Intraocular sarcoidosis: association of clinical characteristics of uveitis with findings from 18F-labelled fluorodeoxyglucose positron emission tomography
- Ahmed Rahmi1,5,
- Emmanuel Deshayes2,
- Delphine Maucort-Boulch3,5,
- Loig Varron4,5,
- Jean Daniel Grange1,5,
- Laurent Kodjikian1,5,
- Pascal Seve4,5
- 1Department of Ophthalmology, Hôpital Croix-Rousse, 103 Grande Rue de la Croix-Rousse, Lyon, France
- 2Department of Nuclear Medicine, Hôpital Edouard Herriot, Place d'Arsonval, 69437 Lyon, France
- 3Laboratoire Biostatistique Santé, Centre Hospitalier Lyon-Sud, Pierre-Bénite, Lyon, France
- 4Department of Internal Medicine, Hôpital Croix-Rousse, 103 Grande Rue de la Croix-Rousse, 69317 Lyon Cedex 04, France
- 5Université Claude Bernard Lyon 1, 43 Boulevard 11 Novembre 1918, Villeurbanne, France
- Correspondence to Pascal Sève, Department of Internal Medicine, Hôtel Dieu, 1 place de l'Hôpital, 69288 Lyon Cedex 02, France;
- Accepted 24 February 2011
- Published Online First 3 April 2011
Aims To assess specific clinical criteria in patients with uveitis that are related to signs of sarcoidosis on 18F-labelled fluorodeoxyglucose positron emission tomography (18F-FDG-PET).
Methods Retrospective study of 54 consecutive patients with chronic uveitis for whom a PET scan was done because of suspected sarcoidosis, between July 2004 and December 2009. All the patients underwent a clinical examination, biological tests and a high-resolution CT of the chest.
Results 17 of the 54 patients (31.5%) presented hypermetabolic foci on 18F-FDG-PET scan consistent with sarcoidosis. Among them eight patients (14.8%) underwent biopsy showing non-caseating granuloma. At the end of the study, 10 patients (18.5%) were considered as having a presumed sarcoidosis and seven patients (12.9%) as having indeterminate sarcoidosis. The increasing age at the diagnosis of uveitis (p=0.01), the presence of posterior synechiae (p=0.01) and a positive high-resolution CT of the chest (p=0.01) were significantly related to an abnormal PET scan.
Conclusions Increasing age at the diagnosis of uveitis, the presence of posterior synechiae and the positivity of high-resolution chest CT are associated with 18F-FDG PET scan signs consistent with sarcoidosis.
- Uveitis, sarcoidosis
- PET scan
- fluorine-18 fluorodeoxyglucose positron emission tomography
- diagnostic tests/investigation
Ocular sarcoidosis is one of the commonest causes of uveitis and ocular involvement is often seen in patients with systemic sarcoidosis.1 In uveitis patients, sarcoidosis is probably underestimated depending on the extent of the work-up diagnosis. The evaluation of patients with uveitis and suspected sarcoidosis can be staged from non-invasive laboratory and radiological tests to invasive ones, depending on the ease or difficulty of the diagnosis.2 Initial assessment consists of intradermal skin tests, serum angiotensin converting enzyme (ACE) measurement, chest radiograph or chest CT. At this stage, the biopsy of clinically suspicious, easily accessible tissues (conjunctiva, skin) is advised. If these initial studies are negative, 67Ga-labelled gallium citrate scintigraphy (67Ga scan) should be performed. If the chest radiograph, chest CT or 67Ga scan are characteristic of sarcoidosis, then bronchoalveolar lavage (BAL) and transbronchial lung biopsy are advised. An open lung biopsy is reserved for patients with radiological evidence of sarcoidosis in whom histological proof is lacking. The definitive diagnosis of sarcoidosis requires histological confirmation. The biopsy of such non-ocular tissues, however, is often unacceptable to many patients with uveitis. Therefore, some authors have developed criteria for the diagnosis of suspected sarcoidosis-associated uveitis in the absence of histological proof.3 An international workshop has recently established criteria to make a diagnosis of ‘intraocular sarcoidosis’ (sarcoidosis uveitis) on the basis of a combination of suggestive ophthalmological findings and laboratory tests when a biopsy is not carried out or is negative.4
