Primary Intraocular LymphomaImproving the Diagnostic Procedure
Anni Karma, MD, PhD,1 Eva O. von Willebrand, MD, PhD,2 Petri V. Tommila, MD, PhD,1
Anders E. Paetau, MD, PhD,3 Pertti S. Oskala, MD,1 Ilkka J. Immonen, MD, PhD1
Objective: To analyze the clinical features of primary intraocular lymphoma (PIOL) and to describe cyto-chemical and immunocytochemical findings of the vitreous specimens as well as the reasons for delayeddiagnosis of PIOL.
Design: Prospective noncomparative study.Participants: Eleven patients referred to the uveitis or medical retina units, Department of Ophthalmology,
University of Helsinki, were diagnosed as having PIOL between 2000 and 2005. The median follow-up of thepatients was 32 months.
Methods: Clinical features and diagnostic workup of uveitis were described. Twelve vitrectomies wereperformed on 9 patients. The first 5 biopsies were fixed in an equal volume of 50% alcohol. The specimens ofthe next 7 vitrectomies were handled without alcohol, and tissue culture medium was added to the samples.
Main Outcome Measures: Clinical features of PIOL, intervals from ocular symptoms and from first ophthalmo-logical examination to diagnosis, and the role of a proper handling of the vitreous sample in the diagnosis of PIOL.
Results: Six females (54%) and 5 males (46%) (median age, 61 years) were included. Ten patients had ocularsymptoms for 1 to 30 months (median, 8) before the first contact with an ophthalmologist. Uveitis was bilateralin 9 patients. Vitreitis was seen in all patients, and it was severe in 8. Fundus lesions dominated in 3 patients. Sixpatients lost useful vision in one eye before the diagnosis of PIOL. Cytologic and immunohistochemical stainingsprepared of the unfixed vitreous specimens showed PIOL in 6 patients. The samples fixed in alcohol werenondiagnostic in 4 patients, and in them, verification of diagnosis was based on brain biopsy (3) or cerebrospinalfluid (1) findings. Seven patients died due to primary nervous system lymphoma.
Primary intraocular lymphoma (PIOL) is a subset of B-cellnon-Hodgkin’s lymphoma of the nervous system, primarynervous system lymphoma (PNSL). Although it is still arare disease, the incidence of PNSL has tripled over 3decades.1 Systemic spread of PNSL outside the nervoussystem occurs rarely. However, recent studies suggestthat the malignant event occurs in a lymphocyte outsidethe nervous system, in an extraneural germinal center,but the nervous system and the eye are the only sites ofdetectable disease, according to currently available diagnos-tic modalities.2–4
Originally received: April 3, 2006.Accepted: November 1, 2006. Manuscript no. 2006-389.1 Department of Ophthalmology, University of Helsinki, Helsinki, Finland.2 Transplantation Laboratory, University of Helsinki, Helsinki, Finland.3 Department of Pathology, University of Helsinki, Helsinki, Finland.
Correspondence to Anni Karma, MD, PhD, Department of Ophthalmology,University of Helsinki, FIN-00029 Helsinki, Finland. E-mail: [email protected].
Reprint requests to Petri V. Tommila, MD, PhD, Department of Ophthal-
mology, University of Helsinki, FIN-00029 Helsinki, Finland.
Ocular involvement is found in 20% to 25% of patientswith PNSL.5,6 Primary intraocular lymphoma masqueradesas different uveitis entities. In uveitis surveys, it has rarelybeen recognized: 2 cases of PIOL were diagnosed among1300 consecutive uveitis patients in a tertiary survey fromthe northeastern United States,7 and 9 cases were foundamong 828 consecutive uveitis patients referred to a tertiaryophthalmologic center in The Netherlands.8
Because of the rarity of the disease, vague ocular symp-toms, diverse clinical picture, and the demanding procedurefor a successful vitreous specimen, the diagnosis of PIOL isa great challenge to ophthalmologists and cytopathologists.The preparation of vitreous specimens for cytological eval-uation varies,9–14 and multiple specimens may be requiredbefore an unequivocal diagnosis of PIOL can be made.11–13
More than 40% of the vitrectomy specimens may remainnondiagnostic.13,15
The disease has a poor prognosis. Patients die of cerebraltumor growth after a mean survival of 20 months from thediagnosis of the ocular disorder.15 However, as advances in
the management of PNSL and PIOL have resulted in im-
ISSN 0161-6420/07/$–see front matterdoi:10.1016/j.ophtha.2006.11.009
Karma et al � Primary Intraocular Lymphoma
proved survival and preservation of vision, prompt diagno-sis will become increasingly critical.16–19
We report our experience with 11 patients with PIOLdiagnosed since March 2000 in the Department of Ophthal-mology, Helsinki University Central Hospital. We describethe reasons for delayed diagnosis of PIOL and the processof improving the diagnosis.
