| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC
27710.
2 Department of Pediatrics, Duke University Medical Center, Durham, NC
27710.
3 Department of Pathology, Duke University Medical Center, Durham, NC
27710.
Received December 1, 2003;
accepted after revision February 2, 2004.
Presented in part at the 1998 annual meeting of the Radiological Society of
North America, Chicago, IL.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Clinical records of eight children with trilateral retinoblastoma were reviewed for the patient's age at the time of diagnosis of the ocular tumor, time interval from diagnosis of ocular retinoblastoma to discovery of the intracranial tumor, time interval from diagnosis of retinoblastoma to death, and time interval from diagnosis of the intracranial tumor to death. CT or MRI studies were reviewed for the appearance of the primary intracranial neoplasm, intracranial metastases, and spinal metastases.
RESULTS. The mean age of the patients at diagnosis of bilateral retinoblastoma was 4.5 months, and the mean age at diagnosis of the intracranial midline tumor was 26 months. The mean interval from the time of diagnosis of retinoblastoma to discovery of the intracranial tumor was 21.5 months. Two children had spinal leptomeningeal metastases at the time of discovery of the midline intracranial mass although no intracranial metastases were seen on imaging. In the other children, intracranial and spinal leptomeningeal metastases frequently developed within months of the diagnosis of retinoblastoma despite lack of progression in the midline intracranial lesion. Six children died of leptomeningeal spread of tumor. The mean interval from diagnosis of the ocular tumor to death was 46 months and from diagnosis of the intracranial tumor to death was 17 months. One child developed metastatic retinoblastoma in the ulna 10 years after the diagnosis of the intracranial tumor.
CONCLUSION. Children typically died of leptomeningeal tumor dissemination despite lack of progression in the midline intracranial mass. Effective treatment of trilateral retinoblastoma may require close evaluation of these children for leptomeningeal dissemination.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The few radiology articles on trilateral retinoblastoma that have appeared in the literature have typically been individual case reports, and to our knowledge, only one series emphasizing radiologic findings has been published [4]. The purpose of our study was to assess the imaging findings in a series of patients with trilateral retinoblastoma to determine the rates and patterns of intracranial and spinal leptomeningeal tumor spread because these aspects of central nervous system dissemination do not seem to have been documented.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Record Review
Clinical records were reviewed for the patient's age at time of the
diagnosis of the ocular tumor, length of the latent period from diagnosis of
retinoblastoma to discovery of the intracranial tumor, time interval from
diagnosis of retinoblastoma to death, and time interval from diagnosis of the
pineoblastoma to death. Laboratory records were also analyzed for the presence
of tumor cells in cerebrospinal fluid (CSF) acquired via lumbar puncture,
which was performed in all patients.
Image Review
Patients underwent various imaging protocols for the assessment of the
intracranial or intraspinal tumors. Two patients (patients 3 and 4) underwent
both CT of the brain and CT myelography before MRI became available at our
institution. Hence, they did not have spinal imaging with either CT or MRI.
The other six patients were evaluated using both CT and MRI of the brain. All
CT scans of the brain were obtained after an IV infusion of iodinated contrast
material using contiguous 5-mm slices through the posterior fossa and 10-mm
slices through the supratentorial compartment. All CT myelograms were obtained
using 5-mm slices through the entire spine with a 5-mm gap between slices. All
MR images were obtained on a 1.5-T system (Signa, GE Healthcare) using
T1-weighted images (TR/TE range, 800/20-28) with and without contrast material
infusion and T2-weighted images (TR range/TE range, 2,000-2,400/30-80). Four
patients also underwent contrast-enhanced spinal MRI.
