Diagnostic delay in primary central nervous system lymphoma

Acta Oncologica, 2005; 44: 728
/734

ORIGINAL ARTICLE

Diagnostic delay in primary central nervous system lymphoma

INGFRID S. HALDORSEN1, ANSGAR ESPELAND1, JOHN LUDVIG LARSEN1 & OLAV MELLA2 Department of Radiology, Haukeland University Hospital, Bergen, Norway, 2Department of Oncology and Medical Physics, Haukeland University Hospital, Bergen, Norway Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

1

Abstract This study investigates delay in diagnosing primary central nervous system lymphoma (PCNSL), which has a variable clinical and radiological presentation. Early diagnosis and treatment may improve survival and cause less sequela in PCNSL. Medical records of all new cases of PCNSL morphologically verified while alive or by autopsy in Norway in 1989
/1998 were reviewed (n/74). The time from initial symptom to final morphological diagnosis of PCNSL had a median (mean, range) of 70 (106, 22
/330) days in 16 AIDS patients and 75 (157, 8
/1285) days in 58 non-AIDS patients. Among nonAIDS patients, the time to diagnosis was longer in patients with no tumour in the first neuroimaging report after initial symptom (p /0.001). Median (mean, range) time from initial symptom to neuroimaging was 14 (25, 1
/60) days in AIDS patients and 21 (88, 1
/1095) days in non-AIDS patients. In the non-AIDS group, those presenting with personality change or visual disturbance had more delayed imaging than the others. The time from first neuroimaging examination to final diagnosis in non-AIDS patients had a median (mean, range) of 28 (69, 1
/845) days, and was longer when no tumour was indicated in the imaging report (p/0.005) and if first biopsy did not confirm the diagnosis (p /0.02). All AIDS patients had their diagnosis of PCNSL first established by autopsy. The time from first neuroimaging to autopsy had a median (mean, range) of 48 (81, 10
/270) days. There is a considerable delay in the diagnosis of PCNSL and strategies for earlier diagnosis are thus needed. Physicians should consider early neuroimaging in patients with personality changes or visual disturbance, early renewed imaging in patients with persistent neurological symptoms but no tumour on initial imaging, and early/repeated biopsy of focal brain lesions in both AIDS patients and non-AIDS patients.

Primary central nervous system lymphomas (PCNSL) are extranodal malignant lymphomas arising in the brain, the spinal cord, the leptomeninges or in the eyes, with the absence of lymphoma outside the nervous system at the time of diagnosis. In untreated patients, the prognosis of PCNSL is generally poor with a reported median survival of about 1 month for AIDS patients [1,2] and 2
/3 months for non-AIDS patients [3,4]. However, the median survival is 4 months in AIDS patients receiving radiotherapy [1,2] and 13 months in AIDS patients having combined radio- and chemotherapy [1]. In non-AIDS patients a median survival of up to 50 months and a five-year survival of 22
/43% has been reported in patients receiving only chemotherapy [5] or combined radio- and chemotherapy [6
/9]. To ensure early treatment of PCNSL, early morphological diagnosis is essential. There is broad

consensus that if PCNSL is suspected, stereotactic biopsy is the diagnostic procedure of choice since open neurosurgical resection does not add to survival and may increase functional deficits [10
/12]. The biopsy should be taken before giving steroid medication, which may induce shrinkage of tumour rendering the diagnostic yield of biopsy rather low [13,14]. Good performance status at time of diagnosis is associated with better survival [8,15]. Early diagnosis of PCNSL may facilitate early treatment before the patient’s performance status has declined. Treatment while the tumour-size is relatively small may also result in less sequela and functional deficits. Among AIDS patients, improved systemic control of the AIDS-disease has made early diagnosis and local control of concomitant PCNSL more important [16,17]. However, diagnosing PCNSL may be difficult in both AIDS patients and non-AIDS patients. Com-

Correspondence: Ingfrid S. Haldorsen, Department of Radiology, Haukeland University Hospital, 5021 Bergen, Norway. Tel: 47 55972327. Fax: 47 55975140. E-mail: [email protected]

