Keywords
chordoma - clivus - skull base - radiation - stereotactic radiosurgery
Introduction
Chordomas follow a malignant course defined by local invasiveness and destruction,
a tendency to recur, and occasional distant metastases. Although a consensus approach
on the multimodal treatment of newly diagnosed skull base chordomas exists,[1] there is a paucity of outcomes data directing management of locally and systemically
recurrent/progressive skull base chordomas. The surgical treatment of skull base chordomas
has advanced with improvements in endoscopic and open cranial approaches, and the
development of salvage reconstruction techniques has led to the consideration of repeat
surgery at the time of tumor recurrence. Treatment of progressive disease, however,
is currently complicated by a lack of effective systemic therapeutic options and the
unknown efficacy of re-irradiation. In general, the ability to offer curative treatment
at the time of recurrence is rare for skull base malignancies, and selecting an optimal
therapy requires a multidisciplinary approach focused on striking a balance between
local control, quality of life, and potential morbidity of further interventions.
We recently described factors affecting overall disease control and survival in patients
with progressive skull base chordoma at our institution.[2] In this study, it was noted that disease progression after radiation therapy was
more difficult to control than after surgery alone, and the presence of distant metastases
or leptomeningeal disease conferred a poorer disease-specific survival and freedom
from progression (subsequent progression at any local or systemic site). Moreover,
repeat surgical resection followed by adjuvant radiotherapy (RT) at the time of progression
significantly improved both local and systemic disease control in patients presenting
after previous surgical resection alone. In the case of postradiation disease progression,
no single treatment paradigm was found to be effective in reducing the risk of local
or distant disease progression.
In this study, we focused on our ability to offer palliative interventions for local
and distant recurrences in patients previously treated with surgery and radiation
therapy. The nature of the advanced disease in these patients prevents a cure. To
our knowledge, this is the only report systematically examining rates of site-specific
control by treatment modality in the treatment of local or distant chordoma recurrence.
The aim of this study was to identify factors influencing the ability to control individual
areas of disease independent of whether or not there is progression at either local
or distant sites. We hypothesized that treatment with stereotactic radiosurgery (SRS)
provides effective tumor control in skull base and spine recurrences, with acceptable
morbidity in patients with few remaining treatment options. To our knowledge, this
is the first report specifically addressing the role of SRS in the management of locally
progressive skull base chordomas in the post-RT setting.
Methods
Study Population
A retrospective review of all patients treated at our institution for skull base chordomas
between 1993 and 2016 was performed. The study was performed under an Institutional
Review Board approved protocol in compliance with regulations set by our institution
for the study of human subjects and met all HIPAA (Health Insurance Portability and
Accountability Act of 1996) standards. For this type of retrospective study, patient
consent was not required. Patients were identified through a search of a prospectively
collected registry. From 1993 to 2016, 91 newly diagnosed chordomas of the skull base
have been treated at our institution. Based on the study design shown in [Fig. 1], 16 patients with 40 episodes of post-RT (proton or photon therapy) disease progression
(median follow-up length: 28 months) were eligible for inclusion. Clinical charts,
imaging, pathology, and treatment regimens were reviewed.
Fig. 1 Study flow chart.
Our previous work assessed the impact of treatment in preventing further progression
in the overall disease status (either local or systemic progress). This outcome was
defined as the freedom from progression. The primary outcome in this study was freedom
from treatment site progression (FFTSP) defined as the time between treatment for
an individual site of progressive disease to subsequent progression at that treated
site with radiographic demonstration of unequivocal increase in residual tumor or
tumor recurrence based on the RECIST (response evaluation criteria in solid tumors)
criteria ([Fig. 2]). For patients with multiple recurrences, each unique site of recurrent tumor (local
or distant) was assessed as a separate event. At each time point, the overall pattern
of progression was classified as local (skull base) progression only, distant metastasis
only, or combined local and distant disease progression. Of note, our previous work has demonstrated no impact
of any particular treatment modality on disease-specific survival; hence, this was
not used as a secondary outcome in the presented study.
