Is PCNSL resection safe? Evidence from clinical trials in which enrollment follows surgery-such as G-PCNSL-SG-1-is not valid proof of the safety of resection.
Primary central nervous system (CNS) lymphoma (PCNSL) is an aggressive intracranial malignant neoplasm in adults whose incidence in the United States is about 15% of that of glioblastoma. Tumor neurosurgeons have typically preferred to diagnose this tumor by biopsy, rather than resection, for several reasons: first, the frequent early wide dissemination of PCNSL (as reflected by the multiplicity of lesions on imaging studies, common ocular and cerebrospinal fluid involvement, and a largely distant pattern of failure after treatment); second, the striking responsiveness of PCNSL to radiation and especially chemotherapy, with frequent prolonged survival after nonsurgical therapy alone; and third, evidence from retrospective reports involving patients with PCNSL treated in the 1980s and 90s in whom surgical resection was associated with shorter survival compared with biopsy. Some surgeons may also remember a highly publicized anecdotal case of PCNSL: William J. Casey, Ronald Reagan’s Director of Central Intelligence, who became symptomatic in December 1986 from a left frontal PCNSL just hours before he was scheduled to testify on his knowledge of the Iran-Contra affair before a US Senate panel. After a left frontal tumor resection, he never appeared publicly again and lived only 4 months in what was widely reported to be poor neurologic condition.
This negative view of resection has recently been challenged by Weller et al. Their retrospective secondary analysis of a multicenter randomized trial (the German Primary CNS Lymphoma Study Group 1 [G-PCNSL-SG-1] trial, which compared high-dose methotrexate followed by cytarabine vs high-dose methotrexate followed by whole-brain radiotherapy) showed significantly longer median survival in PCNSL patients who underwent subtotal or gross total resection at diagnosis (31 and 32 months, respectively) compared with patients diagnosed by biopsy (18 months). Aside from this single retrospective analysis, evidence favoring surgical resection, as well summarized by Bierman, is scant.
From an oncologic standpoint, retrospective reports comparing survival after tumor resection vs biopsy are examples of analysis of survival by treatment response. In a classic 1983 article, Anderson et al pointed out reasons why longer survival in treatment responders may not reflect treatment efficacy. For surgical studies, the most important bias in a responder analysis is that response (ie, resectability) may simply identify patients who would have lived longer in the absence of the treatment.
Surgeons tend to resect intracerebral lesions that are single and superficial and that do not affect eloquent cortical areas. Patients with known ocular or leptomeningeal lymphomatous involvement would not normally be candidates for resection. Additional clinical factors, such as patient age, frailty, comorbidity, and performance status, are also considered in a surgeon’s decision. Many of these are important prognostic factors for survival in PCNSL, in patient cohorts principally diagnosed using needle biopsy. Weller et al were able to adjust for some of these factors in their multivariate analysis-namely age, multiplicity of lesions, Karnofsky Performance Status (KPS) score (measured after surgery), and a rough specification of lesion location (supratentorial, cortical, subcortical, spinal, or cerebellar/brainstem). None of these factors correlates perfectly with resectability, and we should expect this type of post hoc analysis to be a partial adjustment at best for resectability’s true prognostic value. In particular, by analogy with recent advances in our knowledge of the molecular biology of gliomas, we should suspect that resectability might reflect important differences in the primary tumor’s natural history and response to other treatments. In malignant glioma, isocitrate dehydrogenase (IDH)-mutant tumors are now known to be both highly resectable and unusually treatment-responsive, meaning that some of the apparent “benefit” of malignant glioma resection is actually the result of favorable intrinsic tumor characteristics. Similarly, Iwadate et al have recently shown that deep-seated PCNSL tumors, which are typically unresectable, are much less likely to be methotrexate-responsive (25% complete response, compared with 50% for superficial lesions), and that deep and superficial PCNSL tumors have been found to have different gene expression profiles in a hierarchical clustering analysis of microarray data. Several molecular prognostic factors, such as BCL6 expression, are known to influence survival in PCNSL, but their correlation with resectability is unknown.
Is PCNSL resection safe? Evidence from clinical trials in which enrollment follows surgery-such as G-PCNSL-SG-1-is not valid proof of the safety of resection. Patients who are hurt by surgery may not survive to enter a prospective trial, and those who do survive may not choose aggressive treatment, or may not meet minimum performance status criteria for clinical trial enrollment. Population-based data would be a better resource for checking surgical safety, but to date such studies in PCNSL have not reported surgical outcomes; some databases, such as that of the Surveillance, Epidemiology and End Results (SEER) program, are poor sources for extent-of-resection studies because of coding anomalies. Single-center studies showing the safety of resection in PCNSL are lacking, and we are left with information such as that DeAngelis et al provided from one major US cancer center in 1990: 3 of 10 resected patients had major, permanent neurologic deficits from surgery. Even whether resection is superior to immediate chemotherapy or radiation in relieving preoperative neurologic deficits in PCNSL is not proven. In a 2005 series of 32 PCNSL patients of whom 25 underwent resection, postoperative KPS scores were not improved (mean KPS scores: 37.5 before surgery and 38.7 after surgery).
