s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/survophthal
Diagnostic and surgical technique
The potential of 3T high-resolution magnetic resonance imaging for diagnosis, staging, and follow-up of retinoblastoma Marcus C. de Jong, MD, MSca,*, Pim de Graaf, MD, PhDa, Herve´ J. Brisse, MD, PhDb, Paolo Galluzzi, MDc, Sophia L. Go¨ricke, MDd, Annette C. Moll, MD, PhDe, Francis L. Munier, MDf, Maja Beck Popovic, MDg, Alexandre P. Moulin, MDh, Stefano Binaghi, MDi, Jonas A. Castelijns, MD, PhDa, Philippe Maeder, MDi, On behalf of the European Retinoblastoma Imaging Collaboration (ERIC)1 a
Department of Radiology and Nuclear Medicine, VU University Medical Center, Amsterdam, The Netherlands Department of Radiology, Institut Curie, Paris, France c Unit of Diagnostic and Therapeutic Neuroradiology, Department of Neurosciences, Azienda Ospedaliera Universitaria Senese, Siena, Italy d Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Germany e Department of Ophthalmology, VU University Medical Center, Amsterdam, The Netherlands f Jules-Gonin Eye Hospital, Lausanne, Switzerland g Pediatric Hematology-Oncology Unit, University Hospital, Lausanne, Switzerland h Pathology Unit, Jules-Gonin Eye Hospital, Lausanne, Switzerland i Department of Radiology, University Hospital, Lausanne, Switzerland b
article info
abstract
Article history:
We demonstrate the value of high-resolution magnetic resonance imaging (MRI) in diag-
Received 18 July 2014
nosing, staging, and follow-up of retinoblastoma during eye-saving treatment. We have
Received in revised form
included informative retinoblastoma cases scanned on a 3T MRI system from a retro-
19 January 2015
spective retinoblastoma cohort from 2009 through 2013. We show that high-resolution MRI
Accepted 20 January 2015
has the potential to detect small intraocular seeds, hemorrhage, and metastatic risk factors
Financial support: Grant support: M.C.J. is financially supported by a grant from the ODAS Foundation, Delft, The Netherlands; The European Retinoblastoma Imaging Collaboration (ERIC) is financially supported by a grant from the ODAS Foundation, Delft, The Netherlands. The authors’ work was independent of the funding organizations. The funding organizations had no involvement in the design or conduct of this study, data management and analysis, or manuscript preparation and review or authorization for submission. * Corresponding author: Marcus C. de Jong, MD, MSc, Department of Radiology and Nuclear Medicine, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands. E-mail address: [email protected] (M.C. de Jong). 1 The European Retinoblastoma Imaging Collaboration: Jonas A. Castelijns, MD, PhD, BSc (Chair), Pim de Graaf, MD, PhD, and Marcus C. de Jong, MD, MSc (Department of Radiology and Nuclear Medicine, VU University Medical Center, Amsterdam, The Netherlands), Herve´ J. Brisse, MD, PhD (Department of Radiology, Institut Curie, Paris, France), Paolo Galluzzi, MD (Unit of Diagnostic and Therapeutic Neuroradiology, Department of Neurosciences, Azienda Ospedaliera Universitaria Senese, Siena, Italy), Sophia L. Go¨ricke, MD (Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Germany), and Philippe Maeder, MD (Department of Radiology, Centre Hospitalier Universitaire Vaudois [CHUV], and University of Lausanne, Lausanne, Switzerland). 0039-6257/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.survophthal.2015.01.002
2
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Neelakshi Bhagat and David Chu,
not visible with fundoscopy (e.g., optic nerve invasion and choroidal invasion), and treat-
Editors
ment response. Unfortunately, however, the diagnostic accuracy of high-resolution MRI is not perfect, especially for subtle intraocular seeds or minimal postlaminar optic nerve
Keywords:
invasion. The most important application of MRI is the detection of metastatic risk factors,
magnetic resonance imaging
as these cannot be found by fundoscopy and ultrasound. ª 2015 Elsevier Inc. All rights reserved.
retinoblastoma imaging high resolution diagnostic accuracy staging treatment
1.
Introduction
Retinoblastoma is a malignant tumor of the retina that represents approximately 3% of all pediatric malignancies. The incidence is one in 17,000, typically presenting in the first five years of life.31 In most cases, retinoblastoma can be distinguished accurately from other ocular lesions with fundoscopy and ultrasound (US). When there is still doubt about the diagnosis (intratumoral calcifications set retinoblastoma apart from other lesions) and/or in case of an unclear ocular medium (e.g., massive vitreous hemorrhage), magnetic resonance imaging (MRI) can aid in diagnosing, staging, and follow-up of intraocular retinoblastoma. In staging, MRI plays an important role, as fundoscopy and US cannot assess tumor extent beyond a certain point. Retinoblastoma can spread beyond the eye in different ways. The tumor can invade the highly vascularized choroid, giving an increased risk of hematogenous spread. As the tumor continues to grow outwards, it can invade the sclera, carrying the risk of local recurrences. The tumor can also extend into the optic nerve through the lamina cribrosa, which may result in metastases in the brain and cerebrospinal fluid. Whether involvement of the anterior segment is also a risk factor for metastasis is still subject to discussion.5,22 Most commonly, the International Classification of Retinoblastoma is used to classify the metastatic risk of intraocular retinoblastoma.25 For extraocular extension of retinoblastoma, the International Retinoblastoma Staging System has been developed.11,41 Increasing use of eye-saving treatment strategies emphasize the importance of non-invasive diagnostic tools because often decisions cannot be confirmed by histopathologic analysis. The diagnostic accuracy of MRI in staging retinoblastoma tumor extent largely depends on image quality and resolution. False positives and false negatives are often borderline cases; for example, optic nerve invasion posterior to the lamina cribrosa sclerae could be falsely diagnosed if the tumor has invaded the optic nerve pushing the lamina outwards, but has not actually passed itdthat can give the impression of postlaminar invasion on MRI.8 The diagnostic accuracy of MRI will continue to increase as technology advances. Most modern hospitals have at least a 1.5T MRI system. Such a system with a dedicated surface coil, or a 3T system with a multi-channel head coil, can produce highresolution images of the eyes. We demonstrate the value of
high-resolution MRI in diagnosing, staging, and follow-up of retinoblastoma by showing high-resolution MRI scans of a wide spectrum of the disease.
2.
Methods
We evaluated MRI scans from 56 retinoblastoma patients between 2009 and 2013 who were scanned 119 times in total, ranging from 1 to 10 scans per person. The median age at which the first MR imaging was performed was 25 months, ranging from 1e130 months. Forty percent (22 of 55) of the patients had unilateral disease. About a third (18 of 56) of the retinoblastoma patients underwent enucleation. All were scanned on a 3T MRI machine (Verio; Siemens, Erlangen, Germany) combined with a 32-channel receiver coil (running Syngo VB 15-17 software; Erlangen, Germany) at the University Hospital of Lausanne. During the positioning of patients, the distance between the eyes and the coil was minimized by lifting the patient. All were scanned under general anesthesia. As contrast agent we used an intravenous injection of 0.1 mmol/kg of gadoteric acid (Dotarem; Guerbet, Villepinte, France). Details about the used high-resolution sequences for ocular and orbital imaging can be found in Appendix A. Additionally, routine brain imaging was performed in all patients. The institutional review board approved the study with waiver of informed consent.
2.1.
Image evaluation
During two consensus meetings in September 2012 and June 2013 of the European Retinoblastoma Imaging Collaboration (ERIC), five experienced radiologists (P.M., P.d.G., H.J.B., P.G. and S.G., all with more than 10 years’ experience in ocular MR imaging) made a selection from the available MRI scans to give an extensive overview of the retinoblastoma disease spectrum. We included detailed examples of various disease stages. In all patients, the diagnosis of retinoblastoma was confirmed with fundoscopy and US. Two radiologists (P.G. and S.G.) assessed the quality of the images (axial T1 and T2) independently on a scale from 1 (poor), 2 (average), 3 (good), to 4 (excellent). During the first ERIC consensus meeting, images dating from January 2011 until September 2012 (73 out of 119 MRI scans) were included in the quality scoring. Older images were excluded, because image quality improved over time.
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
3
Fig. 1 e Axial T2-weighted images showing a small endophytic tumor (A) that can freely grow into the vitreous unconstrained by the retina, an exophytic tumor (B) constrained by the retina giving the tumor a more lens-like shape, and a combined endophitic (arrow on C and exophytic tumor (arrow on C). Calcification shows up as dark areas in this exophytic tumor (arrow on D; this image also shows extensive retinal detachment with subretinal seeding (arrow on D). A buphthalmic right eye (E) compared to a normal sized left eye (F) in the same patient. Images G and H depict the shrinkage of an eye before and 1 year after conservative eye-salvage treatment. A tumor with cavitary changes presents as clearly delineated hyperintense areas within the tumor (arrows on I).
2.2.
Image quality
There were some differences in the quality scores between the two readers, but this was mostly due to differences between readers when in terms of scoring 3 (good) or 4 (excellent). Appendix B shows that most of the images had at least good quality.
3.