The CT scans or MR images were reviewed for the appearance of a primary or metastatic neoplasm in the brain as well as for spinal metastases. The brain was evaluated mainly for mass lesions in the pineal and suprasellar regions and for curvilinear or nodular areas of contrast enhancement suggestive of leptomeningeal tumor dissemination. For metastases to the subarachnoid space of the spinal cord, MR images of the spine and CT myelograms were assessed. Particular attention was paid to four features for which we believe little documentation exists in the literature: the presence of an intracranial tumor at the time that retinoblastoma was diagnosed, patterns of change in intracranial tumor size, presence of spinal leptomeningeal tumor at discovery of the brain tumor, and the length of the apparently intracranial disease-free interval (based on imaging criteria).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
Table 1 summarizes the spinal leptomeningeal tumor spread in six children (confirmed in four children with MRI and in two children with CSF analysis alone). The patients could be classified into four groups. In one group (patients 1 and 2), spinal metastases were evident at the time that the pineoblastoma was diagnosed. In both of these children, intracranial leptomeningeal metastases became evident radiologically within 8 months of diagnosis but were initially not evident. In patient 1, the spinal leptomeningeal tumor temporarily regressed after chemotherapy, as evidenced on MRI, but recurred within 7 months. A second group (patients 3 and 4) did not undergo spinal imaging but were found to have neoplastic cells in the CSF at lumbar puncture 6 months after the intracranial tumor was detected. In children in the third group (patients 5 and 6), spinal imaging revealed spinal metastases approximately 2 years after the pineal tumor was found (Fig. 2A, 2B, 2C). Finally, the children in the fourth group (patients 7 and 8, the only survivors) have not developed a leptomeningeal tumor.
|
|
|
Intracranial Tumor Location and Appearance
The primary intracranial neoplasms were in the pineal region in seven
children and in a suprasellar location in the other patient (patient 4). In
seven patients, the tumor was found on surveillance imaging, and in one
patient (patient 5), the tumor was discovered during imaging evaluation for
increased intracranial pressure. The lesions measured 5-15 mm in diameter at
diagnosis, and when compared with gray matter, they were hypointense or
isointense on unenhanced T1-weighted images and hyperintense on T2-weighted
images. After contrast administration, the enhancement was homogeneous in five
tumors and nonhomogeneous in three tumors. Hydrocephalus was present in three
children when the intracranial tumor was diagnosed and was treated with either
a ventriculoperitoneal shunt (two patients) or fenestration of the third
ventricle (one patient).
Presence of an Intracranial Tumor at Diagnosis of Retinoblastoma
In two children (patients 4 and 7), an intracranial tumor was present when
the retinoblastomas were first recognized
(Table 1). An intracranial
tumor may also have been present in another infant (patient 8) who at 4 months
old was found to have a 1-cm pineal mass on MRI 2 months after the diagnosis
of bilateral retinoblastoma.
Change in the Size of the Intracranial Tumor After Therapy
Within 1 month of detection, three pineoblastomas (in patients 1, 2, and 8)
were surgically resected after the patients had undergone chemotherapy in
addition to bone marrow transplantation or radiation therapy
(Table 2). The pineal tumor in
another child (patient 4) underwent biopsy but was not resected. The remaining
four intracranial tumors (patients 3, 4, 5, and 6) were initially treated
solely with a combination of chemotherapy and radiotherapy. One of the latter
tumors (patient 5) was excised 21 months after a course of chemotherapy and
radiation therapy.
|
Among the three children who were initially treated with surgery, one child (patient 1) was found to have only a small residual tumor on postoperative MR images, and it remained stable during further therapy. In another child (patient 2), MR images obtained immediately after surgery also disclosed a small portion of unexcised pineal tumor that enlarged to 15 mm within 3 months. Over the next 2 years, the diameter of this tumor varied from 8 to 15 mm after chemotherapy and radiation therapy. In the third child (patient 8), the intracranial tumor was totally excised and has not recurred; this is the sole patient with a good clinical outcome. Patients 1 and 2 both had spinal leptomeningeal metastases at the time of the cranial surgery and subsequently manifested intracranial leptomeningeal disease.
One pineal mass (in patient 5) initially treated with chemotherapy and radiation therapy enlarged to 2 cm after 21 months of therapy and was then partially resected. Residual tumor with a 5-mm diameter slowly resolved over 17 months in response to three courses of high-dose cyclophosphamide and craniospinal irradiation. At that point, the child underwent bone marrow transplantation. However, the diameter of the pineoblastoma grew to 6 cm within 10 months, and the child died shortly thereafter.
