Академический Документы
Профессиональный Документы
Культура Документы
DOI 10.1007/s11060-010-0512-2
123
546 J Neurooncol (2011) 104:545551
type of hydrocephalus, ventricular opening We placed VPSs in a routine manner and used a pressure
pressure, symptom improvement after VPS, adjustable programmable valve (The CODMAN HAKIM
Karnofsky performance status (KPS) scale Programmable Valve System or The MIETHKE proGAV
change, procedure-related complications, cancer Shunt System) or a fixed-pressure valve (The CODMAN
treatments before and after VPS, systemic HAKIM Precision Fixed Pressure valve). High pressure
disease status at time of death, and post-VPS and valves were used in all cases (Fig. 1).
overall survival. We measured the height of CSF in the catheter from
CNS metastasis was diagnosed based on external auditory canal when we inserted the proximal
magnetic resonance image (MRI) findings, catheter into the ventricle. We defined this height as the
pathologic findings, or both. The presence of opening pressure. In our institution, endotracheal general
LMS at the time of VPS was diagnosed by MRI, anesthesia was maintained with neck position enabling good
by the presence of malignant cells in CSF, or venous drainage, no positive end expiratory pressure
both. Our definition of positive MRI finding is (PEEP), end tidal CO2 around 3035 mmHg, using
contrastenhancement of diffuse or localized osmotherapy (e.g., mannitol), and using propofol instead of
areas such as meninges, ependyma, tentorium, gas anesthetics as much as possible.
basal cistern and sulci. Since CSF cytology test The outcomes of VPS were analyzed with respect to the
for malignant cells has a 4050% falsenegative following risk factors; age ([55 versus B55 years), sex, type
rate, three serial spinal taps were performed to of primary malignancy (lung cancer versus breast cancer
reduce such a high rate. After cytospin versus others), onset of CNS involvement (synchronous
preparations, cells were evaluated, and if versus metachronous), systemic disease status (progressive
malignant cells were confirmed in CSF, we could versus non-progressive), hydrocephalus type
define cytologically positive LMS. (communicating versus obstructive), ventricular opening
Hydrocephalus was classified as obstructive pressure ([30 versus 1530 versus B15 cmH2O), and
hydrocephalus with a large parenchymal or additional cancer treatment after VPS. Univariate analysis
intraventricular mass obstructing the CSF was used to identify potential risk factors associated with
pathway or as communicating hydrocephalus post-VPS survival, and the KaplanMeier method was used
with leptomeningeal enhancement by MRI or to evaluate survival rates for different risk factors. The Log
positive CSF cytology (Fig. 1). We also rank test was used to compare group survival curves, and
documented the time intervals between initial multivariate analysis was performed using the Cox
primary tumor diagnosis, CNS involvement, and proportional hazards model to identify independent risk
VPS. In cases positive for LMS, the time factors.
intervals between brain parenchymal metastasis,
Fig. 1 Image findings of a breast
cancer patient with CNS
involvement. Left axial CT scan
shows ventriculomegaly,
periventricular low density, and
bilateral sulci obliteration. Post-
VPS CT scan (right) shows
much relieved
ventriculomegaly, loss of
periventricular low density, and
reappearance of bilateral sulci
123
J Neurooncol (2011) 104:545551 547
mental change (24%), gait disturbance (14%), consciousness. Cognitive function and urinary incontinence
cognitive dysfunction (e.g., memory did not respond as well to VPS (only 2 of 5 patients
impairment) (10%), urinary incontinence (10%), improved) (Table 1). Most of the patients were relieved from
seizure (4%), and cranial nerve palsy (2%). symptoms within 12 days after the operation. Complication
CNS metastasis was from a hematologic rate was 10%. Four patients experienced a valve pressure
malignancy (diffuse large B-cell lymphoma) in 1 adaptation failure, increased ICP, an overdrain, or
patient and from solid tumors in 49. Regarding uncontrolled hydrocephalus and underwent revision to
the solid tumors, the primary diagnosis was lung change the valve type. One death was attributed to
cancer in 32 patients (non-small cell lung cancer intracerebral hematoma probably provoked by multiple
in 29, small cell lung cancer in 2, and mixed in 1 ventricular puncture because of an inadequate initial
patient), breast cancer (invasive ductal trajectory. There were no coagulation disorders, abnormal
carcinoma) in 8, colorectal cancer in 3, renal cell vascular structures, or tumor which was blocking the
carcinoma in 2, and other cancers (gastric cancer, catheter trajectory for this patient. And no other shunt-
cervix cancer, ovarian cancer, and cancer of related complication, such as infection, valve malfunction,
unknown origin) in 5. or peritoneal carcinomatosis, was observed during the study
Thirty-seven patients were found to have CNS period.