Fig. 2 Depiction of primary outcomes. Freedom from treatment site progression defined as
the time between the treatment for an individual site of progressive disease and subsequent
progression at that treated site, with radiographic demonstration of unequivocal increase
in residual tumor or tumor recurrence independent of progression elsewhere locally
or systemically. Dashed square highlights local progression independent of the original treated site.
Histological Subtype
All tissue samples were reviewed at the time of treatment by pathologists experienced
in the evaluation of bone and soft tissue sarcomas and were classified into conventional,
chondroid, and dedifferentiated chordomas based on accepted histological criteria.
Statistical Analysis
Frequency distribution and summary statistics were calculated for all variables. Kaplan–Meier
estimates of FFTSP were calculated; survival curves were compared by using the log-rank
test. Univariate and multivariate predictors of FFTSP were assessed using the Cox
proportional hazards model. Relative risk ratios (RRs) and their 95% confidence intervals
were computed. When appropriate, a two-tailed Student's t-test was performed to assess correlation. A p-value of <0.05 was considered significant for all analyses. All analyses were performed
using SPSS (version 2015, IBM Inc.).
Results
Demographic and Descriptive Data
During the study period, 16 patients were treated for recurrence or progression of
their chordoma following prior radiation therapy. Multiple recurrences were common,
and each site of progression was counted as a unique event or progressive episode.
In total, there were 40 instances of disease progression, with a median of 2.4 episodes
(range: 1–5) per patient. The maximum number of progressive events recorded in a single
patient was 5. The majority of patients received proton therapy (n = 11) at the time of initial treatment, and the remaining 5 patients had been treated
with intensity-modulated radiation therapy ([Table 1]). Progression was classified as (1) local disease confined to the skull base, (2)
distant disease, or (3) both local and distant. The most frequent site of chordoma
recurrence was local and was present in 78% of the episodes. Distant locations affected
included the spinal column (n = 7), leptomeninges (n = 7), and other organ sites (n = 13). For cases of local tumor progression, the pattern of recurrence was further
classified as occurring within the prior radiation field (infield failure), at the
margin (marginal failure), or entirely outside of the irradiated field (outfield failure).
Nearly half of the radiation failures were infield treatment failures ([Table 2]).
Table 1
Baseline data at initial episode of disease progression postradiation therapy
|
Characteristic
|
Total number (% of cohort)
|
|
Total number of progression episodes
|
40 tumors (16 patients)
|
|
Sex
|
|
Male
|
7
|
|
Female
|
9
|
|
Median age
|
46 y (range: 3–74)
|
|
Radiation modality at initial definitive treatment
|
|
Proton therapy
|
11 patients
|
|
Intensity-modulated radiation therapy (photon)
|
5 patients
|
|
Histological subtype
|
|
Chondroid
|
7 patients (43.7%)
|
|
Conventional
|
8 patients (50%)
|
|
Dedifferentiated
|
1 patient (6.3%)
|
|
Median number of progressions per patient
|
2.4
|
Table 2
Baseline data regarding all episodes of disease progression
|
Characteristic
|
Total number (% of cohort)
|
|
Total number of progression episodes
|
40
|
|
Patterns of progression
|
|
Local only
|
23 (57.5)
|
|
Distant only
|
9 (22.5)
|
|
Local + distant
|
8 (20)
|
|
Distant metastases
|
17 episodes total
|
|
Spinal column metastases