After chemotherapy, radiation therapy, or both, the size of the four intracranial tumors not initially treated with surgery decreased by 50-90%. The mass in patient 3 initially decreased by approximately 50% after chemotherapy but regained its original dimensions within 3 months. Subsequent radiation therapy resulted in the transformation of the mass into a few radiodense flecks seen on CT, consistent with calcification. However, during the following 3 weeks, evidence of intracranial and spinal leptomeningeal metastatic pineoblastoma appeared.
At the age of 1 month, patient 6, the offspring of a survivor of a unilateral retinoblastoma, was found to have bilateral retinoblastoma; a 1.5-cm-diameter pineal tumor was discovered 7 months later. After chemotherapy and radiation therapy, the diameter of the intracranial lesion decreased to 0.8 cm and became calcified. The child then underwent bone marrow transplantation. No contrast enhancement within the pineal region was detected on CT or MRI until approximately 2 years after diagnosis of the intracranial mass, when extensive intracranial and spinal leptomeningeal metastases developed, leading to the child's death 2 months later (Fig. 2A, 2B, 2C).
A pineal mass found in a third child (patient 7) at the age of 10 months was treated with chemotherapy and whole brain-spine irradiation. The tumor resolved within 18 months, and the child has been free of intracranial and spinal disease for 10 years but developed a small cell malignant neoplasm in the ulna that is histopathologically consistent with metastatic retinoblastoma or a new undifferentiated malignant tumor (Fig. 3A, 3B, 3C, 3D).
|
|
|
|
Patient 4 had an undifferentiated suprasellar tumor composed of small cells with hyperchromatic nuclei and scant cytoplasm (small blue cell tumor) that reacted positively with synaptophysin and neurofilament protein and underwent biopsy. The tumor was not accompanied by other intracranial lesions. The size of this neoplasm diminished by 50% in response to chemotherapy and radiation therapy. However, spinal metastases were detected 6 months after discovery of the intracranial tumor, and 6 months later, the child died.
Time Interval from Discovery of the Intracranial Tumor to Detection of Intracranial Metastases
Intracranial leptomeningeal dissemination was not apparent in any child at
the time that the primary intracranial mass was discovered. However, MRI
subsequently documented such dissemination in four patients (patients 1, 2, 3,
and 6) at a mean interval of 13 months after diagnosis of the intracranial
tumor (Table 1). In two other
children (patients 4 and 5), such spread presumably occurred some time before
the development of intraspinal metastases 6 months after diagnosis of the
intracranial tumor.
Time Interval from Discovery of the Intracranial Tumor to Detection of Intraspinal Metastases
The time interval from discovery of the intracranial mass to detection of
intraspinal metastasis in the six children who had such metastasis ranged from
0-26 months (mean, 10 months). Intraspinal metastases were detected in four of
the children (patients 1, 2, 5, and 6) because of spinal surveillance MRI. In
the other two children (patients 3 and 4), intraspinal leptomeningeal
metastases were detected at a cytologic examination of the CSF obtained at
lumbar puncture. In patient 3, who did not undergo myelography, the neoplastic
cells were evident when the child was nearly 4 years old (47 months). Patient
4 had normal findings on a myelogram at 16 months old when neoplastic cells
were found in the CSF. All six children who developed spinal metastases
died.
Apparent Disease-Free Interval
After receiving therapy, two children (patients 7 and 8), the only
survivors among the eight patients, had periods of 1 year or longer without
evidence of an intracranial tumor. A pineal neoplasm diagnosed when patient 7
was 10 months old resolved after chemotherapy and whole-brain and total-spine
irradiation. Patient 8 was initially treated by surgical resection,
chemotherapy, and bone marrow transplantation and has been followed up
postoperatively for 28 months.
