Abstract
Objective: To evaluate the usefulness of the punctate pattern (PP) for the diagnosis and follow-up of patients with progressive multifocal leukoencephalopathy (PML).
Methods: A cohort of 20 consecutive patients with PML, related to natalizumab (NTZ) (n = 14) or not (n = 6), underwent 3T MRI (147 MRI examinations). MRI was available at presymptomatic (n = 9 patients), symptomatic (n = 15), immune reconstitution inflammatory syndrome (IRIS), and chronic stages (n = 20). A pathologic control group of patients without PML (n = 80), with clinically definitive multiple sclerosis or a clinically isolated syndrome suggestive of CNS demyelination, underwent the same MRI protocol. Number and appearance of punctate lesions were assessed by 3 blinded readers using T2-weighted, fluid-attenuated inversion recovery (FLAIR), and postcontrast T1-weighted images.
Results: Interobserver agreement was good (κ = 0.79) (0.72–0.87). Of the 20 patients with PML, 18 had PP, including the 14 patients with NTZ-PML; none in the pathologic control group. Of the 9 presymptomatic patients with NTZ-PML, PP was observed in 7 (78% sensitive and 100% specific). Nonenhancing PP on T2-weighted/FLAIR images was detected in 13 patients with PML, exclusively at the presymptomatic or symptomatic stages (including 7 NTZ-PML), whereas enhancing PP occurred in 16 patients with PML, including 13 of the 14 patients with NTZ-PML at the IRIS stage.
Conclusions: PP is a highly specific feature of PML and may be the first imaging feature at the presymptomatic stage with potential implications in patient care.
Classification of evidence: This study provides Class II evidence that a PP on MRI accurately identifies patients with NTZ-PML.
GLOSSARY
- CI=
- confidence interval;
- CIS=
- clinically isolated syndrome;
- DWI=
- diffusion-weighted imaging;
- FLAIR=
- fluid-attenuated inversion recovery;
- FOV=
- field of view;
- IRIS=
- immune reconstitution inflammatory syndrome;
- IST=
- immunosuppressive therapies;
- MS=
- multiple sclerosis;
- NTZ=
- natalizumab;
- PML=
- progressive multifocal leukoencephalopathy;
- PP=
- punctate pattern;
- RRMS=
- relapsing-remitting multiple sclerosis;
- TE=
- echo time;
- TR=
- repetition time
Natalizumab (NTZ), an effective treatment in patients with relapsing-remitting multiple sclerosis (RRMS), is associated with a risk of progressive multifocal leukoencephalopathy (PML).1 Approximately 134,600 patients received NTZ in the postmarketing setting worldwide.2 As of March 2015, the overall PML incidence is 3.87 per 1,000 patients (95% confidence interval [CI] 3.55–4.21 per 1,000 patients).2 Factors that increase the risk of PML have been identified: high level of anti–John Cunningham virus antibody, receiving an immunosuppressant prior to receiving NTZ, and treatment duration, especially >2 years.2,3 Diagnosis of NTZ-associated PML (NTZ-PML) at the presymptomatic stage, with more localized brain involvement, improves the survival and functional outcome and mainly relies on MRI.4,5
MRI is crucial for the recognition of PML.6,7 The known imaging findings for presymptomatic NTZ-PML include the following: subcortical location involving U-fibers, sharp lesional border toward the gray matter contrasting with ill-defined border toward the white matter, and increased signal intensity on T2-weighted and diffusion-weighted images.7 However, the diagnosis of presymptomatic or early NTZ-PML remains difficult due to the coexistence of multiple sclerosis (MS) lesions and the different imaging patterns of NTZ-PML lesions.7,–,9 In patients diagnosed with PML at early stages, the detection of immune reconstitution inflammatory syndrome (IRIS) also relies on MRI using postcontrast images, with important implications for patient care.7,8,10
The punctate pattern (PP) refers to T2-weighted hyperintense or enhancing brain punctate lesions.6,–,8 Previous case reports have suggested that the PP is related to inflammation with CD8-positive T-cells within the perivascular spaces.11,12 PP was reported in patients with PML,12,–,14 neurosarcoidosis,15 hematologic diseases,16 chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids,17 or CNS vasculitis.18 The clinical relevance as well as the underlying histopathology of this inflammation pattern is not completely understood. Particularly, there are no data available on its diagnostic value in patients with PML. Our purpose was to evaluate the usefulness of the PP for the diagnosis and follow-up of patients with PML.
