Clinical Study Received: June 6, 2014 Accepted after revision: February 5, 2015 Published online: April 16, 2015
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Proton Radiotherapy for Midline Central Nervous System Lesions: A Class Solution Neil C. Estabrook Mark W. McDonald Ted A. Hoene Greg K. Bartlett Peter A.S. Johnstone Kevin P. McMullen Jeffrey C. Buchsbaum Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, Ind., USA
Abstract Objective: Midline and central lesions of the brain requiring conventional radiotherapy (RT) present complex difficulties in dose avoidance to organs at risk (OAR). In either definitive or adjuvant settings, proper RT coverage of these lesions involves unnecessary treatment of large volumes of normal brain. We propose a class solution for these lesions using proton radiotherapy (PrT). Materials and Methods: The records of the Indiana University Health Proton Therapy Center were reviewed for patients presenting between January 1, 2005 and October 1, 2013 with midline central nervous system (CNS) lesions. Twenty-four patients were identified. After Institutional Review Board approval was granted, their dosimetry was reviewed for target volume doses and OAR dose avoidance. Results: For these cases, meningiomas were the most common histology (8 cases), and next most prevalent were craniopharyngiomas (6 cases). The others were various different deep midline brain tumors (10 cases). In all cases, fields formed by vertex and/or anterior/posterior superior oblique PrT beams along the midsagittal plane were used to provide coverage with minimal dose to the brain stem or to the cerebral hemispheres. The median prescribed
© 2015 S. Karger AG, Basel 0030–2414/15/0892–0111$39.50/0 E-Mail [email protected] www.karger.com/ocl
dose to the planning target volume for treating these patients was 54.0 Gy RBE (range 48.6–62.5) with a mean dose of 53.5 Gy RBE. The average of the mean doses to the brain stems using these fields in the 24 plans was 18.4 Gy RBE (range 0.0–44.7). Similarly, the average of the mean doses to the hippocampi was 15.8 Gy RBE (range 0.0–52.6). Conclusions: We consider these patients to be optimally treated with PrT. The use of modified midsagittal PrT schemas allows for the treatment of midline CNS lesions with sparing of most of the uninvolved brain. © 2015 S. Karger AG, Basel
Introduction
Conventional photon radiotherapy (RT) allows for beam modification in a plane perpendicular to its entry to the target. This is the mechanism of custom blocks, electron cutouts, and multileaf collimation: each shapes the beam to include the target at depth. The concept is best visualized using beam’s eye view, where the target is optimally in the center of the field and surrounded by a margin. The issue with this technique relates to the fact that RT travels completely through the patient, so structures beyond the target receive the dose as well. This solution is different with proton radiotherapy (PrT). Perpendicular modifications remain necessary Dr. Jeffrey C. Buchsbaum Department of Radiation Oncology Indiana University School of Medicine 535 Barnhill Drive, RT 041, Indianapolis, IN 46202 (USA) E-Mail jbuchsba @ iupui.edu
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Key Words Proton radiotherapy · Pediatric tumors · Meningioma · Craniopharyngioma
and are accomplished most commonly with brass apertures. A proton beam, distinctively, has a finite depth in tissue before the Bragg peak. Briefly, the range cannot be any further than allowed by beam energy but may be shorter if desired. The beam at the distal edge of the target can be shaped using (acrylic/Lucite) compensators. Apertures and compensators are collectively termed patientspecific devices. Pencil-beam scanning is another method for shaping proton treatment fields, which also has the advantage of finite depth for dose delivery inherent with PrT, but without the need of patient-specific devices. These unique aspects of the PrT technique make it very attractive for midline lesions of the central nervous system (CNS) when complete surgical resection is not safe or possible due to a myriad of patient characteristics (e.g. performance status, age, or comorbidities) or in those cases when adjuvant radiation is indicated and tumor location will lead to a dose that is harmful to normal structures. For parasagittal and falcine meningiomas, external beam RT or stereotactic radiosurgery are viable treatment modalities when such lesions are not safely resectable due to invasion of the venous sinus or when they involve the cerebral veins [1]. RT has also been shown in multiple studies to have a role for long-term control with minimal long-term toxicities for other centrally located benign lesions such as craniopharyngiomas [2, 3]. However, in such cases, many authors voice concern over RT-related IQ loss and neurocognitive dysfunction, among other sequelae [4]. The purpose of this report is to demonstrate that treating such midline structures with PrT offers sparing of many organs at risk (OARs), specifically the brain stem and hippocampi. This would be difficult with conventional RT. Furthermore, we define a class solution PrT midsagittal field design to accomplish this. Methods