Survival
The diagnosis of pineoblastoma (as opposed to retinoblastoma) was an
ominous sign. Six of the children died within 7-32 months (mean, 17 months;
median, 13 months). The two survivors are patient 7 (> 13 years since the
diagnosis of the bilateral retinoblastomas and an intracranial tumor) and
patient 8 (currently > 40 months old). Survival after discovery of the
spinal metastases in two children (patients 1 and 2) whose spinal metastases
were discovered at the same time as their intracranial mass was longer
(surviving an average of 16 months after discovery of spinal metastases) than
survival after discovery of the spinal metastases in four children (patients
3, 4, 5, and 6) whose intracranial tumor was detected before the spinal
metastases (surviving an average of 3 months after discovery of spinal
metastases). However, length of survival after discovery of the brain tumor
was similar in the two groups: an average of 16 months in the children in whom
an intracranial tumor and spinal metastases were found at the same time and an
average of 19 months in the four children in whom spinal metastases were found
after discovery of the intracranial tumor. If one defines early spinal
leptomeningeal dissemination as occurring within 6 months of discovery of the
primary intracranial tumor, children with early spinal metastases (patients 1,
2, 3, and 4) survived an average of approximately 13 months after discovery of
the brain tumor, and those with later spinal metastases (patients 5 and 6)
survived approximately 28 months after discovery of the brain tumor. Autopsies
were not performed on the children who died.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
25%) is higher than that documented in the
literature, perhaps because our institution is a tertiary care referral center
for retinoblastoma and brain tumors. Recent data suggest that chemotherapy for
bilateral retinoblastoma may decrease the likelihood that an intracranial
tumor will develop [8]. Although the associated intracranial tumor in persons with trilateral retinoblastoma is most frequently located in the pineal region, some tumors occur in the suprasellar or parasellar region [9]. The pineoblastoma rarely develops in the presence of unilateral retinoblastoma. In one series of 440 consecutive patients with retinoblastoma (238 unilateral, 202 bilateral cases), only one patient with a unilateral ocular tumor developed an intracranial neoplasm, compared with 11 children with bilateral retinoblastoma who developed the neoplasm [9]. In previous reports, trilateral retinoblastoma has almost invariably been fatal, and the mean length of survival after discovery of the intracranial tumor of untreated and treated patients has been 1.3 months and 9.7 months, respectively [10].
Tumor Location and Appearance
MR images of the intracranial tumor in our series typically disclosed a
midline mass that was relatively isointense compared with gray matter on
unenhanced T1-weighted images and densely enhanced, usually in a homogeneous
manner, after administration of contrast material. On CT, the intracranial
neoplasm was usually isodense relative to gray matter and had dense contrast
enhancement. Findings on contrast-enhanced T1-weighted images of the spine
typically consisted of small enhancing foci over the surface of the spinal
cord and nerve roots.
Previous studies have provided information about the timing of a few clinical milestones in trilateral retinoblastoma [7, 9, 10]. Because these milestones were usually similar to those in our series, our patients appear to be representative of patients with trilateral retinoblastoma in general. In previous reports, the mean age of patients when the ocular tumor was diagnosed was 7 months [7, 9, 10], which is comparable to that seen in our patients (5.1 months). The mean age at diagnosis of the intracranial neoplasm in our patients (26 months) was also similar to the mean age reported in recently published series (23-24 months) [7, 9]. The mean time from diagnosis of retinoblastoma to discovery of the intracranial tumor in our patients was 21 months, which is identical to that found in a composite of reported patients [11]. However, our patients survived longer than those in previous studies. Even excluding our 13-year survivor, the mean interval from diagnosis of the intracranial tumor to death (17 months) was moderately longer than the 6 months reported in a review of the world literature on trilateral retinoblastoma [12].
Presence of the Intracranial Tumor When Retinoblastoma Was Diagnosed
It is difficult to use the results of previous studies to determine the
frequency with which an intracranial midline tumor and bilateral
retinoblastoma are diagnosed simultaneously, but in our series, this
simultaneous diagnosis was made in two children (patients 4 and 7) and may
have also been made in a third child (patient 8). None of these patients had
overt neurologic manifestations, but subtle abnormalities may have been missed
because of the patients' young ages. The small size and sampling bias of our
tertiary-care university-based medical center population may account for the
high frequency of a coexisting intracranial tumor at the time of discovery of
the retinoblastomas in our patients. Nevertheless, our findings indicate that
the simultaneous diagnosis of bilateral retinoblastomas with an intracranial
tumor is not rare, suggesting that cranial imaging for a midline tumor should
be performed early.