(Received 3 January 2005; accepted 6 July 2005) ISSN 0284-186X print/ISSN 1651-226X online # 2005 Taylor & Francis DOI: 10.1080/02841860500256272

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

Diagnostic delay in primary central nervous system lymphoma 729 mon debut symptoms such as personality change or symptoms of raised intracranial pressure [3,12] are non-specific and findings at computerized tomography (CT) and magnetic resonance imaging (MRI) are variable and may not suggest the diagnosis [18]. The time from initial symptoms to final diagnosis of PCNSL had a median of 1.5
/5 months in foreign studies of non-AIDS patients [19
/21] and a mean of 3 months for non-AIDS and 2 months for AIDS patients in a review of studies published 1980
/1992 [3]. We hypothesised that a considerable diagnostic delay also could occur in Norway. Based on a material including all patients in Norway having PCNSL morphologically verified while alive or by autopsy in the ten-year period 1989
/1998 [8], we have studied delay in the diagnosis of PCNSL in an unselected series of patients from an entire country. The first aim of our investigation was to record the following timespans: a) from the debut of the initial clinical symptom to the final morphological diagnosis of PCNSL, b) from the initial symptom to the first following neuroimaging examination, and c) from this examination to the final morphological diagnosis. The second aim was to assess potential associations between these time spans and patient age and gender, initial clinical symptom, the neuroimaging report, whether MRI was performed, steroid medication given prior to biopsy, whether first biopsy established the diagnosis and year of diagnosis. Material and methods In Norway (population 4.4 million in 1998) there is mandatory cancer reporting to The Norwegian Cancer Registry from which we were kindly permitted data of all patients recorded with PCNSL in the period 1989
/1998. These patients numbered 101 of whom 95 were diagnosed and treated at one of the 5 major university hospitals in Norway. We excluded 27 of the 101 patients because 6 had lymphoma of the spine with secondary involvement of the spinal canal and the meninges, 10 had lymphoma in other locations at the time of diagnosis, 3 had plasmocytoma, 3 had Langerhans cell histiocytosis X and 5 had clinical and radiological signs of PCNSL but no morphological verification of the diagnosis. Thus, a total of 74 patients had morphologically verified PCNSL in the decade studied and were included in the study. These were 16 HIV-infected patients (1 woman, 15 men) with AIDS-related PCNSL and 58 non-AIDS patients (30 women, 28 men), of which 3 received immuno-suppressive medication prior to diagnosis. Age at diagnosis of PCNSL ranged from 24 to 54 (median 35, mean 37)

years in the AIDS group and from 11 to 83 (median 68, mean 63) years in the non-AIDS group. The first author personally reviewed the hospital records of all patients and registered clinical features, neuroimaging reports and time span from recorded debut of the initial clinical symptom to the first neuroimaging examination and from this examination to the final morphological diagnosis. The time of morphological diagnosis was recorded as date of diagnostic procedure (operation, biopsy, cytology) or death in cases diagnosed at autopsy. If two or more initial symptoms could be attributed to PCNSL, the most dominating/disabling was recorded. Each neuroimaging report was classified as indicating tumour versus no tumour (non-AIDS patients) or focal lesion(s) (hematoma or lesion(s) with contrast enhancement) versus no focal lesion(s) (AIDSpatients). The method of Kaplan and Meier [22] was applied to analyse potential associations between different factors and time span. The Log-Rank test was used for univariate comparisons of time spans between groups. The Cox proportional Hazard Model was used to study the effect on time span of several factors simultaneously and to estimate hazard ratios [23]. In the analyses of time to final diagnosis, patients diagnosed post-mortem were censored. Fischer’s exact test was used to analyse possible associations between non-diagnostic biopsy and other factors (age, stereotactic versus open biopsy, steroid treatment prior to biopsy). All p-values in the analyses are two-sided. The analyses were performed with SPSS. Results The initial symptom of PCNSL was personality change in 25 (34%) of the 74 patients, headache in 15 (20%), focal neurological deficit in 15 (20%), cerebellar symptom in 11 (15%), visual disturbance in 3 (4%), epilepsy in 3 (4%) and infectious disease in 2 (3%). Neuroimaging (CT and /or MRI) was performed in all 74 patients. The first examination was CT in 73 (with i.v. contrast in 56) and MRI with i.v. Gadolinium in one. The morphological diagnosis of PCNSL in the 58 non-AIDS patients was established by open biopsy in 29, stereotactic biopsy in 14, cytology of cerebrospinal fluid (CSF) in 2 and by autopsy in 13. In 8 non-AIDS patients the first diagnostic procedure (stereotactic biopsy in 5 and open biopsy in 3) did not establish the diagnosis, and the diagnosis was ascertained by renewed biopsy in 6, cytology of CSF in 1 and autopsy in 1. Biopsy was performed in 45 of the 58 non-AIDS patients. By contrast, none of the

