- Email: [email protected]
Stereotactic catheter placement for Ommaya reservoirs
Journal of Clinical Neuroscience 27 (2016) 44–47
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Stereotactic catheter placement for Ommaya reservoirs Benjamin C. Kennedy a,⇑, Lauren T. Brown a, Ricardo J. Komotar b, Guy M. McKhann II a a b
Department of Neurological Surgery, The Neurological Institute, Columbia University, 710 West 168th Street, 4th floor, New York, NY 10032, USA Department of Neurological Surgery, Miami University, Miami, FL, USA
a r t i c l e
i n f o
Article history: Received 8 November 2015 Accepted 21 November 2015
Keywords: Frame-based Frameless Ommaya Stereotaxy
a b s t r a c t Ommaya reservoirs are an important surgical therapy for the chronic intrathecal administration of chemotherapy for patients with leptomeningeal carcinomatosis. Surgical accuracy is paramount in these patients with typically normal sized ventricles, and may be improved with stereotactic guidance. This paper aimed to review a large series of stereotactic Ommaya catheter placements, examining accuracy and complications. We conducted a retrospective review of 109 consecutive adult patients who underwent stereotactic Ommaya catheter placement for leptomeningeal carcinomatosis or central nervous system lymphoma at Columbia University Medical Center, USA, from 1998–2013. The rate of accurate placement in the ventricular system was 99%, with the only poor catheter position due to postplacement migration. The rate of peri-operative complications was 6.4%. Hemorrhagic complications occurred in patients with thrombocytopenia or therapeutic anti-coagulation pre-operatively or during the post-operative period. Use of stereotaxy for catheter placement of Ommaya reservoirs is safe and effective, and should be considered when placing a catheter into non-hydrocephalic ventricles. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The use of Ommaya reservoirs is a common method of intrathecal anti-neoplastic medication delivery for patients with malignancies with diffuse leptomeningeal involvement. Ventricular catheters are often placed in patients with ventriculomegaly, while most patients with leptomeningeal metastasis or central nervous system lymphoma requiring Ommaya placement are not hydrocephalic, and their normal or small ventricular size demands a more spatially accurate catheter placement. For this patient population with diminished ventricular volume, use of stereotaxy for placement of these devices may improve accuracy and safety [1,2]. Improved accuracy may also reduce the number of catheter passes necessary to achieve proper placement. Fewer passes reduces disruption of parenchyma, and in this population with frequent bone marrow dysfunction and thrombocytopenia, may decrease hemorrhagic complications. Series of non-stereotactically guided Ommaya reservoir placement have reported complication rates of 7–10% [3–5]. Complications include malposition of the catheter, hemorrhage, neurologic deficit, and bacterial meningitis and catheter infection. Small series of experience with stereotaxy for placement of Ommaya reservoirs have reported complication rates of 0–16% ⇑ Corresponding author. Tel.: +1 847 962 0696; fax: +1 212 305 2026. E-mail address: [email protected] (B.C. Kennedy). http://dx.doi.org/10.1016/j.jocn.2015.11.005 0967-5868/Ó 2015 Elsevier Ltd. All rights reserved.
[6–11]. In the present study, which represents one of the largest single center series of stereotactic Ommaya catheter placement, we demonstrate that stereotactic Ommaya placement is extremely accurate and associated with a low complication rate. 2. Methods After Institutional Review Board approval, all patients undergoing frame-based and frameless image-guided stereotactic Ommaya reservoir placement procedures at Columbia University Medical Center, USA, between 1998 and 2013 were identified. Patient demographics, diagnoses, clinical presentations, radiological studies, laboratory values, and clinical outcomes were reviewed. End points reviewed included rate of successful placement, revision, removal, procedure-related neurological morbidity, infection, intracranial hemorrhage, conversion to ventriculoperitoneal shunt (VPS), direct parenchymal toxicity, and 30 day in-hospital mortality. Generally, all patients had a pre-operative volumetric MRI or CT scan of the head performed within 36 hours of the procedure. For frameless stereotactic procedures, general anesthesia or intravenous sedation was induced, depending on surgeon and patient preference. The patient’s head was subsequently fixed in a three-point Mayfield clamp secured to the operating table. Pre-operative images were transferred to the operating room workstation, and intra-operative image guidance was performed using a wand-based navigation system. The scalp anatomy was
45
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