Change in the Intracranial Tumor After Therapy
Compared with the number of reports of the clinical milestones outlined
above, few imaging reports have documented rates of tumor progression in
children with trilateral retinoblastoma. One analysis documented the location
and imaging appearance of the intracranial tumor without commenting on changes
in tumor size over time or tumor dissemination throughout the central nervous
system [4]. In our series, the
size of the intracranial mass diminished to variable degrees after
chemotherapy or radiation therapy in children who did not undergo early
surgical resection, and tumor size correlated poorly with clinical outcome.
For instance, the size of the pineal tumor decreased markedly in patient 3
after therapy, whereas the tumor size remained stable in patients 4 and 6.
Nonetheless, these three children had fatal outcomes with leptomeningeal
dissemination. These findings again emphasize that tumor dissemination, rather
than local recurrence, is the most important prognostic feature. The
intracranial tumor in one of the surviving children (patient 7) regressed over
a 4-year period after treatment. After a total surgical resection, the tumor
in the other survivor (patient 8) did not recur. However, perhaps more
important, neither patient developed leptomeningeal metastases.
Physicians assessing patients with trilateral retinoblastoma need to be aware that tumor progression occurs more often in the subarachnoid space than at the initial site of the intracranial tumor. Furthermore, if these children are to have a reasonable hope for survival, early and repeated evaluation for leptomeningeal metastases using cytological CSF analysis and spinal MRI is needed.
Time Interval from Discovery of the Intracranial Tumor to Detection of Intracranial Metastases
Very little information exists in the previously published medical
literature regarding the timing of the development of intracranial metastases
in patients with trilateral retinoblastoma. Nonetheless, the importance of
this possibility is confirmed by the fact that most patients with trilateral
retinoblastoma are treated with whole-brain irradiation rather than with local
radiotherapy. The children in our series did not show intracranial metastases
at the time that the primary intracranial tumor was discovered. Instead,
intracranial metastases typically developed within 6-12 months after the
discovery of the primary intracranial mass.
Time Interval from Discovery of the Intracranial Tumor to Detection of Intraspinal Metastases
The importance of spinal leptomeningeal dissemination as an indicator of
poor prognosis has been noted in the medical literature
[9]. However, to our knowledge,
the time interval from diagnosis of the intracranial tumor to the development
of spinal metastases has not been described in any published series. In our
small series, spinal metastases were found at the time of diagnosis of the
intracranial tumor in 25% of patients. In our six patients who developed
spinal leptomeningeal metastases, the metastases always developed within
approximately 2 years of discovery of the primary intracranial tumor. In both
children (patients 7 and 8) who did not develop spinal metastases, good local
control of the primary brain tumor was achieved and maintained. However,
spinal metastases developed in some patients in whom only a small amount of
residual tumor was present (e.g., patient 4).
None of our patients with disseminated intracranial and spinal leptomeningeal metastases were long-term survivors. Surprisingly, children in whom discovery of spinal metastases coincided with that of the primary intracranial tumor (patients 1 and 2) had a longer survival time after discovery of spinal metastases than children in whom spinal metastases were found later (patients 3, 4, 5, and 6). However, the length of survival after discovery of the primary intracranial tumor did not differ substantially between the two groups.
Controversy Over the Timing of Brain Screening in Patients with an Ocular Tumor
Some argue that the infrequency of trilateral retinoblastoma does not
warrant routine neuroimaging in patients in whom ocular disease is diagnosed
[13]. However, one recent
meta-analysis of reported retinoblastomas found not only that trilateral
retinoblastoma was detected earlier in patients undergoing routine serial
imaging but also that those patients survived longer than those who did not
undergo such imaging (an average survival of 16 months vs an average of 8
months, respectively) [11].