|
7 (17.5)
|
|
Leptomeningeal disease
|
7 (17.5)
|
|
Systemic metastases
|
13 (32.5)
|
|
Patterns of local progression
|
38 episodes total
|
|
Infield failure
|
18 (47.4)
|
|
Marginal failure
|
5 (13.2)
|
|
Outfield failure
|
15 (39.4)
|
|
Treatment paradigms
|
|
Local progression
|
38 episodes
|
|
Repeat surgery
|
9 (24)
|
|
Chemotherapy
|
12 (32)
|
|
Stereotactic radiosurgery
|
13 (34)
|
|
Systemic metastases
|
17 episodes total
|
|
Surgery
|
7 (41)
|
|
Chemotherapy
|
13 (76)
|
|
Radiation therapy
|
8 (47)
|
Treatment Paradigms
At the time of local disease progression, patients were offered repeat surgical resection,
chemotherapy, or SRS. The majority of local tumor recurrences were treated with SRS
(n = 13; 34%) and various cytotoxic or targeted chemotherapy agents (n = 12; 32%). Additional surgical resection was less frequently used in this cohort
of patients who had already undergone an extensive skull base surgical resection and
adjuvant radiation therapy at the time of their initial diagnosis and treatment. Nine
patients underwent repeat surgical resection followed by adjuvant SRS approximately
3 to 4 weeks postoperatively. The two most common reasons for the low rate of repeat
of surgical resection were as follows: if the recurrent disease was considered low
volume amenable to SRS or if it was considered unresectable (i.e., circumferential
encasement of a major vascular structure despite a previous resection through a well-executed
surgical approach). However, for infield or marginal failure adjacent to the radiation
sensitive structures (i.e., brainstem), repeat surgery was pursued if it was felt
it would help reduce the risk of radiation toxicity. In the setting of distant or
systemic metastases, patients were also treated with a combination of surgical resection
(n = 7), chemotherapy (n = 17), and/or radiation therapy (n = 8). Ultimately, six of the spinal metastases were treated with radiation therapy.
The 13 locally progressive tumors treated with re-irradiation were treated with gamma
knife SRS (GKSRS). The median treatment volume was 23.7 cm3, with a range of 14 to 73 cm3. The median margin and maximum doses were 15 and 28 Gy, respectively ([Table 3]). Following SRS, seven of the tumors remained stable in size and six exhibited evidence
of tumor regression. Re-irradiation with SRS was not associated with any adverse radiation
events (AREs) or radiation necrosis, and there were no instances of tumor progression
seen during the median follow-up of 28 months.
Table 3
Re-irradiation for local skull base progression: treatment parameters and outcomes
|
Variable
|
Total number (13 treated tumors)
|
|
GKSRS parameters
|
|
Median radiation volume (range), cm3
|
23.7 (14–73)
|
|
Margin dose, Gy
|
|
|
Median (range)
|
15 (12–20)
|
|
Maximum dose, Gy
|
|
|
Median (range)
|
28 (17–32)
|
|
Tumor response
|
|
Regression
|
6
|
|
Stable
|
7
|
|
Progression
|
0
|
|
Radiation-induced changes
|
|
AREs
|
0
|
|
Radiation necrosis
|
0
|
Abbreviations: AREs, adverse radiation events; GKSRS, gamma knife stereotactic radiosurgery.
Main Results: Factors Affecting Freedom from Treatment Site Progression
The primary outcome measured was FFTSP after salvage therapy. For the entire cohort,
the mean FFTSP was 47.2 months (range: 1–65 months) ([Table 4]). The median disease-specific survival was 36 months (range: 2–114 months). Analysis
of histological subtype revealed a relationship between histology and FFTSP. Patients
treated for the recurrence of chondroid chordoma had significantly longer mean FFTSP
(71.2 months) than conventional (29.8 months) or dedifferentiated subtypes (20 months).