Intraocular retinoblastoma
MRI has limited additional benefit to fundoscopy and US in assessing the appearance, diagnosis, and staging of intraocular retinoblastoma. In case of doubt, however, MRI may play an important additional role.
4
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 2 e T2-weighted images showing examples of tumor seeds (arrows on A and B). In eyes with retinal detachment, the subretinal fluid shows a high signal intensity on these T2-weighted images (C and E) and a fat-suppressed T1-weighted image (D), indicating a high protein content. Intraocular hemorrhages are identifiable as distinct fluidefluid levels on these T2-weighted images: a subretinal hemorrhage (arrow on F), intravitreal hemorrhage (arrow on G), and 10 months later in the same patient a subretinal hemorrhage (arrow on H).
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
3.1.
Growth pattern
In terms of growth pattern, retinoblastoma can be divided into four groups: endophytic tumors, exophytic tumors, a combination of endo- and exophytic growth patterns, and diffuse infiltrating tumors.21 Endophytic tumors develop in the intraretinal layers and have broken through the inner limiting membrane and hyaloid into the vitreous (Fig. 1A). Eventually, vitreous seeding may develop when little chunks of cells break off the tumor mass and float in the vitreous. Exophytic tumors develop in the intraretinal layer and grow into the subretinal space and may result in subretinal fluid and retinal detachment (Fig. 1B). Retinoblastoma with an exophytic growth pattern has an increased probability of growing into the choroid.24 When chunks of cells break free in the subretinal space, we speak of subretinal seeding. Often, retinoblastoma presents as a combination of an exophytic and endophytic growth pattern (Fig. 1C). Rarely, retinoblastoma can also present as diffusely infiltrating patches in the retina, without a clear tumor mass, that often lack calcifications and present in older children.9
3.2.
Tumor presentation
Typically, retinoblastoma presents as an area with moderately high signal intensity on T1-weighted images (T1WI) and low signal intensity on T2-weighted images (T2WI) compared with the vitreous body.14 On post-gadolinium T1WI, retinoblastoma usually shows a heterogeneous enhancement pattern (visible in a figure that is referenced later in this article for contrast enhancement of the iris: Fig. 3A]). Calcifications are a characteristic of the tumor, setting retinoblastoma apart from other intraocular lesions like Coats disease, persistent fetal vasculature, and medulloepithelioma.10 MRI has not been considered sensitive for the detection of calcifications, although two studies show that MRI has good diagnostic accuracy with a sensitivity of 92e93% and a specificity of 89e100%.20,24 Calcifications show up as areas of low intensity on T2WI (Fig. 1D). As the tumor grows, the risk of tumor extension increases, and as a result more aggressive and invasive treatment options will be necessary. Brisse et al7 found a significant difference in maximum tumor diameter on MRI between eyes with (17.6 mm) and without (15.3 mm) histopathologically proven postlaminar optic nerve invasion (P ¼ 0.001). De Graaf et al15 demonstrated that, on average, retinoblastoma eyes were slightly smaller than normal eyes and that tumor volume was negatively correlated with eye volume. When diagnosed in a late stage, retinoblastoma can present with increased intraocular pressure and buphthalmia.1 Figure 1E and 1F show an example of a normal-sized left eye and an enlarged right eye with globe deformation, a shallow anterior chamber, and a tumor-induced impairment of the aqueous humor drainage. In a late stage or (more commonly) after extensive conservative treatment, the affected eye can start to shrink, leading to a nonfunctional shrunken eye (phthisis bulbi; Fig. 1G and 1H). On fundoscopy, retinoblastoma sometimes presents with cavitary spaces in the tumor (Fig. 1I). There is some evidence that cavitary retinoblastoma is an indication of a more
5
differentiated tumor that responds less well to chemotherapy, but long-term outcomes do not seem to be worse.29,38
3.3. Vitreous and subretinal seeding, and retinal detachment Intra-arterial chemotherapy may be effective for the treatment of subretinal seeding, and with the introduction of intravitreal injection of chemotherapy, vitreous seeds can also be treated, avoiding external beam radiotherapy.2,33 Tumor seeding can often easily be seen at fundoscopy and was usually missed by MRI. Several studies showed that the diagnostic accuracy of MRI in detecting vitreous seeding compared with histopathologic analysis was not much better than a coin flip.4,23,42 Histopathologic analysis is not, however, an ideal method to detect seeding. Sometimes part of the vitreous is lost during histopathological preparation and harvesting of fresh tumor tissue for DNA mutation analysis. As image quality of MRI continuously improves over time, these numbers underestimate the current diagnostic accuracy for the detection of seeding. To be able to detect small subretinal or vitreous tumor seeds, the in-plane resolution of the images needs to be high enough and the slice thickness small enough. Figure 2A shows an example of a small vitreous seeds directly posterior to an artificial lens. Figure 2B shows a small seed in the anterior chamber. Fundoscopy confirmed the presence of tumor seeding in these two examples. In case seeding cannot be seen with fundoscopy (e.g., an unclear ocular medium due to massive hemorrhage), high-resolution MRI could be a useful addition. Optical coherence tomography has also shown promising results with regard to imaging of tumor seeding.39 Exophytic tumors combined with subretinal fluid accumulation can cause the retina to detach. Seeding can then develop in the subretinal space (Fig. 1D). Three studies have reported sensitivities of 50e89% and specificities of 88e100% on the diagnostic accuracy of MRI at detecting retinal detachment compared to clinical findings or histopathologic analysis.4,13,23 Retinal detachment shows up as a distinct hypointense line on T2WI and can be seen in Figure 2Ce2E. Detachment of the retina may be focally around the tumor, but can eventually evolve into a total retinal detachment, with a typical V-shaped appearance.
3.4.
Hemorrhage
Besides intratumoral hemorrhage that can cause falsepositive diagnoses of tumor calcification, hemorrhage also may occur in the vitreous, subretinal space, or the anterior chamber.20 Hemorrhage in any of these spaces is a predictor of a worse outcome for eye salvage. Intraocular hemorrhage can be identified on MRI scans as fluidefluid levels.14 Figure 2F shows subretinal hemorrhage in an eye with total retinal detachment. Figure 2G and 2H show an example of an eye initially with vitreous hemorrhage, but with subretinal hemorrhage after 10 months.
3.5.
Anterior eye chamber involvement
Tumor invasion into the anterior segment is considered by some to be a risk factor for extraocular relapse, but this is still
6
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 3 e A contrast-enhanced fat-suppressed axial T1-weighted image (A) shows a heterogeneously enhancing tumor and enhancement of the iris (arrows), indicative of neoangiogenesis in the iris. A dislocated lens is depicted on a T2-weighted image (arrow in B). This contrast-enhanced axial fat-suppressed T1-weighted image (C) shows focal enhancement in the distal optic nerve (arrow) suggestive of postlaminar optic nerve invasion and a sagittal subtraction image (E) of a
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
debated.5 Besides tumor invasion, other tumor-related changes in the anterior segment can be depicted on MRI. An increase in contrast enhancement on T1WI has been correlated with rubeosis iridis with a sensitivity ranging from 68e93% and a specificity ranging from 43e82%.18 Figure 3A shows an example of clinically and histopathologically confirmed rubeosis iridis. Rarely, advanced intraocular retinoblastoma can cause a dislocated lens, which usually occurs in the presence of massive tumor necrosis (Fig. 3B). Vitreous tumor seeding can eventually also invade the anterior segment, and from there the seeds can end up in the trabecular meshwork of the anterior chamber angle. The diagnostic accuracy of MRI compared with histopathologic analysis at detecting ciliary body invasion has a sensitivity ranging from 71e100% and a specificity ranging from 65e100%.18 Recently, the use of US biomicroscopy has demonstrated a sensitivity of 81% and specificity of 100% in the assessment of anterior segment invasion compared with histopathology.32
4.
Metastatic risk factors
Cross-sectional imaging, like MRI, is essential for the evaluation of tumor invasion into the ocular wall or optic nerve, as fundoscopy and US are of little value.40
4.1.