Another series found a 1-year survival rate of 40% among patients who were
asymptomatic when the intracranial tumor was discovered compared with a rate
of 10% among patients who were symptomatic
[12]. Nonetheless, as other
authors have noted, such data may simply represent lead-time bias rather than
a real prolongation of survival
[9]. In any event, it seems
reasonable to assume that long-term prognosis depends on diagnosing the
intracranial tumor while the patient is asymptomatic. Hence, we recommend that
brain imaging be performed within weeks of diagnosing bilateral retinoblastoma
to allow early treatment because of the ramifications for outcome.
Various protocols for timing of cross-sectional imaging studies for screening of the brain of patients with ocular retinoblastoma have been advocated [4, 7, 9, 13, 14]. Previous investigators have noted that by the time neurologic symptoms from the intracranial tumor arise, the tumor may already be too large for treatment to be effective [9]. This fact has led to the recommendation that routine screening with cross-sectional brain imaging is indicated for patients with bilateral ocular retinoblastoma [15]. At our institution, we routinely screen patients with bilateral retinoblastoma for the presence of a midline intracranial mass but also routinely image patients with such a mass to detect tumor progression. Our protocol has been to routinely follow up patients with bilateral retinoblastoma using CT of the orbits and brain and patients with trilateral retinoblastoma using MRI of the brain and spine. The protocol calls for routine contrast-enhanced CT of the brain and thin-section axial CT of the orbits in patients with bilateral retinoblastoma every 3 months during the first year after diagnosis and at 4-month intervals during the second year. Thereafter, patients are screened every 6 months until they are 5 years old. For patients with trilateral retinoblastoma, MRI of the brain and spine is performed every 2 months during the first year after diagnosis, at 4-month intervals during the second year (assuming no leptomeningeal spread is seen), and thereafter at 6-month intervals. If leptomeningeal tumor spread is found, MRI is continued at 2-month intervals during therapy.
Screening the Spine in Patients with Trilateral Retinoblastoma
The issue of when to perform spinal imaging in patients with trilateral
retinoblastoma has received little attention in the literature. Furthermore,
data regarding rates of development of spinal leptomeningeal metastases are
not available. Intraspinal dissemination can be present when trilateral
retinoblastoma is diagnosed, as our series shows (patients 1 and 2).
Furthermore, spinal metastases can be present in the absence of obvious
intracranial leptomeningeal dissemination. Hence, routine performance of both
cranial and spinal MRI appears reasonable for patients with trilateral
retinoblastoma, even when an intracranial mass is initially diagnosed.
However, multiple studies have shown that MRI findings are frequently normal
in patients with metastases proven by positive results on CSF cytology
analysis [16,
17]. On the basis of the
results of these studies, CSF analysis should remain a routine part of the
assessment of patients with trilateral retinoblastoma.
Our study has several limitations. The study is retrospective, and a routine method of imaging was not performed at standard points during treatment, limiting our ability to definitively determine the timing of tumor progression. In fact, two patients did not undergo spinal imaging and had spinal tumor spread diagnosed at CSF cytology analysis alone. Furthermore, the small sample size limits our ability to make generalizations. Finally, the patients in our study underwent a variety of treatments rather than a consistent single treatment regimen that might have influenced the rate of disease progression.
In summary, in our series of eight patients with trilateral retinoblastoma, the intracranial tumor often responded to chemotherapy, but the tumor spread to distant sites. Discovery of intracranial and spinal leptomeningeal metastases frequently followed within months of treatment despite a lack of progression in the intracranial neoplasm. Furthermore, 25% of children in this small sample had spinal leptomeningeal metastases when the pineoblastoma was discovered and before the detection of intracranial metastases on neuroimaging. Cranial and spinal leptomeningeal metastases were invariably followed by a lethal outcome.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Kadom and R. W. Sze Radiological Reasoning: Leukocoria in a Child Am. J. Roentgenol., September 1, 2008; 191(3_Supplement): S40 - S44. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= ocular bilateral retinoblastoma, 
= intracranial retinoblastoma,
= confirmed intracranial leptomeningeal disease, 
= confirmed spinal
leptomeningeal disease, 
= death, 
= alive.