co-registered to the volumetric imaging study using either surface registration with BrainLAB Z-touch software (BrainLAB AG, Munich, Germany) or fiducial registration with the Stealth system (Medtronic Sofamor Danek, Memphis, TN, USA). A linear incision was outlined in the right frontal area centered over the middle frontal gyrus and trajectory planning carried out down to the lateral ventricle. Specifically, the trajectory was planned to have all distal catheter holes within the ventricular system, minimizing the possibility of the more superficial holes ending in the periventricular white matter. To do this, the burr hole was usually lateral to Kocher’s point to allow the ventricular entry point to be the anterolateral corner of the frontal horn of the lateral ventricle, with a maximized intraventricular catheter course toward the foramen of Monro. Intra-operatively, the Ommaya catheter was advanced the stereotactically predetermined length into the frontal horn of the lateral ventricle, and cerebrospinal fluid (CSF) flow was confirmed. The catheter was then soft passed without the stylet to its final premeasured length to place the perforations within the ventricular system. When technically available, a registered stereotactic stylet provided real-time visual feedback during the catheter pass. The Ommaya reservoir was then placed medial or posterior to the incision in the subgaleal space. The ventricular catheter was then secured to the reservoir and anchored to the pericranium. The first dose of intrathecal chemotherapy was often administered if desired by the treating oncologist. Patients were closed in routine fashion, and patency of the apparatus was again confirmed by pumping the reservoir. The basics of the frame-based procedure were similar, with the following differences. The Cosman–Roberts–Wells (CRW) head frame was affixed to the skull under local or general anesthesia, depending on surgeon and patient preference, and a volumetric head CT scan was obtained. This was used to plan a trajectory for catheter insertion at the BrainLAB workstation. The trajectory specifics were similar to those described above for the frameless procedures. Coordinates for the target trajectory as well as ring and arc parameters for trajectory were obtained, and the placement of the catheter was performed using the head frame rather than the frameless system. A long ventricular stylet was required to place the Ommaya catheter using the CRW frame in order to accommodate the target length of the CRW system of 175 mm, including block and collar. 3. Results 3.1. Demographics One hundred nine stereotactic Ommaya reservoir placement procedures were performed at Columbia University Medical Center between 1998 and 2013 for central nervous system involvement of various malignancies. The mean patient age was 51 years. The study cohort consisted of 47 male patients (43%). 3.2. Accuracy Thirty-five of 109 (32%) patients had mild ventricular dilation, and the other 74 patients (68%) had normal or small ventricles. Two (1.8%) patients required two passes of the catheter to confirm CSF flow. Ninety-three (85%) patients had post-operative axial imaging. The initial scan showed good placement in 92/93 (99%) scans. The one initial scan showing suboptimal placement showed the catheter too deep with a kink in the catheter in the frontal white matter and the reservoir just outside the burr hole, suggesting post-placement migration prior to the CT scan. All other 92
catheters were judged by the attending neurosurgeon to be in good position with all the catheter holes within the ventricular system, without need to be replaced, pulled back, or advanced (Table 1). As a result, no patient demonstrated evidence of toxicity due to direct delivery of chemotherapy into brain parenchyma. 3.3. Revisions Eight (7.3%) patients required further surgery due to Ommaya malfunction, infection, or poor placement; four catheters required removal for infection and four were revised. The infections were diagnosed on post-operative day (POD) 5, 30, 31, and 44, respectively, followed by immediate removal of the entire system without replacement. One of the four revisions was the aforementioned immediate post-operative catheter migration, and the other three all demonstrated good position on post-operative CT scan, and malfunctioned for different reasons. One worked well for over a month but was noted on POD 40 to be malfunctioning, at which point a CT scan showed migration into the thalamus, and the catheter was revised on POD 41. One malfunctioned on POD 12, and CT scan at that time showed stable position but a small amount of pericatheter hemorrhage, and was found to be clotted at revision. This patient’s platelet level had fallen to 2 on POD 3. The fourth revised catheter malfunctioned on POD 4 with no change on CT scan, and was revised. In addition to these eight patients, four patients developed symptomatic hydrocephalus and their catheters were revised to VPS (Table 1). 3.4. Hemorrhagic complications Seven (6.4%) patients experienced hemorrhages, four (3.7%) symptomatic and three (2.8%) asymptomatic. One patient was the aforementioned small pericatheter hemorrhage resulting in clotting and malfunctioning of the catheter. One patient had platelet level of 3 on the morning of surgery and had previously responded to platelet transfusions with an appropriate increase. However, despite intra-operative transfusion of 12 units of platelets, the patient experienced intra-operative bleeding after an initially clear CSF pass. Severe thrombocytopenia remained refractory to transfusions, and the surgery was aborted without leaving any implant. The hemorrhage grew slowly over the next several days with the platelet level remaining refractory to transfusions. The patient was eventually listed as do not resuscitate and died. One patient had a post-operative CT scan without acute blood, started therapeutic enoxaparin, and developed altered mental status on POD 3 with tract hemorrhage and intraventricular
Table 1 Accuracy and complications in stereotactic catheter placement for Ommaya reservoirs Number of patients (%) Total patients Accuracy Normal sized ventricles Only one catheter pass Had post-operative scan Good radiographic placement Complications Total peri-operative complications Hemorrhagic complications Symptomatic hemorrhage Peri-operative malfunctions Delayed malfunctions Peri-operative infections Delayed infections Conversion to VPS VPS = ventriculoperitoneal shunt.