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

730

I. S. Haldorsen et al.

16 AIDS patients underwent biopsy and all of them had the diagnosis of PCNSL established by autopsy. The total time span from initial symptom to final morphological diagnosis of PCNSL ranged from 8 to 1285 days and did not differ significantly between non-AIDS patients (median 75 days) and AIDS patients (median 70 days) (Table I). Time from initial symptom to the first neuroimaging examination tended to be longer in the non-AIDS group than in the AIDS group (p /0.06) (Table I). Time from first neuroimaging examination to final diagnosis was similar in the two groups (Table I). Among non-AIDS patients, the time span from initial symptom to neuroimaging was significantly related to initial symptom (p /0.03) (Table II). Patients presenting with personality change or visual disturbance had more delayed imaging than patients with epilepsy, cerebellar symptoms, focal neurological symptoms or headache. In the non-AIDS group, the time span from first neuroimaging examination to final diagnosis was longer among patients having no indication of tumour in the imaging report (p /0.005) and when first biopsy did not establish the diagnosis (p /0.02) (Table II). A multivariate Cox Proportional Hazard Model including these two variables (imaging report, non-diagnostic/diagnostic biopsy) showed that they were independently related to the time span. The hazard ratio for no tumour/ tumour in the imaging report was 3.07 (95% CI: 1.17
/8.07, p/0.02) and for non-diagnostic/diagnostic first biopsy 3.73 (95% CI: 1.54
/9.03, p/0.004). First biopsy was non-diagnostic in 32% of patients aged B/65 years versus 4% aged ]/65 years (p /0.02), in 29% of stereotactic versus 11% of open biopsies (p /0.23), and in 22% of patients given steroids prior to the biopsy versus none of those not given steroids (p /0.18). It took 9 to 290 (median 100) days from first biopsy to diagnosis in the 8 patients with non-diagnostic first biopsy. All these patients had received steroid treatment prior to the biopsy. Patients with no indication of tumour in the imaging report also had a longer total time span from initial symptom to final diagnosis (p /0.001), as did non-AIDS patients diagnosed in 1989
/1993

versus 1994
/1998 (p /0.04) (Table II). However, multivariate Cox Proportional Hazard Model including these variables (imaging report, year of diagnosis in two categories) showed that only imaging report was independently related to total time span. The hazard ratio for year of diagnosis was 1.49 (95% CI: 0.75
/2.96, p /0.25) and for no tumour/tumour in the imaging report 4.26 (95% CI: 1.46
/12.43, p/0.01). For the 45 non-AIDS patients diagnosed while alive, univariate analysis showed no relation between year of diagnosis and total time span (p /0.44). The AIDS group was rather small and no significant associations were found in this group between the studied time spans and factors. However, both imaging and final diagnosis (i.e., death) tended to be more delayed for patients aged /35 years (Table III). A total time span of more than 6 months from initial symptom to final morphological diagnosis was found in 16 (22%) of 74 patients (13 non-AIDS, 3 AIDS). Table IV depicts clinical and diagnostic features of these patients. Discussion In this population-based study, diagnostic delay was assessed for all verified cases of PCNSL in an entire country during a specific period (1989
/1998). As far as we know, no such study has yet been published. Since our study was retrospective, it did not influence the diagnostic process we wished to evaluate. However, the retrospective data on debut of initial symptom may have been inaccurate in some cases, and some of the hospital records may have been incomplete. The material was also insufficient to study potential associations between diagnostic delay and some possibly relevant factors, such as the patients’ education and socio-economic status. The total time span from initial symptom to final morphological diagnosis of PCNSL in our study (non-AIDS patients: median 75 days, mean 157 days; AIDS patients: median 70 days, mean 106 days) is within the upper range of findings in other studies [3,21,24]. This time span had a particularly wide range in non-AIDS patients (8
/1285 days) and exceeded a year in five (9%) of