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Twenty-four patients with a falcine/parasagittal meningioma, craniopharyngioma, or another type of deep midline CNS tumor were identified in our chart review. All cases were treated with an arrangement of proton fields along the midsagittal plane. Eight of the patients in this series had midline meningiomas involving the falx cerebri. For these patients, the mean prescription dose was 56.9 Gy RBE (range 54.0– 62.5). The average of the mean doses delivered to the brain stems of these patients was 3.8 Gy RBE (range 0.0– 27.1). The average of the mean doses to the right and left hippocampi in these patients was 2.8 Gy RBE (range 0.0– Estabrook/McDonald/Hoene/Bartlett/ Johnstone/McMullen/Buchsbaum
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Results
After approval from the Indiana University Institutional Review Board was obtained, we reviewed the medical records of patients seen at the Indiana University Health Proton Therapy Center between January 1, 2005 and October 1, 2013 who had been treated for midline CNS lesions. Our retrospective search found 24 patients planned and treated for meningiomas involving the falx cerebri, craniopharyngiomas, or other various midline CNS tumors (table 3) where midline sagittal fields were used for target coverage. Data from the treatment planning system were collected for these patients including primary tumor location and size, prescription dose, field design, and minimum/maximum/mean dose to the brain stem and to the bilateral hippocampi. Five patients required the bilateral hippocampi and 1 patient required the brain stem to be retrospectively contoured by a radiation oncologist so that dosimetric information could be calculated and recorded for these structures.
112
Lesion treatment targets were generated via physician contouring using 3-dimensional sequence MRI scans collected at 1 mm fused with CT scans collected at 1 mm. Clinical target volume (CTV) was designed on a case-by-case basis as delineated by the treating physician. In this series, all lesions had a 5-mm expansion for CTV that was modified to respect natural anatomic boundaries when appropriate. Planning target volume (PTV) was uniformly defined as 2 mm from an expansion point of view, but smearing was also employed to make the treatment robust. Smearing varied from 4 to 7 mm in general and was employed based on the location of the tumor and clinician judgment regarding the immobilization status of the patient and the anatomy being traversed. In the case of craniopharyngiomas, patients were evaluated with weekly MRI scans to confirm tumor shape and normal anatomy stability. Beam stopping around optical and other critical structures was done in much the same way as if photon therapy were employed. The maximal dose to the optic chiasm was kept under 50 Gy. That to the cochleae was kept under 5 Gy, and that to the hippocampi was kept as low as a case would allow. In order to perform a comparison between PrT and photon plans using both intensity-modulated RT (IMRT) and 3-dimensional conformal RT (3DCRT), 3 patients were re-planned using IMRT and 3DCRT techniques. For these photon-based plans, the CTV and PTV expansions of the original gross tumor volume (GTV) were 0.5 and 0.3 cm, respectively. Dosimetric data for this subset of patients were collected on other OARs including whole brain, temporal lobes, cochleae, and cerebellum. These data as well as the dose to the brain stem and hippocampi were used for comparing the plans. Nine patients, 6 with craniopharyngiomas and 3 with other deep midline tumors (none with meningiomas), received a dose to the target volume of 1–3 fractions of PrT delivered with right and left lateral fields with custom blocking to spare the optic chiasm. To account solely for the midline fields in this analysis, each of those cases was re-planned in our treatment planning software to remove the dose contribution of the lateral fields. For the 16 patients with either craniopharyngiomas or deep midline tumors, Spearman rank order and linear correlation coefficients were calculated to measure the association between GTV and mean doses to both the brain stem and bilateral hippocampi.
Patient No.
GTV, cm3
Rx, Gy
BS mean dose, Gy
R/L hippo mean dose, Gy
1 2 3 4 5 6 7 8 Average
6.9 32.6 5.6 24.0 5.0 24.8 25.0 88.0
59.4 61.2 54.0 54.0 54.0 54.0 62.5 55.8 56.9
0.0 0.1 0.5 0.5 1.8 0.0 27.1 0.0 3.8
0.0 0.0 0.0 0.5 0.0 0.0 21.9 0.0 2.8
Rx = Prescription; BS = brain stem; R/L = right and left; hippo = hippocampi.