109 (100) 74/109 (68) 107/109 (98) 93/109 (85) 92/93 (99) 7/109 7/109 4/109 3/109 1/109 1/109 3/109 4/109
(6.4) (6.4) (3.7) (2.8) (0.9) (0.9) (2.8) (3.7)
46
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
hemorrhage noted on CT scan, developed status epilepticus, and eventually died. The fourth symptomatic hemorrhage occurred in another patient with a post-operative CT scan negative for acute blood, became somnolent on POD 9 with a CT scan showing a right-sided subdural hematoma, which worsened on POD 10 prompting surgical evacuation. This patient was eventually discharged home. The patient’s platelet nadir was 41 on POD 10. Three patients had small amounts of asymptomatic acute blood on post-operative CT scan (Table 1). 3.5. Peri-operative morbidity and mortality Within 30 days of surgery, there was one infection, four symptomatic hemorrhages, one further revision for malfunction, and one revision for poor placement, for a total of seven (6.4%) perioperative morbidities. Seventy-one (65%) patients had their death documented in our hospital records. Nineteen of those 71 died within 30 days of their Ommaya surgery. 4. Discussion Though a variety of uses have been described for Ommaya reservoirs, their primary application is the administration of chronic intrathecal chemotherapy for patients with leptomeningeal carcinomatosis, and as such, they can be an invaluable aspect of some patients’ cancer treatment. Series of nonstereotactically guided Ommaya reservoir placement have reported complication rates of 7–10% [3–5], including poor placement. Even with the use of fluoroscopic guidance for this indication, up to 4.7% of catheters can be malpositioned [7]. Stereotactic systems offer an opportunity to improve on this. While use of stereotaxy to place intraventricular catheters for VPS in hydrocephalus patients has been studied in large series, the study of stereotactic catheter placement in leptomeningeal carcinomatosis patients warrants separate study, as the ventricles are commonly not dilated as in the hydrocephalus population. In our series only 32% had mild ventricular dilation, with no patients exhibiting more than mild dilation, rendering the accuracy of the catheter trajectory yet more important. Further, accuracy may reduce the number of catheter passes, and in this patient population with bone marrow dysfunction, at high risk of bleeding and infection, fewer passes could reduce risks of hemorrhage and infection. Further reduced risk of bleeding is possible due to planning a stereotactic trajectory to avoid cortical vessels. Improved accuracy can also avoid direct delivery of chemotherapy to parenchyma and revision surgery to replace the catheter. The published experience with stereotactic placement of catheters for Ommaya reservoirs has consisted of a few small series, with complication rates up to 16% [6–11]. In this largest reported series of stereotactic placement of catheters for Ommaya reservoirs in 109 patients with leptomeningeal carcinomatosis, we have shown excellent accuracy (99%) and low complication rate (6.4%). Regarding accuracy, only two (1.8%) patients required more than one pass despite 68% of the patients having normal or small-sized ventricles, and no patient required more than two passes. Eighty-five percent of patients had post-operative imaging with 99% of catheters in good position on the first post-operative scan, similar to rates reported in prior series of stereotactic ventricular catheter placement [10,11]. Furthermore, the only catheter not in good position, by virtue of the previously extracranial kink seen in the parenchyma and the reservoir seen close to the burr hole, had clearly demonstrated a post-placement migration to a deeper location. This suggests not an issue with the stereotaxy but with a failure of sufficient anchoring to extracranial tissue. There was only one other revision for a catheter being in a suboptimal position, and this was a catheter that had good placement on
post-operative CT scan and migrated into the thalamus and malfunctioned on POD 40, further highlighting the importance of securing the system well to extracranial tissue. It is important to note that our goals of catheter placement in this surgery are for the catheter to be placed safely without hemorrhage, infection, or misplacement, and to ensure that all catheter holes are within the ventricular system. Prior to the institution of stereotactic placement of Ommaya catheters at our institution, one of our patients had a catheter placed with some of the catheter holes in the ventricle and some remaining proximally in the parenchyma, and when the more distal ventricular holes became occluded, the patient experienced a severe toxicity due to parenchymal delivery of chemotherapy. This experience helped clarify these goals of surgery. We always strive to have all of the 2 cm of catheter holes within the ventricular system to avoid the potential toxicity of chemotherapy being delivered directly to the parenchyma. For this reason, the criterion used by our neurosurgeons clinically and for the purposes of these analyses for accurate catheter placement for Ommaya surgery was that all catheter holes were in the ventricular system, rather than that the catheter tip was within an arbitrary distance from the foramen of Monro, as has been used in other reports [8]. Further, no patient