Patients and Methods
Between March 2000 and December 2005, the diagnosis of PIOLwas established in 11 patients initially sent to the uveitis and/ormedical retina units of the Department of Ophthalmology, HelsinkiUniversity Central Hospital. Nine patients were referred fromeither practicing ophthalmologists (6 patients) or other eye hospi-tals in Finland (3 patients). Two patients (nos. 1 and 11; Table 1),in whom PNSL already had been diagnosed, were referred fromthe Department of Oncology, Helsinki University Central Hos-pital. One patient (no. 4; Table 1) referred by a practicingophthalmologist had gone through a neurological examinationbecause of facial palsy, hemiparesis, and periodic severe head-aches suspected to be due to recurrent herpetic meningitis. Allpatients were Caucasian, and all were immunocompetent. Hu-man immunodeficiency virus infection was not an exclusioncriterion in our study.
The ophthalmological examination included slit-lamp biomi-croscopy and fundus examination with indirect ophthalmoscopyand 3-mirror lens examination after pupillary dilatation. Uveitiswas categorized anatomically according to the Uveitis Interna-tional Study Group recommendations,20 and vitreous reaction wasgraded according to the scale described by Nussenblatt and Pal-estine.21 Laboratory and ancillary tests were taken on the basis ofthe differential diagnosis compiled for each case. In most patients,the tests included complete blood count, angiotensin-convertingenzyme, antiborrelial and antinuclear antibodies, rheumatoid fac-tor, liver function tests, and chest radiography. Magnetic resonance
ARN � acute retinal necrosis; CSF � cerebrospinal fluid; ND � not do*Died due to progressive primary nervous system lymphoma of the brain.
First vitrectomy had been performed in another hospital.
imaging of the head and examination of cerebrospinal fluid as apart of the diagnostic workup of uveitis were performed on 5patients. Six patients (nos. 1 and 3–7; Table 1) received topicaland/or systemic corticosteroids for a period of 1 to 5 months.Additionally, patients 4 and 5 received intravenous and oral acy-clovir or valavir therapy, and patient 8 had received intravenousceftriaxone followed by oral amoxicillin before referral to us. Twopatients’ uveitis responded temporarily to oral corticosteroid ther-apy (patients 3 and 4), but the remaining patients were nonrespon-sive to corticosteroids and antimicrobials (Table 1).
Twelve vitrectomies were performed on 9 patients. On 1 patient(no. 8; Table 1) the vitrectomy had been performed in another eyehospital, but the result was nondiagnostic. For sampling of thevitreous specimen, a 3-port pars plana vitrectomy setup was used.One milliliter of undiluted vitreous core sample was aspiratedmanually through a 20- or 25-gauge vitreous cutter, with thecutting mode activated at 600 to 1200 cuts per minute. Theinfusion line was opened after sampling. The aspirated vitreousspecimens of patients 1, 3 (performed on both eyes), 4, and 5 (firstvitrectomy) were fixed in an equal volume of 50% alcohol andtaken immediately to the cytopathologist. The specimens wereprepared using a centrifuge (Cyto-Tek, Miles Scientific, Elkhart,IL).