21.9; table 1). All patients with midline meningiomas required only fields in a midsagittal orientation. Including their lateral ‘off chiasm’ fields, the 6 patients with craniopharyngiomas were prescribed a total dose of 54 Gy RBE. The average of the mean doses to the brain stems in these plans was 28.9 Gy RBE (range 5.5–44.7), and the average of the mean doses to the bilateral hippocampi was 25.5 Gy RBE (range 10.5–48.6; table 2). For these patients, when considering only the contribution from the midsagittal fields (∼50 Gy RBE/patient, range 48.6–50.4), the average of the mean doses to the brain stems was 27.1 Gy RBE (range 5.3–40.1) and the average of the mean doses to the bilateral hippocampi was 24.0 Gy RBE (range 2.7–45.9; table 2). Finally, including 3 cases who also required lateral fields to spare the chiasm, there were 10 patients with other types of deep midline CNS tumors whose plans were analyzed. For these patients, when accounting for the total dose delivered including that received from lateral fields, the mean prescription dose was 54.2 Gy RBE (range 50.4–59.4). The average of the mean doses to the brain stems in these plans was 23.0 Gy RBE (range 0.6–45.0), and the average of the mean doses to the bilateral hippocampi was 22.3 Gy RBE (range 0.0–55.1; table 3). For these patients, when considering only the contribution from the midsagittal fields (average prescription 53.1 Gy RBE/patient, range 48.6–59.4), the average of the mean doses to the brain stems was 24.8 Gy RBE (range 0.6– 44.2), and the average of the mean doses to the bilateral hippocampi was 21.3 Gy RBE (range 0.0–49.7; table 3). For 3 of the patients in our cohort, comparative dosimetry was done by re-planning their cases with IMRT and Proton Radiotherapy for Midline CNS Lesions
Table 2. Craniopharyngiomas including lateral fields versus no lateral fields, midsagittal fields only
Patient No.
Rx, Gy
BS mean dose, Gy
R/L hippo mean dose, Gy
54.0 54.0 54.0 54.0 54.0 54.0 54.0
17.7 42.1 39.5 44.7 5.5 23.8 28.9
10.5 43.0 48.6 37.9 3.0 10.4 25.5
No lateral fields, midsagittal fields only 10.0 50.4 17.0 9 10 43.0 48.6 39.9 11 24.0 50.4 38.0 12 32.0 48.6 40.1 13 50.4 4.5 5.3 14 31.5 50.4 22.4 Average 49.8 27.1
9.3 39.3 45.9 37.9 2.7 9.2 24.0
GTV, cm3
Including lateral fields 10.0 9 10 43.0 11 24.0 12 32.0 13 4.5 14 31.5 Average
Rx = Prescription; BS = brain stem; R/L = right and left; hippo = hippocampi.
3DCRT plans. We chose 1 patient with a falcine meningioma (table 1, patient No. 4) and 2 patients with craniopharyngiomas, 1 with a small to midsized tumor from our population (10 cm3 GTV) and 1 with the largest tumor (43 cm3 GTV; table 2, patients No. 9 and 10, respectively). The mean dose to multiple intracranial OARs as well as the brain stem volume receiving ≥54 Gy (V54) were compared for these 3 patients (table 4). For the meningioma case where the tumor is more superficial, the proton plan is superior to the photon plans with respect to the mean dose to all the OARs. Similarly, the PrT plan for the smaller-sized craniopharyngioma (10 cm3 GTV) has lower mean doses to all OARs compared to the 2 photon plans. However, when comparing the mean dose to OARs for the large craniopharyngioma (43 cm3 GTV), the results are mixed. The proton plan has a lower mean dose to the whole brain, temporal lobes, and cochleae, while the IMRT plan has a lower mean dose to the hippocampi and cerebellum. The mean dose and V54 to the brain stem are similar between the proton and IMRT plans. The 3DCRT plan has higher mean doses to all structures except the cerebellum and a higher brain stem V54 when compared to the PrT plan. When analyzing the correlation between the GTV in cubic centimeters and the doses to the brain stem and Oncology 2015;89:111–117 DOI: 10.1159/000377727
113
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Table 1. Meningiomas treated with midsagittal fields only
Table 3. Meningiomas involving other deep midline CNS tumors including lateral fields versus no lateral fields, midsagittal fields only
Patient No.
