Histologic Response Following Pre-Operative Radiotherapy for Soft Tissue Sarcoma (STS) in a French Reference Center and Review
A B S T R A C T
Background and Purpose: Limb sparing surgery and radiotherapy is the main treatment of patients harboring soft tissue sarcoma of the extremity. There is limited data regarding the prognostic impact of histologic response after pre-operative radiotherapy.
Patients and Methods: Between 2010 and 2018, 123 patients were treated with a pre-operative radiotherapy for soft tissue sarcoma at Leon Berard Centre (Lyon, France) and were retrospectively reviewed. All patients received a dose of 50 Gy in 25 fractions. The histologic response has been analysed by considering the following factors: necrosis ≥ 90%, percentage of viable tumor cells ≤ 10% and fibrosis ≥ 10%. Overall survival (OS), local recurrence-free survival (LRFS), distant recurrence-free survival (DRFS) and event-free survival (EFS) were evaluated.
Results: Median follow up was 33.2 months (range 2.3-128.1 months). Local recurrence occurred in 9 patients (7.5%) and 40 patients (33%) presented a distant recurrence. The 2 and 5-year OS was 84% and 63%. Histologic response factors (necrosis ≥ 90%, viable tumor cells ≤ 10% and fibrosis ≥ 10%) were not predictive in DRFS and EFS. In multivariate analysis, grade was the only significant prognostic factor for EFS P=0.0087. Among the 14 patients with ≤ 10% viable cells after irradiation 13 presented a metastatic evolution within 6 months.
Conclusion: This study showed that current histological response evaluation based on necrosis, fibrosis and viable cells could not predict clinical outcomes after radiotherapy for extremity soft tissue sarcoma. A significant proportion of patients with a good response after pre-operative radiotherapy present a metastatic recurrence.
Keywords
Sarcoma, pathologic response, pre-operative radiotherapy
Introduction
Soft tissue sarcomas (STS) are rare tumors and their management is still controversial because of more than 100 histological subtypes. The combination of conservative surgery and radiotherapy (RT) has widely replaced amputations [1]. At the localized stage, surgery with negative margins associated with radiotherapy has shown a very good local control rate of 85% for patients with intermediate or high-grade STS [2]. RT can be delivered pre-operatively or postoperatively [3]. Local control seems to be equivalent, but doses volumes and complications rates are different [4-6]. Neoadjuvant treatment has substantial effects on the histological evaluation, the histological response, particularly for necrosis, has been established as a prognostic factor in osteosarcoma and Ewing sarcoma but not for soft tissue sarcomas (STS) [7, 8]. Identifying predictors of outcome following neoadjuvant treatment could be useful to identify patients’ subgroups for treatment intensification and/or to look for arguments to better adapt the timing of the RT.
The objective of this study is to evaluate the impact of histologic response to RT in terms of overall survival (OS), local recurrence-free survival (LRFS), distant recurrence-free survival (DRFS) and event-free survival (EFS) in a series of patients treated in a sarcoma expert French Center.
Materials and Methods
I Study Population
One hundred twenty-three patients with localized STS of the extremity, trunk and retroperitoneal area treated with a pre-operative RT at Leon Berard Centre (Lyon, France) between January 2010 and December 2018 were included in this retrospective study. We excluded patients who received nanoparticles in Nanobiotix, and those who received concomitant trabectedin in TRASTS, patients with Ewing sarcoma or patients with progressive disease after RT leading to surgery contraindication [9, 10]. Patients were extracted from the radiotherapy database cross with CIM10 code C499: malignant neoplasm of connective tissue and other soft tissue, unspecified.
Diagnostic core needle biopsies of the tumors were obtained in all patients before radiotherapy. Clinical, histopathological, radiological, and treatment characteristics as well as outcomes, were collected retrospectively from our computerized database. All the histology has been reviewed by expert pathologist from our center. RT technics: RT delivered a dose of 50 Gy in 25 fractions to all patients. The technique was IMRT or 3D technique. Patients were treated with Tomotherapy (proximal extremity of limbs and trunk) whereas, 3D technique was preferred for distal limbs due to the difficulty in saving uneradicated strip of normal tissue with IMRT. GTV was defined as tumor volume defined on pre radiotherapy MRI, a longitudinal margin of 3 cm and a radial margin of 1.5 cm, was added to obtain the CTV except in case of natural barriers such as fascia, bone, or skin surface and additional 5mm for the PTV.