Seven most recent vitreous specimens of 6 patients (patients5–10; Table 1) were handled in a different way: the aspiratedmaterial was emptied into a 10-ml glass tube, in which 5 to 6 mlof Rosewell Park Memorial Institute media 1640 (Life Technolo-gies Inc., Grand Island, NY) was added to improve cell viability.The vitreous specimen was carried immediately to the cytologylaboratory for processing. In 7 cases, the diluted specimen in thevitrector cassette was delivered without delay to the cytologylaboratory as well. The vitreous specimen first was centrifuged at200 rounds per minute for 10 minutes. After that, the cells werewashed and resuspended in the culture medium and counted. Thenthe cells were cytocentrifuged at 500 rounds per minute for 10minutes onto glass slides. One slide was stained with May–Grün-wald–Giemsa, and the rest were used for the immunostainings.Immunohistochemical staining was performed using a sensitive
indirect 3-layer immunoperoxidase method22 and monoclonal an-tibodies. The following monoclonal mouse immunoglobulin G1antibodies were used: CD2, CD20, CD30, � light chains, � lightchains, and, in some cases, CD68 (DakoCytomation Denmark A/S,Glostrup, Denmark). The antibodies were used at an optimal dilutionin tris(hydroxymethyl)aminomethane/1% bovine serum albumin. Af-ter incubation for 30 minutes in a humid chamber at room tempera-ture, the preparates were washed in tris(hydroxymethyl)aminometh-ane buffer and incubated for 30 minutes with peroxidase-labeledrabbit antimouse immunoglobulin (Dako P161, DakoCytomationDenmark), washed again, then incubated for 30 minutes with perox-idase-labeled goat antirabbit immunoglobulin (Caltag Laboratories,San Francisco, CA). The reaction was revealed by chromogen 3-ami-no-9-ethylcarbatzole containing hydrogen peroxide. The specimenswere counterstained with Mayer’s hemalum and mounted (Aqua-mount, BDH Ltd., Poole, United Kingdom).
In addition to the cytological and immunohistochemical anal-ysis, fractions of the aspirated fluid from patients 1, 3 to 7, and 10were sent to the microbiology laboratory. Gram’s stain and cul-tures for bacteria, fungi, and viruses were performed as well asgenomic amplification by polymerase chain reaction (PCR) to testfor the presence of herpes simplex virus, herpes zoster virus,herpes simplex 6 virus, cytomegalovirus, Toxoplasma gondii pro-tozoa, Borrelia burgdorferi bacteria, and mycobacteria.
Results
Patients’ demographic data, suggestions for the etiology of uveitisat the referral visit in our clinic, and the delay in the diagnosticworkup as well as the procedures on which the final diagnosis wasbased are presented in Table 1. There were 6 (54%) females and 5(46%) males. The median age of the patients at the first referralvisit was 61 years (range, 51–81).
Ten patients had had ocular symptoms for 1 to 30 months (median,8) before the first contact with an ophthalmologist, and 1 patient (no.11) had no ocular symptoms. In 6 patients, the symptoms were mildand consisted of blurred vision and floaters or shadows in the visualfields (VFs). In 3 patients, the main complaint was decreased vision inone eye during previous months. Two patients complained of rapidalterations of visual acuity (VA), and additionally, 1 patient com-plained of deterioration of color vision.
In 2 patients, VA was 20/20 in both eyes at the initial referralvisit. Visual acuity of the remaining patients was 20/40 or less inone or both eyes; in 4 of them, VA was counting fingers (CF) inone eye. During the diagnostic workup period, 2 additional pa-tients lost useful vision in one eye.
Uveitis was bilateral in 9 patients (Table 1). The anteriorsegment was quiet in 5 patients, and 6 patients had a mild aqueousflare (�1). All patients had vitreitis, which was extensive (�3 or�4) in 8 patients in either one (4 patients) or both eyes. Vitreitisconsisted of sheets and opacities in addition to cells. Vitreous cells
4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™Figure 1. Case 1. Clear-cut round lesions at the level of the outer retinaFigure 2. Case 5. Ill-defined confluent yellowish deep retinal and subreti
Figure 3. Case 4. Arteriovenous phase of a fluorescein angiogram of thehypofluorescent dots intermingling with hyperfluorescent dots; and a largeof tumor cells under the retinal pigment epithelium (arrows).Figure 4. Cytologic staining of the cytocentrifuged vitreous specimen. Larhave cloverleaf nuclei. Nucleoli are prominent (stain, May–Grünwald–GFigure 5. Immunocytologic staining of the specimen with a B-cell marker�400).
Figure 6. � light clonality in the B cells (original magnification, �400).
were immobile yellowish or brownish clumps and attached to thedetached posterior vitreous membrane.