15* 16* 17* 18 19 20 21 22 23 24 Average
Histology
GTV, cm3
JPA – thalamus/basal ganglion hypothalamic glioma optic glioma third ventricle – astrocytoma tectal plate – glioma pineoblastoma third ventricle – chordoid glioma pituitary – macroadenoma ependymoma – recurrent JPA – brain stem
82.5 38.0 10.0 7.9 25.7 1.9 5.3 10.2 16.1 3.4
Including lateral fields
No lateral fields, midsagittal fields only
Rx, Gy mean dose, Gy
Rx, Gy
54.0 54.0 54.0 54.0 54.0 54.0 50.4 54.0 59.4 54.0 54.2
brain stem
R/L hippo
45.0 42.2 18.2 24.8 44.2 20.6 15.7 17.7 0.6 26.1 25.5
54.4 55.1 40.6 10.8 30.6 5.7 17.1 8.7 0.1 0.0 22.3
52.2 48.6 50.4 54.0 54.0 54.0 50.4 54.0 59.4 54.0 53.1
mean dose, Gy brain stem
R/L hippo
42.9 38.2 16.9 24.8 44.2 20.6 15.7 17.7 0.6 26.1 24.8
52.6 49.7 38.2 10.8 30.6 5.7 17.1 8.7 0.1 0.0 21.3
Rx = Prescription; R/L = right and left; hippo = hippocampi; JPA = juvenile pilocytic astrocytoma. * Three patients (No. 15, 16, 17) had lateral fields for 1 – 3 fractions to spare the optic apparatus.
Table 4. Dosimetric comparison between proton and photon plans
Patient No. 4 54 Gy, 24 cm3 GTV
Patient No. 9 50.4 Gy, 10 cm3 GTV
Patient No. 10 54 Gy, 43 cm3 GTV
PrT
IMRT
3DCRT
PrT
IMRT
3DCRT
PrT
IMRT
3DCRT
9.3 0.5 0.5 0.0 0.0 2.0
14.3 3.0 3.7 1.5 0.9 6.9
21.7 6.9 13.2 4.4 2.1 8.0
8.1 17.7 10.5 6.7 4.9 0.0
8.9 25.3 13.3 14.4 8.0 5.4
11.8 40.9 32.5 16.5 14.6 19.9
15.3 42.1 43.0 16.3 3.5 16.7
18.6 42.3 38.9 38.0 8.9 12.0
21.6 45.3 49.8 26.4 43.9 14.3
Volume receiving ≥54 Gy, cm3 Brain stem 0.0
0.0
0.0
0.0
0.0
0.0
2.3
2.3
8.7
Mean dose, Gy Whole brain Brain stem R/L hippocampi R/L temporal lobes R/L cochleae Cerebellum
R/L = Right and left.
114
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Discussion
When surgical intervention is not possible or potentially too morbid, significant data exist for treating benign midline CNS tumors, like meningiomas and craniopharyngiomas, with either RT alone or with partial safe resection plus radiation [1–3]. For other malignant midline CNS tumors, RT is often indicated for adjuvant therapy, whether they are excised or not. It was our goal here to demonstrate a refinement to previously used external beam RT techniques. Specifically, we propose a PrT midEstabrook/McDonald/Hoene/Bartlett/ Johnstone/McMullen/Buchsbaum
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
hippocampi for all the deep midline tumors (including the 6 craniopharyngiomas and the 10 other deep midline cases), the Spearman rank order was found to be statistically significant (rs = 0.64, p = 0.004) when comparing GTV to mean brain stem dose. Linear correlation was also significant for this comparison (r = 0.67, p = 0.002). When comparing the association between GTV and mean dose to the hippocampi, the correlation was again statistically significant for both analyses, with a Spearman rank order coefficient rs = 0.71 (p = 0.001) and linear correlation coefficient r = 0.73 (p = 0.0007).
Fig. 1. Midsagittal fields for craniopharyngioma. Sagittal, coronal, and axial views.