involvement of metachronous onset and the After VPS, further systemic chemotherapy, WBRT,
remaining 13 patients synchronous onset. At the radiosurgery, and tumor resection performed in 11, 6, 4, and
time of hydrocephalus diagnosis and VPS, 10 1 patients, respectively. Intrathecal chemotherapy after VPS
patients had parenchymal metastases only and 40 was not considered because an onoff valve was not
patients had coexistent LMS. In these 40 LMS available and we did not believe it would work at that stage
positive patients, 18 patients had parenchymal of disease with an extremely high ICP and a blockage of
metastases prior to LMS and 22 patients were CSF flow. Five patients remained alive at the end of this
found to have brain parenchymal involvement study in May 2010. At the time of death, we found that
and LMS simultaneously. Primary diseases of systemic malignant tumor status had progressed in 27, but
LMS positive patients were 25 non-small cell no definitive progression was observed in 23. Deaths were
lung cancer, 8 breast cancer, 2 colorectal cancer, attributed to an increased ICP due to CNS lesion progression
and 5 other cancers. (gastric cancer, cervix in 10 patients, and to an aggravated general condition or an
cancer, ovarian cancer, and cancer of unknown organ failure in 35.
origin) And among them, synchronous Median time from initial primary malignancy diagnosis
development of LMS and intraparenchymal to CNS involvement was 10.5 months (range 0106), and
metastases was from 14 non-small cell lung from CNS involvement to VPS was 3.5 months (023). In
cancer, 1 small cell lung cancer, 3 breast cancer, 40 patients with LMS, median time from brain parenchymal
and other cancers. (1 gastric cancer, 1 colon metastasis to LMS was 0 months (023), and from LMS to
cancer, 1 diffuse large B-cell lymphoma, and 1 VPS was 1.0 month (2 days to 13 months). There were 2
unknown origin) In terms of hydrocephalus type, patients who had developed hydrocephalus and underwent
37 were of the communicating type and 13 were VPS before LMS appeared, 1 and 2 months
of the obstructive type, and of these 13, 6 were
blocked by a supratentorial and 7 by an
infratentorial lesion.
Before VPS, patients had been treated for
CNS lesions using various modalities; 23 by
whole-brain radiotherapy (WBRT), 24 by
radiosurgery, 23 by intrathecal chemotherapy,
and 5 by tumor resection. During surgery,
ventricular opening pressure was B15 cmH2O in
5 patients, 1530 cmH2O in 22, and [30 cmH2O
in 23. After VPS, symptoms improved in 40
patients, and most of these improvements
involved relief from headache, nausea, and/ or
vomiting, and an improved level of
123
548 J Neurooncol (2011) 104:545551
123
J Neurooncol (2011) 104:545551 549
Table 2 Characteristics of the patients Primary malignancy
Lung cancer [NSCLC/SCLC/Mixed] 32 (64) [29 (58)/2 (4)/1 (2)]
Breast cancer 8 (16)
Others 5 (10)
P value
123
550 J Neurooncol (2011) 104:545551
Fig. 3 Post-VPS survival in patients who received additional cancer Fig. 4 Post-VPS survival in patients according to ventricular opening
treatment versus those who received conservative management only. pressure. After VPS, the hazard ratio for patients with a ventricular
After VPS, the hazard ratio for death of patients who received further opening pressure[30 cmH2O versus those with a pressure B15 cmH2O
cancer treatment versus those who received supportive care alone was was 6.44 (95% CI 1.2632.9, P = 0.02)
0.17 (95% CI 0.070.42, P\0.0001)
associations between LMS and hydrocephalus or anesthesia and of various shunt-related complications [9
intracranial hypertension. In the present study, 80% of 11], and improves survival and favorably affects quality of
hydrocephalus patients who underwent VPS were found to life
have coexistent LMS. The dominant type of hydrocephalus [4].