18F-labelled fluorodeoxyglucose positron emission tomography (18F-FDG-PET) is a functional imaging modality that, in contrast to anatomical imaging modalities such as CT, is based on increased glucose metabolism. In sarcoidosis, 18F-FDG-PET can be of value in solving two vexing clinical problems: (1) identifying occult sites of disease for diagnostic biopsy; and (2) assessing activity of disease, especially in patients with pulmonary fibrosis. Indeed, 18F-FDG-PET has shown its usefulness for the diagnosis of isolated extrapulmonary sarcoidosis, including neurosarcoidosis and cardiac sarcoidosis, by demonstrating suggestive 18F-FDG-PET uptake in the patients despite normal chest x-ray or chest CT.5–8 Moreover, several studies have shown that 18F-FDG-PET appeared to be more accurate than 67Ga scan, and contributes to a better evaluation of both extrapulmonary and pulmonary involvement in sarcoidosis patients.9–12 In a recently published prospective study, Keijsers et al measured sensitivity of both 67Ga imaging and 18F-FDG-PET in the assessment of sarcoidosis in 34 newly, diagnosed histologically proven patients.13 Overall sensitivity to detect active sarcoidosis was 88% for 67Ga scan and 97% for 18F-FDG-PET. 18F-FDG-PET detected more lesions in the mediastinum, hila, lymph nodes and extrapulmonary regions. Given these data, 18F-FDG-PET is considered as the nuclear imaging technique of choice in sarcoidosis assessment.13–15 18F-FDG-PET seems also to be promising to follow treatment efficacy in patients with sarcoidosis.16 Only very few publications are available reporting results of diagnosis in ocular sarcoidosis.17–19 In an expert panel asked to address some of the most important issues related to the diagnosis and treatment of ocular sarcoidosis, 24% had some experience of this method.20 We have previously shown that 18F-FDG-PET may show focal uptake suggestive of sarcoidosis in patients with unexplained uveitis and normal chest CT.17
In the present study we assessed which clinical and laboratory characteristics, in patients with uveitis, are related to signs consistent with sarcoidosis on 18F-FDG-PET scans.
Patients and methods
This was a retrospective study performed at the University Hospital of Lyon. We have included all the 54 18F-FDG-PET scans of patients ordered either for evaluation of unexplained chronic uveitis or for a suspicion of ocular sarcoidosis, at our tertiary ophthalmology centre, between July 2004 and December 2009. The criteria that defined the cases as being suspected of sarcoidosis were one or more of the followings: ophthalmological features suggestive of sarcoidosis, such as granulomatous uveitis and peripheral multifocal choroiditis in the absence of other diagnosis, elevated serum ACE and chest CT findings suggestive of sarcoidosis.
All the patients underwent a standard screening protocol that included a medical history, a medical examination, tuberculosis skin test, and laboratory tests for syphilis and HIV, Borrelia burgordeferi, Bartonella henselae and human T-lymphotropic virus type 1 (HTLV-1) in patients at risk. Some patients underwent an interferon α release assay for tuberculosis (QuantiFERON Tuberculosis Gold-Cellestis Ltd, Chadstone, Victoria, Australia) from January 2007, when it was available at our institution. Uveitis was classified anatomically according to the International Uveitis Study Group criteria.21
For all the patients, the diagnostic battery for sarcoidosis included an serum ACE assay, measurement of liver enzymes, a salivary gland biopsy and chest CT. Some patients had conjunctiva or skin biopsy if clinically suspicious features were present. Some patients underwent a transbronchial lung biopsy, BAL and 67Ga scan.
The criteria that defined CT as positive for sarcoidosis were: lymph nodes at left and right hilum and mediastinum with a short axis diameter >10 mm, peri-lymphatic pulmonary nodules and other parenchymal lung abnormalities.3 22 23 BAL was considered consistent with pulmonary sarcoidosis if lymphocytes represented >15% of the total cells and were predominantly CD4+, and if the CD4:CD8 cell ratio was >3.5.24
None of the patients was receiving systemic corticosteroids at the time of the 18F-FDG-PET examination or had been given them in the previous 3 months. Five 18F-FDG-PET examinations were performed with an ECAT EXACT HR+ camera (Siemens/CTI Inc, Malvern, PA, USA.) and 49 with a dedicated PET/CT scanner (Philips, Gemini, Andover, MA, USA). The images were reviewed retrospectively by a nuclear medicine physician (ED), blinded to the patient's clinical history other than presence of uveitis. We used a visual qualitative method to consider the PET examinations as positive or not for sarcoidosis. The 18F-labelled-FDG uptake in the following sites was specifically reported on a special sheet as normal or abnormal (focal or diffuse non-physiological 18F-labelled-FDG uptake exceeding that of the adjacent normal tissue): mediastinal lymph nodes, extra-mediastinal lymph nodes, lung, spleen, liver, muscle, skin, parotidal or lacrymal glands, bone.
Theses sites were scored either positive or negative (positive = increased metabolic activity, negative = no increased metabolic activity).13 For 18F-FDG-PET scans with combined CT acquisition, a first reading was performed blind to CT findings and a second with reference to them.