This was significant (p = 0.018) in multivariate analysis ([Fig. 3A]). The pattern of overall disease progression and distant metastases was associated
with FFTSP in multivariate analysis. Patients presenting with simultaneous local and
distant disease progression had significantly shorter durations of FFTSP (7.1 months)
versus those with either local recurrence only (56.3 months) or distant metastases
only (51.2 months) disease ([Fig. 3B]). In the case of distant disease, spinal column metastases were not associated with
worse FFTSP. The presence of leptomeningeal disease, however, was found to correlate
with shorter FFTSP (10.4 vs. 43.5 months; p = 0.022). Other factors at presentation that were not found to have a significant
association with FFTSP in multivariate analysis included the presence or absence of
distant metastasis and if the local treatment failure was infield, marginal, or outfield
([Table 4]). Of note, radiation modality at initial treatment (proton or photon therapy) was
initially found to be significant in univariate analysis but, ultimately, not in multivariate
analysis. Of the factors at presentation that were included in the analysis, only
third or higher recurrence, histological subtype, and local and distant disease progression
had a meaningful influence on FFTSP. Lastly, while a previous radiation dose of >74 Gy
was significantly indicative of improved site control in univariate analysis (24.5
vs. 57.7 months, p < 0.05; RR: 3.8; p < 0.05), this was not deemed to be significant in further multivariate analysis.
Fig. 3 (A) Kaplan–Meier curve demonstrating the impact of histology on freedom from treatment
site progression (FFTSP). (B) Kaplan–Meier curve demonstrating the impact of pattern of recurrence on FFTSP. (C) Impact of re-irradiation with stereotactic radiosurgery on FFTSP.
Table 4
Factors affecting FFTSP
|
Mean FFTSP (months)
|
p-Value
|
Univariate Cox regression analysis (95% CI)
|
p-Value
|
Multivariate Cox regression analysis (95% CI)
|
p-Value
|
|
Overall cohort 47.2
|
|
Factors at presentation
|
|
Progression episode number
|
|
|
1
|
54.9
|
0.036
|
|
|
|
|
|
2
|
54.5
|
|
|
|
|
|
3
|
19.3
|
2.8 (1.2–8.1)[a]
|
0.047
|
10.9 (2–58.8
|
0.006
|
|
4
|
15.6
|
|
|
|
|
|
5
|
5.5
|
|
|
|
|
|
Histological subtype
|
|
Chondroid
|
71.2
|
0.06
|
|
|
|
|
|
Conventional
|
29.8
|
3.2 (1.1–9.3)[b]
|
0.029
|
5.4 (1.3–21.7)[b]
|
0.018
|
|
Dedifferentiated
|
20
|
|
|
|
|
|
Radiation modality received at initial treatment
|
|
Proton therapy
|
53.5
|
0.006
|
|
|
|
|
|
Photon therapy
|
14.7
|
3.8 (1.3–10.8)
|
0.012
|
2.6 (0.26–27.6)
|
0.407
|
|
Radiation dose received at initial treatment
|
|
73 Gy or Less
|
24.5
|
|
|
|
|
|
|
74 Gy or More
|
57.7
|
0.021
|
3.8 (1.1–14.1)
|
0.041
|
2.8 (0.49–16.1)
|
0.288
|
|
Presence of distant metastasis
|
|
No
|
55.6
|
0.06
|
|
|
|
|
|
Yes
|
31.9
|
2.5 (0.9–6.9)
|
0.076
|
|
|
|
Pattern of overall disease progression
|
|
Local only
|
56.3
|
<0.005
|
|
|
|
|
|
Distant only
|
51.2
|
10 (3.14–31.3)
|
<0.005
|
5.4 (1.1–28.6)
|
0.045
|
|
Local + distant
|
7.1
|
10 (2.5–62.5)
|
<0.005
|
9.1 (1.2–66.6)
|
0.029
|
|
Pattern of local failure
|
|
Infield failure
|
|
Yes
|
36.7
|
0.124
|
|
|
|
|
|
No
|
43.4
|
2.65 (0.76–9.7)
|
0.141
|
1.068 (0.226–5.040)
|
0.934
|
|
Marginal failure
|
|
Yes
|
26.8
|
0.902
|
|
|
|
|
|
No
|
47.3
|
1.09 (0.24–5.02)
|
0.903
|
|