Optic nerve invasion
Patients with postlaminar optic nerve tumor invasion without adjuvant systemic chemotherapy are at great risk of regional extraocular recurrences or distant metastases.12 Optic nerve invasion can be seen as contrast enhancement on T1WI (Fig. 3C and 3E). Histopathologic analysis shows postlaminar optic nerve invasion (2.2 mm) in the first example (Fig. 3D). Histopathologically, the second example also proves to be a postlaminar optic nerve invasion (Fig. 3F), and depending on the assessment of the dubious postlaminar enhancement pattern in the second MRI scan (Fig. 3E), this case could either be classified as a false negative or a true positive. Appendix C shows an example of extensive postlaminar optic nerve invasion and the effect of subsequent chemotherapy (four cycles of etoposide and carboplatin) over the course of 3.5 months. Ten months prior to this, this patient was diagnosed with unilateral retinoblastoma (left eye was classified as group E) in another institution for which the patient had been treated with intra-arterial melphalan. In a recent meta-analysis with histopathology as the gold standard, De Jong et al18 showed that MRI could detect postlaminar optic nerve invasion with a sensitivity of 59% (95% confidence interval [CI]: 37e78%) and a specificity of 94% (95%
7
CI: 84e98%). Meta-regression showed no difference between low- and high-quality MRI scans. After exclusion of one potential outlier, the sensitivity of high-quality MRI rose from 58% (95% CI: 27e84%) to 76% (95% CI: 45e93%). This was not, however, statistically different from a sensitivity of 59% (95% CI: 32e82%) for low-quality MRI (P ¼ 0.26). This raises the question whether the resolution of the images is the only problem of test sensitivity, or it may be that depiction of tumor extension by gadolinium contrast enhancement is not an ideal marker for optic nerve invasiondthat is, tumor invasion does not always show contrast enhancement. Furthermore, contrast enhancement also is not 100% specific for tumor invasion, with reactive inflammation being the main cause of false-positive findings.16 On the other handdemphasizing the importance of highresolution imagingdBrisse et al8 showed that false-positive and false-negative findings were often borderline cases. Before the tumor invades the optic nerve beyond the lamina cribrosa, it can cause bulging of the laminadresulting in contrast enhancement on MRI to reach beyond where the lamina cribrosa is expected to be, producing a false-positive. A false-negative result can be caused by minimal postlaminar invasion (only a few cells on histopathologic analysis; Fig. 3F), which is still without the tumor neovascularization that causes the typical contrast enhancement visible on MRI. The only alternative to MRI for the detection of optic nerve invasion is computed tomography (CT), but this appears to have a low diagnostic value.18
4.2.
Choroidal and scleral invasion
Massive choroidal and scleral invasion are seen as risk factors for local retinoblastoma recurrence and distant metastases. According to Sastre et al,41 choroidal invasion of more than 3 mm or reaching the sclera should be classified as massive choroidal invasion. On T1-weighted contrast-enhanced MRI, choroidal invasion can be identified as a focal inhomogeneous enhancement and/or choroidal thickening. Figure 3G and 3H show an example of irregular thickening and abnormal enhancement of the choroid on MRI that correspond with massive choroidal invasion on histopathology (Fig. 3I). In case of scleral invasion, the contrast-enhanced area extends beyond the choroid into the normally unenhanced sclera. De Jong et al18 found a sensitivity of 74% (95% CI: 52e88%) and a specificity of 72% (95% CI: 31e94%) for choroidal invasion, and a sensitivity of 88% (95% CI: 20e100%) and a specificity of 99% (95% CI: 86e100%) for scleral invasion. Sensitivity and specificity were higher for higher-quality MRI, but these differences were not statistically significant. These numbers are in line
=T1-weighted image with and without contrast enhancement shows dubious contrast enhancement of the distal optic nerve; the arrow indicates an area of enhancement that matches bulging of the lamina cribrosa on histologic examination and the arrowhead indicates dubious postlaminar optic nerve enhancement; histopathologically both patients had postlaminar invasion indicated by hyperchromatic cells posterior to the lamina cribrosa (D and F: arrowheads show the prelaminar tumormass and the arrows point at postlaminar tumor cells). A coronal T2-weighted image (G) and a coronal reconstruction of axial VIBE (volumetric interpolated brain examination) images (H) show multifocal irregular thickening and enhancement of the choroid suspicious of choroidal invasion; massive choroidal invasion in the nasal part was confirmed (arrow on I, corresponding with arrows on G and H).
8
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 4 e This axial VIBE (volumetric interpolated brain examination) image (A) shows a broad area of reduced enhancement of the choroid (arrows) corresponding with temporal choroidal ischemia clearly visible on fundoscopy (B) and fluorescent angiography (C).
with expectations, as choroidal invasion is visible as subtle changes on MRI, whereas scleral invasion is more manifest. No studies have evaluated the diagnostic accuracy of CT for the detection of choroidal invasion, and only one study looked at scleral invasion, showing a reasonable sensitivity (71%; 95% CI: 42e92%) and specificity (98%, 95% CI: 93e100%) for the ability of CT to diagnose scleral invasion.
4.3.
Extraocular spread
When the tumor passes though the ocular wall or into the postlaminar optic nerve, it can extend directly into the orbit or spread via the leptomeninges and cerebrospinal fluid. Appendix D shows an example of extensive invasion of the optic nerve that, because the parents refused any treatment, developed into a large intracranial mass within 4 months. As these lesions are large and therefore usually clearly visible on MR images, diagnostic accuracy of MRI of extrascleral growth in retinoblastoma patients is expected to be high.35 There only is one study (with a small sample size) presenting a sensitivity of 100% for MRI and one study showing a sensitivity of 100% for CT. According to three studies, specificity of MRI ranged from 94e100%.18
information to fundoscopy and US. As mentioned in the Introduction, there are a number of conservative treatment options for retinoblastoma. External beam radiotherapy is falling out of favor because it causes orbital bone growth restrictions and cosmetic problems on the face. Furthermore, it increases the risk of second primary tumors within the radiation field later in life in hereditary retinoblastoma patients.28 Depending on tumor size, location, and shape, local treatment (laser therapy, cryotherapy, and radioactive plaque therapy); intra-arterial, intravitreal, and peribulbar chemotherapy; and systemic chemotherapy are currently the conservative treatment options of choice.3,26,27,44 Appendix E shows a number of post-treatment findings on MRI. Appendix F shows the tumor response to chemotherapy over the course of a year. Choroidal ischemia is a known complication of intra-arterial chemotherapy, which can be seen on MRI as the lack of choroidal contrast enhancement. Figure 4Ae4C show the fundoscopic and fluorescent angiography and MRI findings of a patient with treatment-induced choroidal ischemia. Over the course of 6 months, this patient underwent multiple melphalan and topotecan injections. Perfusion and diffusion MRIdstill under development as an application for retinoblastomadhave shown promising initial results in terms of follow-up for the differentiation between viable and necrotic tumor tissue.17,37
5. Post-treatment status and treatment response
6.
Post-treatment follow-up with MRI is a rapidly developing discipline and shows promising results, providing additional
There are different approaches to generate high-quality MRI scans of eyes: with a 1.5T system43 and a surface coil or, as in
Discussion
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
this study, with a 3T system and a 32-channel head coil. The advantage of 1.5T imaging is that these systems are more readily available. On the other hand, positioning of a head coil is less cumbersome than positioning a surface coil. Head coils also generate better images of deeper structures, including the entire brain. At the time of retinoblastoma diagnosis, MRI of the brain is recommended to exclude trilateral retinoblastoma (ocular retinoblastoma combined with an intracranial midline primitive neuroectodermal tumor).19 The cases included in this article were specifically selected to show certain aspects of retinoblastoma. This approach carries the risk of introducing selection bias and potentially overestimating the diagnostic value of MRI (e.g., imaging of vitreal and subretinal seeds remains difficult, even though we were able to show successful examples). Also, the included patients are from a highly specialized center; therefore, most of these patients have more advanced retinoblastoma. Before the introduction of intra-arterial chemotherapy, patients with retinoblastoma and the presence of metastatic risk factors such as massive choroidal invasion or postlaminar optic nerve invasion would have undergone enucleation with systemic chemotherapy after histopathologic proof of the risk factor, or have received systemic chemotherapy in the form of chemoreduction followed by local treatment. Systemic chemotherapy, then, might have successfully prevented systemic metastasis. Nowadays, systemic chemotherapy, and its potential negative long-term effects, can be avoided in certain cases by using intra-arterial or intravitreal chemotherapy, although this does require a reasonable certainty that there are no metastatic risk factors, adding to the importance of reliable non-invasive imaging like MRI.34 To give an idea and provide a reference of the diagnostic accuracy of conventional MRI, we have reported diagnostic accuracy estimates when available. Nevertheless, with these limitations in mind, MRI becomes increasingly important with increasing eye-saving treatment, and we believe that highresolution MRI can contribute to better clinical decision-making. CT may also be used for the diagnosis, staging, and followup of retinoblastoma, but even though evidence for the use CT is scarce, available evidence does suggest that MRI is superior to CT for almost all previously described applications, most importantly the metastatic risk factors.18 Even for the calcifications, CT has no added value as US in combination with MRI accurately detects calcifications.20,36 Most importantly though, CT should be avoided because of the potential harmful effect of radiation in small children, especially those with hereditary retinoblastoma.6,28,30 A disadvantage of MRI is that it might not be accessible in developing countries; therefore, we believe that only when MRI is not available should CT be considered as an alternative method for imaging.
7.
Conclusion
We have provided an overview of the clinical application of retinoblastoma imaging with MRI. MRI in its current technological state can be of great help in retinoblastoma patient care. Because more and more eye-saving treatment options have entered clinical practice, other indications for MRI in
9
retinoblastoma patients have emerged, such as imaging during conservative treatmentdfor example, for the detection of tumor response, tumor recurrence, and adverse effects of treatment. Unfortunately, MRI does not have perfect diagnostic accuracy for the detection of metastatic risk factors like massive choroidal and postlaminar optic nerve invasion, but it currently is the best available tool, and its performance is constantly improving.
8.