in our series developed clinical or radiographic evidence of toxicity due to direct delivery of chemotherapeutic to brain parenchyma. In this patient population at high risk for infection, we observed only one (0.9%) peri-operative infection, on POD 5. The other three infections occurred at least 30 days post-operatively, and these delayed infections are expected in this population [8,12]. With the exception of the three small, asymptomatic bleeds, all hemorrhagic complications in the present series occurred with either thrombocytopenia pre-operatively or within the first 10 days post-operatively, or initiation of therapeutic anti-coagulation in the first few days post-operatively. The patient with pre-operative severe thrombocytopenia was refractory during surgery to multiple platelet transfusions by clinical and laboratory measures. Thrombocytopenia is naturally a relative contraindication to brain surgery, but in these patients with refractory platelet levels who require chronic intrathecal chemotherapy, it may still be prudent to place an Ommaya reservoir despite the risks. Our findings of delayed hemorrhagic complications in the setting of post-operative thrombocytopenia or anti-coagulation should prompt neurosurgeons and oncologists to maintain vigilance in monitoring platelet counts during this period and maintain a high degree of concern when weighing the risks and benefits of starting therapeutic anti-coagulation. We now require that thrombocytopenic patients respond to transfusion on the day of surgery prior to beginning the surgical procedure. Furthermore, we have moved to schedule our Ommaya placements to correspond with chemotherapy nadirs to avoid thrombocytopenia and leukopenia. This study is limited by its retrospective nature and lack of a control group. Further, the heterogeneity among patients regarding factors that may be important in determining accuracy and safety, including among stereotactic systems used, surgeons, ventricular sizes, and comorbidities including thrombocytopenia can make it difficult to generalize conclusions. 5. Conclusion Neurosurgeons will continue to innovate to make surgeries safer, more effective, and easier, and with this innovation comes the responsibility to study our methods in order to continue to improve the outcomes of our patients. Through our experience in the pre-stereotaxy era at our institution, we have clarified what we believe accuracy means for this particular procedure, that is, that all catheter holes rest within the ventricular system. We have further learned within this series after a hemorrhagic complication
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
that, for these patients on chemotherapy, the transfusionresponsiveness of severe thrombocytopenia, even if demonstrated previously, should be verified on the morning of surgery, and we now schedule Ommaya placement for the chemotherapy nadir to minimize thrombocytopenia and leukopenia. Future studies in this area should include newer stereotaxy systems, such as those that are electromagnetically guided, and direct comparisons among systems, including the frameless and frame-based systems. Conflicts of Interest/Disclosures Portions of the current work have been presented at the American Association of Neurological Surgeons Annual Meeting, April 2008, Chicago, IL, USA. The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. References [1] Hagen NA, O’Neill BP, Kelly PJ. Computer assisted stereotactic placement of Ommaya reservoirs for delivery of chemotherapeutic agents in cancer patients. J Neurooncol 1987;5:273–6. [2] Al-Anazi A, Bernstein M. Modified stereotactic insertion of the Ommaya reservoir. Technical note. J Neurosurg 2000;92:1050–2.
47
[3] Obbens EA, Leavens ME, Beal JW, et al. Ommaya reservoirs in 387 cancer patients: a 15-year experience. Neurology 1985;35:1274–8. [4] Lishner M, Perrin RG, Feld R, et al. Complications associated with Ommaya reservoirs in patients with cancer. The Princess Margaret Hospital experience and a review of the literature. Arch Intern Med 1990;150:173–6. [5] Bleyer WA, Pizzo PA, Spence AM, et al. The Ommaya reservoir: newly recognized complications and recommendations for insertion and use. Cancer 1978;41:2431–7. [6] Dickerman RD, Eisenberg MB. Preassembled method for insertion of Ommaya reservoir. J Surg Oncol 2005;89:36–8. [7] Sandberg DI, Bilsky MH, Souweidane MM, et al. Ommaya reservoirs for the treatment of leptomeningeal metastases. Neurosurgery 2000;47:49–54 [discussion 54–5]. [8] Greenfield JP, Schwartz TH. Catheter placement for Ommaya reservoirs with frameless surgical navigation: technical note. Stereotact Funct Neurosurg 2008;86:101–5. [9] Chamberlain MC, Kormanik PA, Barba D. Complications associated with intraventricular chemotherapy in patients with leptomeningeal metastases. J Neurosurg 1997;87:694–9. [10] Takahashi M, Yamada R, Tabei Y, et al. Navigation-guided Ommaya reservoir placement: implications for the treatment of leptomeningeal metastases. Minim Invasive Neurosurg 2007;50:340–5. [11] Sampath R, Wadhwa R, Tawfik T, et al. Stereotactic placement of ventricular catheters: does it affect proximal malfunction rates? Stereotact Funct Neurosurg 2012;90:97–103. [12] Szvalb AD, Raad II, Weinberg JS, et al. Ommaya reservoir-related infections: clinical manifestations and treatment outcomes. J infect 2014;68:216–24.
