Dose Prescription, Tolerances, Side Effects and Safety and Efficacy of SBRT of the Spine IAEA Singapore SBRT Symposium Josh Yamada MD FRCPC Department of Radiation Oncology Memorial Sloan Kettering Cancer Center
Feb 24, 2016
Dose Prescription, Tolerances, Side Effects and Safety and Efficacy of SBRT of the Spine
IAEA Singapore SBRT Symposium
Josh Yamada MD FRCPCDepartment of Radiation Oncology
Memorial Sloan Kettering Cancer Center
Spinal Cord Radiation InjuryType Timing
after XRTClinical
Findings Pathogenesis Outcome
Acute During XRT None -- --
Early-Delayed 2-37 Weeks Lhermitte’s Demyelination RecoveryLate Delayed Months-Years
Transverse myelopathy
Para/QuadriplegiaBrown-SequardSpastic paraparesis
Necrosis Irreversible
Motor Neuron Dysfunction
Leg Weakness Ventral roots Irreversible
Hemorrhagic myelopathy
8-30 years Acute paraparesis
Telangectasia Reversible
From: Posner J, Neurologic Complications of Cancer, p 525
Progressive Myelopathy
• Demyelination, necrosis, BBB disruption• 12-50 months post XRT• Slowly progressive symptoms
– Brown Sequard syndrome with paraethesia and weakness in one side and decrease in pain/temp in side, progressing to transverse myelitis
– Progressive weakness, hyperactive reflexes, loss of position and vibration, pain and temp intact
– Decreased motor conduction velocity– CSF usually N, or increased protein.– MRI: Cord swelling and patchy enhancement
Spinal Cord Tolerance
Institution Dose ConstraintMSKCC 14 Gy Dmax
UPMC 10 Gy Dmax
HENRY FORD V10Gy < 10%
MDACC 12 Gy Dmax TO 0.1 CC
PMH 12 Gy Dmax TO THECAL SAC OR CORD + 2mm
CLEVELAND CLINIC 14Gy Dmax AND V10Gy <10%
STANFORD 14 Gy Dmax, V12Gy < 0.3 CC, V10Gy < 0.5 CC V8Gy < 1 CC
DALLAS 14 GY Dmax, V10Gy < 0.35CC, V8Gy < 1.2CC
Background
• Dose-volume tolerances of the spinal cord in spinal stereotactic radiosurgery (SRS) have been difficult to define– Complication rates required to be very low– Published reports of myelopathy do not account
for the total number of patients treated at given dose-volume combinations
Purpose
• Report spinal cord toxicity from single fraction spinal SRS
• Provide a comprehensive atlas of complication incidence to identify dose-volume predictors of spinal cord toxicity
Materials and Methods• Prospective database of all patients treated
with single fraction SRS between 2003-2010• Retrospective review for spinal cord toxicity• No prior radiation to region allowed• Spinal Cord Toxicity
– Clinical Myelitis– MRI spinal signal changes not attributable to
tumor progression or other causes
Materials and Methods• DVH atlases were created• Complication rates with 95% confidence limits • Probabilities that complication rates were <
1% for myelitis and < 10% for signal changes were determined as a function of dose and absolute volume.