sagittal field design for treating such lesions. Furthermore, we present dosimetric data from our institution for some critical structures when such midsagittal beam arrangements are used. Simply, this is typically a configuration of 2 or 3 proton fields usually including a vertex with an anterior and/or a posterior superior oblique. To be clear, all of these fields are in the midsagittal plane of the patient (fig. 1, 2). While vertex and parasagittal superior oblique fields may be used with photon therapy, this field arrangement is uniquely suited to PrT, since there is no exit dose through the body, with sparing of all structures caudal to the target. We reviewed the charts of 24 patients treated with a midsagittal technique for either meningioma (n = 8), craniopharyngioma (n = 6), or any other various type of deep
midline CNS tumor (n = 10) to evaluate dosing to the brain stem and the hippocampi. These OARs were chosen for dosimetric evaluation due to their proximity to the midline tumors being treated and well-documented toxicities to high doses of radiation. In this regard, long-term toxicity from a dose to the brain stem includes permanent cranial neuropathy or necrosis, with a dose constraint to the whole organ set at
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Proton Radiotherapy for Midline Central Nervous System Lesions: A Class Solution Neil C. Estabrook Mark W. McDonald Ted A. Hoene Greg K. Bartlett Peter A.S. Johnstone Kevin P. McMullen Jeffrey C. Buchsbaum Department of Radiation Oncology, Indiana University School of Medicine, Indianapolis, Ind., USA
Abstract Objective: Midline and central lesions of the brain requiring conventional radiotherapy (RT) present complex difficulties in dose avoidance to organs at risk (OAR). In either definitive or adjuvant settings, proper RT coverage of these lesions involves unnecessary treatment of large volumes of normal brain. We propose a class solution for these lesions using proton radiotherapy (PrT). Materials and Methods: The records of the Indiana University Health Proton Therapy Center were reviewed for patients presenting between January 1, 2005 and October 1, 2013 with midline central nervous system (CNS) lesions. Twenty-four patients were identified. After Institutional Review Board approval was granted, their dosimetry was reviewed for target volume doses and OAR dose avoidance. Results: For these cases, meningiomas were the most common histology (8 cases), and next most prevalent were craniopharyngiomas (6 cases). The others were various different deep midline brain tumors (10 cases). In all cases, fields formed by vertex and/or anterior/posterior superior oblique PrT beams along the midsagittal plane were used to provide coverage with minimal dose to the brain stem or to the cerebral hemispheres. The median prescribed
© 2015 S. Karger AG, Basel 0030–2414/15/0892–0111$39.50/0 E-Mail [email protected] www.karger.com/ocl
dose to the planning target volume for treating these patients was 54.0 Gy RBE (range 48.6–62.5) with a mean dose of 53.5 Gy RBE. The average of the mean doses to the brain stems using these fields in the 24 plans was 18.4 Gy RBE (range 0.0–44.7). Similarly, the average of the mean doses to the hippocampi was 15.8 Gy RBE (range 0.0–52.6). Conclusions: We consider these patients to be optimally treated with PrT. The use of modified midsagittal PrT schemas allows for the treatment of midline CNS lesions with sparing of most of the uninvolved brain. © 2015 S. Karger AG, Basel
Introduction
Conventional photon radiotherapy (RT) allows for beam modification in a plane perpendicular to its entry to the target. This is the mechanism of custom blocks, electron cutouts, and multileaf collimation: each shapes the beam to include the target at depth. The concept is best visualized using beam’s eye view, where the target is optimally in the center of the field and surrounded by a margin. The issue with this technique relates to the fact that RT travels completely through the patient, so structures beyond the target receive the dose as well. This solution is different with proton radiotherapy (PrT). Perpendicular modifications remain necessary Dr. Jeffrey C. Buchsbaum Department of Radiation Oncology Indiana University School of Medicine 535 Barnhill Drive, RT 041, Indianapolis, IN 46202 (USA) E-Mail jbuchsba @ iupui.edu
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Key Words Proton radiotherapy · Pediatric tumors · Meningioma · Craniopharyngioma
and are accomplished most commonly with brass apertures. A proton beam, distinctively, has a finite depth in tissue before the Bragg peak. Briefly, the range cannot be any further than allowed by beam energy but may be shorter if desired. The beam at the distal edge of the target can be shaped using (acrylic/Lucite) compensators. Apertures and compensators are collectively termed patientspecific devices. Pencil-beam scanning is another method for shaping proton treatment fields, which also has the advantage of finite depth for dose delivery inherent with PrT, but without the need of patient-specific devices. These unique aspects of the PrT technique make it very attractive for midline lesions of the central nervous system (CNS) when complete surgical resection is not safe or possible due to a myriad of patient characteristics (e.g. performance status, age, or comorbidities) or in those cases when adjuvant radiation is indicated and tumor location will lead to a dose that