was the communicating type associated with LMS, and the A median survival time of 1.85.8 months after a
median time interval between the diagnosis of LMS and diagnosis of LMS [2, 4, 5, 1214] and of 2.0 months
VPS was only 2.2 months. Accordingly, we recommend a
thorough investigation for LMS be undertaken when a
cancer patient develops neurological symptoms and signs of
hydrocephalus and/or images of ventriculomegaly. If LMS
is confirmed, its treatment should be actively pursued
without delay before VPS. The early diagnosis of LMS is
important because the opportunity to administer intrathecal
chemotherapy might be missed after VPS for
hydrocephalus. Furthermore, if hydrocephalus persists,
intrathecal chemotherapy cannot provide a viable treatment
option because CSF flow is inhibited, that is, CSF
compartmentalization and prohibits intrathecal
chemoagents being delivered effectively [2, 5, 7, 8].
In the present study, 80% of patients experienced
immediate post-operative symptom relief. A headache was
the most common symptom, and was resolved by surgery in
85.7% of cases. On the other hand, cognitive function and
urinary incontinence were improved in less than a half of
cases by VPS (40%). Unlike cancer treatment itself, the
treatment of intracranial hypertension is not optional
because of its debilitating consequences. When relief of the
symptoms and signs of increased ICP are of major concern,
VPS helps enormously despite the risks of general
123
J Neurooncol (2011) 104:545551 551
Table 4 Results of multivariate Cox Hazard ratio (95% CI) P value
regression analysis
Primary malignancy
LCA versus others
0.65 (0.172.48) 0.94
BCA versus others 0.45 (0.102.05) 0.48
Ventricular opening pressure: [30 versus B15 cmH2O 6.44 (1.2632.9) 0.02
(2 days to 3.6 years) after VPS [3] have been reported, which pressure, which reflects an increased ICP, probably
concur with our findings (3.5 months of median survival indicates severe distortion of CSF dynamics due to a long-
after a diagnosis of LMS and 3.0 months after VPS). standing, heavy tumor burden in the CNS when it is from a
Approximately 2434% of CNS metastasis patients communicating hydrocephalus, and although VPS can
succumb to LMS alone, 2225% to LMS and progressive effectively relieve an elevated ICP, a poor prognosis is
systemic cancer, 1944% to systemic progressive disease, expected due to the progression of the underlying disease.
and up to 10% to other causes [1518]. In the present study, When the hydrocephalus is obstructive, the lesion is usually
the majority of patients succumbed to aggravated systemic deep-seated around the third or fourth ventricle. Surgical
disease progression (70%), which included systemic cancer resection could be incomplete and even impossible, and this
progression and an aggravated general condition mainly due affects the survival unfavorably.