Evaluation: All patients had 2 pre-treatment imaging (before RT and before surgery) of their primary tumors with magnetic resonance imaging (MRI). Surgery was planned 4-9 weeks after the end of RT. Surgery was performed and the pathological analysis carried out by the sarcoma specialist pathologists of the Leon Berard Centre. The size of the tumor was evaluated from the largest diameter, then we compared the evolution of this larger diameter before and after RT.
The histological response was assessed based on standardized histological reports. The pathologists evaluated the percentage of necrosis, fibrosis and viable cells for each resection specimen. One block per centimeter of the tumor and one full transverse slice of the resected tumor were mapped and sampled for histologic processing. The percentage of necrosis, fibrosis and viable cells assessment was based on the histological examination of the entire tumor area sampled. Factors related to a better prognostic have been evaluated, such as histologic response factors (necrosis ≥ 90%, % viable tumor cells ≤ 10% and fibrosis ≥ 10%). In this study, good responders are defined as patients who had viable tumor cells ≤ 10% regardless of the rate of fibrosis and necrosis, which is analysed independently.
II Statistical Analysis
Continuous variables were described as median with range and categorical variables as numbers with percentages. Median follow-up was determined using the reverse Kaplan Meier method. Survival curves and survival rates were obtained using the Kaplan Meier method. Curves were compared using the log-rank tests.
Univariate Cox model regressions were used to find factors associated with disease recurrence-free survival. Hazard ratios were given with their 95% confidence interval (95% CI). All variables with a p-value of less than 0.15 in univariate analyses were entered in the multivariate model. Variable with more than 20% of missing data were not considered for the multivariate model. The full multivariate model is presented. Tests were two-sided and p-values lower than 0.05 were considered significant. All analyses were performed using SAS 9.4.
Table 1: Patients characteristics.
Sex |
Women |
60(49%) |
|
Men |
63(51%) |
Age |
Median |
63 years (15-86y) |
|
Extremity |
81(66%) |
Localisation |
Tronc |
16(13%) |
|
Head &Neck |
1(1%) |
|
Pelvis |
25(20%) |
Histologic type |
UPS |
46 (37%) |
|
Myxofibrosarcoma |
16 (13%) |
|
Liposarcoma dedif / Pleo |
14(11,5%) |
|
Liposarcoma Well diff |
10(8%) |
|
Leiomyosarcoma |
9(7,5%) |
|
Synovialosarcoma |
8(6,5%) |
|
Liposarcoma myxoide |
8(6,5%) |
|
MPNST |
2(2%) |
|
Others |
10(8%) |
Grade |
1 |
13(12%) |
|
2 |
50(47%) |
|
3 |
43(41%) |
Size> 10 cm |
yes |
70(60%) |
|
no |
47(40%) |
Mitosis count >19 |
yes |
78(73,5%) |
|
no |
28(26,5%) |
KI67>50% |
yes |
54(78%) |
|
no |
15(22%) |
Chemotherapy |
yes |
88(71,5%) |
|
no |
35(28,5%) |
Medical staff |
yes |
6(5%) |
|
no |
117(95%) |
Necrosis 95% |
yes |
84(88%) |
|
no |
11(12%) |
|
Missing data |
28() |
Complication |
yes |
82(69,5%) |
|
no |
36(30,5%) |
Reconstruction |
yes |
79(66%) |
|
no |
41(34%) |
Complications |
infectious |
3 |
|
pain |
1 |
|
necrosis |
9 |
|
haemorrhage |
3 |
|
fistula |
2 |
|
Healing problems |
12 |
|
Renal failure |
1 |
|
NR |
5 |
Results
I Patients’ Characteristics
The cohort included 123 patients. Patient, tumor and treatment characteristics are summarized in (Table 1). The median age was 63 years (range, 15-86 years). The main localization was the extremity in 88 patients (66%). The largest diameter was > 5cm for 76% of patients and > 10 cm for 40% of patients. The most common histologic groups were UPS (37%) and 88% of sarcoma were grade 2-3. Multidisciplinary consultation meetings for pre-operative RT was performed in 95% of patients. Neoadjuvant chemotherapy was performed in 28% of patients. The T1 Gadolinium median size change before and after radiotherapy was + 3 mm (range, -77 mm, +95 mm). The margins status was R0 by 76%. Thirty-four percent of the patients needed a flap reconstruction and 30% of the patients have had post-operative complications.