In the final analysis, we used criteria similar to those of Abad et al3 for the diagnosis of sarcoidosis associated with uveitis excluding 67Ga scintigraphy and adding the 18F-FDG-PET. Sarcoidosis was considered proven when the histological examination of the biopsy material showed a non-caseating granuloma, presumed when the patients met at least two of the three following criteria (BAL, 67Ga scan and/or 18F-labelled-FDG, and ACE) but their biopsy results were normal or unavailable, and indeterminate when only one criterion was met.
The outcome was the result of the 18F-FDG-PET classified as normal or abnormal as described previously. The characteristics of the patients with a positive PET scan suggestive of sarcoidosis were compared with the characteristics of the patients with a negative PET scan using χ2 or t tests. Univariate and multivariate Cox regressive analysis were subsequently used to determine the association between signs consistent with sarcoidosis on 18F-FDG-PET scans and the clinical and laboratory characteristics of the patients. p Values <0.05 were considered significant. Data were analysed using the statistical software R (R development core team. R: A Language for Environmental and Statistical Computing. Vienna: R Foundation, 2008.)
A summary of the general characteristics of the 54 patients involved is shown in table 1. Twenty-three patients underwent 18F-FDG-PET for unexplained uveitis while 31 had suspicion of ocular sarcoidosis because of either ocular features (n=22), elevated ACE and/or lysozyme (n=9) or CT findings (n=8) suggestive of sarcoidosis. All the patients received a chest CT within the 3 months before the PET scan. Sixteen of these patients have been reported elsewhere.17 Three patients had positive PPD tuberculin skin tests (>10 mm). QuantiFERON tuberculosis testing was performed in 17 patients and was negative in all but one case. No patients proved to have active tuberculosis as their cause of their ophthalmic findings during the follow-up period.
Whole-body 18F-FDG-PET revealed hypermetabolic foci in the mediastinal, lung hilar lymph nodes, salivary or lachrymal glands suggestive of sarcoidosis in 17 patients (31.5%) (figure 1). Table 2 shows the results of these 17 extracardiac 18F-FDG-PET scans according to inflammatory regions. There was only one discrepancy between the PET scan re-assessment and the original report, demonstrating that the signs suggestive of sarcoidosis were relatively easy to assess with our PET scan protocol. Ten of the patients with a positive PET scan suggestive of sarcoidosis had a normal chest CT.
The characteristics of the patients with negative and positive PET scan suggestive of sarcoidosis are presented in table 3. In the univariate analysis (table 4), a positive PET scan suggestive of sarcoidosis was related to patient age >56 years (p=0.01), an abnormal chest CT (p=0.003) and a posterior synechia (p=0.04). There was a trend for an association between the presence of peripheral multifocal choroiditis and a positive PET scan (p=0.07). Multivariate regression analysis (table 5) showed that patients >56 years old (p=0.01), with abnormal chest CT (p=0.01) and posterior synechia (p=0.01) were likely to show significantly abnormal PET scan. Age >56 years yielded a hazard ratio of 1.09 (95% CI 1.02 to 1.165), abnormal high-resolution CT yielded a hazard ratio of 17.83 (95% CI 1.66 to 191.90) and posterior synechia yielded a hazard ratio of 11.50 (95% CI 1.73 to 76.49).
In eight of 17 patients with a positive PET scan suggestive of sarcoidosis (table 1), the histological examination (obtained by hilar lymph node biopsy after mediastinoscopy in four cases, bronchial biopsy in one case, supraclavicular lymph node biopsy in one case, skin biopsy in one case and conjunctival granuloma biopsy appeared during follow-up in one case) revealed a non-caseating granuloma without evidence of infection or neoplasm, and these eight patients were therefore classified as having proven sarcoidosis. Ten patients were classified as presumed sarcoidosis and seven patients were classified as indeterminate sarcoidosis.
The final diagnosis in 33 patients with a negative PET scan included ocular lymphoma (n=1), toxoplasmosis (n=1), acute tubulointerstitial syndrome and uveitis and not anterior uveitis syndrome (n=1) and Birdshot's disease (n=1). For one patient with a normal PET scan, a renal biopsy revealed a non-caseating granuloma, confirming the diagnosis of sarcoidosis. Six patients with a negative PET scan were considered as having an indeterminate sarcoidosis according to the Abad's et al 3. criteria because of the positivity of one of other diagnostic tests. The 22 remaining patients were finally considered as having idiopathic uveitis.
Of the 17 patients with a positive PET scan consistent with sarcoidosis, 9 received systemic treatment with corticosterois (oral, intravenous), 3 had corticosteroids and methotrexate (due to a severe ocular involvement and/or steroid dependence), while the 5 remaining patients were treated locally and/or by periocular steroid injections.