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Stereotactic catheter placement for Ommaya reservoirs Benjamin C. Kennedy a,⇑, Lauren T. Brown a, Ricardo J. Komotar b, Guy M. McKhann II a a b
Department of Neurological Surgery, The Neurological Institute, Columbia University, 710 West 168th Street, 4th floor, New York, NY 10032, USA Department of Neurological Surgery, Miami University, Miami, FL, USA
a r t i c l e
i n f o
Article history: Received 8 November 2015 Accepted 21 November 2015
Keywords: Frame-based Frameless Ommaya Stereotaxy
a b s t r a c t Ommaya reservoirs are an important surgical therapy for the chronic intrathecal administration of chemotherapy for patients with leptomeningeal carcinomatosis. Surgical accuracy is paramount in these patients with typically normal sized ventricles, and may be improved with stereotactic guidance. This paper aimed to review a large series of stereotactic Ommaya catheter placements, examining accuracy and complications. We conducted a retrospective review of 109 consecutive adult patients who underwent stereotactic Ommaya catheter placement for leptomeningeal carcinomatosis or central nervous system lymphoma at Columbia University Medical Center, USA, from 1998–2013. The rate of accurate placement in the ventricular system was 99%, with the only poor catheter position due to postplacement migration. The rate of peri-operative complications was 6.4%. Hemorrhagic complications occurred in patients with thrombocytopenia or therapeutic anti-coagulation pre-operatively or during the post-operative period. Use of stereotaxy for catheter placement of Ommaya reservoirs is safe and effective, and should be considered when placing a catheter into non-hydrocephalic ventricles. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The use of Ommaya reservoirs is a common method of intrathecal anti-neoplastic medication delivery for patients with malignancies with diffuse leptomeningeal involvement. Ventricular catheters are often placed in patients with ventriculomegaly, while most patients with leptomeningeal metastasis or central nervous system lymphoma requiring Ommaya placement are not hydrocephalic, and their normal or small ventricular size demands a more spatially accurate catheter placement. For this patient population with diminished ventricular volume, use of stereotaxy for placement of these devices may improve accuracy and safety [1,2]. Improved accuracy may also reduce the number of catheter passes necessary to achieve proper placement. Fewer passes reduces disruption of parenchyma, and in this population with frequent bone marrow dysfunction and thrombocytopenia, may decrease hemorrhagic complications. Series of non-stereotactically guided Ommaya reservoir placement have reported complication rates of 7–10% [3–5]. Complications include malposition of the catheter, hemorrhage, neurologic deficit, and bacterial meningitis and catheter infection. Small series of experience with stereotaxy for placement of Ommaya reservoirs have reported complication rates of 0–16% ⇑ Corresponding author. Tel.: +1 847 962 0696; fax: +1 212 305 2026. E-mail address: [email protected] (B.C. Kennedy). http://dx.doi.org/10.1016/j.jocn.2015.11.005 0967-5868/Ó 2015 Elsevier Ltd. All rights reserved.
[6–11]. In the present study, which represents one of the largest single center series of stereotactic Ommaya catheter placement, we demonstrate that stereotactic Ommaya placement is extremely accurate and associated with a low complication rate. 2. Methods After Institutional Review Board approval, all patients undergoing frame-based and frameless image-guided stereotactic Ommaya reservoir placement procedures at Columbia University Medical Center, USA, between 1998 and 2013 were identified. Patient demographics, diagnoses, clinical presentations, radiological studies, laboratory values, and clinical outcomes were reviewed. End points reviewed included rate of successful placement, revision, removal, procedure-related neurological morbidity, infection, intracranial hemorrhage, conversion to ventriculoperitoneal shunt (VPS), direct parenchymal toxicity, and 30 day in-hospital mortality. Generally, all patients had a pre-operative volumetric MRI or CT scan of the head performed within 36 hours of the procedure. For frameless stereotactic procedures, general anesthesia or intravenous sedation was induced, depending on surgeon and patient preference. The patient’s head was subsequently fixed in a three-point Mayfield clamp secured to the operating table. Pre-operative images were transferred to the operating room workstation, and intra-operative image guidance was performed using a wand-based navigation system. The scalp anatomy was
45
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
co-registered to the volumetric imaging study using either surface registration with BrainLAB Z-touch software (BrainLAB AG, Munich, Germany) or fiducial registration with the Stealth system (Medtronic Sofamor Danek, Memphis, TN, USA). A linear incision was outlined in the right frontal area centered over the middle frontal gyrus and trajectory planning carried out down to the lateral ventricle. Specifically, the trajectory was planned to have all distal catheter holes within the ventricular system, minimizing the possibility of the more superficial holes ending in the periventricular white matter. To do this, the burr hole was usually lateral to Kocher’s point to allow the ventricular entry point to be the anterolateral corner of the frontal horn of the lateral ventricle, with a maximized intraventricular catheter course toward the foramen of