|
|
Outfield failure
|
|
Yes
|
34.6
|
0.697
|
|
|
|
|
|
No
|
47.4
|
1.24 (0.41–3.7)
|
0.699
|
|
|
|
Pattern of distant metastases
|
|
Spinal column metastases
|
|
Yes
|
47.6
|
0.05
|
|
|
|
|
|
No
|
15.5
|
3.4 (0.8–18.9)
|
0.082
|
|
|
|
Leptomeningeal disease
|
|
Yes
|
10.4
|
0.012
|
4.6 (1.2–16.9)
|
0.022
|
|
|
|
No
|
43.5
|
|
|
|
|
|
Local progression treatment
|
|
Surgery
|
|
|
|
|
|
|
|
Yes
|
40.8
|
0.552
|
|
|
|
|
|
No
|
37.8
|
1.4 (0.41–5.4)
|
0.6
|
1.021 (0.220–4.731)
|
0.98
|
|
Chemotherapy
|
|
|
|
|
|
|
|
Yes
|
20.8
|
0.008
|
4.2 (1.3–13.3)
|
0.015
|
2.4 (0.32–18.7)
|
0.39
|
|
No
|
70.8
|
|
|
|
|
|
Radiation therapy
|
|
|
|
|
|
|
|
Yes
|
77.3
|
<0.005
|
|
|
|
|
|
No
|
22.4
|
6.1 (1.7–21.7)
|
0.006
|
11.9 (1.3–109.1)
|
0.029
|
Abbreviations: CI, confidence interval; FFTSP, freedom from treatment site progression.
a First/second progression versus third or higher progression.
b Chondroid versus conventional and dedifferentiated histological subtype.
The type of treatment modality used in treating progressive disease had an impact
on local control or FFTSP. There was no improvement in primary outcome measures for
patients offered repeat surgical resection of their local recurrence. The mean FFTSP
for patients undergoing surgery versus nonoperative management was 40.8 and 37.8 months,
respectively (p = 0.98). Patients treated with chemotherapy received either targeted therapies or
other cytotoxic medications. Surprisingly, the mean FFTSP was actually shorter for
patients receiving chemotherapy (20.8 months) compared with those who did not receive
chemotherapy (70.8 months). This difference was significant in univariate analysis
(p = 0.015) but did not reach significant in multivariate analysis (p = 0.39). Patients receiving re-irradiation for their local skull base recurrence
or spinal disease experienced significantly longer mean times of freedom from progression
(77.3 months) compared with those treated with nonradiation-based modalities (22.4
months) ([Fig. 3C]). This correlation was significant in multivariate analysis (p = 0.029). Of the treatments offered to patients in this series, SRS was the only
paradigm providing a significant improvement in local control of advanced or recurrent
chordoma. As previously described, radiation treatments were well tolerated with no
reported adverse radiation outcomes including radiation necrosis.
Discussion
Although there is consensus regarding the use of multimodal treatment in the definitive
management of chordomas at diagnosis, there is a paucity of data related to the management
of advanced disease. The management of tumor recurrence after previous radiation clearly
poses several management challenges; the ability to achieve a gross total resection
from surgery is often limited,[3]
[4]
[5]
[6] and surgical intervention alone without any adjuvant therapy is unlikely to provide
any durable tumor response. Previous radiation, with either photon- or proton-based
modalities, has previously been considered a contraindication to repeat radiation;
studies on other skull base malignancies is now demonstrating a role of SRS as a re-irradiation
modality.[7]
[8]
[9] Determining the optimal therapy for local or distant recurrence is complex and requires
consideration of tumor location, functional status, histology, disease burden, and
prior interventions.