Results: Cohort Characteristicsn Percent
Patients 221 100%Total number of lesions
251 100%
≥ 2 treatment sites 30 15%Signal information p tx
203 92%
GenderFemale 87 40%Male 134 60%
Age (years)Median 60.2Range 20-86
Baseline KPSMedian 90% Range 50% - 90%
Spinal RegionCervical 46 18%Thoracic 196 79%Lumbar 9 3%
Follow-Up (months)Median 15 Range 3-81
Status at Last Follow-upAlive 98 44%Deceased 123 56%
nPercen
tPrimary Tumor SiteBreast 13 6%Upper GI (esoph, pancreas, gallbladder) 4 2%Lower GI (anal, rectal, colon) 14 7%Hepatocellular 9 4%Sarcoma 30 14%Melanoma 13 6%Lung 22 10%Prostate 24 11%Renal Cell 40 18%Other GU (penile, testicular, bladder) 5 2%Thyroid 15 7%H&N SCC 7 2%CNS 10 4%Other 15 7%
Histological CategoryAdenocarcinoma 33%Carcinoma 37%Melanoma 7%Sarcoma 14%Other 9%
Prescribed Dose1800 cGy 6 2%2100 cGy 21 8%2200 cGy 4 2%2300 cGy 1 < 1%2400 cGy 219 87%Median (cGy) 2400Mean (cGy) 2356
Spinal Cord Toxicity with Single Fraction Paraspinal SRS
• One case of clinical myelitis (0.4%)• 5 cases of signal changes without clinical signs
or symptoms of myelopathy (2.4%)
Spinal Cord Toxicity with Single Fraction Paraspinal SRS
• All myelitis or signal changes:– Maximum cord dose > 13.33 Gy– Minimum doses to the hottest:
• 0.1 cc > 10.66 Gy• 0.2 cc > 10 Gy• 0.5 cc > 9 Gy• 1 cc > 8 Gy
Myelitis Signal Changes
Vol v(cc)
Dose
d(Gy)
#
comp
#
tot
99% conf
lim on r
prob
r < 1%
#
comp
#
tot
99% conf
lim on r
prob
r < 10% 0 13.33 0 64 0.068 0.49 0 55 0.079 1.00
0.1 10.66 0 60 0.073 0.47 0 51 0.085 1.000.2 10 0 56 0.078 0.45 0 46 0.093 0.990.5 9 0 45 0.095 0.38 0 36 0.117 0.98
1 8 0 26 0.157 0.25 0 30 0.138 0.96
Statistics for treatments with DVHs passing below the locations (v,d), chosen just below the Myelitis DVH.
95% confidence limit on signal changes ratefor DVHs passing above plot point
Dose (Gy)0 2 4 6 8 10 12 14
Volu
me
(cc)
0
2
4
6
8
10
12
14
160.05 0.10
0.15 0.20
0.25 0.30
0.35 0.40
0.45 0.50
0.55 0.60
0.65 0.70
0.75 0.80
0.85 0.90
0.95
signal changesmyelitis case
95% confidence limit on signal changes rate for DVHs passing above plot point
myelitisprobability that true complication rate > 2%
dose (Gy)0 2 4 6 8 10 12 14
cord
vol
ume
(cc)
0
2
4
6
8
10
12
14
160.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
myelitis case
Myelitis:Probability that True Complication
Rate > 2%
signal changesprobability that true complication rate > 5%
dose (Gy)0 2 4 6 8 10 12 14
cord
vol
ume
(cc)
0
2
4
6
8
10
12
14
160.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95
signal changesmyelitis case
Cord Signal Changes:Probability that True Complication
Rate > 5%
Conclusions• High dose paraspinal SRS has a low rate of
clinical apparent myelopathy (<1%)• Asymptomatic spinal cord signal changes are
more common (2%)• The following dose limits minimize the
potential for spinal cord toxicity after SRS– Maximum cord dose < 13.3 Gy– Minimum doses to the hottest 0.1, 0.2, 0.5, and 1
cc < 10.66, 10, 9, and 8 Gy respectively
Future Directions
• Pooled multi-institutional effort to overcome limitations– Single institution cohort with homogeneously
treated population– Limited events
Cord Myelopathy DosimetryCase 1 Case 2 Case 3 Case 4 MSKCC
Total Dose (Gy) 25 24 16 16 NA
Prescription Line 80% 70% 90% 90% NA
Fraction Number 2 3 1 1 NA
PTV Volume (cc) 14.6 7.6 66 10.8 NA
V100% 96 98 95 99 NA
D100 (Gy) 22 21 11 15 NA
D90 (Gy) 26 25 16 15 NA
Level T Spine C Spine Clivus Clivus NA
Time to Sx (Mon) 9 9 13 ? NA
Cord Dmax BED Gy2 190 190 122 136 112
Cord Dmax BED Gy2(1) 18.5 18.5 14.6 15.6 14.0
Yucatan Mini Pig ReirradiationMedin et al. IJROBP 2010
• 23 mature mini pigs received 3000cGy/10• Single Fraction Spine SRS one year later
Dose N Deficit
FU
14 Gy 2 0 40 weeks16 Gy 3 0 52 weeks18 Gy 5 2 48-52 weeks20 Gy 5 4 52 weeks22 Gy 5 5 20 weeks24 Gy 3 3 14-19 weeks
Pig Cord Reirradiation Histopathology
• No changes at 14-16 Gy• 18-20 Gy changes limited to small foci of
demyelination• 22-24 Gy extensive tissue damage including
grey matter infarction• Pigs reirradiated with SRS one year after
3000cGy/10 no different that pigs receiving de novo SRS.
Pig Cord ED50
• 96% calculated recovery after 3000cGy/10 after one year.