is harmful to normal structures. For parasagittal and falcine meningiomas, external beam RT or stereotactic radiosurgery are viable treatment modalities when such lesions are not safely resectable due to invasion of the venous sinus or when they involve the cerebral veins [1]. RT has also been shown in multiple studies to have a role for long-term control with minimal long-term toxicities for other centrally located benign lesions such as craniopharyngiomas [2, 3]. However, in such cases, many authors voice concern over RT-related IQ loss and neurocognitive dysfunction, among other sequelae [4]. The purpose of this report is to demonstrate that treating such midline structures with PrT offers sparing of many organs at risk (OARs), specifically the brain stem and hippocampi. This would be difficult with conventional RT. Furthermore, we define a class solution PrT midsagittal field design to accomplish this. Methods
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Twenty-four patients with a falcine/parasagittal meningioma, craniopharyngioma, or another type of deep midline CNS tumor were identified in our chart review. All cases were treated with an arrangement of proton fields along the midsagittal plane. Eight of the patients in this series had midline meningiomas involving the falx cerebri. For these patients, the mean prescription dose was 56.9 Gy RBE (range 54.0– 62.5). The average of the mean doses delivered to the brain stems of these patients was 3.8 Gy RBE (range 0.0– 27.1). The average of the mean doses to the right and left hippocampi in these patients was 2.8 Gy RBE (range 0.0– Estabrook/McDonald/Hoene/Bartlett/ Johnstone/McMullen/Buchsbaum
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Results
After approval from the Indiana University Institutional Review Board was obtained, we reviewed the medical records of patients seen at the Indiana University Health Proton Therapy Center between January 1, 2005 and October 1, 2013 who had been treated for midline CNS lesions. Our retrospective search found 24 patients planned and treated for meningiomas involving the falx cerebri, craniopharyngiomas, or other various midline CNS tumors (table 3) where midline sagittal fields were used for target coverage. Data from the treatment planning system were collected for these patients including primary tumor location and size, prescription dose, field design, and minimum/maximum/mean dose to the brain stem and to the bilateral hippocampi. Five patients required the bilateral hippocampi and 1 patient required the brain stem to be retrospectively contoured by a radiation oncologist so that dosimetric information could be calculated and recorded for these structures.
112
Lesion treatment targets were generated via physician contouring using 3-dimensional sequence MRI scans collected at 1 mm fused with CT scans collected at 1 mm. Clinical target volume (CTV) was designed on a case-by-case basis as delineated by the treating physician. In this series, all lesions had a 5-mm expansion for CTV that was modified to respect natural anatomic boundaries when appropriate. Planning target volume (PTV) was uniformly defined as 2 mm from an expansion point of view, but smearing was also employed to make the treatment robust. Smearing varied from 4 to 7 mm in general and was employed based on the location of the tumor and clinician judgment regarding the immobilization status of the patient and the anatomy being traversed. In the case of craniopharyngiomas, patients were evaluated with weekly MRI scans to confirm tumor shape and normal anatomy stability. Beam stopping around optical and other critical structures was done in much the same way as if photon therapy were employed. The maximal dose to the optic chiasm was kept under 50 Gy. That to the cochleae was kept under 5 Gy, and that to the hippocampi was kept as low as a case would allow. In order to perform a comparison between PrT and photon plans using both intensity-modulated RT (IMRT) and 3-dimensional conformal RT (3DCRT), 3 patients were re-planned using IMRT and 3DCRT techniques. For these photon-based plans, the CTV and PTV expansions of the original gross tumor volume (GTV) were 0.5 and 0.3 cm, respectively. Dosimetric data for this subset of patients were collected on other OARs including whole brain, temporal lobes, cochleae, and cerebellum. These data as well as the dose to the brain stem and hippocampi were used for comparing the plans. Nine patients, 6 with craniopharyngiomas and 3 with other deep midline tumors (none with meningiomas), received a dose to the target volume of 1–3 fractions of PrT delivered with right and left lateral fields with custom blocking to spare the optic chiasm. To account solely for the midline fields in this analysis, each of those cases was re-planned in our treatment planning software to remove the dose contribution of the lateral fields. For the 16 patients with either craniopharyngiomas or deep midline tumors, Spearman rank order and linear correlation coefficients were calculated to measure the association between GTV and mean doses to both the brain stem and bilateral hippocampi.
Patient No.
GTV, cm3
Rx, Gy
BS mean dose, Gy
R/L hippo mean dose, Gy
1 2 3 4 5 6 7 8 Average
6.9 32.6 5.6 24.0 5.0 24.8 25.0 88.0
59.4 61.2 54.0 54.0 54.0 54.0 62.5 55.8 56.9
0.0 0.1 0.5 0.5 1.8 0.0 27.1 0.0 3.8
0.0 0.0 0.0 0.5 0.0 0.0 21.9 0.0 2.8
Rx = Prescription; BS = brain stem; R/L = right and left; hippo = hippocampi.
