches, nausea and vomiting.

On examination, he had diminished verbal output, flat affect and reduced prosody.

Neurologic examination was otherwise unrevealing. Referenced electroencephalography demonstrated mild generalized slowing without epileptiform activity.

Imaging and pathology findings

A contrast-enhanced brain MRI demonstrated extensive confluent periventricular T2 signal abnormality extending into the corpus callosum with numerous small T2-hyperintense cavitary lesions.

Post-contrast images demonstrated nodular enhancement in the corpus callosum and surrounding the lateral ventricles, abutting the ependymal lining of the ventricular system (Figure 1).

 Figure 1. Axial T2-weighted (left) and post-contrast T1-weighted (right) images demonstrate confluent periventricular areas of hyperintense signal abnormality, with associated cavitation and patchy foci of accompanying nodular enhancement on post-contrast images.
Figure 1. Axial T2-weighted (left) and post-contrast T1-weighted (right) images demonstrate confluent periventricular areas of hyperintense signal abnormality, with associated cavitation and patchy foci of accompanying nodular enhancement on post-contrast images.

Source: M. Ghesani, MD reprinted with permission.

There was subtle enhancement along the pial surface of the medulla and possibly the pons, raising the possibility of leptomeningeal spread.

Total spine MRI was notable for diffuse leptomeningeal enhancement along the surface of the spinal cord, including thickening and nodular enhancement of the cauda equina nerve roots with associated mass-like enhancement of the conus and distal thecal sac (Figure 2).

 Figure 2. Sagittal (left) and axial (right) post-contrast T1-weighted images of the lumbar spine demonstrate enhancing, thickened cauda equina nerve roots with associated abnormal mass-like enhancement of the conus and distal thecal sac.
Figure 2. Sagittal (left) and axial (right) post-contrast T1-weighted images of the lumbar spine demonstrate enhancing, thickened cauda equina nerve roots with associated abnormal mass-like enhancement of the conus and distal thecal sac.

Cerebrospinal fluid analysis demonstrated increased white blood cell count (lymphocyte predominance) with cytology negative for malignant cells. Infectious markers — including cerebrospinal fluid encephalitis panel, Epstein-Barr virus, herpes simplex virus, hepatitis C virus, hepatitis B surface antigen, and HIV-1 and HIV-2 antibody screen — as well as cerebrospinal fluid gram stain/cultures all were negative.

Of note, serum alpha-fetoprotein was less than 1.3 ng/mL (reference range, 0-10) and cerebrospinal fluid serum alpha-fetoprotein was less than 1 ng/mL (reference range, 0-1).

Figure 3. Biopsy specimen (Figure 3a) demonstrates hypercellular CNS tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli.
Figure 3. Biopsy specimen (Figure 3a) demonstrates hypercellular CNS tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli. Immunostain (Figure 3b) demonstrates positive staining of tumor cells for OCT-4, CD117 (c-Kit) and placental alkaline phosphatase.

Serum beta-human chorionic gonadotropin was elevated, with a level of 2.5 IU/L (reference range, less than 1 IU/L).

The patient underwent a brain biopsy of the anterior frontal horn lesion.

Pathologic sections demonstrated hypercellular tissue with rare atypical cells with large nuclei, slightly irregular nuclear contours, open chromatin and often multiple prominent nucleoli (Figure 3a).

These large atypical cells stained immunopositive for CD117 (c-Kit), OCT-4 and placental alkaline phosphatase (Figure 3b). The results suggested a histopathologic diagnosis of malignant germ cell tumor.

PAGE BREAK

Further diagnostic workup included a scrotal ultrasound and a contrast-enhanced CT of the chest, abdomen and pelvis, all of which were normal. Given the absence of disease outside of the central nervous system, the patient was diagnosed with primary intracranial germ cell tumor.

He began chemotherapy with carboplatin and etoposide. After completing two cycles, he underwent an MRI that showed cavitary change of the periventricular enhancing nodules, and minimal residual enhancement along the ependymal surface.

This suggested a near-complete treatment response.

MRI of the spine showed near resolution of the enhancing lesions in the distal thecal sac of the lumbar spine, with minimal residual enhancement along the ventral dura.

He completed two more cycles of carboplatin and etoposide as an outpatient, and restaging MRI brain and spine was negative for malignancy.

Given the diffuse leptomeningeal spread at presentation and his ultimate complete response to chemotherapy, the decision was made to proceed with craniospinal irradiation to 30.6 Gy with proton therapy.

Case discussion

Intracranial germ cell tumors account for approximately 3% of neoplasms among children in North America aged 0 to 19 years. They account for nearly 15% of intracranial pediatric tumors in Asia.

Ninety percent of patients with pure germinomas who undergo appropriate therapy survive 10 years after diagnosis.

Pure germinomas historically have been treated with craniospinal irradiation.

Whole-ventricular irradiation or whole-brain irradiation with a local boost to the tumor bed have demonstrated similarly high local control rates and gradually are replacing craniospinal irradiation as a localized, reduced-radiation approach given concern about the late toxicities associated with craniospinal irradiation.