Vertebral Body Fracture RiskVertebral body involvement is a significant risk for fracture (p=0.02)
• Not found to be significant risk factors:
• Obesity, posterior element involvement, local kyphosis, pre-exisiting endplate infraction or fracture,
• XRT dose
• N = 114 patients
• FU = 10.9 months (median)
• Grade 2 N= 5, Grade 3 N= 1, grade 4 N= 1
Esophagus Constraints Level 1 < 15 Gy/2cc
Level 2 < 20 Gy/2cc
Progressive risk of vertebral body fractures post high dose IGRT
10-O6 3-O7 3-O7
63 year old female with NSCLC acute onset of back pain post 2400cGy to T5
Methods
• 71 treated sites in 62 consecutive patients with solid tumor spine metastases
• 1800-2400 cGy single fraction IGRT• Serial MRI every 3-4 months• All images reviewed by same 3 spine surgeons and
neuroradiologist• Primary outcome: New fracture or progression of exisiting
fracture• Secondary outcomes:
– ASIA score– VAS pain score, narcotic use
Results
• Fracture/progression noted in 27 sites (39%) • 65% lytic, 17% mixed, 18% sclerotic• Lytic lesions were 6.8 times more likely to fracture vs.
mixed/sclerotic lesions (p<0.001)
Lytic vs. Non Lytic
Function and Symptoms
• Fracture/progression not correlated with – BMI– XRT dose
• ASIA score not impacted• Median VAS score in fracture patients 5 vs 2 in non
fracture patients (p=0.051)• Fracture patients more likely to use narcotics (70% vs.
41% p = 0.005)
Conclusions
• A high risk of radiographic vertebral body fracture was found after high dose single fraction radiation – Dose not a significant predictor– Lytic lesions and greater tumor involvement were found to be
risk factors– ASIA score not affected– Patients with fracture were found to report more pain and
require more narcotics• High dose radiation may contribute to the development of
vertebral body fracture• Currently investigating the role of prophylactic
kyphoplasty in highest risk patients
Materials & Methods
• 204 consecutive spinal metastases abutting the esophagus in 182 patients were treated with single fraction paraspinal SRS at MSKCC between 2003-2010
• Esophageal toxicity scored with NCI CTCAE 4.0
• Atlases of complication incidences were generated
• Clinical factors were correlated with toxicity
Cohort Characteristicsn Percen
tPatients 182 100%Lesions 204 100%Patients ≥ 2 treatment sites
28 15%
GenderFemale 73 40%Male 109 60%
Age (years)Median 61Age 21-88
Baseline KPSMedian 90% Range 50% -
90%
Spinal RegionCervical 26 13%Thoracic 178 87%
Follow-Up (months)Median 12 Range 3-81
Status at Last Follow-upAlive 102 56%Deceased 80 44%
nPercen
tPrimary Tumor SiteBreast 11 6%Upper GI (esoph, pancreas, gallbladder) 4 2%Lower GI (anal, rectal, colon) 13 7%Hepatocellular 8 4%Sarcoma 25 14%Melanoma 12 6%Lung 18 10%Prostate 20 11%Renal Cell 33 18%Other GU (penile, testicular, bladder) 3 2%Thyroid 13 7%H&N SCC 3 2%CNS 7 4%Other 12 7%
Prescribed Dose1600 cGy 1 < 1%1800 cGy 24 12%2100 cGy 10 5%2200 cGy 3 1%2300 cGy 1 < 1%2400 cGy 165 81%Median cGy 240
0
Esophageal Toxicity with Single Fraction Paraspinal SRS
n Percent
Acute Toxicities
Overall 31 15 %
Grade 1-2 28 14%Grade 3 1 < 1%Grade 4 2 1%Grade 5 0 0%
Esophagitis/Pain 28 90%Esophageal ulcer 2 7%Esophageal edema 1 3%
n Percent
Late Toxicities
Overall 24 12%
Grade 1-2 13 6%Grade 3 6 3%Grade 4 4 2%Grade 5 1 < 1%
Esophagitis/Pain 12 50%Esophageal stenosis
4 17%
Esophageal fistula