Monro. Intra-operatively, the Ommaya catheter was advanced the stereotactically predetermined length into the frontal horn of the lateral ventricle, and cerebrospinal fluid (CSF) flow was confirmed. The catheter was then soft passed without the stylet to its final premeasured length to place the perforations within the ventricular system. When technically available, a registered stereotactic stylet provided real-time visual feedback during the catheter pass. The Ommaya reservoir was then placed medial or posterior to the incision in the subgaleal space. The ventricular catheter was then secured to the reservoir and anchored to the pericranium. The first dose of intrathecal chemotherapy was often administered if desired by the treating oncologist. Patients were closed in routine fashion, and patency of the apparatus was again confirmed by pumping the reservoir. The basics of the frame-based procedure were similar, with the following differences. The Cosman–Roberts–Wells (CRW) head frame was affixed to the skull under local or general anesthesia, depending on surgeon and patient preference, and a volumetric head CT scan was obtained. This was used to plan a trajectory for catheter insertion at the BrainLAB workstation. The trajectory specifics were similar to those described above for the frameless procedures. Coordinates for the target trajectory as well as ring and arc parameters for trajectory were obtained, and the placement of the catheter was performed using the head frame rather than the frameless system. A long ventricular stylet was required to place the Ommaya catheter using the CRW frame in order to accommodate the target length of the CRW system of 175 mm, including block and collar. 3. Results 3.1. Demographics One hundred nine stereotactic Ommaya reservoir placement procedures were performed at Columbia University Medical Center between 1998 and 2013 for central nervous system involvement of various malignancies. The mean patient age was 51 years. The study cohort consisted of 47 male patients (43%). 3.2. Accuracy Thirty-five of 109 (32%) patients had mild ventricular dilation, and the other 74 patients (68%) had normal or small ventricles. Two (1.8%) patients required two passes of the catheter to confirm CSF flow. Ninety-three (85%) patients had post-operative axial imaging. The initial scan showed good placement in 92/93 (99%) scans. The one initial scan showing suboptimal placement showed the catheter too deep with a kink in the catheter in the frontal white matter and the reservoir just outside the burr hole, suggesting post-placement migration prior to the CT scan. All other 92
catheters were judged by the attending neurosurgeon to be in good position with all the catheter holes within the ventricular system, without need to be replaced, pulled back, or advanced (Table 1). As a result, no patient demonstrated evidence of toxicity due to direct delivery of chemotherapy into brain parenchyma. 3.3. Revisions Eight (7.3%) patients required further surgery due to Ommaya malfunction, infection, or poor placement; four catheters required removal for infection and four were revised. The infections were diagnosed on post-operative day (POD) 5, 30, 31, and 44, respectively, followed by immediate removal of the entire system without replacement. One of the four revisions was the aforementioned immediate post-operative catheter migration, and the other three all demonstrated good position on post-operative CT scan, and malfunctioned for different reasons. One worked well for over a month but was noted on POD 40 to be malfunctioning, at which point a CT scan showed migration into the thalamus, and the catheter was revised on POD 41. One malfunctioned on POD 12, and CT scan at that time showed stable position but a small amount of pericatheter hemorrhage, and was found to be clotted at revision. This patient’s platelet level had fallen to 2 on POD 3. The fourth revised catheter malfunctioned on POD 4 with no change on CT scan, and was revised. In addition to these eight patients, four patients developed symptomatic hydrocephalus and their catheters were revised to VPS (Table 1). 3.4. Hemorrhagic complications Seven (6.4%) patients experienced hemorrhages, four (3.7%) symptomatic and three (2.8%) asymptomatic. One patient was the aforementioned small pericatheter hemorrhage resulting in clotting and malfunctioning of the catheter. One patient had platelet level of 3 on the morning of surgery and had previously responded to platelet transfusions with an appropriate increase. However, despite intra-operative transfusion of 12 units of platelets, the patient experienced intra-operative bleeding after an initially clear CSF pass. Severe thrombocytopenia remained refractory to transfusions, and the surgery was aborted without leaving any implant. The hemorrhage grew slowly over the next several days with the platelet level remaining refractory to transfusions. The patient was eventually listed as do not resuscitate and died. One patient had a post-operative CT scan without acute blood, started therapeutic enoxaparin, and developed altered mental status on POD 3 with tract hemorrhage and intraventricular
Table 1 Accuracy and complications in stereotactic catheter placement for Ommaya reservoirs Number of patients (%) Total patients Accuracy Normal sized ventricles Only one catheter pass Had post-operative scan Good radiographic placement Complications Total peri-operative complications Hemorrhagic complications Symptomatic hemorrhage Peri-operative malfunctions Delayed malfunctions Peri-operative infections Delayed infections Conversion to VPS VPS = ventriculoperitoneal shunt.
