(P 052) PATHOLOGICAL AND RADIOLOGICAL RESPONSE AFTER NEOADJUVANT TREATMENTS IN PRIMARY RESECTABLE MYXOFIBROSARCOMA AND UNDIFFERENTIATED PLEOMORPHIC SARCOMA OF THE EXTREMITIES AND TRUNK WALL

Objective: Surgical resection is the mainstay of treatment for patients (pts) with primary localized resectable soft tissue sarcomas of the extremities and trunk wall. Among them undifferentiated pleomorphic sarcoma (UPS) and myxofibrosarcoma (MFS) represent a spectrum of diseases with variable local and distant risks, often addressed by the use of multimodal therapy. Recent studies have shown a possible role of immune check-point inhibitors in these sarcoma types, with preliminary evidence of pathologic response in an exploratory single institution phase 2 study on localized disease. While a larger international and multi-institutional phase 2 randomized trial comparing preoperative radiotherapy (RT) alone and preoperative RT + pembrolizumab in localized UPS and MFS is presently ongoing, we aimed at defining the pathologic response to preoperative chemotherapy (ChT) and RT alone, and concurrent chemo-radiotherapy (ChT-RT) in a contemporary series of UPS and MFS surgically treated at a single institution, to provide some strong historical data to serve as historical control. Correlation between pathological and radiological response and outcome was also investigated.

Methods: All consecutive pts with primary localized MFS and UPS of the extremities and trunk wall surgically treated with curative intent at Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy between Jan 2005 and Jul 2021 were recorded into a prospectively maintained database and retrospectively identified. Among these, pts who received neoadjuvant treatments were identified and divided in three groups: neoadjuvant ChT (naChT), neoadjuvant RT (naRT), and both (naChT-RT). Pts and tumor data (including pathological response defined as the % of residual viable tumor (%VT) on surgical specimen) were collected. Radiological response according to RECIST criteria was retrospectively evaluated. Kaplan-Meier (KM) curves for overall survival (OS) and disease-free survival (DFS) and curves for crude cumulative incidence of local relapse (CCI-LR) and distant metastasis (CCI-DM) were estimated in a competing risk framework according to RECIST response (partial response, PR, stable disease, SD, and progressive disease, PD) and according to pathological response, overall and by treatment group. Cox and Fine and Gray multivariable models were performed in the overall cohort and UPS subgroup to assess the factors’ association with the outcomes. A similar analysis was conducted for the UPS subgroup.

Results: Out of 693 pts (312 MFS and 381 UPS), 233 (66 MFS and 167 UPS) received neoadjuvant treatments and were considered for the present analysis. Median follow-up (FU) was 43.8 months overall (interquartile range, IQR: 14.1- 94.5), being 93.3, 17.3, and 62.8 for naChT, naRT, and naChT-RT groups respectively (IQRs: 41.4-108.6; 4.7-43.8; 32.5-114.3). Pts and tumor characteristics overall and according to treatment group are shown in Table 1.
Median VT was 45% after naChT, 20% after naRT and 10% after naChT-RT (IQRs: 3-70, 10-60, and 5-20, respectively). VT was ≤5% and 0% in 13/46 (28%) and 10/46 (22%), 24/99 (24%) and 5/99 (5%), 40/88 (45%) and 7/88 (8%) of pts treated by naCht, naRT and naChT-RT, respectively. No clinically significant correlation between %VT and any of the end-point analyzed was found, either when analyzed as a continuous variable or as discrete variable with 5% threshold.
Five-yr OS was 82.3, 64.6 and 61.9% for PR, SD, and PD respectively (95% CIs: 68.5-98.9; 55.2-75.5; and 41.9-91.3; p = 0.031); 5-yr DFS was 71.5, 53.4, and 21.8% for PR, SD, and PD respectively (95% CIs: 58.4-87.6; 44.7-63.8; 9.9-47.9; p < 0.001); 5-yr CCI-LR was 2.6, 3.4 and 22 for PR, SD, and PD respectively (95% CIs: 0.4-18.5; 1.3-9.1; 9.5-50.8; p = 0.006); and 5-yr CCI-DM was 25.9, 39.3 and 56.2 for PR, SD, and PD respectively (95% CIs 14.9-45.0; 31.0-49.7; 39.6-79.8; p = 0.005) (Figure 1).
Pts and tumor characteristics of the UPS subgroup are shown in Table 2. Median VT% was 40%, 10% and 10% (IQRs: 3-70, 5-40, and 5-20, respectively) and VT was ≤5% and 0% in 11/38 (29%) and 8/38 (21%), 17/60 (28%) and 3/60 (5%), and 34/69 (49%) and 7/69 (10%), after naCht, naRT and naChT-RT, respectively. Similarly to overall results, while RECIST response was significantly related to the outcome, VT% was not.
Table 3 shows results of the multivariable analysis for OS, DFS, CCI-LR and DM of the entire cohort and UPS subgroup.

Conclusion: In this large series of primary, localized, resectable MFS and UPS of the extremities and trunk wall treated by multimodal therapy at a single institution, pathologic response was not associated to the overall outcome, while radiologic response by RECIST was. Pathologic response to preoperative treatments in these diseases may not serve as a surrogate marker for treatment efficacy, unless proven otherwise in formal comparative prospective studies. Radiologic response by RECIST remains the most reproducible and reliable surrogate of treatment efficacy.

Table 1.docx

Table 2.docx

Table 3.docx

Figure 1.docx