METHODS
Our primary research questions were (1) to investigate the occurrence of PP in patients with PML (related to NTZ or not, at the time of the diagnosis and during the follow-up period) as well as in patients with clinically isolated syndrome (CIS) and patients with RRMS without PML; and (2) to assess the diagnostic precision of PP in NTZ-PML patients at the presymptomatic stage. This study provides Class II evidence.
Participants.
From February 2011 to April 2015, 20 consecutive patients (9 women, mean age 49.1 years, minimum 26 years, maximum 63 years) were diagnosed with PML, including 14 patients with RRMS treated with NTZ (9 women, mean age 47.1 years, minimum 26 years, maximum 63 years) and 6 with PML from other causes (6 men, mean age 53.7 years, minimum 43 years, maximum 62 years), including 2 patients with leukemia, 3 treated with immunosuppressive therapies (IST) after liver or renal transplant, and 1 with sarcoidosis. In patients with NTZ-PML, the mean NTZ exposure duration at the time of PML diagnosis was 48.7 months (minimum 20, maximum 79), mean Expanded Disability Status Scale score was 3.3 (minimum 2.5, maximum 5.5), and mean disease duration was 11 years (SD 7 years). Patient 2 with NTZ-PML underwent postmortem brain neuropathologic examination.
The diagnosis of PML was made on the basis of the following:
Suggestive clinical and imaging findings associated with positive JC virus DNA PCR in the CSF for 17 patients (definite PML according to American Academy of Neurology criteria19)
Highly suggestive imaging and clinical follow-up for 3 patients for whom iterative CSF examinations were negative: 2 RRMS treated with NTZ (patients 10 and 12) and 1 patient treated with IST (patient 18)
The 9 patients suspected of PML at the presymptomatic stage developed progressive neurologic symptoms, positive JC virus DNA PCR in the CSF, and suggestive imaging findings that confirmed definite PML.
The presence of PML-IRIS was determined on the basis of clinical findings: sudden worsening of neurologic signs following NTZ withdrawal associated or not with swelling, mass effect, or contrast enhancement on MRI.20
At the time of initial diagnosis and during the follow-up period, all patients were consecutively referred to our radiology department for diagnostic evaluation on a 3T MRI scanner (Achieva, Philips, Best, the Netherlands).
A pathologic control group of 80 consecutive patients without PML (56 female, mean age 48.3 years, minimum 23 years, maximum 65 years), including 49 patients with RRMS treated with NTZ and 31 patients with CIS not treated, was also included to evaluate the accuracy of the PP. We included patients with CIS to test the hypothesis whether the PP may occur at the early stage of MS. We used a systematic MRI protocol in all patients including those with PML (related or not to NTZ), CIS, or RRMS.
All the controls came from our institution and were scanned according to a systematic MRI protocol on the same 3T scanner and during the same period as the patients with NTZ-PML. The inclusion criteria retained for the patients with RRMS were as follows: patients with RRMS with at least 2 years of NTZ exposure, JC virus seronegative, and absence of new clinical symptoms during the duration of NTZ exposure. The mean disease duration was 10.6 years (SD 6.1 years) in patients with RRMS treated with NTZ and without PML and 0.5 years (SD 0.3 years) in patients with CIS.
There was no active lesion detected in patients with RRMS treated with NTZ without PML. A total of 47 enhanced active inflammatory lesions (with a typical appearance including rim-like or nodular shapes) were detected in 19 of the 31 patients with CIS.
MRI studies.