4 17%
Esophageal ulcer 4 17%
Patients with Grade ≥ 3 Toxicityn Time Grad
eSite Dose
(cGy)Toxicity Class
Time to Maximum Toxicity (days)
1 Acute
3 T2-3 2100 Esophagitis 54
2 Acute
4 T4 2400 Esophagitis 65
3 Acute
4 C7 2400 Ulcer 84
Median 2400 65
4 Late 3 C5 2200 Stenosis 2455 Late 3 T2-T4 2400 Stenosis 1786 Late 3 T2-T3 2400 Ulcer 3507 Late 3 C7-T1 2400 Stenosis 3958 Late 3 C6-T1 2400 Stenosis 3469 Late 3 T6 2400 Ulcer 132
10 Late 4 T3 2400 Fistula 44411 Late 4 T3 2400 Fistula 14912 Late 4 T7 2400 Fistula 38013 Late 4 C4-T2 2400 Ulcer 12814 Late 5 T2 2100 Fistula 584
Median 2400 346
Dosimetric and Volumetric Predictors of Grade ≥ 3
Esophageal ToxicityDosimetric Variable
Median Split
Toxicity Incidence
Below Median
Split
Toxicity Incidence
Above Median
Split
RR Grade ≥ 3 Toxicity
p value
n % n %
D2.5 cc 14.02 Gy 2/102 2% 12/102 12% 12/2 = 6 0.01
V10 Gy 4.77 cc 4/102
4% 10/102 10% 10/4 = 2.5 0.16
V12 Gy 3.78 cc 3/102 3% 11/102 11% 11/3 = 3.7 0.05V15 Gy 1.87 cc 1/102 1% 13/102 13% 13/1 = 13 0.0013V20 Gy 0.11 cc 2/102 2% 12/102 12% 12/2 = 6 0.01V22 Gy 0.0 cc 1/102 1% 13/102 13% 13/1 = 13 0.0013
Atlas of Complication Incidence for Grade ≥ 3 Acute or Late
Esophageal ToxicityProbability the true complication rate > 10%
Dose (Gy)0 5 10 15 20 25 30
Esop
hagu
s Vo
lum
e (c
c)
0
2
4
6
8
10
12
140.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Dose Response Model for Grade ≥ 3 Esophageal Toxicity
Dose Response for >= Grade 3 EsophagitisSingle Fraction Treatments
Dose to hottest 2.5 cc (Gy)
0 5 10 15 20 25 30
Prob
. of >
= gr
ade
3 es
oph.
com
p.
0.0
0.1
0.2
0.3
logistic fit observed rate (quartiles)
p < 0.0006
Clinical Risk Factors for Developing Grade ≥4 Esophageal
Toxicityn Site Dose
(cGy)Time Grad
eToxicity Class
Probable Radiation
Recall Reaction (Agent)
Iatrogenic Manipulatio
n Before Maximum Toxicity
Time to Maximum Toxicity (days)
1 T4 2400 Acute 4 Esophagitis Adriamycin -- 652 C7 2400 Acute 4 Ulcer Liposomal
AdriamycinBiopsy 84
3 T3 2400 Late 4 Fistula -- Dilation 4444 T3 2400 Late 4 Fistula -- Biopsy 1495 T7 2400 Late 4 Fistula Gemcitabin
eStent 380
6 C4-T2 2400 Late 4 Stenosis Adriamycin Dilation 128
7 T2 2100 Late 5 Fistula -- Stent 584
Grade IV Esophageal Fistula• 45 year old male• Oligometastatic RCC• Symptomatic T3 lesion• 2400 cGy• Cord Dmax < 14 Gy• Esophagus 15 Gy / 2 cc
Grade IV Esophageal Fistula• 4 months: Grade 2 esophageal pain
• 4.5 months: EGD• 3 cm non bleeding ulcer @ 22 cm• Cold forceps biopsy
• 6 months: Worsening pain • Increased ulceration with
superinfection• ¾ circumference with moderate
stenosis• Dilation and cold forceps biopsy
• 6.5 months: Acute development of TEF• Multiple repairs and stent procedures
• 11 months: Expired from distant progression
Conclusions• High dose, single fraction paraspinal SRS has a low
rate of grade ≥ 3 esophageal toxicity• Careful attention to esophageal doses minimizes
toxicity
• MSKCC: 2.5 cc of esophagus ≤ 14 Gy• Radiation recall reactions and iatrogenic manipulation
of the irradiated esophagus predispose for development of grade ≥ 4 toxicity
Toxicity Summary
• Spinal cord injury at current dose levels is extremely rare– Poor statistics because of limited events
• Vertebral body injury is common after spine SBRT– Radiographic 40%– Symptomatic 15-20%
• Esophageal injury is most common and very worrisome complication