In all patients except one, a few or multitudinous punctatewhite spots or clear-cut round lesions up to one quarter of a discdiameter in size were seen at the level of the outer retina and/or thepigment epithelium in the posterior pole and/or peripheral retina(Fig 1). Fundus lesions dominated the clinical picture in 1 eyein 3 patients. In 3 eyes, a poorly defined confluent yellowishretinal or subretinal mass was seen in the posterior pole (Fig 2).Because of poor visibility of the fundus in 8 patients, fluores-cein angiography initially was performed on only 4 patients. Itwas assessed as normal in 2 cases, and in 2 patients, hypofluo-rescent dots intermingling with hyperfluorescent round areaswere disclosed (Fig 3).
The interval from the onset of ocular symptoms to diagnosis ofPIOL ranged from 8 to 48 months (Table 1). The duration betweenthe first contact in our clinic and diagnosis was considerablyshorter among the 6 most recent patients than among the first 5patients: 0 to 6 months (median, 1) versus 6 to 40 months (median,9), respectively. The follow-up of our patients ranged from 10 to99 months (median, 32). Seven patients died of PNSL of the brain.In 3 patients, the disease is still restricted to the eyes.
Findings of the Vitreous Biopsies
Diagnosis of PIOL by vitreous biopsy was reached in 6 patients(54%). The vitreous specimens of patients 1, 3 (both eyes), 4, and5 (first biopsy) were nondiagnostic. From these samples, only 1 to2 Cyto-Tek preparations could be made. They showed a fewinflammatory cells and debris or were totally acellular. The re-peated vitreous biopsy from the fellow eye of patient 5 and 5biopsies thereafter of patients 6 to 10 were diagnostic (Table 2).The number of the cells ranged from 0.02�106 to 0.2�106,enabling 10 to 21 slides to be prepared. In 2 cases, the specimensprepared from the cassette fluid were as representative as that fromthe core vitreous. Despite that, the material was sufficient forassessment of clonality in only 4 cases. The PCR technique todetect B- and T-cell gene rearrangements was tried in 2 cases andflow cytometry in 1, but the cell count was too scanty for that.The cytologic picture of the successful specimens of patients 5to 10 was as follows: 5% to 80% of the cells were largebasophilic pleomorphic blasts that intermingled with lympho-cytes (10%– 63%) and monocytes (4%–32%), apoptotic cells,and debris (Fig 4). The blasts had basophilic cytoplasm and largeround, oval, bean-shaped, or cloverleaf nuclei with coarse andprominent nucleoli. In all specimens, 40% to 80% of the blastsshowed B-cell markers in immunohistochemical staining (Table 2,Fig 5). In 2 specimens, the blasts expressed � light chain clonality(Table 2, Fig 6).
Microbiological analysis of the vitreous fluid showed no pos-itive results in any patients.
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™the pigment epithelium.sions in the posterior pole of the right eye.
ye. There are areas of granularity at the level of the pigment epithelium;of blockage of staining temporal to the macula, suggestive of an aggregate
ophilic pleomorphic blast cells with scanty cytoplasm are seen. Some cells; original magnification, �400).0). Nearly all blast cells expressed CD20 positivity (original magnification,
™™™
and
nal le
left er area
ge bas
iemsa(CD2
1375
Ophthalmology Volume 114, Number 7, July 2007
Discussion
Primary intraocular lymphoma is a subset of non-Hodgkin’sB-cell nervous system lymphoma. In Finland, with a pop-ulation of 5.2 million, 118 new cases of PNSL were regis-tered by the Finnish Cancer Registry during 1999 to 2003,3 times more than during a 5-year period 2 decades earlier(unpublished data). The same tendency has been de-scribed in the U.S. as well.1 Eleven patients with in-traocular–primary nervous system lymphoma were diag-nosed at the Helsinki University Central Hospital duringa 6-year period. Nine patients presented with oculardisease. In 4 of them, only the development of the ner-vous system malignancy led to the right diagnosis. To ourknowledge, they are the first patients with PIOL reportedin Finland.