to an infection, like pneumonia. In a previous study, it was found that primary tumor
Furthermore, overall outcome was in line with that found histology has a significant impact on survival after a
previously, although it is noteworthy that five patients diagnosis of LMS. The longest median survival was
(10%) survived after VPS for more than 1 year and one observed in breast cancer (3.1 months), and the shortest in
patient on continuing systemic cancer treatment survived 54 lung cancer (0.8 months) [5]. However, we failed to find any
months. These patients who achieved long-term survival significant relationship between primary malignancy type
after VPS support the notion that LMS development and and survival after VPS; for example, median survival times
VPS do not necessarily mean the end of the further cancer after VPS in lung cancer, breast cancer, and other primary
treatment. Aggressive treatments in patients with CNS malignancies were 3.0, 2.0, and 2.5 months, respectively
metastasis have been reported to substantially increase (P= 0.93). Our findings could be explained by the fact that
survival [2, 46, 1315, 1921]. We also found that further CNS involvement with LMS represents the most advanced
aggressive cancer treatment after VPS was a prognostic stage of cancer irrespective of the nature of the primary
factor of better survival time by univariate and multivariate malignancy, and that short life expectancies do not allow
analysis. The most reasonable explanation for this result is survival differences to be attributed to different histologies.
that we selected patients with a better performance status To improve survival, further research should be focused
and a lower tumor burden, which increased the likelihood of on improving diagnostic and therapeutic tools. More
eligibility for further systemic or local cancer treatment. sensitive diagnostic tools are needed to overcome the high
Furthermore, this finding strengthens the hypothesis that the falsenegative rates of imaging and cytology findings.
bloodbrain barrier is made more permeable by infiltrative Furthermore, personalized and tumor-specific therapeutic
tumor cells and that systemic chemotherapy agents can regimens should be established. At present, systemic or
penetrate into the CSF space more easily, which may intrathecal chemotherapies based on methotrexate,
favorably affect survival after VPS. Multivariate analysis cytarabine, liposomal cytarabine, or thiotepa and on several
also identified an independent risk factor that negatively emerging drugs, such as mafosfamide, diaziquone,
affected survival after VPS, namely a ventricular opening topotecan, gemcitabine, and temozolomide, are being
pressure of [30 cmH2O. An elevated ventricular opening investigated to identify regimens best suited for particular
123
552 J Neurooncol (2011) 104:545551
types of malignant tumors [2, 6, 15]. In addition, novel with leptomeningeal metastases of different primary tumors: a
therapeutic modalities for LMS are required, such as agents retrospective analysis. J Cancer Res Clin Oncol (in press)
6. Taillibert S, Laigle-Donadey F, Chodkiewicz C, Sanson M,
that can overcome CSF flow stagnation and a VPS valve Hoang-Xuan K, Delattre JY (2005) Leptomeningeal metastases
system that enables both CSF shunting and intrathecal from solid malignancy: a review. J Neurooncol 75:8599
chemotherapy. 7. Glantz MJ, Cole BF, Recht L, Akerley W, Mills P, Saris
As post-operative survival gets longer, unanticipated side S,Hochberg F, Calabresi P, Egorin MJ (1998) High-dose
intravenous methotrexate for patients with nonleukemic
effects may be unveiled such as shunt infection, malfunction leptomeningeal cancer: is intrathecal chemotherapy necessary? J
and even rarer peritoneal carcinomatosis. Considering Clin Oncol 16:15611567
reports from primary malignant brain tumors disseminating 8. Lassman AB, Abrey LE, Shah GD, Panageas KS, Begemann
through peritoneal catheter, it is likely that the same risk M,Malkin MG, Raizer JJ (2006) Systemic high-dose intravenous
methotrexate for central nervous system metastases. J
exist in the patients with LMS who undergo CSF diversion. Neurooncol 78:255260
The main reason why we could not find the patient with 9. Berger MS, Baumeister B, Geyer JR, Milstein J, Kanev
peritoneal carcinomatosis seems to be the short survival PM,LeRoux PD (1991) The risks of metastases from shunting in
period of the patients after VPS. Recently, we experienced a children with primary central nervous system tumors. J
Neurosurg 74:872877
case not included in analysis of this manuscript in which
10. Lobotesis K, UK-I JM, Cross JJ, Gillard JH, Antoun NM (2009)
malignant ascites developed 1 month after VPS for the Gliomatosis peritonei associated with a ventriculo-peritoneal
hydrocephalus with LMS. If survival of the patient is shunt. Clin Radiol 64:9599
prolonged after VPS, it is expected that the risk of peritoneal 11. Zemack G, Romner B (2000) Seven years of clinical
experiencewith the programmable Codman Hakim valve: a
seeding will get higher.