Method of literature search
We searched the literature in PubMed (Medline) for English, Dutch, or German full-text articles published through 2014. Alternatively found articles (e.g., through checking reference lists of retrieved articles, including from De Jong et al18) were also considered. No articles were translated and we discarded articles in any other language even if they were accompanied by an English abstract. The search strategy included MeSH terms and Text Words to retrieve articles on retinoblastoma, optic nerve invasion, choroidal invasion, scleral invasion, metastasis, magnetic resonance imaging, computed tomography, diagnosis, staging, and follow-up.
9.
Disclosures
No conflicting relationship exists for any author.
Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.survophthal.2015.01.002.
references
1. Abramson DH, Frank CM, Susman M, et al. Presenting signs of retinoblastoma. J Pediatr. 1998;132(3 Pt 1):505e8 2. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma in eyes with vitreous and/ or subretinal seeding: 2-year results. Br J Ophthalmol. 2012;96:499e502 3. Aerts I, Lumbroso-Le Rouic L, Gauthier-Villars M, et al. Retinoblastoma. Orphanet J Rare Dis. 2006;1:31 4. Barkhof F, Smeets M, van der Valk P, et al. MR imaging in retinoblastoma. Eur Radiol. 1997;7:726e31 5. Baroni LV, Sampor C, Fandin˜o A, et al. Anterior segment invasion in retinoblastoma: is it a risk factor for extraocular relapse? J Pediatr Hematol Oncol. 2014;36:e509e12 6. Brenner D, Elliston C, Hall E, et al. Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol. 2001;176:289e96 7. Brisse HJ, Graaf P, de, Galluzzi P, et al. Assessment of earlystage optic nerve invasion in retinoblastoma using highresolution 1.5 Tesla MRI with surface coils: a multicentre, prospective accuracy study with histopathological correlation. Eur Radiol 2014; http://dx.doi.org/10.1007/s00330-014-3514-1 8. Brisse HJ, Guesmi M, Aerts I, et al. Relevance of CT and MRI in retinoblastoma for the diagnosis of postlaminar invasion with normal-size optic nerve: a retrospective study of 150
10
9.
10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
patients with histological comparison. Pediatr Radiol. 2007;37:649e56 Brisse HJ, Lumbroso L, Fre´neaux PC, et al. Sonographic, CT, and MR imaging findings in diffuse infiltrative retinoblastoma: report of two cases with histologic comparison. AJNR Am J Neuroradiol. 2001;22:499e504 Bullock JD, Campbell RJ, Waller RR. Calcification in retinoblastoma. Invest Ophthalmol Vis Sci. 1977;16:252e5 Chantada G, Doz F, Antoneli CBG, et al. A proposal for an international retinoblastoma staging system. Pediatr Blood Cancer. 2006;47:801e5 Chantada GL, Casco F, Fandin˜o AC, et al. Outcome of patients with retinoblastoma and postlaminar optic nerve invasion. Ophthalmology. 2007;114:2083e9 De Graaf P, Barkhof F, Moll AC, et al. Retinoblastoma: MR imaging parameters in detection of tumor extent. Radiology. 2005;235:197e207 De Graaf P, Go¨ricke S, Rodjan F, et al. Guidelines for imaging retinoblastoma: imaging principles and MRI standardization. Pediatr Radiol. 2012;42:2e14 De Graaf P, Knol DL, Moll AC, et al. Eye size in retinoblastoma: MR imaging measurements in normal and affected eyes. Radiology. 2007;244:273e80 De Graaf P, Moll AC, Imhof SM, et al. Retinoblastoma and optic nerve enhancement on MRI: not always extraocular tumour extension. Br J Ophthalmol. 2006;90:800e1 De Graaf P, Pouwels PJW, Rodjan F, et al. Single-shot turbo spin-echo diffusion-weighted imaging for retinoblastoma: initial experience. AJNR Am J Neuroradiol. 2012;33:110e8 De Jong MC, de Graaf P, Noij DP, et al. Diagnostic performance of magnetic resonance imaging and computed tomography for advanced retinoblastoma: a systematic review and metaanalysis. Ophthalmology. 2014;121:1109e18 De Jong MC, Kors WA, de Graaf P, et al. Trilateral retinoblastoma: a systematic review and meta-analysis. Lancet Oncol. 2014;15:1157e67 Galluzzi P, Hadjistilianou T, Cerase A, et al. Is CT still useful in the study protocol of retinoblastoma? AJNR Am J Neuroradiol. 2009;30:1760e5 Houston SK, Murray TG, Wolfe SQ, et al. Current update on retinoblastoma. Int Ophthalmol Clin. 2011;51:77e91 Khelfaoui F, Validire P, Auperin A, et al. Histopathologic risk factors in retinoblastoma: a retrospective study of 172 patients treated in a single institution. Cancer. 1996;77:1206e13 Khurana A, Eisenhut CA, Wan W, et al. Comparison of the diagnostic value of MR imaging and ophthalmoscopy for the staging of retinoblastoma. Eur Radiol. 2012;23:1e10 Lemke A-J, Kazi I, Mergner U, et al. RetinoblastomadMR appearance using a surface coil in comparison with histopathological results. Eur Radiol. 2007;17:49e60 Linn Murphree A. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin North Am. 2005;18:41e53 Lumbroso L, Doz F, Urbieta M, et al. Chemothermotherapy in the management of retinoblastoma. Ophthalmology. 2002;109:1130e6 Lumbroso-Le Rouic L, Aerts I, Le´vy-Gabriel C, et al. Conservative treatments of intraocular retinoblastoma. Ophthalmology. 2008;115:1405e10, 1410.e1e2. Marees T, Moll AC, Imhof SM, et al. Risk of second malignancies in survivors of retinoblastoma: more than 40 years of follow-up. J Natl Cancer Inst. 2008;100:1771e9
29. Mashayekhi A, Shields CL, Eagle RC, et al. Cavitary changes in retinoblastoma: relationship to chemoresistance. Ophthalmology. 2005;112:1145e50 30. Mathews JD, Forsythe AV, Brady Z, et al. Cancer risk in 680 000 people exposed to computed tomography scans in childhood or adolescence: data linkage study of 11 million Australians. BMJ. 2013;346:f2360 31. Moll AC, Kuik DJ, Bouter LM, et al. Incidence and survival of retinoblastoma in The Netherlands: a register-based study 1862e1995. Br J Ophthalmol. 1997;81:559e62 32. Moulin AP, Gaillard M-C, Balmer A, et al. Ultrasound biomicroscopy evaluation of anterior extension in retinoblastoma: a clinicopathological study. Br J Ophthalmol. 2012;96:337e40 33. Munier FL, Gaillard M-C, Balmer A, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol. 2012;96:1078e83 34. Ong SJ, Chao A-N, Wong H-F, et al. Selective ophthalmic arterial injection of melphalan for intraocular retinoblastoma: a 4-year review. Jpn J Ophthalmol 2014; http://dx.doi.org/10.1007/s10384-014-0356-y 35. Radhakrishnan V, Sharma S, Vishnubhatla S, et al. MRI findings at baseline and after neoadjuvant chemotherapy in orbital retinoblastoma (IRSS stage III). Br J Ophthalmol. 2013;97:52e8 36. Rodjan F, de Graaf P, van der Valk P, et al. Detection of calcifications in retinoblastoma using gradient-echo MR imaging sequences: comparative study between in vivo MR imaging and ex vivo high-resolution CT. Am J Neuroradiol. 2015;36:355e60 37. Rodjan F, de Graaf P, van der Valk P, et al. Retinoblastoma: value of dynamic contrast-enhanced MR imaging and correlation with tumor angiogenesis. AJNR Am J Neuroradiol. 2012;33:2129e35 38. Rojanaporn D, Kaliki S, Bianciotto CG, et al. Intravenous chemoreduction or intra-arterial chemotherapy for cavitary retinoblastoma: long-term results. Arch Ophthalmol. 2012;130:585e90 39. Rootman DB, Gonzalez E, Mallipatna A, et al. Hand-held highresolution spectral domain optical coherence tomography in retinoblastoma: clinical and morphologic considerations. Br J Ophthalmol. 2013;97:59e65 40. Roth DB, Scott IU, Murray TG, et al. Echography of retinoblastoma: histopathologic correlation and serial evaluation after globe-conserving radiotherapy or chemotherapy. J Pediatr Ophthalmol Strabismus. 2001;38:136e43 41. Sastre X, Chantada GL, Doz F, et al. Proceedings of the consensus meetings from the International Retinoblastoma Staging Working Group on the pathology guidelines for the examination of enucleated eyes and evaluation of prognostic risk factors in retinoblastoma. Arch Pathol Lab Med. 2009;133:1199e202 42. Schueler AO, Hosten N, Bechrakis NE, et al. High resolution magnetic resonance imaging of retinoblastoma. Br J Ophthalmol. 2003;87:330e5 43. Sirin S, Schlamann M, Metz KA, et al. Diagnostic image quality of gadolinium-enhanced T1-weighted MRI with and without fat saturation in children with retinoblastoma. Pediatr Radiol. 2013;43:716e24 44. Smith SJ, Smith BD, Mohney BG. Ocular side effects following intravitreal injection therapy for retinoblastoma: a systematic review. Br J Ophthalmol. 2013;98:292e7
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/survophthal
Diagnostic and surgical technique
The potential of 3T high-resolution magnetic resonance imaging for diagnosis, staging, and follow-up of retinoblastoma Marcus C. de Jong, MD, MSca,*, Pim de Graaf, MD, PhDa, Herve´ J. Brisse, MD, PhDb, Paolo Galluzzi, MDc, Sophia L. Go¨ricke, MDd, Annette C. Moll, MD, PhDe, Francis L. Munier, MDf, Maja Beck Popovic, MDg, Alexandre P. Moulin, MDh, Stefano Binaghi, MDi, Jonas A. Castelijns, MD, PhDa, Philippe Maeder, MDi, On behalf of the European Retinoblastoma Imaging Collaboration (ERIC)1 a