Table I. Time spans (days) between debut of initial symptom, first neuroimaging examination and final morphological diagnosis of PCNSL. Non-AIDS patients (n /58) Time span from Symptom to imaging Imaging to diagnosis Symptom to diagnosis

AIDS patients (n /16)

Mean (95% CI)

Median (95% CI)

Range

Mean (95% CI)

88 (36
/140) 69 (36
/101) 157 (93
/222)

21 (6
/36) 28 (24
/32) 75 (49
/101)

1
/1095 1
/845 8
/1285

25 (13
/37) 81 (35
/127) 106 (59
/153)

Median (95% CI) 14 (11
/17) 48 (21
/75) 70 (58
/82)

Range 1
/60 10
/270 22
/330

Diagnostic delay in primary central nervous system lymphoma 731 Table II. Time span (days) between initial symptom, first neuroimaging examination and final morphological diagnosis of PCNSL in 58 non-AIDS patients according to various factors. From symptom to imaging

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

Factor Patient age (years) B/65 ]/65 Patient sex Male Female Initial symptom Personality changes Epilepsy Cerebellar symptoms Focal neurological Headache Visual First imaging No tumour Tumour MRI performed prior to final diagnosis Yes No Steroids prior to first biopsy Yes No Diagnostic first biopsy Yes No Year of diagnosis 1989
/1993 1994
/1998

n

Median (95% CI)

From imaging to diagnosis

Log-Rank test Estimated (p-value) median* (95% CI) 0.13

From symptom to diagnosis

Log-Rank Estimated Log-Rank Test* (p-value) median* (95% CI) Test* (p-value) 0.59

35 (13
/57) 30 (25
/35)

0.31

24 34

21 (4
/38) 21 (3
/39)

28 30

21 (11
/31) 14 (5
/23)

21

49 (33
/65)

25 (18
/32)

88 (0
/203)

2 10

1** 14 (8
/20)

21** 90 (0
/186)

36** 111 (71
/151)

12

14 (5
/23)

24 (4
/44)

42 (11
/73)

0.88

0.61 35 (12
/58) 30 (22
/38)

0.03

10 3

14 (11
/17) 270 (126
/414)

30 (11
/49) 35 (3
/67)

21 (8
/34) 21 (6
/36)

21 (7
/35) 14 (7
/21)

0.005

35 (15
/55) 24 (14
/34)

21 (17
/25) 60 (16
/104)

30 (14
/46) 14 (10
/18)

23 35

14 (9
/19) 21 (8
/34)

0.30

124 (64
/184) 85 (56
/114) 0.62

25 (16
/34) 21 (0
/44) 0.62

37 8

0.001 1130 (181
/2079) 79 (52
/106)

0.25

0.37 36 9

0.34

56 (0
/156) 1130**

210 (0
/440) 25 (16
/34) 0.44

37 21

0.74 111 (59
/163) 85 (63
/107)

0.92

0.10 14 44

124 (79
/169) 85 (69
/101)

0.87 70 (27
/113) 85 (56
/114)

0.02 21 (18
/24) 98 (49
/147)

0.47

0.14 63 (24
/102) 149 (51
/247)

0.13 35 (20
/50) 30 (20
/40)

0.04 195 (27
/363) 85 (39
/131)

* Patients diagnosed post mortem were censored. ** CI not given because of small number of patients.