All MRI studies were performed on a 3T MRI system (Achieva). The following standardized protocol was used in all participants (patients with PML and pathologic control group without PML): T2-weighted (2D, repetition time [TR] 6,600 ms, echo time [TE] 100 ms, field of view [FOV] 230 mm, slice thickness 3 mm), fluid-attenuated inversion recovery (FLAIR) (3D, TR 8,000 ms/TE 220 ms, voxel size 1 mm isotropic), and postcontrast T1-weighted (3D, TR 400 ms/TE 26 ms, FOV 230 mm, voxel size 1 mm isotropic) MRI sequences. The T1-weighted postcontrast sequence was performed 15 minutes after manual injection of single-dose gadolinium chelates (Dotarem; Guerbet, Villepinte, France).
The 20 consecutive patients with PML underwent repeated MRI assessments (147 MRI scans, mean MRI scans per patient: 6.9, minimum 3, maximum 12): from presymptomatic to chronic stages (i.e., at least 3 months after the IRIS) (n = 9 patients with NTZ-PML); from symptomatic (i.e., MRI performed within 2 weeks of clinical PML diagnosis) to chronic stages (n = 6 patients including 4 with NTZ-PML); and from IRIS (i.e., within 2 weeks of diagnosis of IRIS) to chronic stages (n = 5 patients including 1 with NTZ-PML). Of the 147 MRI examinations performed in patients with PML, 11 were performed at the presymptomatic stage, 48 at the symptomatic stage, 53 at the IRIS stage, and 35 at the chronic stage (figure 1). The mean time interval between 2 consecutive MRIs was 2 weeks.
IRIS = immune reconstitution inflammatory syndrome; NTZ = natalizumab; RRMS = relapsing-remitting multiple sclerosis.
Retrospective review of MRIs.
MRI scans of patients with PML were randomly interspersed with those of controls with a total of 227 MRI examinations available for blinded review. All the MRI examinations (n = 227) were examined in a fully randomized order by 3 independent radiologists (J.H., C.D., and X.L.). The readers were unaware of the clinical or biological data (patient with PML or not).
First, for each patient, they were asked to assess the presence or the absence of the PP (0: absent and 1: present) defined as at least 3 adjacent punctiform (less than 5 mm) brain lesions, hyperintense using T2-weighted and FLAIR MRI sequences or enhancing on T1-weighted postcontrast images. Nonenhancing PP was defined as punctate lesions only visible on T2-weighted and FLAIR images, whereas enhancing PP was defined as punctate lesions enhancing on postcontrast T1-weighted images. Number of punctate lesions visible was assessed according to a 4-point scale (0: absence of punctate lesions, 1: from 3 to 10 punctate lesions, 2: from 10 to 20 punctate lesions, 3: more than 20 lesions) as well as the presence of punctate lesions located in the immediate vicinity of a larger PML lesion (i.e., the milky way appearance) (figure 2).
Numerous punctate lesions, hyperintense on T2-weighted images, are visible in the immediate vicinity of the PML lesion (A, arrows) with a specific star-like distribution (milky way appearance). Punctate lesions were also observed within the pons (B, arrows).
The blinded readers resolved discrepancies in consensus regarding the presence of PP and assessed the longitudinal changes on consecutive MRI of patients with PML using all the clinical data available.
Statistical analysis.
Statistical analyses were performed using SAS software version 9.3 (SAS Institute Inc., Cary, NC). Interobserver agreement among the 3 blinded readers was calculated for the presence of PP (enhancing or not) using all MRI available using the Cohen κ test. The κ values were interpreted as follows: κ value of 0 indicates poor agreement; κ value of 0.01–0.20, minor agreement; κ value of 0.21–0.40, fair agreement; κ value of 0.41–0.60, moderate agreement; κ value of 0.61–0.80, good agreement; and κ value of 0.81–1, excellent agreement. We calculated the sensitivity and specificity of the PP for the diagnosis of presymptomatic NTZ-PML considering the 9 patients with PML scanned at the presymptomatic stage as cases and the 80 patients with RRMS and patients with CIS without PML as controls.