Our study is the first to assess specific clinical criteria in patients with uveitis that are related to findings consistent with sarcoidosis on 18F-FDG-PET. In a previous study, we offered 18F-FDG-PET as a second-step examination in most patients with unexplained chronic uveitis seen in our uveitis clinics.17 All the patients had previously had high-resolution CT with findings not indicative of sarcoidosis. PET showed significant abnormal 18F-labelled FDG uptake in 11 of 19 patients. In the end, subsequent mediastinal biopsies proved sarcoidosis in three patients, two patients were considered as presumed sarcoidosis and six as indeterminate sarcoidosis. Interestingly, seven of the nine middle-aged or elderly women (>55 years old) had positive 18F-FDG-PET.
In the present study, we found that elderly age, posterior synechia and positive chest CT were significantly associated with PET scan findings suggestive of sarcoidosis. Several studies showed that chest CT was predominantly positive in elderly patients, especially in women.22 23 In our series, 15 of 17 patients (88.2%) with an abnormal PET scan suggestive of sarcoidosis were >50 years at the diagnosis of chronic uveitis and 12 of the 17 patients were female. It is hypothesised that young patients with ocular sarcoidosis, who often present an acute anterior uveitis, are diagnosed by different means and without the use of chest CT, whereas sarcoidosis at an elderly age is more often associated with posterior uveitis and might be accompanied by a less frequent pulmonary involvement.25 This hypothesis is in accordance with our findings since all our patients underwent both chest x-ray and CT before the PET scan.
Interestingly, by using methods very similar to ours, Clement et al found that increasing age, the presence of multifocal chorioretinitis and of posterior synechia were associated signs of sarcoidosis on chest CT in 50 patients with uveitis.23 Peripheral multifocal chorioretinitis is an ocular disorder characterised by peripheral punched-out lesions associated with intraocular inflammation and has already been previously linked to sarcoidosis.3 23 In our study, we found a trend for an association between multifocal chorioretinitis and abnormal PET scan that was not confirmed in multivariate analysis. Posterior synechia are typical for chronic anterior uveitis and are frequently found in chronic types of sarcoidosis.23 26 As in the study of Clément et al,23 some laboratory tests that were reported to be typical for sarcoidosis, such as serum ACE, calcium levels and abnormal BAL, were not associated with an abnormal PET scan. The study of Clément et al was limited by the fact that they did not use systematic criteria for the diagnosis of sarcoidosis.23
Despite its probable higher accuracy for the diagnosis of sarcoidosis, 18F-FDG-PET cannot be advocated as the primary screening test in patients with uveitis because of its higher cost and dosimetric considerations. Moreover, so far, a systematic study comparing the diagnostic value of chest x-ray, CT and PET scan in a large group of patients with uveitis is not available and specific recommendations on the role of 18F-FDG-PET cannot be made. Our retrospective study is a first step in recognising specific criteria associated with a high chance of positive PET scan. We are currently conducting an open, randomised (by clusters of hospitals) study named ULISSE (Uveitis: medico-economical and cLInical evaluation of a Standardised Strategy for an Etiological diagnosis) whose primary purpose is to assess the efficiency of a standardised strategy for uveitis diagnosis in comparison with the usual one named ‘free’ (http://clinicaltrials.gov/ct2/show/NCT01162070?term=uveitis&rank=3). The standardised diagnostic algorithm is a specific procedure adapted to the anatomo-clinical type of the uveitis and assesses the role of chest X-ray (CXR) and CT in granulomatous uveitis and multifocal choroiditis, and subsequently the role of PET scanning in patients with normal or non-contributive procedures.
Our study has several limitations. The first is the lack of histological gold standard for the diagnosis of sarcoidosis. Only eight of the 17 patients with an abnormal PET scan suggestive of sarcoidosis had a histological proof of sarcoidosis. However, it was difficult to offer mediastinoscopy with lymph node biopsy for all patients, especially for patients without severe ocular inflammation or for elderly patients with significant comorbidities. In addition, the independent risk factors identified apply only to patients similar to those defined in the inclusion criteria and seen in the type of clinic of our study.
In conclusion, our study shows that in patients with uveitis at an old age, and the presence of posterior synechia and of abnormal chest CT increase the probability of 18F-FDG-PET findings suggestive of sarcoidosis. Given the high cost of 18F-FDG-PET and dosimetric considerations, we think that it could be first offered to elderly patients with normal high-resolution CT and suggestive ophthalmological findings. Further controlled studies involving a larger number of patients are now required to clarify the clinical usefulness of 18F-FDG-PET in patients with suspected sarcoidosis associated with unexplained uveitis.
Study performed at Hospices Civils de Lyon, Quai des Célestins, 69288 Lyon Cedex 02, France.
Competing interests None
Ethics approval This study was conducted with the approval of the Ethical Review Board of the University of Lyon.
Provenance and peer review Not commissioned; externally peer reviewed.