these patients, which corresponds to findings by others [21]. The time from initial symptom to neuroimaging has received little attention in previous studies of PCNSL. Among non-AIDS patients, we found especially delayed imaging for those presenting with personality change or visual disturbance. Others have found that personality change may be the only symptom of PCNSL for a long time and may be misinterpreted as dementia or neurodegenerative disease [12]. It has also been shown that patients with personality changes as dominating symptoms have delayed diagnosis of malignant glioma [25]. Symptoms and signs of visual disturbance have been reported in 8
/10% of PCNSL patients [12,26] and eye involvement by uveal or

vitreal lymphoma often antedates the development of clinically evident brain lymphoma [26]. Personality changes or eye symptoms involving uvea or corpus vitreum should thus raise awareness of possible PCNSL. In our study the first imaging report indicated tumour in 5 of 8 non-AIDS patients imaged more than 3 months after the debut of personality change or visual disturbance. These patients might perhaps have profited from earlier imaging. Hardly surprising, among non-AIDS patients, we found more delayed final diagnosis of PCNSL when the report of the first imaging examination (CT in 57 of 58 patients) did not indicate tumour. This finding also remains if based on any of all neuroimaging examinations in the first 2 weeks of

732

I. S. Haldorsen et al.

Table III. Time span (days) between initial symptom, first neuroimaging examination and final morphological diagnosis of PCNSL in 16 AIDS patients according to various factors.

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

Factor Patient age (years) B/35 ]/35 Patient sex Male Female Initial symptom Personality changes Epilepsy Cerebellar symptoms Focal neurological Headache Infectious disease First imaging No focal lesion Focal lesion(s) MRI performed prior to final diagnosis Yes No Year of diagnosis 1989
/1993 1994
/1998

n

From symptom to imaging

From imaging to diagnosis

From symptom to diagnosis

Median (95% CI)

Median (95% CI)

Median (95% CI)

Log-Rank test (p-value) 0.08

8 8

7 (0
/19) 30 (9
/51)

15 1

14 (12
/16) 30*

4 1 1 3 5 2

28 (12
/44) 1* 30* 12 (0
/30) 14 (8
/20) 3*

6 10

14 (7
/21) 28 (9
/47)

0.21 21 (0
/68) 56 (13
/99)

0.82

9 7

12 (0
/38) 30 (0
/71)

0.53 71 (42
/100) 65*

0.33 14 (0
/38) 56* 90* 150 (0
/374) 49 (8
/90) 10*

0.40 65 (37
/93) 57* 120* 151 (0
/357) 90 (17
/163) 51*

0.47 56 (34
/78) 21 (0
/54)

0.17 1* 14 (0
/34)

0.11

0.52

0.44

0.59 80 (40
/120) 65 (43
/87)

0.24 150* 35 (12
/68)

0.09

0.24 151* 65 (41
/89)

0.44 56 (33
/79) 35 (22
/48)

Log-Rank test (p-value)

51 (32
/70) 80 (54
/106)

49 (16
/82) 35* 0.17

2 14

Log-Rank test (p-value)

0.90 57 (39
/75) 71 (68
/74)

All 16 AIDS patients were diagnosed post mortem. * CI not given because of small number of patients.

the initial imaging work-up. This included MRI in 27 (all received i.v. contrast) of the 58 non-AIDS patients and indicated tumour in 47 (i.e., only 3 more than already indicated in the first imaging report). Delayed diagnosis following ‘‘negative’’ imaging is also seen in other malignancies, such as in breast cancer following mammography reports indicating benign findings [27]. Non-diagnostic first biopsy also prolonged the time span from first neuroimaging to final diagnosis in non-AIDS patients. The unfavourable effect of steroids on the yield of biopsy seems well documented [10,14,28], although it was not statistically significant in our small series. The importance of avoiding steroid medication prior to biopsy if PCNSL is suspected must be emphasised. Interestingly, non-AIDS patients diagnosed in 1994
/1998 had shorter total time span to final diagnosis than non-AIDS patients diagnosed in 1989
/1993 (p /0.04). However, in the multivariate analysis only tumour versus no tumour in the report of first imaging (CT in all but one patient), not year of diagnosis, was independently associated with total time span. Thus, shorter total time span might reflect improved CT imaging during the decade.