Standard protocol approvals, registrations, and patient consents.
This retrospective study was approved by our institutional review board, which waived informed consent.
RESULTS
Interobserver agreement.
The interobserver agreement among the 3 blinded readers for the presence of PP (enhancing or not) was good, κ = 0.79 (0.72–0.87). There was no disagreement between readers using T1-weighted postcontrast images. Disagreements were related to PP misdiagnosed by 1 or 2 of the blinded readers using 3D FLAIR or T2-weighted images.
Overall MRI findings.
Results are summarized in the table. The PP was exclusively observed in PML patients (18 of the 20 included) involving 24 brain areas: frontal (right n = 4, left n = 5) and parietal (right n = 3, left n = 6) lobes, pons (n = 3), middle cerebellar peduncles (right n = 1, left n = 2). The PP was present in all the patients with NTZ-PML (n = 14) while there was no PP detected in the pathologic control group (i.e., patients with CIS or RRMS without PML). Of the 6 patients with PML not related to NTZ, 4 had PP including 3 treated with IST. There was no PP in 2 patients with PML for whom MRI were not available at the presymptomatic and symptomatic stages: 1 treated with IST (patient 15) and 1 with leukemia (patient 20).
Punctate pattern according to PML stage
The number of punctate lesions varied according to the PML stage: maximal during IRIS, minimal at presymptomatic and chronic stages. Fourteen of the 18 patients with PML with PP exhibited more than 20 punctate lesions including 11 during IRIS.
The milky way appearance (i.e., punctate lesions at the vicinity of a larger brain PML lesion) was exclusively observed at the presymptomatic and symptomatic stages in 5 and 11 patients, respectively (figure 2).
Precision of PP for the diagnosis of NTZ-PML at the presymptomatic stage.
Results are summarized in the table. At the presymptomatic stage, the PP was observed in 7 of the 9 patients scanned: 5 with nonenhancing PP and 2 with enhancing PP. PP was the unique imaging manifestation for 2 presymptomatic patients: patient 10 with enhancing PP within the right corona radiata (figure 3) and patient 11 with nonenhancing PP within the pons (figure 4). In the 5 other presymptomatic patients with NTZ-PML, the PP was associated with a larger PML lesion (milky way appearance) involving the subcortical region (n = 4, figure e-1 on the Neurology® Web site at Neurology.org) or the left middle cerebellar peduncle (n = 1, figure 2). Using PP for the diagnosis of presymptomatic NTZ-PML, there were 7 true-positives, 2 false-negatives, 80 true-negatives, and no false-positives (78% sensitive, 95% CI 40%–96% and 100% specific, 95% CI 94%–100%).
At the presymptomatic stage, 3 enhancing punctate lesions were observed within the right corona radiata and were the unique imaging manifestation of NTZ-PML (A, arrows). At the PML–immune reconstitution inflammatory syndrome stage, the number of enhancing punctate lesions increased significantly and some of the lesions tended to coalesce (B, arrows). New punctate lesions appeared in the left contralateral hemisphere (B, arrowheads). Note the improved detection of enhancing lesions using maximal intensity projection view (C, arrows and arrowheads). At the chronic stage, the number of enhancing punctate lesions decreased significantly (D, arrows).
In December 2012, a T2-weighted hyperintense punctate lesion appeared within the pons (A, arrow). A new MRI was performed 3 weeks later (January 2013), showing 2 additional T2-weighted hyperintense punctate lesions without enhancement (i.e., nonenhancing PP, arrows) while the patient was asymptomatic (B, arrows). PCR of the JC virus was positive within the CSF, confirming the diagnosis of NTZ-PML. MRI performed during immune reconstitution inflammatory syndrome (March 2013) revealed numerous punctate and nodular lesions on fluid-attenuated inversion recovery images (C, arrows).
Occurrence of PP at the symptomatic, IRIS, and chronic PML stages.
