Letters to the Editor

Hemorrhagic Pituitary Adenoma versus Rathke Cleft Cyst: A Frequent Dilemma


J.-F. Bonneville
Department of Radiology and Endocrinology
University Hospital of Liège, University of Liège
Liège, Belgium

Iread with interest the article of Park et al1 published in the October issue of theAmerican Journal of Neuroradiology. Differentiation with MR imaging of a cystic or hemorrhagic pituitary adenoma from a Rathke cleft cyst (RCC) remains a common issue. In daily practice, this situation may be particularly confusing in a young woman with mild hyperprolactinemia whose symptoms are frequently hidden by taking contraceptive pills. Moreover, parallelism between the prolactin level and tumoral volume is missing in hemorrhagic microprolactinomas. Then, diagnosis of hemorrhagic microprolactinoma versus T1 hyperintense intrasellar mucoid RCC is challenging.

Park et al reported that the main differentiating features of pituitary adenomas are off-midline location, tilting of the pituitary stalk, fluid-fluid level, T2 hypointense hemosiderin rim, and septations, while Rathke cleft cysts are more likely located on the midline and frequently present with a T2 hypointense characteristic nodule.

Nevertheless, pituitary adenomas may be on the midline, for instance corticotroph adenomas. The fluid-fluid level is inconstant, particularly with fresh hemorrhage and a peripheral hemosiderin rim because of the absence of blood-brain barrier in the pituitary gland; septations are inconstant. On the other hand, Rathke cleft cysts may be, rarely, in an off-midline location, and their T2 hypointense waxy nodules are detected in no more than 70% of cases.

Moreover, the diagnostic tree model proposed by Park et al seems difficult to apply to the strictly intrasellar infracentimetric lesions, which are more and more frequently seen with high- resolution 3T scanners, either discovered fortuitously or in the assessment of hyperprolactinemia.

Ancillary signs are then welcome to differentiate RCCs from cystic or hemorrhagic pituitary adenomas.

The axial T1-weighted sequence is optimal for making the diagnosis of such RCCs: Strict midline location, …

Cerebral Amyloid Angiopathy as an Etiology for Cortical Superficial Siderosis: An Unproven Hypothesis


H.X. Bai
Department of Radiology
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania

H. Zhou
Department of Neurology
Xiangya Hospital, Central South University
Hunan, China

X. Tan, X. Huang and L. Yang
Department of Neurology
The Second Xiangya Hospital, Central South University
Hunan, China

We read with great interest a recent article by Inoue et al1 on the diagnostic significance of cortical superficial siderosis (cSS) for Alzheimer disease in patients with cognitive impairment. The article focused on presymptomatic cases of cSS diagnosed by MR imaging.

The major finding of the article was that cSS was associated with a lobar location of microbleeds (MBs) and may be an initial radiologic finding of cerebral amyloid angiopathy (CAA) in patients with cognitive impairment. Lobar location includes both the cortical gray matter and the subcortical white matter. The imaging manifestations of hemosiderin deposition from cSS and MBs from CAA can be indistinguishable on gradient-echo T2-weighted images,2 especially when MBs are seen on the surface of the cerebral cortex. There is even a suggestion that CAA can be an underlying cause of cSS.3 In the current study, there was 72% correspondence between the location of cSS and that of MBs. The definition of cSS was only based on the shape of the signal abnormality on SWI (ie, linear). Because there are currently no widely recognized criteria to distinguish hemosiderin deposition from MBs on imaging, it may be helpful to show interobserver variability in the assignment of individual lesions to ensure agreement on the nature of the hypointensity seen on T2-weighted MR imaging.

Although its detection has increased with the advances in MR imaging technology, cSS is still a rare disease.4 The most accepted hypothesis for its etiology has been chronic iron deposition in neuronal tissues associated with CSF.5 Chronic …

Re: The Benefits of High Relaxivity for Brain Tumor Imaging (The BENEFIT Study)


Re: The Benefits of High Relaxivity for Brain Tumor Imaging: Results of a Multicenter Intraindividual Crossover Comparison of Gadobenate Dimeglumine with Gadoterate Meglumine (The BENEFIT Study)

E. Lancelot, B. Piednoir, P. Desché
Roissy CdG Cedex, France

We have read with interest the publication by Vaneckova et al1 reporting the results of a clinical study that assessed the diagnostic performances of 2 gadolinium-based contrast agents (GBCAs) used for brain tumor imaging. The authors performed a multicentric, prospective, randomized, intraindividual, crossover, 2-arm study. The objective of Arm 1 was to demonstrate the superiority of a full dose (0.1 mmol/kg) of gadobenate dimeglumine over the same dose of gadoterate meglumine, whereas in Arm 2, the aim was to ascertain whether a half dose (0.05 mmol/kg) of gadobenate provides diagnostic information similar to that of a full dose of gadoterate. GBCA administrations and image analyses were performed in a blinded manner. The primary end point was the overall diagnostic preference of the readers for one GBCA over the other. In Arm 1, a significant superiority was shown in favor of gadobenate, and in Arm 2, no significant differences could be found between the 2 GBCAs. The authors concluded that when administered at the approved dose of 0.1 mmol/kg, gadobenate is superior to gadoterate for qualitative and quantitative assessment of brain lesions, and that a half dose of the former agent is equivalent to a full dose of the latter. However, we consider that some biases limit the interpretation of the results and even lead to wrong assertions.