Additionally, preradiation chemotherapy has successfully allowed a reduction in radiation dose to decrease the risk for long-term toxicity.

An intriguing aspect of intracranial germ cell tumors is their predilection for midline structures surrounding the third ventricle; specifically, the pineal, neurohypophyseal and suprasellar regions of the brain.

Germinomas with malignant characteristics — as was the case with our patient — will infiltrate into the subependymal lining of the ventricular system, spreading in a periventricular pattern.

A modern challenge is identifying germinomas that contain malignant nongerminomatous elements (eg, mixed malignant germ cell tumors) that may benefit from more intensive treatment aimed at reducing the risk for leptomeningeal dissemination.

Primary tumor location is an important consideration.

Germinomas centered in the basal ganglia and thalamus account for only 5% to 10% of all intracranial germ cell tumors and are thought to have an increased risk for tumor recurrence.

Tumors in these locations traditionally have been treated with craniospinal irradiation. Off-midline structures are less commonly involved, with only nine reported cases of primary frontal lobe germinomas in the literature.

PAGE BREAK

It is unclear whether bifocal disease represents true synchronous lesions vs. early locoregional metastatic spread. Regardless, craniospinal irradiation has been associated with excellent disease control in such cases.

Serum and cerebrospinal fluid beta-human chorionic gonadotropin and alpha-fetoprotein, now commonly used as surrogate diagnostic tools in lieu of pathological confirmation of disease, are additional prognostic considerations.

Classically, pure germinomas are associated with normal serum and cerebrospinal fluid alpha-fetoprotein and low levels of beta-human chorionic gonadotropin.

Germinomas that secrete beta-human chorionic gonadotropin and/or alpha-fetoprotein may contain nongerminomatous elements, even among patients with histological confirmation of a pure germinoma.

An emerging challenge is setting uniform diagnostic criteria based on biomarkers alone, given highly variable levels in cerebrospinal fluid and serum of patients with pathologically confirmed pure germinomas.

Contemporary European and North American trials have used beta-human chorionic gonadotropin cutoff values of 50 IU/L to 100 IU/L for diagnosis and initiation of chemotherapy, yet a research group in Japan found no overall difference in 5-year survival or recurrence rates among a pediatric population when serum and cerebrospinal fluid beta-human chorionic gonadotropin were less than 200 IU/L.

Another study recommended new diagnostic beta-human chorionic gonadotropin cutoff points of more than 8.2 IU/L in the cerebrospinal fluid and more than 2.5 IU/L in the serum confirmed by pathology and cerebrospinal fluid cytology, of which our patient exceeded.

With the emerging role of serum biomarkers in the diagnosis of intracranial germ cell tumors, it is possible to identify mixed malignant germinomas without the need for surgical biopsy. Accordingly, all patients suspected to have an intracranial germ cell tumor should have serum and cerebrospinal fluid tumor marker assays performed.

Tumor location — specifically involvement of off-midline structures and the presence of bifocal disease — may be useful prognostic indicators.

Evidence-based and uniform clinical diagnostic criteria using serum and cerebrospinal fluid beta-human chorionic gonadotropin and alpha-fetoprotein are needed before neuroimaging and tumor assays alone can guide treatment decisions.

References:

Allen JC, et al. Cancer. 1994;74(3):940-944.

Cho J, et al. Radiother Oncol. 2009;doi:10.1016/j.radonc.2008.10.012.

Hu M, et al. Eur J Med Res. 2016;doi:10.1186/s40001-016-0204-2.

Khatua S, et al. Pediatr Blood Cancer. 2010;doi:10.1002/pbc.22468.

Kim JY and Park J. J Korean Neurosurg Soc. 2015;doi:10.3340/jkns.2015.57.5.315.

Luo Z, et al. World Neurosurg. 2017;doi:10.1016/j.wneu.2016.12.004.

Matsutani M, et al. J Neurooncol. 2001;54(3):311-316.

Phi JH, et al. J Neurosurg Pediatr. 2013;doi:10.3171/2012.10.PEDS11487.

Weksberg DC, et al. Int J Radiat Oncol Biol Phys. 2012;doi:10.1016/j.ijrobp.2011.04.033.

For more information:

Munir Ghesani, MD, FACNM, is assistant professor of radiology and director of PET/CT fellowship at NYU Langone Medical Center in New York. He also is a HemOnc Today Editorial Board Member. He can be reached at munir.ghesani@nyumc.org.

Benjamin T. Cooper, MD, is an attending radiation oncologist at NYU Langone Medical Center.

Ana M. Franceschi, MD, is a neuroradiology fellow at NYU Langone Medical Center.

Jonathan Howard, MD, is assistant professor of neurology and psychiatry at NYU Langone Medical Center, and associate director of the neurology service at Bellevue Hospital Center.

Juhi M. Purswani is a medical student at New York University School of Medicine.

Nikhil A. Sahasrabudhe, MD, is a neurosurgeon at Loma Linda University Health.

David Zagzag, MD, is professor of pathology and neurosurgery, as well as chief of the division of neuropathology, at NYU Langone Medical Center.

Disclosures: The authors report no relevant financial disclosures.