First imaging report indicated tumour in as much as 29 (83%) of 35 patients in 1994
/1998 versus 15 (65%) of 23 in 1989
/1993, and use of i.v. contrast in 83% versus 70%. Similarly, 30 (71%) of 42 patients were diagnosed with PCNSL while alive in 1994
/1998 versus only 15 (47%) of 32 in 1989
/1993. This may suggest an improved overall diagnostic process. In AIDS patients the time span from initial symptom to neuroimaging was relatively short (median 14 days, Table I), which could reflect an active attitude towards diagnosing concomitant disease in AIDS. However, the AIDS patients had strikingly long time spans from first neuroimaging examination to final diagnosis of PCNSL (median 48 days, Table I) and none of them had this diagnosis established while alive. Imaging reports in the course of the disease suggested AIDS-related opportunistic infection in 15 of the 16 AIDS patients and 13 of these received toxoplasma therapy. In spite of lacking effect of this therapy in 11 of the 13 patients, and consideration of possible lymphoma in the medical records of 8 of the 16 AIDS patients, none of the AIDS patients had stereotactic or open biopsy performed and none received radio- or chemotherapy. By contrast, in an Australian study of HIV-related PCNSL diagnosed in 1987
/1998,

Diagnostic delay in primary central nervous system lymphoma 733

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

Table IV. Clinical and diagnostic features of 16 patients having final morphological diagnosis of PCNSL established more than 6 months after debut of initial symptom.

Age

Year of Diagnosed Sex AIDS diagnosis while alive

79 29 59 55 55 46 53 68 74 32 39 37 56 68 44 56

F M F F M M F F F F M M M M F F

No Yes No No No Yes No No No No Yes No No No No No

1995 1989 1990 1993 1998 1994 1998 1993 1997 1989 1990 1989 1996 1993 1993 1993

Yes No Yes Yes Yes No Yes Yes Yes No No No Yes No Yes Yes

Initial symptom Personality changes Headache Visual Headache Focal neurological Headache Personality changes Personality changes Focal neurological Focal neurological Focal neurological Visual Personality changes Cerebellar symptoms Visual Headache

Days from symptom to first neuroimaging 180 14 180 14 180 14 105 240 120 8 60 270 300 5 1095 1030

Imaging report indicating Tumour No focal lesion Tumour Tumour Tumour No focal lesion Tumour Tumour Tumour Tumour Focal lesion No tumour No tumour No tumour No tumour No tumour

MRI prior to final Diagnostic diagnosis first biopsy Yes No No No Yes Yes Yes No Yes Yes No Yes Yes No Yes Yes

Yes * No No No * Yes Yes Yes No * * Yes * Yes Yes

Days from symptom to final diagnosis 193 194 195 224 255 269 270 275 300 308 330 420 510 850 1130 1285

* No biopsy performed before death.

more than 40% had biopsy and 49% received radiotherapy [17]. The large difference is difficult to explain, but might suggest a more reluctant attitude towards invasive diagnostic procedures and treatment in our material. However, most of the AIDS patients in our study had a poor general condition, which could complicate diagnostic procedures and active therapy. Many studies have documented that AIDS patients with PCNSL receiving radiotherapy or combined radio- and chemotherapy have significantly longer survival than patients receiving no therapy [1,2,17] and they tend to die of opportunistic infections rather than tumour progression [2]. Highly active antiretroviral therapy (HAART) in AIDS patients with PCNSL has also been shown to significantly prolong survival [16]. It has been argued that toxoplasma therapy is no longer appropriate as a standard first-line approach to focal brain lesions in AIDS patients, because of changing disease patterns of such lesions [29]. It seems that AIDS patients today, when there is better management of the systemic AIDS, may profit from earlier differential diagnosis of focal brain lesions, in order to identify the underlying disease and give specific treatment. In conclusion, we have demonstrated considerable delays in the diagnosis of PCNSL. In the non-AIDS group, patients with personality change or visual disturbance as initial clinical symptom had particularly delayed first neuroimaging. Non-AIDS patients with no tumour reported on initial neuroimaging or with non-diagnostic first biopsy had