109 (100) 74/109 (68) 107/109 (98) 93/109 (85) 92/93 (99) 7/109 7/109 4/109 3/109 1/109 1/109 3/109 4/109
(6.4) (6.4) (3.7) (2.8) (0.9) (0.9) (2.8) (3.7)
46
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
hemorrhage noted on CT scan, developed status epilepticus, and eventually died. The fourth symptomatic hemorrhage occurred in another patient with a post-operative CT scan negative for acute blood, became somnolent on POD 9 with a CT scan showing a right-sided subdural hematoma, which worsened on POD 10 prompting surgical evacuation. This patient was eventually discharged home. The patient’s platelet nadir was 41 on POD 10. Three patients had small amounts of asymptomatic acute blood on post-operative CT scan (Table 1). 3.5. Peri-operative morbidity and mortality Within 30 days of surgery, there was one infection, four symptomatic hemorrhages, one further revision for malfunction, and one revision for poor placement, for a total of seven (6.4%) perioperative morbidities. Seventy-one (65%) patients had their death documented in our hospital records. Nineteen of those 71 died within 30 days of their Ommaya surgery. 4. Discussion Though a variety of uses have been described for Ommaya reservoirs, their primary application is the administration of chronic intrathecal chemotherapy for patients with leptomeningeal carcinomatosis, and as such, they can be an invaluable aspect of some patients’ cancer treatment. Series of nonstereotactically guided Ommaya reservoir placement have reported complication rates of 7–10% [3–5], including poor placement. Even with the use of fluoroscopic guidance for this indication, up to 4.7% of catheters can be malpositioned [7]. Stereotactic systems offer an opportunity to improve on this. While use of stereotaxy to place intraventricular catheters for VPS in hydrocephalus patients has been studied in large series, the study of stereotactic catheter placement in leptomeningeal carcinomatosis patients warrants separate study, as the ventricles are commonly not dilated as in the hydrocephalus population. In our series only 32% had mild ventricular dilation, with no patients exhibiting more than mild dilation, rendering the accuracy of the catheter trajectory yet more important. Further, accuracy may reduce the number of catheter passes, and in this patient population with bone marrow dysfunction, at high risk of bleeding and infection, fewer passes could reduce risks of hemorrhage and infection. Further reduced risk of bleeding is possible due to planning a stereotactic trajectory to avoid cortical vessels. Improved accuracy can also avoid direct delivery of chemotherapy to parenchyma and revision surgery to replace the catheter. The published experience with stereotactic placement of catheters for Ommaya reservoirs has consisted of a few small series, with complication rates up to 16% [6–11]. In this largest reported series of stereotactic placement of catheters for Ommaya reservoirs in 109 patients with leptomeningeal carcinomatosis, we have shown excellent accuracy (99%) and low complication rate (6.4%). Regarding accuracy, only two (1.8%) patients required more than one pass despite 68% of the patients having normal or small-sized ventricles, and no patient required more than two passes. Eighty-five percent of patients had post-operative imaging with 99% of catheters in good position on the first post-operative scan, similar to rates reported in prior series of stereotactic ventricular catheter placement [10,11]. Furthermore, the only catheter not in good position, by virtue of the previously extracranial kink seen in the parenchyma and the reservoir seen close to the burr hole, had clearly demonstrated a post-placement migration to a deeper location. This suggests not an issue with the stereotaxy but with a failure of sufficient anchoring to extracranial tissue. There was only one other revision for a catheter being in a suboptimal position, and this was a catheter that had good placement on
post-operative CT scan and migrated into the thalamus and malfunctioned on POD 40, further highlighting the importance of securing the system well to extracranial tissue. It is important to note that our goals of catheter placement in this surgery are for the catheter to be placed safely without hemorrhage, infection, or misplacement, and to ensure that all catheter holes are within the ventricular system. Prior to the institution of stereotactic placement of Ommaya catheters at our institution, one of our patients had a catheter placed with some of the catheter holes in the ventricle and some remaining proximally in the parenchyma, and when the more distal ventricular holes became occluded, the patient experienced a severe toxicity due to parenchymal delivery of chemotherapy. This experience helped clarify these goals of surgery. We always strive to have all of the 2 cm of catheter holes within the ventricular system to avoid the potential toxicity of chemotherapy being delivered directly to the parenchyma. For this reason, the criterion used by our neurosurgeons clinically and for the purposes of these analyses for accurate catheter placement for Ommaya surgery was that all catheter holes were in the ventricular system, rather than that the catheter tip was within an arbitrary distance from the foramen of Monro, as has been used in other reports [8]. Further, no patient in our