There are many reasons for the delay of the properdiagnosis. Minor ocular symptoms do not urge the patientwith PIOL to seek ophthalmologic examination, the biomi-croscopical features of PIOL mimic many uveitis entities,and the vitreous biopsy procedure may be nondiagnostic.Vague and diffuse ocular complaints in PIOL patients alsohave been pointed out by others.6,9,11,23 Visual acuity wasnormal in both eyes in only 2 patients, and in as many as 4patients, VA was CF in one eye. Despite that, it was theblurring of vision and shadows in the VF most patientscomplained of. We agree with others that PIOL mimicsother uveitis entities.8,11,23–25 In the diagnostic workup ofuveitis, patient age, the anatomic localization of uveitis,bilaterality, biomicroscopic appearance, onset and durationof uveitis, and response to therapy play an important role.21
Typically, PIOL is a bilateral uveitis in elderly patients.Although it may be seen rarely in young patients,26 themedian age is usually 50 to 60 years.6,9,11,15 The median ageof our patients was 61 years.
The yield of the first 5 vitreous biopsies was nondi-agnostic. The main reason for the unsuccessful result wasprobably the 50% alcohol added to the fresh samples.Further, corticosteroids may be cytolytic in PNSL andmay contribute to the difficulty in obtaining viable cellsfrom a vitrectomy specimen.11,25,27 Although alcohol as afixative has been used with success by others,9,12 it didnot work in our laboratory. After we improved the pro-
ND � not done; PIOL � primary intraocular lymphoma.
cess and began to handle the samples unfixed, represen-
1376
tative specimens were received in the remaining 6 pa-tients (6/7 specimens). Freeman et al15 and Coupland etal13 used a cytospin preparation of the unfixed vitreousspecimens and obtained successful biopsy results in 56%and 58% of their PIOL patients, respectively.
Cytologic assessment has been the basic method in thediagnosis of PIOL since the 1970s.9 –11,28 More sophisti-cated techniques in the diagnosis of PIOL have beendeveloped as well: flow cytometric analysis,29 gene re-arrangements by PCR technique,30,31 and measurementof the interleukin 10/interleukin 6 ratio of the vitreoussample.32,33 However, PCR methodology is found towork best in tissue biopsy specimens,30 and on the basisof an increased interleukin 10/interleukin 6 ratio, thediagnosis of PIOL cannot be made.32,33 Because thenumber of viable cells in the vitreous sample is oftenlimited, newer techniques should be used only if a suf-ficient number of slides for cytologic and immunohisto-chemical stainings has been guaranteed.
In one patient with recently diagnosed PNSL, the in-traocular findings were compatible with PIOL. She had noocular symptoms, a fact reported by others as well.6,11
Because of vague ocular symptoms or even their totalabsence, patients with PNSL should be examined for oculardisease. On the other hand, magnetic resonance imaging ofthe brain should be repeated in patients with PIOL to planthe management properly.
In conclusion, our study indicates that diagnosis of PIOLis difficult and that it can be improved. A high index ofclinical suspicion is most important to diagnostic success.Severe bilateral vitreitis in an elderly patient should berecognized as a characteristic finding of PIOL. In such apatient, vitreous biopsy should not be postponed. Optimalhandling of the specimen and evaluation of the slides by anexperienced cytopathologist are critical in the diagnosticworkup of PIOL. To achieve the best possible result, directcommunication between an ophthalmic surgeon and cyto-pathologist is essential. The patients should be observedoncologically with respect to a high risk of developingcerebral PNSL disease subsequently.
Acknowledgment. The authors acknowledge Risto Sankila,MD, PhD, of the Finnish Cancer Registry for provision of the
1. Corn BW, Marcus SM, Topham A, et al. Will primary centralnervous system lymphoma be the most frequent brain tumordiagnosed in the year 2000? Cancer 1997;79:2409–13.
2. Aho R, Ekfors T, Haltia M, Kalimo H. Pathogenesis of pri-mary central nervous system lymphoma: invasion of malig-nant lymphoid cells into and within the brain parenchyma.Acta Neuropathol 1993;86:71–6.
3. Larocca LM, Capello D, Rinelli A, et al. The molecular andphenotypic profile of primary central nervous system lym-phoma. Blood 1998;92:1011–9.
13. Coupland SE, Bechrakis NE, Anastassiou G, et al. Evaluationof vitrectomy specimens and chorioretinal biopsies in thediagnosis of primary intraocular lymphoma in patients withmasquerade syndrome. Graefes Arch Clin Exp Ophthalmol2003;241:860–70.