Department of Radiology and Nuclear Medicine, VU University Medical Center, Amsterdam, The Netherlands Department of Radiology, Institut Curie, Paris, France c Unit of Diagnostic and Therapeutic Neuroradiology, Department of Neurosciences, Azienda Ospedaliera Universitaria Senese, Siena, Italy d Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Germany e Department of Ophthalmology, VU University Medical Center, Amsterdam, The Netherlands f Jules-Gonin Eye Hospital, Lausanne, Switzerland g Pediatric Hematology-Oncology Unit, University Hospital, Lausanne, Switzerland h Pathology Unit, Jules-Gonin Eye Hospital, Lausanne, Switzerland i Department of Radiology, University Hospital, Lausanne, Switzerland b
article info
abstract
Article history:
We demonstrate the value of high-resolution magnetic resonance imaging (MRI) in diag-
Received 18 July 2014
nosing, staging, and follow-up of retinoblastoma during eye-saving treatment. We have
Received in revised form
included informative retinoblastoma cases scanned on a 3T MRI system from a retro-
19 January 2015
spective retinoblastoma cohort from 2009 through 2013. We show that high-resolution MRI
Accepted 20 January 2015
has the potential to detect small intraocular seeds, hemorrhage, and metastatic risk factors
Financial support: Grant support: M.C.J. is financially supported by a grant from the ODAS Foundation, Delft, The Netherlands; The European Retinoblastoma Imaging Collaboration (ERIC) is financially supported by a grant from the ODAS Foundation, Delft, The Netherlands. The authors’ work was independent of the funding organizations. The funding organizations had no involvement in the design or conduct of this study, data management and analysis, or manuscript preparation and review or authorization for submission. * Corresponding author: Marcus C. de Jong, MD, MSc, Department of Radiology and Nuclear Medicine, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands. E-mail address: [email protected] (M.C. de Jong). 1 The European Retinoblastoma Imaging Collaboration: Jonas A. Castelijns, MD, PhD, BSc (Chair), Pim de Graaf, MD, PhD, and Marcus C. de Jong, MD, MSc (Department of Radiology and Nuclear Medicine, VU University Medical Center, Amsterdam, The Netherlands), Herve´ J. Brisse, MD, PhD (Department of Radiology, Institut Curie, Paris, France), Paolo Galluzzi, MD (Unit of Diagnostic and Therapeutic Neuroradiology, Department of Neurosciences, Azienda Ospedaliera Universitaria Senese, Siena, Italy), Sophia L. Go¨ricke, MD (Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Germany), and Philippe Maeder, MD (Department of Radiology, Centre Hospitalier Universitaire Vaudois [CHUV], and University of Lausanne, Lausanne, Switzerland). 0039-6257/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.survophthal.2015.01.002
2
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Neelakshi Bhagat and David Chu,
not visible with fundoscopy (e.g., optic nerve invasion and choroidal invasion), and treat-
Editors
ment response. Unfortunately, however, the diagnostic accuracy of high-resolution MRI is not perfect, especially for subtle intraocular seeds or minimal postlaminar optic nerve
Keywords:
invasion. The most important application of MRI is the detection of metastatic risk factors,
magnetic resonance imaging
as these cannot be found by fundoscopy and ultrasound. ª 2015 Elsevier Inc. All rights reserved.
retinoblastoma imaging high resolution diagnostic accuracy staging treatment
1.
Introduction
Retinoblastoma is a malignant tumor of the retina that represents approximately 3% of all pediatric malignancies. The incidence is one in 17,000, typically presenting in the first five years of life.31 In most cases, retinoblastoma can be distinguished accurately from other ocular lesions with fundoscopy and ultrasound (US). When there is still doubt about the diagnosis (intratumoral calcifications set retinoblastoma apart from other lesions) and/or in case of an unclear ocular medium (e.g., massive vitreous hemorrhage), magnetic resonance imaging (MRI) can aid in diagnosing, staging, and follow-up of intraocular retinoblastoma. In staging, MRI plays an important role, as fundoscopy and US cannot assess tumor extent beyond a certain point. Retinoblastoma can spread beyond the eye in different ways. The tumor can invade the highly vascularized choroid, giving an increased risk of hematogenous spread. As the tumor continues to grow outwards, it can invade the sclera, carrying the risk of local recurrences. The tumor can also extend into the optic nerve through the lamina cribrosa, which may result in metastases in the brain and cerebrospinal fluid. Whether involvement of the anterior segment is also a risk factor for metastasis is still subject to discussion.5,22 Most commonly, the International Classification of Retinoblastoma is used to classify the metastatic risk of intraocular retinoblastoma.25 For extraocular extension of retinoblastoma, the International Retinoblastoma Staging System has been developed.11,41 Increasing use of eye-saving treatment strategies emphasize the importance of non-invasive diagnostic tools because often decisions cannot be confirmed by histopathologic analysis. The diagnostic accuracy of MRI in staging retinoblastoma tumor extent largely depends on image quality and resolution. False positives and false negatives are often borderline cases; for example, optic nerve invasion posterior to the lamina cribrosa sclerae could be falsely diagnosed if the tumor has invaded the optic nerve pushing the lamina outwards, but has not actually passed itdthat can give the impression of postlaminar invasion on MRI.8 The diagnostic accuracy of MRI will continue to increase as technology advances. Most modern hospitals have at least a 1.5T MRI system. Such a system with a dedicated surface coil, or a 3T system with a multi-channel head coil, can produce highresolution images of the eyes. We demonstrate the value of
high-resolution MRI in diagnosing, staging, and follow-up of retinoblastoma by showing high-resolution MRI scans of a wide spectrum of the disease.
2.
Methods
We evaluated MRI scans from 56 retinoblastoma patients between 2009 and 2013 who were scanned 119 times in total, ranging from 1 to 10 scans per person. The median age at which the first MR imaging was performed was 25 months, ranging from 1e130 months. Forty percent (22 of 55) of the patients had unilateral disease. About a third (18 of 56) of the retinoblastoma patients underwent enucleation. All were scanned on a 3T MRI machine (Verio; Siemens, Erlangen, Germany) combined with a 32-channel receiver coil (running Syngo VB 15-17 software; Erlangen, Germany) at the University Hospital of Lausanne. During the positioning of patients, the distance between the eyes and the coil was minimized by lifting the patient. All were scanned under general anesthesia. As contrast agent we used an intravenous injection of 0.1 mmol/kg of gadoteric acid (Dotarem; Guerbet, Villepinte, France). Details about the used high-resolution sequences for ocular and orbital imaging can be found in Appendix A. Additionally, routine brain imaging was performed in all patients. The institutional review board approved the study with waiver of informed consent.
2.1.
Image evaluation
During two consensus meetings in September 2012 and June 2013 of the European Retinoblastoma Imaging Collaboration (ERIC), five experienced radiologists (P.M., P.d.G., H.J.B., P.G. and S.G., all with more than 10 years’ experience in ocular MR imaging) made a selection from the available MRI scans to give an extensive overview of the retinoblastoma disease spectrum. We included detailed examples of various disease stages. In all patients, the diagnosis of retinoblastoma was confirmed with fundoscopy and US. Two radiologists (P.G. and S.G.) assessed the quality of the images (axial T1 and T2) independently on a scale from 1 (poor), 2 (average), 3 (good), to 4 (excellent). During the first ERIC consensus meeting, images dating from January 2011 until September 2012 (73 out of 119 MRI scans) were included in the quality scoring. Older images were excluded, because image quality improved over time.
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
3
Fig. 1 e Axial T2-weighted images showing a small endophytic tumor (A) that can freely grow into the vitreous unconstrained by the retina, an exophytic tumor (B) constrained by the retina giving the tumor a more lens-like shape, and a combined endophitic (arrow on C and exophytic tumor (arrow on C). Calcification shows up as dark areas in this exophytic tumor (arrow on D; this image also shows extensive retinal detachment with subretinal seeding (arrow on D). A buphthalmic right eye (E) compared to a normal sized left eye (F) in the same patient. Images G and H depict the shrinkage of an eye before and 1 year after conservative eye-salvage treatment. A tumor with cavitary changes presents as clearly delineated hyperintense areas within the tumor (arrows on I).
2.2.
Image quality
There were some differences in the quality scores between the two readers, but this was mostly due to differences between readers when in terms of scoring 3 (good) or 4 (excellent). Appendix B shows that most of the images had at least good quality.
3.