II Histologic Response
Necrosis has been evaluated in 95 patients. The Median rate of necrosis was 10% (range, 0-100%). Necrosis ≥ 90% has been highlighted in 11 patients (12%). The percentage of viable cells is available for 100 patients. Viable tumor cells ≤ 10% have been highlighted in 34 patients (34%). The median percentage of viable cells after radiotherapy was 27.5% (range 0-100%). Thirteen patients had ≤ 1% of viable tumor cells and 5 patients had no viable residual tumor (pathologic complete response). All patients with ≤ 10% of viable tumor cells after radiotherapy also had a high rate of necrosis. (≥ 90%). Fibrosis has been evaluated in 70 patients. The median rate of fibrosis was 20% (range 0-100%) and for 54 patients (77%), fibrosis ≥ 10% was described.
Table 2: Median, LRFS, DRFS, EFS and OS.
|
Median |
At 2years |
At 5 years |
LRFS |
Not reached |
94% |
80% |
DRFS |
Not reached |
69% |
54% |
OS |
Not reached |
84% |
63% |
EFS |
44,5 months (min:27-max:not reached) |
63% |
41% |
III Survival
Median follow up was 33.2 months (range 2.3-128.1 months). Nine patients (7,5%) presented a local recurrence and 40 patients (33%) presented a distant recurrence. At the end of the study, 22 patients died (20%) of cancer progression. The 2 and 5-years OS was 84% and 63%. The 2 and 5-year EFS was 63 and 41%. LRFS and DRFS were 94% and 69% at 2 years and 80% and 54% at 5 years. These results are summarized in (Table 2).
IV Prognostics Factors
i Univariate Analysis
In univariate analysis, none of the histologic response factors (necrosis ≥ 90%, viable tumor cells ≤ 10% and fibrosis ≥ 10%) were predictive with the statistically significant differences in DRFS and EFS. R1 margins does not have a significant impact on DRFS and EFS. On the other hand, the factors related to the tumor are predictive of the patient's outcome. Grade, mitosis count ≥ 19 (calculated from the number of mitosis for 10 High power fields corresponding to 1.734 mm2), Ki 67 ≥ 50% have an impact on EFS, as well as size > 10 cm. Grade and mitosis count ≥ 19 and Ki 67 ≥ 50% influence the DRFS. Those results are summarized in (Table 3). Regarding histological classification, no statistically significant difference in terms of histologic response factors was found.
Table 3: Univariate analysis for DRFS and EFS.
|
|
DRFS |
|
|
|
|
EFS |
|
|
|
|
nb |
HR |
95%CI |
|
p |
nb |
HR |
95%CI |
|
p |
Sex Women Men |
120 |
Ref 1.297 |
0,695 |
2,419 |
0,4135 |
123 |
Ref 1,346 |
0,773 |
2,343 |
0,2942
|
Grade 1 2 3 |
103 |
Ref 2.398 8.83 |
0.54 1.367 |
10.649 24.863 |
0.0058 |
106
|
Ref 2.435 4.588 |
0.718 1.380 |
8.256 15.249 |
0.0131 |
Large diameter <10 ≥10 |
114 |
Ref 1.572 |
0.829 |
2.982 |
0.1658 |
117 |
Ref 1.861 |
1.052 |
3.29 |
0.0327 |
Mitoses <19 ≥19 |
103 |
Ref 2.5 |
1.294 |
4.833 |
0.0064 |
105 |
Ref 2.12 |
1.171 |
3.841 |
0.0132 |
Ki 67 <50 ≥50 |
67 |
Ref 2.486 |
1.1068 |
5.575 |
0.0271 |
69 |
Ref 2.467 |
1.154 |
5.272 |
0.0198 |
Fibrosis <10 ≥10 |
68 |
Ref 0.653 |
0.284 |
1.504 |
0.317 |
70 |
Ref 0.668 |
0.314 |
1.42 |
0.2947 |
Chemotherapy Yes No |
120 |
Ref 1.462 |
0.762 |
2.806 |
0.2538 |
123 |
ref 1.447 |
0.813 |
2.574 |
0.2091 |
Margin R0 R1 R2 |
115 |
Ref 1.268 1.42 |
ref 0.598 0.338 |
2.688 5.971 |
0.7614 |
118 |
Ref 1.548 2.685 |
0.797 0.945 |
3.007 7.633 |
0.1106 |
Necroses <90% ≥90% |
94 |
Ref 2.032 |
0.887 |
4.658 |
0.0393 |
95 |
ref 1.569 |
0.699 |
3.524 |
0.2751 |
Viable cells >10% <10% |
98 |
Ref 1.36 |
ref 0.697 |
2.563 |
0.3669 |
100 |
ref 1.14 |
0.599 |
2.169 |
0.6905 |
ii Multivariate Analysis
In multivariate analysis, the only grade was a significant prognostic factor for EFS P=0.0087. For multivariate analyses, Ki67 and fibrosis were not included because of a higher rate of missing data. These results are summarized in (Table 4).