First, the statistical analysis was not adapted to the objectives of the study. To compute the sample size in each arm, the authors assumed that no difference in overall diagnosis preference would be found between the 2 GBCAs in half of the patients. In …

Flow-Diverter Stents for the Treatment of Saccular Middle Cerebral Artery Bifurcation Aneurysms: Is “Unsuitable” the Right Conclusion?

C. Iosif and C. Mounayer
Department of Interventional Neuroradiology
Dupuytren University Hospital
Limoges, France

We read with interest the recent publication by Caroff et al1 on saccular MCA aneurysm treatment by flow diverters (FDs), in regard to the increasing tendency to use FDs in distal, off-label localizations. We appreciate their important contribution to the subject but remain reserved regarding their strong conclusion. The subject is still under research, but some promising clinical and angiographic outcomes have been published.2 We were slightly perplexed by the interpretation of the reported complication rate, taking into account the methodologic limitations of the study due to the small number of subjects.

Various mechanisms of ischemic complications are implicated in flow diversion with covered arteries, which are not always sufficiently represented in a small case series. Thrombosis related to an ineffective anti-aggregation regimen, suboptimal stent apposition on the arterial wall, or distal emboli provoked by the mechanical manipulations inside the arteries are types of complications that should not be confounded with the pure hemodynamic effect of flow diversion on a covered branch. Good discrimination of the various potential ischemic mechanisms is mandatory to obtain a more profound understanding of FD-related complications.

To clear a misunderstanding, we must comment on the complication rates of the case series by Saleme et al3 cited in the discussion. The overall complication rate for the MCA cases in the study was 5% (n = 1/19), with a sole side branch–related complication in the MCA aneurysm sub-series. Apart from reporting complication rates, they performed a classification of jailed branches, dividing them into anastomotic or terminal. In this section, a symptomatic angiographic change was defined as any angiographic remodelling at 6 months in patients who reported transient symptoms at any point during the follow-up of 6 months.

These findings, …

Stents for Idiopathic Intracranial Hypertension: Meta-Analyzed, Hypo-Analyzed, and In Need of a Trial

P. Noonan
Department of Radiology
Scott and White Memorial Hospital
Temple, Texas

How did the recent article by Satti et al1 pass muster? A review of the status of venous sinus stent placement in 20132 included more patients with stents than the so-called meta-analysis review in your October 2015 issue. Seven of the studies cited in the 2013 analysis had only 1 patient each and were not included in the 2015 analysis; stripped of those 7 single patient studies, the 2015 meta-analysis looks much like the 2013 review. Nevertheless, why choose to include in the meta-analysis a study with only 4 patients—or the studies with only 10, 12, 15, or 18 patients? The power in both the 2013 review and the current meta-analysis lies in the 52-patient study by Ahmed et al,3 in which CSF opening pressure, an essential criterion for the diagnosis of idiopathic intracranial hypertension (IIH), was not documented in either 11 or 9 patients depending on which meta-analysis one chooses to read. Likewise, the 2 studies with 15 patients each (Fields et al4 and Albuquerque et al5) did not record the CSF opening pressure for any of their included patients; and both the 4-patient study (Owler et al6) and the 18-patient study (Kumpe et al7) did not record CSF opening pressure in 1 and 4 patients, respectively. A study of 10 patients noted to be without recorded CSF opening pressures in the 2013 analysis (Bussière et al8) is noted in Table 4 of the 2015 meta-analysis as having an opening pressure range of 25–50 cm H20. What was the source of that post hoc information?