Intraocular retinoblastoma
MRI has limited additional benefit to fundoscopy and US in assessing the appearance, diagnosis, and staging of intraocular retinoblastoma. In case of doubt, however, MRI may play an important additional role.
4
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 2 e T2-weighted images showing examples of tumor seeds (arrows on A and B). In eyes with retinal detachment, the subretinal fluid shows a high signal intensity on these T2-weighted images (C and E) and a fat-suppressed T1-weighted image (D), indicating a high protein content. Intraocular hemorrhages are identifiable as distinct fluidefluid levels on these T2-weighted images: a subretinal hemorrhage (arrow on F), intravitreal hemorrhage (arrow on G), and 10 months later in the same patient a subretinal hemorrhage (arrow on H).
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
3.1.
Growth pattern
In terms of growth pattern, retinoblastoma can be divided into four groups: endophytic tumors, exophytic tumors, a combination of endo- and exophytic growth patterns, and diffuse infiltrating tumors.21 Endophytic tumors develop in the intraretinal layers and have broken through the inner limiting membrane and hyaloid into the vitreous (Fig. 1A). Eventually, vitreous seeding may develop when little chunks of cells break off the tumor mass and float in the vitreous. Exophytic tumors develop in the intraretinal layer and grow into the subretinal space and may result in subretinal fluid and retinal detachment (Fig. 1B). Retinoblastoma with an exophytic growth pattern has an increased probability of growing into the choroid.24 When chunks of cells break free in the subretinal space, we speak of subretinal seeding. Often, retinoblastoma presents as a combination of an exophytic and endophytic growth pattern (Fig. 1C). Rarely, retinoblastoma can also present as diffusely infiltrating patches in the retina, without a clear tumor mass, that often lack calcifications and present in older children.9
3.2.
Tumor presentation
Typically, retinoblastoma presents as an area with moderately high signal intensity on T1-weighted images (T1WI) and low signal intensity on T2-weighted images (T2WI) compared with the vitreous body.14 On post-gadolinium T1WI, retinoblastoma usually shows a heterogeneous enhancement pattern (visible in a figure that is referenced later in this article for contrast enhancement of the iris: Fig. 3A]). Calcifications are a characteristic of the tumor, setting retinoblastoma apart from other intraocular lesions like Coats disease, persistent fetal vasculature, and medulloepithelioma.10 MRI has not been considered sensitive for the detection of calcifications, although two studies show that MRI has good diagnostic accuracy with a sensitivity of 92e93% and a specificity of 89e100%.20,24 Calcifications show up as areas of low intensity on T2WI (Fig. 1D). As the tumor grows, the risk of tumor extension increases, and as a result more aggressive and invasive treatment options will be necessary. Brisse et al7 found a significant difference in maximum tumor diameter on MRI between eyes with (17.6 mm) and without (15.3 mm) histopathologically proven postlaminar optic nerve invasion (P ¼ 0.001). De Graaf et al15 demonstrated that, on average, retinoblastoma eyes were slightly smaller than normal eyes and that tumor volume was negatively correlated with eye volume. When diagnosed in a late stage, retinoblastoma can present with increased intraocular pressure and buphthalmia.1 Figure 1E and 1F show an example of a normal-sized left eye and an enlarged right eye with globe deformation, a shallow anterior chamber, and a tumor-induced impairment of the aqueous humor drainage. In a late stage or (more commonly) after extensive conservative treatment, the affected eye can start to shrink, leading to a nonfunctional shrunken eye (phthisis bulbi; Fig. 1G and 1H). On fundoscopy, retinoblastoma sometimes presents with cavitary spaces in the tumor (Fig. 1I). There is some evidence that cavitary retinoblastoma is an indication of a more
5
differentiated tumor that responds less well to chemotherapy, but long-term outcomes do not seem to be worse.29,38
3.3. Vitreous and subretinal seeding, and retinal detachment Intra-arterial chemotherapy may be effective for the treatment of subretinal seeding, and with the introduction of intravitreal injection of chemotherapy, vitreous seeds can also be treated, avoiding external beam radiotherapy.2,33 Tumor seeding can often easily be seen at fundoscopy and was usually missed by MRI. Several studies showed that the diagnostic accuracy of MRI in detecting vitreous seeding compared with histopathologic analysis was not much better than a coin flip.4,23,42 Histopathologic analysis is not, however, an ideal method to detect seeding. Sometimes part of the vitreous is lost during histopathological preparation and harvesting of fresh tumor tissue for DNA mutation analysis. As image quality of MRI continuously improves over time, these numbers underestimate the current diagnostic accuracy for the detection of seeding. To be able to detect small subretinal or vitreous tumor seeds, the in-plane resolution of the images needs to be high enough and the slice thickness small enough. Figure 2A shows an example of a small vitreous seeds directly posterior to an artificial lens. Figure 2B shows a small seed in the anterior chamber. Fundoscopy confirmed the presence of tumor seeding in these two examples. In case seeding cannot be seen with fundoscopy (e.g., an unclear ocular medium due to massive hemorrhage), high-resolution MRI could be a useful addition. Optical coherence tomography has also shown promising results with regard to imaging of tumor seeding.39 Exophytic tumors combined with subretinal fluid accumulation can cause the retina to detach. Seeding can then develop in the subretinal space (Fig. 1D). Three studies have reported sensitivities of 50e89% and specificities of 88e100% on the diagnostic accuracy of MRI at detecting retinal detachment compared to clinical findings or histopathologic analysis.4,13,23 Retinal detachment shows up as a distinct hypointense line on T2WI and can be seen in Figure 2Ce2E. Detachment of the retina may be focally around the tumor, but can eventually evolve into a total retinal detachment, with a typical V-shaped appearance.
3.4.
Hemorrhage
Besides intratumoral hemorrhage that can cause falsepositive diagnoses of tumor calcification, hemorrhage also may occur in the vitreous, subretinal space, or the anterior chamber.20 Hemorrhage in any of these spaces is a predictor of a worse outcome for eye salvage. Intraocular hemorrhage can be identified on MRI scans as fluidefluid levels.14 Figure 2F shows subretinal hemorrhage in an eye with total retinal detachment. Figure 2G and 2H show an example of an eye initially with vitreous hemorrhage, but with subretinal hemorrhage after 10 months.
3.5.
Anterior eye chamber involvement
Tumor invasion into the anterior segment is considered by some to be a risk factor for extraocular relapse, but this is still
6
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 3 e A contrast-enhanced fat-suppressed axial T1-weighted image (A) shows a heterogeneously enhancing tumor and enhancement of the iris (arrows), indicative of neoangiogenesis in the iris. A dislocated lens is depicted on a T2-weighted image (arrow in B). This contrast-enhanced axial fat-suppressed T1-weighted image (C) shows focal enhancement in the distal optic nerve (arrow) suggestive of postlaminar optic nerve invasion and a sagittal subtraction image (E) of a
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
debated.5 Besides tumor invasion, other tumor-related changes in the anterior segment can be depicted on MRI. An increase in contrast enhancement on T1WI has been correlated with rubeosis iridis with a sensitivity ranging from 68e93% and a specificity ranging from 43e82%.18 Figure 3A shows an example of clinically and histopathologically confirmed rubeosis iridis. Rarely, advanced intraocular retinoblastoma can cause a dislocated lens, which usually occurs in the presence of massive tumor necrosis (Fig. 3B). Vitreous tumor seeding can eventually also invade the anterior segment, and from there the seeds can end up in the trabecular meshwork of the anterior chamber angle. The diagnostic accuracy of MRI compared with histopathologic analysis at detecting ciliary body invasion has a sensitivity ranging from 71e100% and a specificity ranging from 65e100%.18 Recently, the use of US biomicroscopy has demonstrated a sensitivity of 81% and specificity of 100% in the assessment of anterior segment invasion compared with histopathology.32
4.
Metastatic risk factors
Cross-sectional imaging, like MRI, is essential for the evaluation of tumor invasion into the ocular wall or optic nerve, as fundoscopy and US are of little value.40
4.1.