14. Coupland SE, Heimann H, Bechrakis NE. Primary intraocularlymphoma: a review of the clinical, histopathological andmolecular biological features. Graefes Arch Clin Exp Oph-thalmol 2004;240:901–13.
15. Freeman LN, Schachat AP, Knox DL, et al. Clinical features,laboratory investigations, and survival in ocular reticulum cellsarcoma. Ophthalmology 1987;94:1631–9.
16. Fishburne BC, Wilson DJ, Rosenbaum JT, Neuwelt EA. In-travitreal methotrexate as an adjunctive treatment of intraoc-ular lymphoma. Arch Ophthalmol 1997;115:1152–6.
17. Smith JR, Rosenbaum JT, Wilson DJ, et al. Role of intravitreal
methotrexate in the management of primary central nervous
system lymphoma with ocular involvement. Ophthalmology2002;109:1709–16.
18. Hormigo A, Abrey L, Heinemann MH, DeAngelis LM. Ocularpresentation of primary central nervous system lymphoma:diagnosis and treatment. Br J Haemotol 2004;126:202–8.
19. Jahnke K, Korfel A, Komm J, et al. Intraocular lymphoma2000–2005: results of a retrospective multicentral trial.Graefes Arch Clin Exp Ophthalmol 2006;244:663–9.
20. Bloch-Michel E, Nussenblatt RB. International Uveitis Studyrecommendations for the evaluation of intraocular inflamma-tory disease. Am J Ophthalmol 1987;103:234–5.
21. Nussenblatt RB, Palestine AG. Uveitis: Fundamentals andClinical Practice. Chicago: Year Book Medical Publishers;1989:75–9.
22. von Willebrand E, Zola H, Häyry P. Thrombocyte aggregatesin renal allografts. Analysis with fine needle aspiration biopsyand monoclonal antithrombocyte antibodies. Transplantation1985;39:258–62.
23. Cassoux N, Merle-Beral H, Leblond V, et al. Ocular andcentral nervous system lymphoma: clinical features and diag-nosis. Ocular Immunol Inflam 2000;8:243–50.
24. Ridley ME, McDonald HR, Sternberg P Jr, et al. Retinalmanifestations of ocular lymphoma (reticulum cell sarcoma).Ophthalmology 1992;99:1153–61.
25. Akpek EK, Ahmed I, Hochberg FH, et al. Intraocular-centralnervous system lymphoma. Clinical features, diagnosis, andoutcomes. Ophthalmology 1999;106:1805–10.
26. Wender A, Adar A, Maor E, Yassur Y. Primary B-cell lym-phoma of the eyes and brain in a 3-year-old boy. Arch Oph-thalmol 1994;112:450–1.
27. Weller M. Glucocorticoid treatment of primary CNS lym-phoma. J Neurooncol 1999;43:237–9.
28. Klingele TG, Hogan MJ. Ocular reticulum cell sarcoma. Am JOphthalmol 1975;79:39–47.
29. Zaldivar RA, Martin DF, Holden JT, Grossniklaus HE. Pri-mary intraocular lymphoma. Clinical, cytologic, and flow cy-tometric analysis. Ophthalmology 2004;111:1762–7.
30. Shen DF, Zhang Z, LeHoang P, et al. Utility of microdissec-tion and polymerase chain reaction for the detection of immu-noglobulin gene rearrangement and translocation in primaryintraocular lymphoma. Ophthalmology 1998;105:1664–9.
31. White VA, Gascoyne R, Paton KE. Use of polymerase chainreaction to detect B- and T-cell gene rearrangements in vitre-ous specimens from patients with intraocular lymphoma. ArchOphthalmol 1999;117:761–5.
32. Chan CC, Whitcup SM, Solomon D, Nussenblatt RB.Interleukin-10 in the vitreous of patients with primary intraoc-ular lymphoma. Am J Ophthalmol 1995;120:671–3.
33. Whitcup SM, Stark-Vancs V, Wittes RE, et al. Association ofinterleukin 10 in the vitreous and cerebrospinal fluid andprimary central nervous system lymphoma. Arch Ophthalmol
![[Secondary malignancies after successful primary treatment of malignant Hodgkin's lymphoma]](https://static.documents.pub/doc/80x56/634c4ac185214dc5fa08edd5/secondary-malignancies-after-successful-primary-treatment-of-malignant-hodgkins.jpg)