Given that an elevated CSF pressure is an essential diagnostic criterion of IIH, for which all of these …

Low Signals on T2* and SWI Sequences in Patients with MS with Progressive Multifocal Leukoencephalopathy

P. Labauge, C. Carra-Dalliere and X. Ayrignac
Department of Neurology

N. Menjot de Champfleur
Department Neuroradiology

Montpellier University Hospital
Centre Hospitalo-Universitair Gui de Chauliac
Montpellier, France

We read with interest the study by Hodel et al1 concerning the occurrence of cortex, U-fiber, and basal ganglia low signals found on T2* and SWI sequences in 12 patients with MS with progressive multifocal leukoencephalopathy (PML). These low signals were frequent in this series and may occur in presymptomatic patients with PML treated with natalizumab (75% of their 8 asymptomatic patients with PML). We previously published such low intensities in a patient with a PML diagnosis.2 We further confirmed the importance of T2* and SWI sequences to detect low signals in a series of 4 patients with PML.3 Pathologic analysis by Hodel et al1 of a patient with low T2* signal intensities identified astrocytic gliosis associated with abundant microglial and macrophage infiltrates, containing myelin-filled vacuoles. The authors hypothesized that low signal on T2* could be related to accumulation of iron in the macrophages. Although T2* and SWI low signal intensities are not constant in PML, their occurrence may differentiate confluent MS lesions from PML. Consequently, an MR imaging survey of patients at risk for PML (notably patients with MS treated with natalizumab or those who are immunosuppressed) should include T2* and SWI sequences.


  1. Hodel J, Outteryck O, Verclytte S, et al. Brain magnetic susceptibility changes in patients with natalizumab-associated progressive multifocal leukoencephalopathy. AJNR Am J Neuroradiol 2015 Aug 27. [Epub ahead of print] » Abstract/FREE Full Text
  2. Carra-Dalliere C, Menjot de Champfleur N, Ayrignac X, et al. Quantitative susceptibility mapping suggests a paramagnetic effect in PML. Neurology 2015;84:1501–02 » Abstract/FREE Full Text
  3. Carra-Dalliere C, Menjot de Champfleur N, Deverdun J, et al. Use of quantitative susceptibility

Risk of Malignancy in Symptomatic Nodular Goiter Treated with Radiofrequency Ablation

L. Yang
Department of Neurology
The Second Xiangya Hospital of Central South University
Changsha, China

H. Tang
Department of Radiology
The First Xiangya Hospital of Central South University
Changsha, China

A.M. Lee
Department of Radiology
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania

Y. Zou, X. Huang and X. Tang
Department of Neurology
The Second Xiangya Hospital of Central South University
Changsha, China

H.X. Bai
Department of Radiology
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania

We read with great interest the recent article by Che et al.1 In this article, the authors compared the efficacy, safety, and cost-effectiveness of surgery versus radiofrequency ablation (RFA) for the treatment of benign thyroid nodules. Two hundred patients were retrospectively selected for each group. The authors assessed procedure-related complications, the length of hospitalization, and cost. In addition, nodule volume, incidence of hypothyroidism, and the rate of residual nodules were assessed at 1-year follow-up. All of these outcomes uniformly favored RFA over surgery; the cost between the 2 procedures was not significantly different. Consequently, the authors advocated RFA as the first-line treatment for benign thyroid nodules.

We commend the authors for conducting the largest retrospective cohort study comparing these 2 treatment options for benign thyroid nodules. However, the result could be significantly confounded by patient-selection bias. The authors used different criterion standards to define benign thyroid nodules in the surgery-versus-RFA groups (surgery pathology and cytology from fine-needle aspiration, respectively). This patient-selection method excluded patients whose nodules were benign by cytology but malignant by surgical pathology in the surgery group. In fact, the potential of mistreating malignant nodules as benign ones is the strongest argument against RFA as the first-line treatment for symptomatic nodular goiter.2

Two studies in the literature compared surgery and RFA as a treatment for benign thyroid nodules. One is …

Recent Advances in Understanding Gadolinium Retention in the Brain

T. Kanda, H. Oba, K. Toyoda and S. Furui
Department of Radiology
Teikyo University School of Medicine
Itabashi-ku, Japan

We wish to comment on the August 2015 article of Adin et al1 in the American Journal of Neuroradiology (AJNR) entitled “Hyperintense Dentate Nuclei on T1-Weighted MRI: Relation to Repeat Gadolinium Administration.” The authors reported the relationship between the hyperintense dentate nucleus on unenhanced T1WI and past gadolinium based–contrast agent (GBCA) administration. This relationship was first reported by our group on December 7, 2013.2 Since then, several important reports have been published, and knowledge regarding gadolinium deposition has increased remarkably.

Our group3 and Radbruch et al4 evaluated the difference in the signal change between patients repeatedly administered linear GBCA and macrocyclic GBCA. A change in the signal intensity of the dentate nucleus was observed in the former, but not in the latter. McDonald et al5 and our group6 evaluated the brain tissue from postmortem specimens, and gadolinium deposition was verified from the brain tissue.7 Robert et al8injected GBCA 20 times into rats and evaluated the signal-intensity change of the dentate nucleus on T1WI and gadolinium concentration in the brain. A hyperintense dentate nucleus was observed in rats with repeat linear GBCA administration, but not with repeat macrocyclic GBCA administration. The gadolinium concentration of the brain with repeat linear GBCA administration was 14 times greater than that with repeat macrocyclic GBCA administration.8 The work of Adin et al1 was confirmed in our first study. It was accepted by AJNR on February 19, 2015, and was published on-line on August 20, 2015. In this short period, studies on gadolinium deposition advance so rapidly, a more prompt publication schedule from AJNR would be desirable.