Optic nerve invasion
Patients with postlaminar optic nerve tumor invasion without adjuvant systemic chemotherapy are at great risk of regional extraocular recurrences or distant metastases.12 Optic nerve invasion can be seen as contrast enhancement on T1WI (Fig. 3C and 3E). Histopathologic analysis shows postlaminar optic nerve invasion (2.2 mm) in the first example (Fig. 3D). Histopathologically, the second example also proves to be a postlaminar optic nerve invasion (Fig. 3F), and depending on the assessment of the dubious postlaminar enhancement pattern in the second MRI scan (Fig. 3E), this case could either be classified as a false negative or a true positive. Appendix C shows an example of extensive postlaminar optic nerve invasion and the effect of subsequent chemotherapy (four cycles of etoposide and carboplatin) over the course of 3.5 months. Ten months prior to this, this patient was diagnosed with unilateral retinoblastoma (left eye was classified as group E) in another institution for which the patient had been treated with intra-arterial melphalan. In a recent meta-analysis with histopathology as the gold standard, De Jong et al18 showed that MRI could detect postlaminar optic nerve invasion with a sensitivity of 59% (95% confidence interval [CI]: 37e78%) and a specificity of 94% (95%
7
CI: 84e98%). Meta-regression showed no difference between low- and high-quality MRI scans. After exclusion of one potential outlier, the sensitivity of high-quality MRI rose from 58% (95% CI: 27e84%) to 76% (95% CI: 45e93%). This was not, however, statistically different from a sensitivity of 59% (95% CI: 32e82%) for low-quality MRI (P ¼ 0.26). This raises the question whether the resolution of the images is the only problem of test sensitivity, or it may be that depiction of tumor extension by gadolinium contrast enhancement is not an ideal marker for optic nerve invasiondthat is, tumor invasion does not always show contrast enhancement. Furthermore, contrast enhancement also is not 100% specific for tumor invasion, with reactive inflammation being the main cause of false-positive findings.16 On the other handdemphasizing the importance of highresolution imagingdBrisse et al8 showed that false-positive and false-negative findings were often borderline cases. Before the tumor invades the optic nerve beyond the lamina cribrosa, it can cause bulging of the laminadresulting in contrast enhancement on MRI to reach beyond where the lamina cribrosa is expected to be, producing a false-positive. A false-negative result can be caused by minimal postlaminar invasion (only a few cells on histopathologic analysis; Fig. 3F), which is still without the tumor neovascularization that causes the typical contrast enhancement visible on MRI. The only alternative to MRI for the detection of optic nerve invasion is computed tomography (CT), but this appears to have a low diagnostic value.18
4.2.
Choroidal and scleral invasion
Massive choroidal and scleral invasion are seen as risk factors for local retinoblastoma recurrence and distant metastases. According to Sastre et al,41 choroidal invasion of more than 3 mm or reaching the sclera should be classified as massive choroidal invasion. On T1-weighted contrast-enhanced MRI, choroidal invasion can be identified as a focal inhomogeneous enhancement and/or choroidal thickening. Figure 3G and 3H show an example of irregular thickening and abnormal enhancement of the choroid on MRI that correspond with massive choroidal invasion on histopathology (Fig. 3I). In case of scleral invasion, the contrast-enhanced area extends beyond the choroid into the normally unenhanced sclera. De Jong et al18 found a sensitivity of 74% (95% CI: 52e88%) and a specificity of 72% (95% CI: 31e94%) for choroidal invasion, and a sensitivity of 88% (95% CI: 20e100%) and a specificity of 99% (95% CI: 86e100%) for scleral invasion. Sensitivity and specificity were higher for higher-quality MRI, but these differences were not statistically significant. These numbers are in line
=T1-weighted image with and without contrast enhancement shows dubious contrast enhancement of the distal optic nerve; the arrow indicates an area of enhancement that matches bulging of the lamina cribrosa on histologic examination and the arrowhead indicates dubious postlaminar optic nerve enhancement; histopathologically both patients had postlaminar invasion indicated by hyperchromatic cells posterior to the lamina cribrosa (D and F: arrowheads show the prelaminar tumormass and the arrows point at postlaminar tumor cells). A coronal T2-weighted image (G) and a coronal reconstruction of axial VIBE (volumetric interpolated brain examination) images (H) show multifocal irregular thickening and enhancement of the choroid suspicious of choroidal invasion; massive choroidal invasion in the nasal part was confirmed (arrow on I, corresponding with arrows on G and H).
8
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
Fig. 4 e This axial VIBE (volumetric interpolated brain examination) image (A) shows a broad area of reduced enhancement of the choroid (arrows) corresponding with temporal choroidal ischemia clearly visible on fundoscopy (B) and fluorescent angiography (C).
with expectations, as choroidal invasion is visible as subtle changes on MRI, whereas scleral invasion is more manifest. No studies have evaluated the diagnostic accuracy of CT for the detection of choroidal invasion, and only one study looked at scleral invasion, showing a reasonable sensitivity (71%; 95% CI: 42e92%) and specificity (98%, 95% CI: 93e100%) for the ability of CT to diagnose scleral invasion.
4.3.
Extraocular spread
When the tumor passes though the ocular wall or into the postlaminar optic nerve, it can extend directly into the orbit or spread via the leptomeninges and cerebrospinal fluid. Appendix D shows an example of extensive invasion of the optic nerve that, because the parents refused any treatment, developed into a large intracranial mass within 4 months. As these lesions are large and therefore usually clearly visible on MR images, diagnostic accuracy of MRI of extrascleral growth in retinoblastoma patients is expected to be high.35 There only is one study (with a small sample size) presenting a sensitivity of 100% for MRI and one study showing a sensitivity of 100% for CT. According to three studies, specificity of MRI ranged from 94e100%.18
information to fundoscopy and US. As mentioned in the Introduction, there are a number of conservative treatment options for retinoblastoma. External beam radiotherapy is falling out of favor because it causes orbital bone growth restrictions and cosmetic problems on the face. Furthermore, it increases the risk of second primary tumors within the radiation field later in life in hereditary retinoblastoma patients.28 Depending on tumor size, location, and shape, local treatment (laser therapy, cryotherapy, and radioactive plaque therapy); intra-arterial, intravitreal, and peribulbar chemotherapy; and systemic chemotherapy are currently the conservative treatment options of choice.3,26,27,44 Appendix E shows a number of post-treatment findings on MRI. Appendix F shows the tumor response to chemotherapy over the course of a year. Choroidal ischemia is a known complication of intra-arterial chemotherapy, which can be seen on MRI as the lack of choroidal contrast enhancement. Figure 4Ae4C show the fundoscopic and fluorescent angiography and MRI findings of a patient with treatment-induced choroidal ischemia. Over the course of 6 months, this patient underwent multiple melphalan and topotecan injections. Perfusion and diffusion MRIdstill under development as an application for retinoblastomadhave shown promising initial results in terms of follow-up for the differentiation between viable and necrotic tumor tissue.17,37
5. Post-treatment status and treatment response
6.
Post-treatment follow-up with MRI is a rapidly developing discipline and shows promising results, providing additional
There are different approaches to generate high-quality MRI scans of eyes: with a 1.5T system43 and a surface coil or, as in
Discussion
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
this study, with a 3T system and a 32-channel head coil. The advantage of 1.5T imaging is that these systems are more readily available. On the other hand, positioning of a head coil is less cumbersome than positioning a surface coil. Head coils also generate better images of deeper structures, including the entire brain. At the time of retinoblastoma diagnosis, MRI of the brain is recommended to exclude trilateral retinoblastoma (ocular retinoblastoma combined with an intracranial midline primitive neuroectodermal tumor).19 The cases included in this article were specifically selected to show certain aspects of retinoblastoma. This approach carries the risk of introducing selection bias and potentially overestimating the diagnostic value of MRI (e.g., imaging of vitreal and subretinal seeds remains difficult, even though we were able to show successful examples). Also, the included patients are from a highly specialized center; therefore, most of these patients have more advanced retinoblastoma. Before the introduction of intra-arterial chemotherapy, patients with retinoblastoma and the presence of metastatic risk factors such as massive choroidal invasion or postlaminar optic nerve invasion would have undergone enucleation with systemic chemotherapy after histopathologic proof of the risk factor, or have received systemic chemotherapy in the form of chemoreduction followed by local treatment. Systemic chemotherapy, then, might have successfully prevented systemic metastasis. Nowadays, systemic chemotherapy, and its potential negative long-term effects, can be avoided in certain cases by using intra-arterial or intravitreal chemotherapy, although this does require a reasonable certainty that there are no metastatic risk factors, adding to the importance of reliable non-invasive imaging like MRI.34 To give an idea and provide a reference of the diagnostic accuracy of conventional MRI, we have reported diagnostic accuracy estimates when available. Nevertheless, with these limitations in mind, MRI becomes increasingly important with increasing eye-saving treatment, and we believe that highresolution MRI can contribute to better clinical decision-making. CT may also be used for the diagnosis, staging, and followup of retinoblastoma, but even though evidence for the use CT is scarce, available evidence does suggest that MRI is superior to CT for almost all previously described applications, most importantly the metastatic risk factors.18 Even for the calcifications, CT has no added value as US in combination with MRI accurately detects calcifications.20,36 Most importantly though, CT should be avoided because of the potential harmful effect of radiation in small children, especially those with hereditary retinoblastoma.6,28,30 A disadvantage of MRI is that it might not be accessible in developing countries; therefore, we believe that only when MRI is not available should CT be considered as an alternative method for imaging.
7.
Conclusion
We have provided an overview of the clinical application of retinoblastoma imaging with MRI. MRI in its current technological state can be of great help in retinoblastoma patient care. Because more and more eye-saving treatment options have entered clinical practice, other indications for MRI in
9
retinoblastoma patients have emerged, such as imaging during conservative treatmentdfor example, for the detection of tumor response, tumor recurrence, and adverse effects of treatment. Unfortunately, MRI does not have perfect diagnostic accuracy for the detection of metastatic risk factors like massive choroidal and postlaminar optic nerve invasion, but it currently is the best available tool, and its performance is constantly improving.
8.