Table 4: Multivariate analysis for DRFS and EFS.
DRFS |
|
Number of patients |
HR |
95 %CI |
|
p |
Grade |
1 |
81 |
ref |
|
|
0,0641 |
|
2 |
|
1,993 |
0,437 |
9,093 |
|
|
3 |
|
4,113 |
0,923 |
18,326 |
|
Nécrose |
|
81 |
|
|
|
0,1392 |
|
<90% |
|
ref |
|
|
|
|
≥90% |
|
1,953 |
0,804 |
4,745 |
|
EFS |
|
nb |
HR |
95% |
CI |
p |
Grade |
|
94 |
|
|
|
0,0087 |
|
1 |
ref |
|
|
|
|
|
2 |
|
1,541 |
0,332 |
7,143 |
|
|
3 |
|
4,239 |
0,977 |
18,388 |
|
Size |
<10cm |
|
|
|
|
0,142 |
|
≥10cm |
|
1,678 |
0,841 |
3,348 |
|
Margins |
R0 |
|
ref |
|
|
|
|
R1 |
|
1,442 |
0,661 |
3,145 |
|
|
R2 |
|
1,925 |
0,443 |
8,37 |
|
“Good responders versus bad responders” (patients who had viable tumor cells ≤ 10). No statistical differences (p= 0,5) were observed in OS between good and bad responders (Figure 1). It should be noted that among patients with distant recurrence (40 patients), 14 patients were good responders and 22 patients were bad responders. The majority of “good responders” were women (10/14), UPS (10/14), extremity (11/14), RO surgery (12/14), 10/11 had not received CT and 6/14 had complications. Among the 14 good responders who had metastatic recurrences, 13 presented this evolution within 6 months.
Figure 1: OS ‘good responders’ versus ‘bad responders’.
Discussion
We presented a cohort of 123 patients treated with pre-operative RT for limb and trunk sarcoma. In our study, tumors are lesions larger than 5 cm in 76% of cases and 88% are grade 2-3 which is in accordance with ESMO recommendations for pre-operative RT [11].
The results in terms of LRFS, OS and DRFS are comparable to those of the important series of the literature. With a median FU of 33 months, 2 years LRFS, OS and DRFS are respectively 94% (9 local recurrences) 69% and 84%. Wang et al., reported a 2-year LRFS at 94%, a 2-year DRFS at 65%, and 2-year OS at 80% very similar to results of the present series [5]. As it has already been published in pre-operative RT series, the majority of tumors tend to increase in size at the end of radiotherapy [12]. The median volume variation intended for this study is +3 mm (range, -77 mm +95 mm). R0 margins were obtained in only 76% of patients, and R1 in 30 cases. The impact of margins on local control does not seem to have the same impact in the case of pre-operative versus post-operative RT [13]. Radiotherapy allows closer surgical margins to preserve critical structures and function [14].
Table 5: Response according histology.
|
<10 % viable cell |
≥90%Necrosis |
≥10% Fibrosis |
Leiomyosarcoma 9pts |
2pts 2NS |
0 pts 2 NS |
3pts 5 NS |
Myxoide liposarcoma 11pt;s |
0 pts 3NS |
0 pts 2 NS |
5pts 3NS |
Liposarcoma 24 pts |
2pts 9NS |
0 pts 2 NS |
9 pts 12NS |
UPS 46 pts |
20 pts 5NS |
8 pts 9 NS |
22 pts 14NS |
Myxofibrosarcoma 16 pts |
5 pts 4 NS |
1 pts 6 NS |
5pts 10NS |
Synovialosarcoma 11pt;s |
3pts 0NS |
1 pts 1 NS |
5pts 2NS |
Others 12 pts |
1pts 3NS |
1 pts 5 NS |
4 pts 7NS |