In PCNSL there are additional drawbacks to open resection even when this is performed safely. Most surgeons advise a period of wound healing after craniotomy before starting radiation or chemotherapy-something that is not necessary after needle biopsy; Rubenstein et al found that treatment delay was the most important adverse prognostic factor in a recent multicenter phase II chemotherapy study. Lastly, an imaging-complete surgical resection precludes assessment of tumor response to subsequent treatments. Most modern PCNSL treatment protocols specify different consolidation phases for patients with complete or less-than-complete responses to initial treatment, raising the possibility of either over- or undertreatment of completely resected patients.[5,12]
Does this question matter? Bierman argues that few surgeons will be confronted with the need to make a conscious decision whether to resect a PCNSL. However, current practice is heavily weighted toward biopsy, even for resectable lesions: in one single-center series, 37 of 38 patients (97%) were diagnosed by biopsy. We reviewed the Nationwide Inpatient Sample database (created by the Agency for Healthcare Research and Quality [AHRQ] for the Healthcare Cost and Utilization Project [HCUP]) for a population-based snapshot of current US practice: from 2008 to 2010, 69% of 82 PCNSL patients had biopsy rather than resection, a much higher proportion than in a glioma surgery study from the same database (26%). Given suitable proof of the value of open resection, it is likely utilization would increase. With no randomized trial of resection in the foreseeable future, single-center series showing that surgery can relieve preoperative deficits from mass effect with an acceptable cost in terms of new deficits, treatment delay, and uncertainty about consolidation treatment would be valuable. Until such information is available, purposeful surgical resection of PCNSL should probably not be adopted outside appropriately designed prospective clinical trials. Meanwhile, adding other locally intensive treatments to consolidation regimens in treatment-refractory PCNSL (such as radiosurgery or interstitial thermal therapy) may also merit testing.
Financial Disclosure:The authors have no significant financial interest in or other relationship with the manufacturer of any product or provider of any service mentioned in this article.
1. Ostrom QT, Gittleman H, Farah P, et al. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2006-2010. Neuro Oncol. 2013;15(suppl 2):1-56.
2. Bataille B, Delwail V, Menet D, et al. Primary intracerebral malignant lymphoma: report of 248 cases. J Neurosurg. 2000;92:261-6.
3. Zoglin R, Peterzell J, van Voorst B. Did a dead man tell no tales? A furor erupts over the disclosures in a book about Bill Casey’s CIA. Time. 1987;130:28.
4. Weller M, Martus P, Roth P, et al. Surgery for primary CNS lymphoma? Challenging a paradigm. Neuro Oncol. 2012;14:1481-4.
5. Thiel E, Korfel A, Martus P, et al. High-dose methotrexate with or without whole brain radiotherapy for primary CNS lymphoma (G-PCNSL-SG-1): a phase 3, randomised, non-inferiority trial. Lancet Oncol. 2010;11:1036-47.
6. Bierman P. Surgery for primary central nervous system lymphoma: Is it time for reevaluation? Oncology (Williston Park). 2014;28:632-7.
7. Anderson JR, Cain KC, Gelber RD. Analysis of survival by tumor response. J Clin Oncol. 1983;1:710-9.
8. Beiko J, Suki D, Hess KR, et al. IDH1 mutant malignant astrocytomas are more amenable to surgical resection and have a survival benefit associated with maximal surgical resection. Neuro Oncol. 2014;16:81-91.
9. Iwadate Y, Suganami A, Ikegami S, et al. Non-deep-seated primary CNS lymphoma: therapeutic responses and a molecular signature. J Neurooncol. 2014;117:261-8.
10. DeAngelis LM, Yahalom J, Heinemann M-H, et al. Primary CNS lymphoma: combined treatment with chemotherapy and radiotherapy. Neurology. 1990;40:80-6.
11. Bellinzona M, Roser F, Ostertag H, et al. Surgical removal of primary central nervous system lymphomas (PCNSL) presenting as space occupying lesions: a series of 33 cases. Eur J Surg Oncol. 2005;31:100-5.
12. Rubenstein JL, Hsi ED, Johnson JL, et al. Intensive chemotherapy and immunotherapy in patients with newly diagnosed primary CNS lymphoma: CALGB 50202 (Alliance 50202). J Clin Oncol. 2013;31:3061-8.
13. Aoki H, Ogura R, Tsukamoto Y, et al. Advantages of dose-dense methotrexate protocol for primary central nervous system lymphoma: comparison of two different protocols at a single institution. Neurol Med Chir. 2013;53:797-804.
14. Barker FG 2nd, Curry WT Jr, Carter BS. Surgery for primary supratentorial brain tumors in the United States, 1988 to 2000: the effect of provider case-load and centralization of care. Neuro Oncol. 2005;7:49-63.
15. Kenai H, Yamashita M, Nakamura T, et al. Gamma Knife surgery for primary central nervous system lymphoma: usefulness as palliative local tumor control. J Neurosurg. 2006;105(suppl):133-8.