21.9; table 1). All patients with midline meningiomas required only fields in a midsagittal orientation. Including their lateral ‘off chiasm’ fields, the 6 patients with craniopharyngiomas were prescribed a total dose of 54 Gy RBE. The average of the mean doses to the brain stems in these plans was 28.9 Gy RBE (range 5.5–44.7), and the average of the mean doses to the bilateral hippocampi was 25.5 Gy RBE (range 10.5–48.6; table 2). For these patients, when considering only the contribution from the midsagittal fields (∼50 Gy RBE/patient, range 48.6–50.4), the average of the mean doses to the brain stems was 27.1 Gy RBE (range 5.3–40.1) and the average of the mean doses to the bilateral hippocampi was 24.0 Gy RBE (range 2.7–45.9; table 2). Finally, including 3 cases who also required lateral fields to spare the chiasm, there were 10 patients with other types of deep midline CNS tumors whose plans were analyzed. For these patients, when accounting for the total dose delivered including that received from lateral fields, the mean prescription dose was 54.2 Gy RBE (range 50.4–59.4). The average of the mean doses to the brain stems in these plans was 23.0 Gy RBE (range 0.6–45.0), and the average of the mean doses to the bilateral hippocampi was 22.3 Gy RBE (range 0.0–55.1; table 3). For these patients, when considering only the contribution from the midsagittal fields (average prescription 53.1 Gy RBE/patient, range 48.6–59.4), the average of the mean doses to the brain stems was 24.8 Gy RBE (range 0.6– 44.2), and the average of the mean doses to the bilateral hippocampi was 21.3 Gy RBE (range 0.0–49.7; table 3). For 3 of the patients in our cohort, comparative dosimetry was done by re-planning their cases with IMRT and Proton Radiotherapy for Midline CNS Lesions
Table 2. Craniopharyngiomas including lateral fields versus no lateral fields, midsagittal fields only
Patient No.
Rx, Gy
BS mean dose, Gy
R/L hippo mean dose, Gy
54.0 54.0 54.0 54.0 54.0 54.0 54.0
17.7 42.1 39.5 44.7 5.5 23.8 28.9
10.5 43.0 48.6 37.9 3.0 10.4 25.5
No lateral fields, midsagittal fields only 10.0 50.4 17.0 9 10 43.0 48.6 39.9 11 24.0 50.4 38.0 12 32.0 48.6 40.1 13 50.4 4.5 5.3 14 31.5 50.4 22.4 Average 49.8 27.1
9.3 39.3 45.9 37.9 2.7 9.2 24.0
GTV, cm3
Including lateral fields 10.0 9 10 43.0 11 24.0 12 32.0 13 4.5 14 31.5 Average
Rx = Prescription; BS = brain stem; R/L = right and left; hippo = hippocampi.
3DCRT plans. We chose 1 patient with a falcine meningioma (table 1, patient No. 4) and 2 patients with craniopharyngiomas, 1 with a small to midsized tumor from our population (10 cm3 GTV) and 1 with the largest tumor (43 cm3 GTV; table 2, patients No. 9 and 10, respectively). The mean dose to multiple intracranial OARs as well as the brain stem volume receiving ≥54 Gy (V54) were compared for these 3 patients (table 4). For the meningioma case where the tumor is more superficial, the proton plan is superior to the photon plans with respect to the mean dose to all the OARs. Similarly, the PrT plan for the smaller-sized craniopharyngioma (10 cm3 GTV) has lower mean doses to all OARs compared to the 2 photon plans. However, when comparing the mean dose to OARs for the large craniopharyngioma (43 cm3 GTV), the results are mixed. The proton plan has a lower mean dose to the whole brain, temporal lobes, and cochleae, while the IMRT plan has a lower mean dose to the hippocampi and cerebellum. The mean dose and V54 to the brain stem are similar between the proton and IMRT plans. The 3DCRT plan has higher mean doses to all structures except the cerebellum and a higher brain stem V54 when compared to the PrT plan. When analyzing the correlation between the GTV in cubic centimeters and the doses to the brain stem and Oncology 2015;89:111–117 DOI: 10.1159/000377727
113
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
Table 1. Meningiomas treated with midsagittal fields only
Table 3. Meningiomas involving other deep midline CNS tumors including lateral fields versus no lateral fields, midsagittal fields only
Patient No.