At the symptomatic stage, the PP was observed in 13 of the 15 patients scanned: 8 with nonenhancing PP (75%) and 5 with enhancing PP (25%).
At the PML-IRIS stage, the number of lesions visible on postcontrast images increased significantly, with enhancing PP being observed in 16 of the 20 patients scanned. The enhancing punctate lesions tended to coalesce with nodular or rim-like enhancement (figure e-2). A postmortem pathologic specimen was obtained in one case (patient 2) at the IRIS stage with enhancing PP revealing a massive perivascular lympho-plasmocytic infiltration (figure e-3).
At the chronic stage, there was no nonenhancing PP detected. Enhancing PP was observed at the chronic stage in 7 patients but the number of punctate lesions tended to decrease, as shown in figures 2 and 3.
Occurrence of PP in patients with PML with negative JC virus PCR in the CSF.
PP was detected in the 3 patients for whom JC virus PCR in the CSF was negative: from presymptomatic to chronic stages for patient 10, at symptomatic and IRIS stages for patient 12, and from symptomatic to chronic stages for patient 18. Indeed, the absence of PP does not seem related to a negative JC virus PCR in the CSF.
DISCUSSION
The value of PP had never been investigated for the diagnosis and follow-up of PML. The PP was visible in all patients with NTZ-PML and not in controls without PML, suggesting its high specificity. The PP may be the first imaging manifestation of PML preceding the emergence of typical PML lesions. It does not appear specific to NTZ-PML as it also occurred in PML from other causes.
NTZ-PML lesions are typically subcortical.7 Interestingly, in our study, PP also involved the deep white matter, suggesting that special attention must be paid to the whole brain when screening MRI for PML lesions. There was a high incidence of the PP in our study compared to previous data. We may hypothesize that this could be an issue of the definition of the PP used in this study (i.e., at least 3 adjacent punctiform brain lesions). Indeed, the first study describing this imaging finding was rather referring to the fact that there are many of these punctate lesions.7 This may also be partly due to the study design. Indeed, all the patients were scanned at 3T with a 32-channel head coil using 3D high-resolution imaging to achieve FLAIR and postcontrast T1-weighted imaging (i.e., millimetric and isotropic voxel).21 There are several potential advantages of using 3D thinner slices for the detection of punctate lesions such as decrease in partial volume effect,22,23 coregistration of the 3D images,24 or isotropic voxel size allowing for maximum intensity projection reformations in various planes.23 The good interobserver agreement for the detection of PP on MRI also suggested its relevance in clinical practice.
The reason why the PP was not detected in 2 patients with PML (patients 15 and 20) remains unclear. In these patients, MRI examinations were not available at early stages, which may contribute to the absence of PP detected, particularly on T2-weighted and FLAIR images. In our study, both nonenhancing and enhancing PP were identified, in agreement with previous studies.7,8 Nonenhancing punctate lesions occurred mainly at the early PML stages, associated with a larger PML lesion (milky way appearance). Such a finding could be particularly relevant to differentiate early PML lesions from MS plaques. Most of these punctate lesions are associated with enhancement during IRIS, suggesting an early immune response in the perivascular spaces that will become more prominent during IRIS7 or alternatively a productive JC virus infection. Further studies are required to assess the prognostic value of PP in presymptomatic patients with PML. Less frequently, enhancing punctate lesions may occur at the early stages, suggesting an inflammatory PML, as NTZ may not completely impair the extravasation of lymphocytes through the blood–brain barrier.25 Such inflammatory PML should be distinguished from PML-IRIS following the NTZ withdrawal, and associated with clinical worsening, mass effect, and nodular/rim-like enhancement on MRI. The underlying cause of PP remains unclear. Our pathologic case revealed a massive lympho-plasmocytic infiltration of perivascular spaces, in agreement with previous reports.11,12 Indeed, the PP is histopathologically characterized by a pronounced inflammatory infiltrate with a predominance of CD8+ T cells and high numbers of plasma cells within PML lesions as well as adjacent gray and white matter.26