delayed diagnosis of PCNSL. In the AIDS group, none had the lymphoma diagnosis established while alive. Strategies for earlier diagnosis of PCNSL seem to be needed in both AIDS patients and nonAIDS patients. Focal brain lesions should be evaluated for early biopsy, also in AIDS patients. To improve the diagnostic yield of biopsy, prior steroid medication should be avoided when PCNSL is suspected. Personality changes or visual disturbance should alert physicians and make them consider early neuroimaging. Persistent or increasing neurological symptoms in patients with no tumour on initial imaging should lead to early renewed imaging. We are now reviewing the present imaging material in detail to identify pitfalls and characteristics of PCNSL to help improve the imaging diagnostics. Acknowledgements This study was supported in part by grants from Amersham Health AS and grants from Haakon and Sigrun Ødegaard‘s Foundation. Thanks are offered to Jan Harald Aarseth (Ph. D., Dep. of Neurology, Haukeland University Hospital) for statistical support and Siri Rye Salvesen for linguistic review. Further thanks are offered to the following institutions that have made possible the review of the medical records of the patients treated at their hospitals: Akershus University Hospital, Asker and Baerum Hospital, Haukeland University Hospital, Innlandet Hospital Trust, Laerdal Hospital, Rikshospitalet University Hospital, Rogaland Central Hospital, St. Olavs

734

I. S. Haldorsen et al.

Hospital, The Norwegian Radiumhospital, Sorlandet Hospital, Ullevaal University Hospital and University Hospital of North Norway.

/

References

/

/

[1] Chamberlain MC, Kormanik PA. AIDS-related central nervous system lymphomas. J Neurooncol 1999;43(3): 269
/76. [2] Baumgartner JE, Rachlin JR, Beckstead JH, Meeker TC, Levy RM, Wara WM, et al. Primary central nervous system lymphomas: natural history and response to radiation therapy in 55 patients with acquired immunodeficiency syndrome. J Neurosurg 1990;73(2):206
/11. [3] Fine HA, Mayer RJ. Primary central nervous system lymphoma. Ann Intern Med 1993;119(11):1093
/104. [4] Henry JM, Heffner RR, Jr, Dillard SH, Earle KM, Davis RL. Primary malignant lymphomas of the central nervous system. Cancer 1974;34(4):1293
/302. [5] Pels H, Schmidt-Wolf IG, Glasmacher A, Schulz H, Engert A, Diehl V, et al. Primary central nervous system lymphoma: results of a pilot and phase II study of systemic and intraventricular chemotherapy with deferred radiotherapy. J Clin Oncol 2003;21(24):4489
/95. [6] Abrey LE, DeAngelis LM, Yahalom J. Long-term survival in primary CNS lymphoma. J Clin Oncol 1998;16(3):859
/63. [7] Shibamoto Y, Tsuchida E, Seki K, Oya N, Hasegawa M, Toda Y, et al. Primary central nervous system lymphoma in Japan 1995
/1999: changes from the preceding 10 years. J Cancer Res Clin Oncol 2004;130(6):351
/6. [8] Salvesen Haldorsen I, Aarseth JH, Hollender A, Larsen JL, Espeland A, Mella O. Incidence, clinical features, treatment and outcome of primary central nervous system lymphoma in Norway. Acta Oncol 2004;43(6):520
/9. [9] Goldkuhl C, Ekman T, Wiklund T, Telhaug R. Age-adjusted chemotherapy for primary central-nervous system lymphoma
/a pilot study. Acta Oncol 2002;41(1):29
/35. [10] Schlegel U, Schmidt-Wolf IG, Deckert M. Primary CNS lymphoma: clinical presentation, pathological classification, molecular pathogenesis and treatment. J Neurol Sci 2000; 181(1
/2):1
/12. [11] Bataille B, Delwail V, Menet E, Vandermarcq P, Ingrand P, Wager M, et al. Primary intracerebral malignant lymphoma: report of 248 cases. J Neurosurg 2000;92(2):261
/6. [12] Herrlinger U, Schabet M, Bitzer M, Petersen D, Krauseneck P. Primary central nervous system lymphoma: from clinical presentation to diagnosis. J Neurooncol 1999; 43(3):219
/26. [13] Buhring U, Herrlinger U, Krings T, Thiex R, Weller M, Kuker W. MRI features of primary central nervous system lymphomas at presentation. Neurology 2001;57(3): 393
/6. [14] Cartmill M, Allibone R, Bessell EM, Byrne PO. Primary cerebral non-Hodgkin’s lymphoma: problems with diagnosis and development of a protocol for management. Br J Neurosurg 2000;14(4):313
/5. [15] Ferreri AJ, Blay JY, Reni M, Pasini F, Spina M, Ambrosetti A, et al. Prognostic scoring system for primary CNS lymphomas: the International Extranodal Lymphoma Study Group experience. J Clin Oncol 2003;21(2):266
/72. /