In previous studies (our study), a hyperintense …

Ipilimumab Therapy for Melanoma: A Mimic of Leptomeningeal Metastases

S. Ali, S.-K. Lee
Department of Radiology, Section of Neuroradiology

University of Chicago Medical Center
Chicago, Illinois

We have recently observed that patients with malignant melanoma on treatment with ipilimumab can demonstrate leptomeningeal enhancement on brain MR imaging mimicking metastatic disease. In Fig 1 is shown one such case of a 40-year-old man presenting with subacute onset of headaches, right-sided facial hemiparesis, and facial paresthesias. He also showed evidence of tongue deviation on neurologic examination suggestive of cranial neuropathy. The patient had a history of stage IIIC melanoma from an unknown primary site with metastasis to the axilla and was undergoing therapy with ipilimumab. MR imaging demonstrated multifocal leptomeningeal and cranial nerve enhancement. Lumbar puncture revealed CSF lymphocytosis and mildly elevated proteins (findings that can be seen with inflammatory processes) and cultures and cytology negative for infection and malignancy, despite repeat examinations. Following drug discontinuation and high-dose steroid therapy, the patient gradually recovered. Lymphocytosis in the CSF also resolved.

Postcontrast T1-weighted images demonstrate nodular enhancement involving the right facial nerve in the meatal (A) and mastoid (B) segments and the bilateral trigeminal nerves (C and D). Leptomeningeal enhancement along the cerebellar surface (E) and abnormal enhancement along the perivascular spaces in the pons and cerebral hemispheres (E and F) are also seen.
Postcontrast T1-weighted images demonstrate nodular enhancement involving the right facial nerve in the meatal (A) and mastoid (B) segments and the bilateral trigeminal nerves (C and D). Leptomeningeal enhancement along the cerebellar surface (E) and abnormal enhancement along the perivascular spaces in the pons and cerebral hemispheres (E and F) are also seen.

To our knowledge, these brain imaging findings have not been shown in the literature. Ipilimumab, also known as MDX-010, is a human monoclonal antibody that augments T-cell-mediated immunity by blocking inhibitory signals that suppress T-cell function (more specifically, it blocks cytotoxic T lymphocyte antigen-4).1 It is approved for the treatment of late-stage melanoma and is currently undergoing clinical trials for other cancers. Manousakis et al2 reported a case of an inflammatory multifocal radiculoneuropathy during ipilimumab therapy …

Standardized Brain Tumor Imaging Protocol for Clinical Trials

G.V. Goldmacher
Department of Medical and Scientific Affairs
ICON Clinical Research
Warrington, Pennsylvania

B.M. Ellingson
David Geffen School of Medicine
University of California, Los Angeles
Los Angeles, California

J. Boxerman
Brown Alpert Medical School
Brown University
Providence, Rhode Island

D. Barboriak
Department of Radiology
Duke University School of Medicine
Durham, North Carolina

W.B. Pope
Department of Radiology
UCLA Medical Center
Los Angeles, California

M. Gilbert
Center for Cancer Research
National Institutes of Health
Bethesda, Maryland

We would like to alert the neuroradiology community to a recent publication in Neuro-Oncology1 that describes a standardized brain tumor MR imaging protocol, which is expected to gain widespread use in multicenter clinical trials of glioblastoma therapy.

Imaging is critical in the assessment of treatment response in glioblastoma, but its use faces a variety of challenges, including the morphologic complexity of these tumors, the confounding effects of therapies such as radiation and antiangiogenic agents, and the modest (at best) incremental benefits that new therapies generate. An important goal of imaging in clinical trials is to minimize assessment variability so that data can be pooled across sites to optimize the detection of treatment effects and provide accurate comparisons with prior trials.

The Imaging End Points in Brain Tumor Clinical Trials workshop, held in January 2014, involved patient advocates, device and pharmaceutical industry leaders, the FDA, the National Cancer Institute (NCI), and academic experts in neuro-oncology, neuroradiology, and imaging physics, with the goal of improving the use of imaging end points in glioblastoma clinical trials. The workshop was predicated on the acknowledgment that large trials using survival end points are costly. Improving the use of imaging would allow smaller, less expensive clinical trials and help accelerate the development of new treatments.

A key recommendation that emerged from this meeting and that received the support of all …