Our study demonstrates that SRS was the only treatment modality associated with improved
site-specific control in patients presenting with locally recurrent disease after
previous surgical resection and radiation. During the follow-up period, there were
no instances of tumor progression following radiosurgery; the treated areas remained
stable or decreased in size. Interestingly, location of the treated recurrence relative
to the prior radiation field did not limit delivery of an effective and safe radiation
dose, with median maximum and marginal doses of 28 and 15 Gy, respectively; these
dosing schemes are similar to those reported by other GKSRS series in the literature.[10]
[11]
[12]
[13]
[14]
[15] As discussed in the following, it is likely that the judicious use of surgery helped
radiation outcomes for patients with infield/marginal recurrences adjacent to the
brainstem. The largest radiosurgery series to date reported outcomes for 71 patients,
20 of whom were treated for recurrence following prior radiation.[14] In all patients, regardless of previous treatment status, the authors demonstrated
that a margin dose of ≥15 Gy significantly improved tumor control and survival compared
with a dose of <15 Gy. Tumor control following radiosurgery in the prior RT cohort
was 98% at 1 year and decreased to 65% at 7 years. In our series, we observed effective
site-specific tumor control across relatively large treatment volumes (14–73 mL).
The role of SRS in the overall management of skull base chordomas—independent of previous
treatments—has certainly been demonstrated in numerous other publications. Other centers
have reported high rates of local control using radiosurgery to treat local residual
or recurrent disease.[10]
[11]
[12]
[13]
[14]
[15] However, these studies do not exclusively address the management of recurrent disease
after a previous charged particle therapy. Most recently, Förander et al demonstrated
the role of GKSRS in the management of residual disease after no previous radiation
and management of progression after previous SRS.[11] With dosing parameters similar to that in our study, a tumor control rate of 50%
at 15 years for first time GKSRS patients was noted in the setting of an ARE rate
of 18%; notably, a majority of the recurrences were considered out of field. In the
repeat SRS subcohort of six patients, a prominent effect on tumor control was also
noted.
The primary concern with re-irradiation of infield or marginal recurrences, regardless
of modality, is the risk of radiation-associated injury to critical structures including
the brainstem, cranial nerves, pituitary gland, and mesial temporal lobes. The North
American Gamma Knife Consortium observed AREs following SRS in 6 of their 71 patients.[14] The reported morbidities were related to worsened cranial nerve or pituitary dysfunction
and occurred exclusively in patients with prior RT. The incidence of grade 2/3 ARE
(i.e., neurologic toxicities requiring management with either steroids or hospitalization)
in the previous RT group was approximately 30%. In our series, there were no AREs,
as per the Radiation Therapy Oncology Group and the European Organization for Research
and Treatment of Cancer criteria, or radiation necrosis; this is particularly notable
for the infield recurrences treated with SRS. The difference in reported outcomes
between our study and prior publications could be related to several factors. The
site of local recurrence and proximity to critical structures may differ between the
two groups, as well as baseline cranial nerve function. Furthermore, we hypothesize
that judicious use of repeat surgical resection to clear disease away from radiation
sensitive structures, such as the brainstem, helped reduce the risk of radiation toxicity
in our cohort. In our experience, however, radiosurgery for local recurrence is an
effective palliative treatment in patients who cannot expect a cure, offering local
tumor control without imposing significant morbidity.
Similar to our prior study, we found that repeat surgical resection as the sole treatment
modality of post-RT local recurrence did not improve FFTSP. This is consistent with
previous publications reporting statistically and significantly reduced local control
rates with posttreatment surgical intervention. Numerous other studies have indicated
a significantly higher risk of subtotal resection in the setting of previous treatments
(either radiation or surgery alone).[3]
[4]
[5]
[6] Presumably, radiation-induced arachnoid and extradural soft tissue scarring limit
the ability to identify appropriate resection margins, further increasing the risk
of a subtotal resection. As with other malignancies, the role of repeat surgical resection
needs to be considered within the context of advancements in surgical technique and
the ability to resect disease previously considered unresectable. It is important
to acknowledge that our cohort may have skewed analysis. Generally, these were patients
who received their initial surgical treatment at our institution and had already undergone
a maximal skull base resection. Hence, the residual disease was typically adherent
to critical neurovascular structures and not due to an inadequate surgical exposure.