series developed clinical or radiographic evidence of toxicity due to direct delivery of chemotherapeutic to brain parenchyma. In this patient population at high risk for infection, we observed only one (0.9%) peri-operative infection, on POD 5. The other three infections occurred at least 30 days post-operatively, and these delayed infections are expected in this population [8,12]. With the exception of the three small, asymptomatic bleeds, all hemorrhagic complications in the present series occurred with either thrombocytopenia pre-operatively or within the first 10 days post-operatively, or initiation of therapeutic anti-coagulation in the first few days post-operatively. The patient with pre-operative severe thrombocytopenia was refractory during surgery to multiple platelet transfusions by clinical and laboratory measures. Thrombocytopenia is naturally a relative contraindication to brain surgery, but in these patients with refractory platelet levels who require chronic intrathecal chemotherapy, it may still be prudent to place an Ommaya reservoir despite the risks. Our findings of delayed hemorrhagic complications in the setting of post-operative thrombocytopenia or anti-coagulation should prompt neurosurgeons and oncologists to maintain vigilance in monitoring platelet counts during this period and maintain a high degree of concern when weighing the risks and benefits of starting therapeutic anti-coagulation. We now require that thrombocytopenic patients respond to transfusion on the day of surgery prior to beginning the surgical procedure. Furthermore, we have moved to schedule our Ommaya placements to correspond with chemotherapy nadirs to avoid thrombocytopenia and leukopenia. This study is limited by its retrospective nature and lack of a control group. Further, the heterogeneity among patients regarding factors that may be important in determining accuracy and safety, including among stereotactic systems used, surgeons, ventricular sizes, and comorbidities including thrombocytopenia can make it difficult to generalize conclusions. 5. Conclusion Neurosurgeons will continue to innovate to make surgeries safer, more effective, and easier, and with this innovation comes the responsibility to study our methods in order to continue to improve the outcomes of our patients. Through our experience in the pre-stereotaxy era at our institution, we have clarified what we believe accuracy means for this particular procedure, that is, that all catheter holes rest within the ventricular system. We have further learned within this series after a hemorrhagic complication
B.C. Kennedy et al. / Journal of Clinical Neuroscience 27 (2016) 44–47
that, for these patients on chemotherapy, the transfusionresponsiveness of severe thrombocytopenia, even if demonstrated previously, should be verified on the morning of surgery, and we now schedule Ommaya placement for the chemotherapy nadir to minimize thrombocytopenia and leukopenia. Future studies in this area should include newer stereotaxy systems, such as those that are electromagnetically guided, and direct comparisons among systems, including the frameless and frame-based systems. Conflicts of Interest/Disclosures Portions of the current work have been presented at the American Association of Neurological Surgeons Annual Meeting, April 2008, Chicago, IL, USA. The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. References [1] Hagen NA, O’Neill BP, Kelly PJ. Computer assisted stereotactic placement of Ommaya reservoirs for delivery of chemotherapeutic agents in cancer patients. J Neurooncol 1987;5:273–6. [2] Al-Anazi A, Bernstein M. Modified stereotactic insertion of the Ommaya reservoir. Technical note. J Neurosurg 2000;92:1050–2.
47
[3] Obbens EA, Leavens ME, Beal JW, et al. Ommaya reservoirs in 387 cancer patients: a 15-year experience. Neurology 1985;35:1274–8. [4] Lishner M, Perrin RG, Feld R, et al. Complications associated with Ommaya reservoirs in patients with cancer. The Princess Margaret Hospital experience and a review of the literature. Arch Intern Med 1990;150:173–6. [5] Bleyer WA, Pizzo PA, Spence AM, et al. The Ommaya reservoir: newly recognized complications and recommendations for insertion and use. Cancer 1978;41:2431–7. [6] Dickerman RD, Eisenberg MB. Preassembled method for insertion of Ommaya reservoir. J Surg Oncol 2005;89:36–8. [7] Sandberg DI, Bilsky MH, Souweidane MM, et al. Ommaya reservoirs for the treatment of leptomeningeal metastases. Neurosurgery 2000;47:49–54 [discussion 54–5]. [8] Greenfield JP, Schwartz TH. Catheter placement for Ommaya reservoirs with frameless surgical navigation: technical note. Stereotact Funct Neurosurg 2008;86:101–5. [9] Chamberlain MC, Kormanik PA, Barba D. Complications associated with intraventricular chemotherapy in patients with leptomeningeal metastases. J Neurosurg 1997;87:694–9. [10] Takahashi M, Yamada R, Tabei Y, et al. Navigation-guided Ommaya reservoir placement: implications for the treatment of leptomeningeal metastases. Minim Invasive Neurosurg 2007;50:340–5. [11] Sampath R, Wadhwa R, Tawfik T, et al. Stereotactic placement of ventricular catheters: does it affect proximal malfunction rates? Stereotact Funct Neurosurg 2012;90:97–103. [12] Szvalb AD, Raad II, Weinberg JS, et al. Ommaya reservoir-related infections: clinical manifestations and treatment outcomes. J infect 2014;68:216–24.