retrospective study of 583 patients. J Neurosurg 92:941948
Lastly, a comparative study with patients who had not 12. Chamberlain MC, Glantz M (2006) Ventriculoperitoneal shuntin
undergone VPS to look for the difference in survival in both patients with leptomeningeal metastasis. Neurology 66:783,
groups will be performed in the near future and be able to author reply 783
define the role of VPS further, if there were no practical 13. DeAngelis LM, Boutros D (2005) Leptomeningeal
metastasis.Cancer Invest 23:145154
limitation nor ethical problems. 14. Pentheroudakis G, Pavlidis N (2005) Management of
leptomeningeal malignancy. Expert Opin Pharmacother 6:1115
1125
Conclusions 15. Groves MD (2010) New strategies in the management of
leptomeningeal metastases. Arch Neurol 67:305312
16. Chamberlain MC, Kormanik P (1998) Carcinoma
Hydrocephalus related to CNS metastasis requiring VPS is meningitissecondary to non-small cell lung cancer: combined
commonly associated with LMS. Though the prognosis of modality therapy. Arch Neurol 55:506512
CNS metastasis and subsequent hydrocephalus is dismal, 17. Chamberlain MC, Kormanik PR (1997) Carcinomatous
VPS is an effective palliative measure, and adequate cancer meningitis secondary to breast cancer: predictors of response to
combined modality therapy. J Neurooncol 35:5564
treatment after VPS may offer the best means of improving 18. Jaeckle KA, Phuphanich S, Bent MJ, Aiken R, Batchelor
survival. Effective diagnostic and therapeutic modalities are T,Campbell T, Fulton D, Gilbert M, Heros D, Rogers L, ODay
required to control LMS and improve outcome. SJ, Akerley W, Allen J, Baidas S, Gertler SZ, Greenberg HS,
LaFollette S, Lesser G, Mason W, Recht L, Wong E,
Chamberlain MC, Cohn A, Glantz MJ, Gutheil JC, Maria B,
Moots P, New P, Russell C, Shapiro W, Swinnen L, Howell SB
(2001) Intrathecal treatment of neoplastic meningitis due to
References breast cancer with a slow-release formulation of cytarabine. Br J
Cancer 84:157163 19. Kesari S, Batchelor TT (2003)
1. Kehrli P (1999) Epidemiology of brain metastases. Leptomeningeal metastases. Neurol Clin 21:2566
Neurochirurgie 45:357363 20. Barker FG 2nd (2005) Surgical and radiosurgical management
2. Herrlinger U, Forschler H, Kuker W, Meyermann R, Bamberg ofbrain metastases. Surg Clin North Am 85:329345
M,Dichgans J, Weller M (2004) Leptomeningeal metastasis: 21. Shapiro WR, Johanson CE, Boogerd W (2009)
survival and prognostic factors in 155 patients. J Neurol Sci 223: Treatmentmodalities for leptomeningeal metastases. Semin Oncol
167178 36:S46 S54
3. Omuro AM, Lallana EC, Bilsky MH, DeAngelis LM (2005)
Ventriculoperitoneal shunt in patients with leptomeningeal
metastasis. Neurology 64:16251627
4. Taillibert S, Hildebrand J (2006) Treatment of central
nervoussystem metastases: parenchymal, epidural, and
leptomeningeal. Curr Opin Oncol 18:637643
5. Oechsle K, Lange-Brock V, Kruell A, Bokemeyer C, de Wit
M(2010) Prognostic factors and treatment options in patients
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.