Method of literature search
We searched the literature in PubMed (Medline) for English, Dutch, or German full-text articles published through 2014. Alternatively found articles (e.g., through checking reference lists of retrieved articles, including from De Jong et al18) were also considered. No articles were translated and we discarded articles in any other language even if they were accompanied by an English abstract. The search strategy included MeSH terms and Text Words to retrieve articles on retinoblastoma, optic nerve invasion, choroidal invasion, scleral invasion, metastasis, magnetic resonance imaging, computed tomography, diagnosis, staging, and follow-up.
9.
Disclosures
No conflicting relationship exists for any author.
Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.survophthal.2015.01.002.
references
1. Abramson DH, Frank CM, Susman M, et al. Presenting signs of retinoblastoma. J Pediatr. 1998;132(3 Pt 1):505e8 2. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma in eyes with vitreous and/ or subretinal seeding: 2-year results. Br J Ophthalmol. 2012;96:499e502 3. Aerts I, Lumbroso-Le Rouic L, Gauthier-Villars M, et al. Retinoblastoma. Orphanet J Rare Dis. 2006;1:31 4. Barkhof F, Smeets M, van der Valk P, et al. MR imaging in retinoblastoma. Eur Radiol. 1997;7:726e31 5. Baroni LV, Sampor C, Fandin˜o A, et al. Anterior segment invasion in retinoblastoma: is it a risk factor for extraocular relapse? J Pediatr Hematol Oncol. 2014;36:e509e12 6. Brenner D, Elliston C, Hall E, et al. Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol. 2001;176:289e96 7. Brisse HJ, Graaf P, de, Galluzzi P, et al. Assessment of earlystage optic nerve invasion in retinoblastoma using highresolution 1.5 Tesla MRI with surface coils: a multicentre, prospective accuracy study with histopathological correlation. Eur Radiol 2014; http://dx.doi.org/10.1007/s00330-014-3514-1 8. Brisse HJ, Guesmi M, Aerts I, et al. Relevance of CT and MRI in retinoblastoma for the diagnosis of postlaminar invasion with normal-size optic nerve: a retrospective study of 150
10
9.
10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
s u r v e y o f o p h t h a l m o l o g y x x x ( 2 0 1 5 ) 1 e1 0
patients with histological comparison. Pediatr Radiol. 2007;37:649e56 Brisse HJ, Lumbroso L, Fre´neaux PC, et al. Sonographic, CT, and MR imaging findings in diffuse infiltrative retinoblastoma: report of two cases with histologic comparison. AJNR Am J Neuroradiol. 2001;22:499e504 Bullock JD, Campbell RJ, Waller RR. Calcification in retinoblastoma. Invest Ophthalmol Vis Sci. 1977;16:252e5 Chantada G, Doz F, Antoneli CBG, et al. A proposal for an international retinoblastoma staging system. Pediatr Blood Cancer. 2006;47:801e5 Chantada GL, Casco F, Fandin˜o AC, et al. Outcome of patients with retinoblastoma and postlaminar optic nerve invasion. Ophthalmology. 2007;114:2083e9 De Graaf P, Barkhof F, Moll AC, et al. Retinoblastoma: MR imaging parameters in detection of tumor extent. Radiology. 2005;235:197e207 De Graaf P, Go¨ricke S, Rodjan F, et al. Guidelines for imaging retinoblastoma: imaging principles and MRI standardization. Pediatr Radiol. 2012;42:2e14 De Graaf P, Knol DL, Moll AC, et al. Eye size in retinoblastoma: MR imaging measurements in normal and affected eyes. Radiology. 2007;244:273e80 De Graaf P, Moll AC, Imhof SM, et al. Retinoblastoma and optic nerve enhancement on MRI: not always extraocular tumour extension. Br J Ophthalmol. 2006;90:800e1 De Graaf P, Pouwels PJW, Rodjan F, et al. Single-shot turbo spin-echo diffusion-weighted imaging for retinoblastoma: initial experience. AJNR Am J Neuroradiol. 2012;33:110e8 De Jong MC, de Graaf P, Noij DP, et al. Diagnostic performance of magnetic resonance imaging and computed tomography for advanced retinoblastoma: a systematic review and metaanalysis. Ophthalmology. 2014;121:1109e18 De Jong MC, Kors WA, de Graaf P, et al. Trilateral retinoblastoma: a systematic review and meta-analysis. Lancet Oncol. 2014;15:1157e67 Galluzzi P, Hadjistilianou T, Cerase A, et al. Is CT still useful in the study protocol of retinoblastoma? AJNR Am J Neuroradiol. 2009;30:1760e5 Houston SK, Murray TG, Wolfe SQ, et al. Current update on retinoblastoma. Int Ophthalmol Clin. 2011;51:77e91 Khelfaoui F, Validire P, Auperin A, et al. Histopathologic risk factors in retinoblastoma: a retrospective study of 172 patients treated in a single institution. Cancer. 1996;77:1206e13 Khurana A, Eisenhut CA, Wan W, et al. Comparison of the diagnostic value of MR imaging and ophthalmoscopy for the staging of retinoblastoma. Eur Radiol. 2012;23:1e10 Lemke A-J, Kazi I, Mergner U, et al. RetinoblastomadMR appearance using a surface coil in comparison with histopathological results. Eur Radiol. 2007;17:49e60 Linn Murphree A. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin North Am. 2005;18:41e53 Lumbroso L, Doz F, Urbieta M, et al. Chemothermotherapy in the management of retinoblastoma. Ophthalmology. 2002;109:1130e6 Lumbroso-Le Rouic L, Aerts I, Le´vy-Gabriel C, et al. Conservative treatments of intraocular retinoblastoma. Ophthalmology. 2008;115:1405e10, 1410.e1e2. Marees T, Moll AC, Imhof SM, et al. Risk of second malignancies in survivors of retinoblastoma: more than 40 years of follow-up. J Natl Cancer Inst. 2008;100:1771e9
29. Mashayekhi A, Shields CL, Eagle RC, et al. Cavitary changes in retinoblastoma: relationship to chemoresistance. Ophthalmology. 2005;112:1145e50 30. Mathews JD, Forsythe AV, Brady Z, et al. Cancer risk in 680 000 people exposed to computed tomography scans in childhood or adolescence: data linkage study of 11 million Australians. BMJ. 2013;346:f2360 31. Moll AC, Kuik DJ, Bouter LM, et al. Incidence and survival of retinoblastoma in The Netherlands: a register-based study 1862e1995. Br J Ophthalmol. 1997;81:559e62 32. Moulin AP, Gaillard M-C, Balmer A, et al. Ultrasound biomicroscopy evaluation of anterior extension in retinoblastoma: a clinicopathological study. Br J Ophthalmol. 2012;96:337e40 33. Munier FL, Gaillard M-C, Balmer A, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol. 2012;96:1078e83 34. Ong SJ, Chao A-N, Wong H-F, et al. Selective ophthalmic arterial injection of melphalan for intraocular retinoblastoma: a 4-year review. Jpn J Ophthalmol 2014; http://dx.doi.org/10.1007/s10384-014-0356-y 35. Radhakrishnan V, Sharma S, Vishnubhatla S, et al. MRI findings at baseline and after neoadjuvant chemotherapy in orbital retinoblastoma (IRSS stage III). Br J Ophthalmol. 2013;97:52e8 36. Rodjan F, de Graaf P, van der Valk P, et al. Detection of calcifications in retinoblastoma using gradient-echo MR imaging sequences: comparative study between in vivo MR imaging and ex vivo high-resolution CT. Am J Neuroradiol. 2015;36:355e60 37. Rodjan F, de Graaf P, van der Valk P, et al. Retinoblastoma: value of dynamic contrast-enhanced MR imaging and correlation with tumor angiogenesis. AJNR Am J Neuroradiol. 2012;33:2129e35 38. Rojanaporn D, Kaliki S, Bianciotto CG, et al. Intravenous chemoreduction or intra-arterial chemotherapy for cavitary retinoblastoma: long-term results. Arch Ophthalmol. 2012;130:585e90 39. Rootman DB, Gonzalez E, Mallipatna A, et al. Hand-held highresolution spectral domain optical coherence tomography in retinoblastoma: clinical and morphologic considerations. Br J Ophthalmol. 2013;97:59e65 40. Roth DB, Scott IU, Murray TG, et al. Echography of retinoblastoma: histopathologic correlation and serial evaluation after globe-conserving radiotherapy or chemotherapy. J Pediatr Ophthalmol Strabismus. 2001;38:136e43 41. Sastre X, Chantada GL, Doz F, et al. Proceedings of the consensus meetings from the International Retinoblastoma Staging Working Group on the pathology guidelines for the examination of enucleated eyes and evaluation of prognostic risk factors in retinoblastoma. Arch Pathol Lab Med. 2009;133:1199e202 42. Schueler AO, Hosten N, Bechrakis NE, et al. High resolution magnetic resonance imaging of retinoblastoma. Br J Ophthalmol. 2003;87:330e5 43. Sirin S, Schlamann M, Metz KA, et al. Diagnostic image quality of gadolinium-enhanced T1-weighted MRI with and without fat saturation in children with retinoblastoma. Pediatr Radiol. 2013;43:716e24 44. Smith SJ, Smith BD, Mohney BG. Ocular side effects following intravitreal injection therapy for retinoblastoma: a systematic review. Br J Ophthalmol. 2013;98:292e7
