Should whole brain radiotherapy be considered standard of care in newly diagnosed primary

central nervous system lymphoma? The G-PCNSL-SG-1 randomized phase IV trial

E.Thiel , A. Korfel, P. Martus, L. Kanz, Griesinger F, P Rauch, L. Fischer, T. Pietsch, M. Bamberg, M. Weller

on behalf of G-PCNSL-SG

Deutsche Studiengruppe Primäre ZNS-LymphomeG-PCNSL-SG

Background

• High-dose methotrexate (HDMTX) is considered standard of care for newly diagnosed PCNSL.

• The role of whole brain radiotherapy (WBRT) is controversial.

• Delayed neurotoxicity limits the acceptance of HDMTX+WBRT as a standard of care.

CR No CR

A1: Cons. WBRT

B1: Rescue WBRT

G-PCNSL-SG-1 trial design

B2: HD-Ara-C

A2: Watch and wait

CR No CR

• newly diagnosed PCNSL • immunocompetent adult patients• creatinine clearance >50 ml/min

Randomization:A1/B1 versus A2/B2

HDMTX 4g/m2 d1 1999-2007* HDMTX 4g/m2 d1 + Ifo 1.5g/m2 d3-5 2007-2009*

* + 3 x 8 mg Dexamethasone (d1-10), 1st cycle only

Statistical considerations

• Hypothesis: omission of WBRT from first-line

treatment does not compromise OS (non-inferiority design). • Omission of WBRT was defined as non-inferior to WBRT if

the lower two-sided 95% confidence limit of the

hazard ratio of WBRT versus no WBRT was not below 0.9.• 60% power to prove non-inferiority of omission of WBRT in

case of a hazard ratio of 1.2 of WBRT versus no WBRT. • Sample size required: 151 patients per group.

n=551 entered

n=24 early drop-out / recruitment error

n=49 dropped out

n=411 entered the post-HDMTX phase (ITT)

n=318 per protocol

Flow diagram of all patients

n=526 fulfilledthe eligibility criteria

and received HDMTX

n=93 protocol violation(49 WBRT arms, 44 no WBRTarms)

n=154 WBRT n=164 no WBRT

CR 182 (34.6%)PR 101 (19.2%)SD/PD 24/123 (4.6/23.4%)

Unknown 30 (5.7%)

Died 66 (12.5%) n=66 died

Patients characteristics (PP population, n=318)

• med. age 61 (19-84)• male/female 183/135• med. Karnofsky 80 (30-100)• diagnostic procedure:

stereotactic biopsy 184 (57.8%)partial/total resection 97 (30.6%)open biopsy 32 (10.1%)CSF cytomorphology 3 (0.9%)no data 2 (0.6%)

• lymphoma cells in the CSF26 (8.2%)• ocular involvement on slit lamp 9 (2.8%)

PFS and OS : PP patientsWBRT n=154, no WBRT n=164

PFS 18.3 vs. 12 mo (P=.13) OS 32.4 vs. 37.1 mo (P=.7)

OS and PFS: PP patients with CR after HDMT

WBRT n=56, no WBRT n=96

PFS 36.3 vs. 21.5 mo (P=.038) OS 38.8 vs. 39.4 (P=.56)

OS and PFS: PP patients without CR after HDMTX

WBRT n=98, no WBRT n=68

PFS 5.6 vs. 3 mo (P=.003) OS 24.3 vs. 18.6 mo (P=.1)

PFS and OS : ITT patientsWBRT = 203, no WBRT = 208

PFS 15.5 vs. 9.9 mo (P=.041) OS 34.4 vs. 32.1 (P=.94)

The influence of age and KPS on OS (n=526)

Age (<60 n=189; ≥60 n=337) KPS (≥70 n=275; <70 n=163)

38.4 vs. 14.2 mo. (P<.005) 31.5 vs. 9.8 mo. (P<.005)

Late neurotoxicity(PP pts. with CR)

48,9

72,5

28

41,7

0

10

20

30

40

50

60

70

80

%

WBRT WBRT

No WBRT No WBRT

Clinical evaluation Neuroradiologic evaluation

Clinical evaluation n=45 WBRT, n=33 no WBRTNeuroradiologic evaluation n=51 WBRT, n=36 no WBRT

P=.054 P=.04

Conclusions

• first randomized phase IV study on PCNSL• no significant difference for OS with versus

without WBRT • PFS prolongation in subgroup analyses confirms

the role WBRT for disease control; lack of OS benefit reflects effectivity of salvage treatments

• age and KPS most important risk factors• late neurotoxicity more frequent with WBRT