Our study had some limitations. Most of the knowledge regarding PML is derived from the HIV-PML population. However, the incidence rate of HIV-related PML dramatically decreased during the last few years owing to the high number of patients treated with highly active antiretroviral therapy. In our institution, during the enrollment period from 2011 to 2015, there was no patient with HIV-related PML identified. The only way would be to include patients with HIV-PML who underwent MRI before the inclusion period, with a MRI protocol including 2D MRI sequences acquired at 1.5T. Such approach may lead to a major bias since we would compare 2D MRI sequences at 1.5T with 3D high-resolution MRI sequences at 3T. As we focused on the diagnosis of NTZ-PML, we did not include patients with other potential causes of punctate patterns such as hematologic diseases or vasculitis. Image interpretation cannot be completely blinded particularly at the delayed stage of the disease when the imaging pattern is highly suggestive of PML. However, we evaluated the diagnostic precision of PP (i.e., sensitivity and specificity) for the diagnosis of PML at the presymptomatic stage when PML lesions may be very similar to MS plaques. Moreover, diagnosis of PML at the presymptomatic stage improves the survival and functional outcome.4,5 Diffusion-weighted imaging (DWI) is helpful for the diagnosis of PML. We performed DWI in all the patients included, with or without PML. On diffusion-weighted images, the PP consisted of hyperintense punctiform lesions, mainly observed during the symptomatic and IRIS stages. However, due to the low spatial resolution inherent to this technique, comparison of DWI with other 3D high-resolution MRI sequences may have introduced a substantial bias. This explains why we did not include these data in the present study. Finally, the control group used in our study did not perfectly match the group of patients with PML in terms of sex, disease severity, and treatment.
The PP is a sensitive imaging feature of NTZ-PML and may be of use to differentiate PML lesions from MS plaques. However, further studies are required to assess its incidence in patients with PML from other causes, particularly in the setting of AIDS or vasculitis.
AUTHOR CONTRIBUTIONS
Jérôme Hodel: data acquisition and analysis, manuscript editing. Olivier Outteryck: data analysis manuscript editing. Christine Darchis, Vincent Deramecourt, Sébastien Verclytte, Marc Zins, Jean-Pierre Pruvo: data acquisition and analysis. Arnaud Lacour: data acquisition. Patrick Vermersch: data analysis. Xavier Leclerc: data analysis, manuscript editing.
STUDY FUNDING
No targeted funding reported.
DISCLOSURE
J. Hodel and C. Darchis report no disclosures relevant to the manuscript. O. Outteryck has received funding for travel from Biogen Idec, Bayer Schering Pharma, Merck Serono, Novartis, Teva Pharmaceutical Industries Ltd., and Sanofi-Aventis and for speaker honoraria from Bayer Schering Pharma, Biogen Idec and Sanofi-Aventis. S. Verclytte and V. Deramecourt report no disclosures relevant to the manuscript. A. Lacour has received funding for travel from Biogen Idec, Genzyme, and Sanofi-Aventis and for speaker honoraria from Genzyme. M. Zins and J. Pruvo report no disclosures relevant to the manuscript. P. Vermersch serves on scientific advisory boards for Biogen Idec, Bayer Schering Pharma, Merck Serono, Novartis, Teva Pharmaceutical Industries Ltd., and Sanofi-Aventis; has received funding for travel and speaker honoraria from Biogen Idec, Bayer Schering Pharma, Novartis, Teva Pharmaceutical Industries Ltd., Sanofi-Aventis, and Merck Serono; and receives research support from Biogen Idec, Merck Serono, Sanofi-Aventis, Teva Pharmaceutical Industries Ltd., and Bayer Schering Pharma. X. Leclerc reports no disclosures relevant to the manuscript. Go to Neurology.org for full disclosures.
Footnotes
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Supplemental data at Neurology.org
- Received June 26, 2015.
- Accepted in final form January 7, 2016.
- © 2016 American Academy of Neurology
REFERENCES
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.