15* 16* 17* 18 19 20 21 22 23 24 Average
Histology
GTV, cm3
JPA – thalamus/basal ganglion hypothalamic glioma optic glioma third ventricle – astrocytoma tectal plate – glioma pineoblastoma third ventricle – chordoid glioma pituitary – macroadenoma ependymoma – recurrent JPA – brain stem
82.5 38.0 10.0 7.9 25.7 1.9 5.3 10.2 16.1 3.4
Including lateral fields
No lateral fields, midsagittal fields only
Rx, Gy mean dose, Gy
Rx, Gy
54.0 54.0 54.0 54.0 54.0 54.0 50.4 54.0 59.4 54.0 54.2
brain stem
R/L hippo
45.0 42.2 18.2 24.8 44.2 20.6 15.7 17.7 0.6 26.1 25.5
54.4 55.1 40.6 10.8 30.6 5.7 17.1 8.7 0.1 0.0 22.3
52.2 48.6 50.4 54.0 54.0 54.0 50.4 54.0 59.4 54.0 53.1
mean dose, Gy brain stem
R/L hippo
42.9 38.2 16.9 24.8 44.2 20.6 15.7 17.7 0.6 26.1 24.8
52.6 49.7 38.2 10.8 30.6 5.7 17.1 8.7 0.1 0.0 21.3
Rx = Prescription; R/L = right and left; hippo = hippocampi; JPA = juvenile pilocytic astrocytoma. * Three patients (No. 15, 16, 17) had lateral fields for 1 – 3 fractions to spare the optic apparatus.
Table 4. Dosimetric comparison between proton and photon plans
Patient No. 4 54 Gy, 24 cm3 GTV
Patient No. 9 50.4 Gy, 10 cm3 GTV
Patient No. 10 54 Gy, 43 cm3 GTV
PrT
IMRT
3DCRT
PrT
IMRT
3DCRT
PrT
IMRT
3DCRT
9.3 0.5 0.5 0.0 0.0 2.0
14.3 3.0 3.7 1.5 0.9 6.9
21.7 6.9 13.2 4.4 2.1 8.0
8.1 17.7 10.5 6.7 4.9 0.0
8.9 25.3 13.3 14.4 8.0 5.4
11.8 40.9 32.5 16.5 14.6 19.9
15.3 42.1 43.0 16.3 3.5 16.7
18.6 42.3 38.9 38.0 8.9 12.0
21.6 45.3 49.8 26.4 43.9 14.3
Volume receiving ≥54 Gy, cm3 Brain stem 0.0
0.0
0.0
0.0
0.0
0.0
2.3
2.3
8.7
Mean dose, Gy Whole brain Brain stem R/L hippocampi R/L temporal lobes R/L cochleae Cerebellum
R/L = Right and left.
114
Oncology 2015;89:111–117 DOI: 10.1159/000377727
Discussion
When surgical intervention is not possible or potentially too morbid, significant data exist for treating benign midline CNS tumors, like meningiomas and craniopharyngiomas, with either RT alone or with partial safe resection plus radiation [1–3]. For other malignant midline CNS tumors, RT is often indicated for adjuvant therapy, whether they are excised or not. It was our goal here to demonstrate a refinement to previously used external beam RT techniques. Specifically, we propose a PrT midEstabrook/McDonald/Hoene/Bartlett/ Johnstone/McMullen/Buchsbaum
Downloaded by: Emory University 198.143.47.65 - 1/27/2016 12:12:37 PM
hippocampi for all the deep midline tumors (including the 6 craniopharyngiomas and the 10 other deep midline cases), the Spearman rank order was found to be statistically significant (rs = 0.64, p = 0.004) when comparing GTV to mean brain stem dose. Linear correlation was also significant for this comparison (r = 0.67, p = 0.002). When comparing the association between GTV and mean dose to the hippocampi, the correlation was again statistically significant for both analyses, with a Spearman rank order coefficient rs = 0.71 (p = 0.001) and linear correlation coefficient r = 0.73 (p = 0.0007).
Fig. 1. Midsagittal fields for craniopharyngioma. Sagittal, coronal, and axial views.
sagittal field design for treating such lesions. Furthermore, we present dosimetric data from our institution for some critical structures when such midsagittal beam arrangements are used. Simply, this is typically a configuration of 2 or 3 proton fields usually including a vertex with an anterior and/or a posterior superior oblique. To be clear, all of these fields are in the midsagittal plane of the patient (fig. 1, 2). While vertex and parasagittal superior oblique fields may be used with photon therapy, this field arrangement is uniquely suited to PrT, since there is no exit dose through the body, with sparing of all structures caudal to the target. We reviewed the charts of 24 patients treated with a midsagittal technique for either meningioma (n = 8), craniopharyngioma (n = 6), or any other various type of deep
midline CNS tumor (n = 10) to evaluate dosing to the brain stem and the hippocampi. These OARs were chosen for dosimetric evaluation due to their proximity to the midline tumors being treated and well-documented toxicities to high doses of radiation. In this regard, long-term toxicity from a dose to the brain stem includes permanent cranial neuropathy or necrosis, with a dose constraint to the whole organ set at