/

/

/

/

/

Acta Oncol Downloaded from informahealthcare.com by 177.98.0.188 on 05/20/14 For personal use only.

[16] Skiest DJ, Crosby C. Survival is prolonged by highly active antiretroviral therapy in AIDS patients with primary central nervous system lymphoma. AIDS 2003;17(12):1787
/93. [17] Newell ME, Hoy JF, Cooper SG, DeGraaff B, Grulich AE, Bryant M, et al. Human immunodeficiency virus-related primary central nervous system lymphoma: factors influencing survival in 111 patients. Cancer 2004;100(12):2627
/ 36. [18] Remick SC, Diamond C, Migliozzi JA, Solis O, Wagner H, Jr, Haase RF, et al. Primary central nervous system lymphoma in patients with and without the acquired immune deficiency syndrome. A retrospective analysis and review of the literature. Medicine (Baltimore) 1990;69(6): 345
/60. [19] Krogh-Jensen M, D’Amore F, Jensen MK, Christensen BE, Thorling K, Pedersen M, et al. Clinicopathological features, survival and prognostic factors of primary central nervous system lymphomas: trends in incidence of primary central nervous system lymphomas and primary malignant brain tumors in a well-defined geographical area. Populationbased data from the Danish Lymphoma Registry, LYFO, and the Danish Cancer Registry. Leuk Lymphoma 1995; 19(3
/4):223
/33. [20] Braus DF, Schwechheimer K, Muller-Hermelink HK, Schwarzkopf G, Volk B, Mundinger F. Primary cerebral malignant non-Hodgkin’s lymphomas: a retrospective clinical study. J Neurol 1992;239(3):117
/24. [21] Herrlinger U, Schabet M, Clemens M, Kortmann RD, Petersen D, Will BE, et al. Clinical presentation and therapeutic outcome in 26 patients with primary CNS lymphoma. Acta Neurol Scand 1998;97(4):257
/64. [22] Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53: 457
/81. [23] Cox DR. Regression models and life tables. J Roy Statist Soc 1972;34:187
/220. [24] Krogh-Jensen M, D’Amore F, Jensen MK, Christensen BE, Thorling K, Pedersen M, et al. Incidence, clinicopathological features and outcome of primary central nervous system lymphomas. Population-based data from a Danish lymphoma registry. Danish Lymphoma Study Group, LYFO. Ann Oncol 1994;5(4):349
/54. [25] Salander P, Bergenheim AT, Hamberg K, Henriksson R. Pathways from symptoms to medical care: a descriptive study of symptom development and obstacles to early diagnosis in brain tumour patients. Fam Pract 1999;16(2): 143
/8. [26] Hochberg FH, Miller DC. Primary central nervous system lymphoma. J Neurosurg 1988;68(6):835
/53. [27] Goodson WH, III, Moore DH. Causes of physician delay in the diagnosis of breast cancer. Arch Intern Med 2002; 162(12):1343
/8. [28] DeAngelis LM, Hormigo A. Treatment of primary central nervous system lymphoma. Semin Oncol 2004;31(5):684
/ 92. [29] Ammassari A, Scoppettuolo G, Murri R, Pezzotti P, Cingolani A, Del Borgo C, et al. Changing disease patterns in focal brain lesion-causing disorders in AIDS. J Acquir Immune Defic Syndr Hum Retrovirol 1998;18(4): 365
/71.

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/

/