The refinement of expanded endoscopic approaches now allows us to consider repeat
surgery for residual disease in areas (i.e., petroclival synchondrosis, clivus) where
the alternative or traditional strategy would have been associated with more soft
tissue morbidity and a lower likelihood of achieving a gross total resection. Hence,
we do still consider repeat surgery in select patients where the area of progressive
disease is amenable to a gross total resection using a well-designed surgical approach.
In particular, this may be an effective strategy to help salvage patients with high
volume marginal field recurrences where there is room for additional higher doses
of radiation therapy. Furthermore, neurologic symptoms related to compression of cranial
nerves or brainstem may warrant conservative surgical resection. Most importantly,
within the context of this study, we hypothesize that the low observed rate of AREs
was a result of employing surgical resection to clear disease away from radiation-sensitive
structures. This was likely an important reason why radiation complications were not
seen with recurrences considered infield relative to previous radiation fields despite
their proximity to the brainstem and other critical neurovascular structures.
Several patients received cytotoxic or targeted therapies for salvage therapy. There
was no improvement in FFTSP in trials of chemotherapy. In fact, tumor control was
worse for patients treated with chemotherapy. This likely reflects the severity and
aggressiveness of disease in patients who were allocated to medical therapy rather
than a result of the treatment itself. In general, these patients are enrolled in
clinical trials due to the limited remaining options for treatment. Related to this
observation is also a finding in our study, indicating that tumor histology, pattern
of recurrence, and number of recurrences impact local control. Our report also identified
that chondroid histology was associated with significantly increased FFTSP in patients
treated for recurrence compared with conventional and dedifferentiated chordoma. The
ability to control local recurrences was notably shorter for patients on their third
or higher episode of recurrence anywhere locally or systemically. At the time of the
first recurrence, FFTSP was 54 months compared with 5.5 months at the fifth recurrence.
Likely indicative of the aggressive biology of some tumors, factors such as history
and patterns of recurrence are typically considered in salvage management strategies
for other solid malignancies. However, this has yet to happen with chordomas due to
the lack of efficacious systemic therapy options. The suboptimal outcomes associated
with commonly employed receptor tyrosine kinase inhibitors emphasize the need for
the development of effective therapies tailored to molecular or histological subtype
and, perhaps, consideration for systemic therapy earlier in the treatment paradigm
for high-risk tumors.
Study Limitations
Despite the fact that this report draws on the experience of a multidisciplinary team
at a large referral center, it is a relatively small series composed of heterogeneous
patients. The analysis included both local and distant recurrences treated with different
modalities, limiting the statistical power of the study. The decision of salvage therapy
was not standardized, and allocation to different treatments was potentially biased,
affecting analysis of each treatment's efficacy. Specifically, the nature in which
surgical resection was applied introduced a selection bias that may have skewed analyses
regarding the role of surgery and the risk of AREs. Additionally, our outcome was
FFTSP. In this patient population, future efforts should include measures of patient-reported
quality of life.
Conclusions
Treatment of recurrent chordomas is a complex and challenging task. Patients frequently
present with multiple recurrences and have undergone several prior procedures including
surgical resection and radiation therapy. The focus of therapy is different than that
at the time of initial diagnosis as an oncological or complete resection is unlikely.
Thus, treatment should maximize local control with minimal morbidity. We advocate
that surgical resection for local recurrence in the setting of prior radiation should
be limited to situations where a gross total resection is considered possible, such
as for decompression of neural structures and for optimization of radiation fields
to decrease AREs. SRS provides patients with local or distant disease progression
improved local control without significant morbidity or complications.
